Document [original]

Assessment of Household Energy Access:

The Progress out of Energy Poverty Index

(PEPI) Toolkit for the Microfinance Sector

vorgelegt von

Dipl. -Ing.

Natalia Realpe Carrillo

geb. in Bogota, Kolumbien

von der Fakult¨at III - Prozesswissenschaften

der Technischen Universit¨at Berlin

zur Erlangung des akademischen Grades

Doktorin der Ingenieurwissenschaften

- Dr.-Ing. -

genehmigte Dissertation

Promotionsausschuss:

Vorsitzender:

Prof. Dr. Tetyana Morozyuk

Technische Universit¨at Berlin, Institut f¨ur Energietechnik

Gutachter:

Prof. Dr.-Ing. Prof. e.h. Dr. h.c. George Tsatsaronis

Technische Universit¨at Berlin, Institut f¨ur Energietechnik

Prof. Dr. Bernd Balkenhol

University of Geneva, Institute of Economics and Econometrics

Tag der wissenschaftlichen Aussprache: 13. Februar 2017

Berlin 2017

Acknowledgements

After this intense and long period combining academia and consultancy, I am

pleased to reflect on the people who made this achievement possible through

their inmense support and kind help.

At first place, I am grateful to Prof. George Tsatsaronis for his guidance and

collaboration in the development of this research, framed under the multidis-

ciplinary Microenergy Systems Doctoral Program. Moreover, I thank Prof.

Bernd Balkenhol for his detailed feedback and his encouragement in the po-

tential of the tool, which shaped the way to sharpen the dissertation. As my

research was integrated in my consultancy work, I wish to thank my colleagues

from MicroEnergy International for their great support and enthusiasm to-

wards this project, Lukas Kahlen, Raluca Dumitrescu, Mariana Daykova, Giu-

lia Corso, Mathias Koepke, thank you. As well, to Hannes Kirchhoff and Fredy

Ca˜nizares’ for your feedback, which helped to tackle the needs of the sector

and to Yeisson Gil and Meerim Ruslanova, for your constant assistance with

the tool. Special thanks go to Sebastian Groh, whose results and thoughtful

advices boosted this learning phase. I would like to single out Noara Kebir, for

empowering me as woman, as mother, professional and researcher, thanks for

your dedicated mentorship. Together with Daniel Philipp, I thank both for the

impulse and all the opportunities I was given to conduct my PhD at MEI and

build a team around the subject.

I wish to acknowledge the financial support of Zentralen Frauenbeauftragten der

TU Berlin, for the scholarship that allowed me in the last stage of my promotion

to exclusively focus in finalizing the dissertation, and of PlaNet Finance, for

the UMM PhD scholarship and for funding the Case Study. Particularly, to

Karla Henning and Azalea Carish for their backing and invitations to share my

results in diverse events. At Contactar, to Gloria Bustos, Herney Mora, Rither

Jaramillo and Juan Carlos Guerrero, for offering the best conditions to conduct

the field research and share the same motivation to bring this project forward.

To the local team, Juan Carlos, Maydi Melo, Daniel Escobar, Mario Pinto

and Carolina L´opez, thanks to all of you for your discipline and commitment

in your work. Many other people have also contributed in many ways along

this project, such as Germ´an Narv´aez from the University of Nari˜no with the

routes design, Scott Graham from FINCA encouraging the translation of the

tool to ValiData and of course, my Colombian and Italian extended family

with their extraordinary encouragement. Certainly, I also thank immensely for

the feedback, corrections and valuable input from Valentin Ade, Jacky Baca,

Lorena D´ıaz, Phil Lindsay, Setu Pelz, Marco Reiser, Daniel Mock and specially

to Rebecca Ch´avez and my dearest friend, Jocelyn Polen.

Finally, I am truly grateful to Alfonso, my life partner, my best entertainer and

coach, with whom we mutated this project in a joyful family adventure. Your

infinite love, Mart´ın and Chiara Lenu, made this journey plenty of delightful

moments collected all around the world. To my parents and my brother, my

engine to grow and challenge myself: such unbearable love and that effortless

way to convert anything in a reason to laugh, make always my day. Thanks!

Abstract

Sustainable energy for all is one of the explicit goals of the United Nation’s

post-2015 development agenda, aiming to overcome the unaffordability, un-

availability and unsustainability of energy supply. It refers to the one third of

the world’s population that still relies on intermittent or non-existent electricity

supply, and on fossil fuels for cooking and heating; and which is mostly located

in rural areas where there is little or no possibility of obtaining cleaner fuels,

grid connection or decentralized energy systems. Microfinance services, among

other mechanisms, offer a vehicle to improve energy supply of these underserved

population. The number of microfinance institutions (MFIs) providing green

lending programs, i.e., offering consumer finance solutions for the acquisition of

modern energy technologies, has steadily increased. However, the success of the

scaling- up process of green lending is still debatable. Thus, by facilitating tools,

the microfinance sector can estimate its potential in committing to meeting the

proposed global goals and strengthen its role in enabling energy access. So far,

a series of indicators exist to measure the environmental performance of MFIs

but none to measure the energy situation of clients. Such indicators would al-

low MFIs to support decision-making processes, monitor clients’ achievements

in energy access or even supply input to assess other basic needs. Hence, this

research addresses the challenges of measuring energy access at household level,

defining the role of the microfinance sector in achieving and improving energy

access and the misleading potential of poverty assessment metrics to predict the

quality of energy supply. During field research in south-west of Colombia at

the MFI Contactar, a thorough energy needs assessment following the ESMAP

multi-tier framework (MTF) [Bhatia and Angelou, 2015] has been conducted on

a representative sample size (384 households-clients) in rural, peri-urban and

urban areas. The results, capturing much broader dimensions of electricity ac-

cess and cooking facilities, unveil the real quality of energy at the households,

which the majority are currently connected to the grid. Based on the results

collected from the field, adopting the underlying structure of the ESMAP MTF

and refining suggestions from [Groh et al., 2016,Bensch, 2013,Stevens et al.,

2015], the Progress out of Energy Poverty Index (PEPI) toolkit has been de-

signed to measure and monitor access to electricity services and supply and

cooking solutions at household level. It also aims to bring energy services to

the core of the metric, facilitating the implementation and understanding of

the multi-tier approach for the microfinance and energy sector, and provides a

meaningful tool for monitoring and assessing progress.

Keywords:

Green microfinance, energy access metrics, multi-tier approach, rural financial

inclusion, rural energy inclusion, two-hand model,

Abstract

Nachhaltige Energie f¨ur Alle ist eines der expliziten Ziele der Entwicklungsagenda

2015 der Vereinten Nationen, um Erschwinglichkeit, Verf¨ugbarkeit und die

Nachhaltigkeit der Energieversorgung zu sichern. Es bezieht sich auf das eine

Drittel der Weltbev¨olkerung, welches immer noch auf intermittierende oder

nicht vorhandene Stromversorgung und auf fossile Brennstoffen f¨ur Kochen

und Heizen angewiesen ist; Und welches sich ¨uberwiegend in l¨andlichen Ge-

bieten befindet, wo es wenig oder gar keine M¨oglichkeit gibt, saubere Kraft-

stoffe, Netzanschl¨usse oder dezentrale Energiesysteme zu erhalten. Mikro-

finanzierungsdienste, neben anderen Mechanismen, bieten ein Werkzeug zur

Verbesserung der Energieversorgung dieser unterversorgten Bev¨olkerung. Die

Zahl der Mikrofinanzinstitute (MFIs), die Programme zur Verf¨ugung stellen,

d.h., die Verbraucherfinanzierungsl¨osungen f¨ur den Erwerb moderner Energi-

etechnologien anbieten, hat stetig zugenommen. Allerdings ist der Erfolg des

Skalierungsprozesses der Green Lending noch umstritten. Folglich kann der

Mikrofinanzsektor sein Potenzial absch¨atzen, um die vorgeschlagenen globalen

Ziele zu erreichen und seine Rolle bei der Erm¨oglichung des Energiezugangs zu

st¨arken. Bisher gibt es eine Reihe von Indikatoren, um die Umweltleistung von

MFIs zu messen, aber keine, um die Energiesituation der Kunden zu messen.

Solche Indikatoren w¨urden es MFIs erm¨oglichen, Entscheidungsprozesse zu un-

terst¨utzen, die Fortschritte der Kunden im Energiezugang zu verfolgen oder

sogar einen Beitrag zur Bewertung anderer Grundbed¨urfnisse zu liefern. Da-

her befasst sich diese Forschung mit den Herausforderungen, den Energiezu-

gang auf Haushaltsebene zu messen und die Rolle des Mikrofinanzsektors bei

der Erreichung und Verbesserung des Energiezugangs und des irref¨uhrenden

Potenzials von Armutsbewertungsmetriken, zur Vorhersage der Qualit¨at der

Energieversorgung, zu definieren. W¨ahrend der Feldforschung im S¨udwesten

von Kolumbien am MFI Contactar wurde eine gr¨undliche Energiebedarfsbew-

ertung nach dem ESMAP-Multi-Tier-Framework (MTF) [Bhatia and Angelou,

2015], auf einer repr¨asentativen Stichprobengr¨oße (384 Haushalte - Kunden), in

l¨andlichen, peri-st¨adtischen und st¨adtischen Gebieten durchgef¨uhrt. Die Ergeb-

nisse, die viel breitere Dimensionen von Elektrizit¨atszugang und Kochgelegen-

heit erfassen, enth¨ullen die reale Energiequalit¨at in den Haushalten, mit welcher

die Mehrheit derzeit mit dem Netz verbunden ist. Das “Index zur Bewer-

tung der Energie-Armuts-Entwicklung (Progress out of Energy Poverty Index -

PEPI)” Toolkit, wurde auf der Grundlage der aus dem Feld gesammelten Ergeb-

nisse, der Annahme der zugrunde liegenden Struktur des ESMAP MTF und

der Verfeinerung von Anregungen von [Groh et al., 2016,Bensch, 2013,Stevens

et al., 2015] entwickelt, um den Zugang zu Elektrizit¨ats- und Versorgungs- und

Kochl¨osungen auf Haushaltsebene zu messen und zu ¨uberwachen. Es zielt auch

darauf ab, Energiedienstleistungen messbar zu machen, die Umsetzung und das

Verst¨andnis des mehrstufigen Ansatzes f¨ur den Mikrofinanz- und Energiesektor

zu erleichtern und ein sinnvolles Instrument zur ¨uberwachung und Bewertung

des Fortschritts zu bieten.

Keywords:

Green microfinance, Energiezugang Messgr¨oßen, Multi-Tier-Ansatz, l¨andliche

Energie-Integration, Zweihand-Modell,

Summary

Energy access has been recognized as an imperative development factor at eco-

nomic, social and human levels. More than two billion people worldwide lack

access to modern energy, including electricity access and modern cooking so-

lutions. Upon the launching of the Sustainable Development Goals (SDGs),

the definition and measurement of energy access has taken on significant im-

portance for governments, development agencies, private and non-governmental

organizations and financial institutions, among others. By aiming to achieve

affordable and reliable universal access to modern energy services by 2030, the

SDG 7, out of the 17 global goals, calls for a standardized monitoring of progress

towards this goal.

The objective of this research is two-fold. The first aim is to present a detailed

analysis of the energy access of microfinance clients in southwestern Colombia,

focusing on the population involved in micro-lending. In collaboration with

the Colombian MFI Contactar, the study is based on the multidimensional

(multi-tiered) energy access framework (MTF), recently launched by the World

Bank’s Energy Sector Management Assistance Program (ESMAP), which con-

siders several attributes associated with the concept of energy access. Results

revealed the poor quality of electricity supply, compared to the above-average

national rate on the country’s energy access figures and an energy stacking be-

havior in terms of the variety of cooking fuels and devices for cooking according

to household affordability and availability. By revealing the real quality of the

population’s current energy access, the analysis calls for interventions in relia-

bility, quality and affordability, providing on the one hand, valuable information

for service improvements to regional electrical utilities to bridge those gaps and,

on the other hand, a detailed overview of energy needs for cooking. In order to

revise the accuracy of metrics at the MFI level, the research also evaluates the

potential of existent poverty metrics such as the Progress out of Poverty Index

(PPI) [Schreiner, 2004] currently used by the MFI to predict quality of energy

access (PPI Scorecard for Colombia). The assessment results showed that the

Colombian PPI fails to properly describe both the level of energy access and the

availability and affordability for cooking solutions. However, significant corre-

lations between consumption and the likelihood to fall under the poverty level

show an inversely proportional correlation.

Driven by the motivation of aligning sectorial strategies to reach SDGs and fur-

ther promote the multi-tier approach, the second focus of this study consisted of

the development of a toolkit for project implementers, aimed at tracking specific

attribute progress at the household level. In particular, based on the ESMAP

MTF tool [Bhatia and Angelou, 2015] and the results of the related survey, the

research proposes the Progress out of Energy Poverty Index (PEPI) as a toolkit

directed, as an initial step, to MFIs willing to identify and to satisfy the energy

needs of their clientele and track their progress. Considering the close, long-

term relationship and MFIs’ infrastructure, and their steady increasing interest

in satisfying client energy needs and increasing their market, the toolkit can

be adopted by MFIs engaging in green lending. Nonetheless, additional orga-

nizations supplying energy access to the base of the pyramid can also use the

tool to monitor the effects of their products, services or programs. The PEPI

supports organizations in (i) identifying the energy needs of their current and

potential clientele in order to efficiently tailor green lending or energy access

programs; (ii) measuring and tracking program impact on household energy ac-

cess; and (iii) assessing and comparing the effects of modern energy technology

on improving the quality of modern energy services and the level of provision

over time for household electricity services and supply and cooking solutions.

The toolkit is based on the adapted multi-tier frameworks [Bhatia and Angelou,

2015], to assess electricity supply and cooking facilities, and it entails (i) a set of

expanded frameworks in attribute detail and differentiations, (ii) a ready-to-use

survey tool filtered according to electricity power source and cooking solution

and (iii) an index to assess the progress towards energy access for all based on

panel data.

The modified set of metrics for energy access aims at aligning the global goal

SDG 7 targets with the redefinition of energy access according to the multi-tier

framework, grouping the energy access attributes into three global attributes

and assessing progress through the variation of their deltas over time. Hence,

the composite index provides deeper information on the quantity and quality of

energy access obtained from the global attributes associated with energy access,

avoiding misinterpretations of a condensed individual index. The developed

surveys are software-based tools for data collection and analysis, in which access

is exclusively assessed based on the services available to the household. The

survey tool contains the electricity supply and services and cooking facilities

frameworks, whose energy metrics can be adjusted to country specific realities.

No electricity consumption is assessed through an individual tier-framework.

Finally, through the repeated implementation of the toolkit, the PEPI index to

measure the progress out of energy poverty across the different dimensions of

energy access is delivered, based on a weighted average of the progress in the

global attributes.

Upon rolling out this approach and highlighting the priority of achieving uni-

versal energy access, the proposed PEPI toolkit enables organizations engaged

in energy access activities to identify the energy needs of their clients (or regions

of work), track improvements in the energy ladder and the associated attributes

of energy access. Hence, providing tools for the microfinance industry will help

stakeholders to design and negotiate the inter-sectorial development of energy

programs, defining the role of the microfinance sector on the achievement of the

SDGs.

Kurzfassung

Zugang zu Energie wurde als einer der entscheidenden Faktoren f¨ur die wirtschaft-

liche soziale und menschliche Entwicklung erkannt. Weltweit haben ¨uber zwei

Milliarden Menschen keinen Zugang zu moderner Energie, weder zu elektrischer

Energie noch zu sauberen Kochm¨oglichkeiten. Durch die Ver¨offentlichung der

Sustainable Development Goals (SDGs), bekamen die Definition und das Er-

fassen von Daten ¨uber den Zugang zu Energie hohe Relevanz, u.a. f¨ur Regierun-

gen, Entwicklungsagenturen, Private- und Nichtregierungsorganisationen sowie

Finanzinstitute. Um das vorgegebene Ziel, bezahlbaren und verl¨asslichen uni-

versellen Zugang zu modernen Energiequellen bis 2030 umzusetzen, verlangt

das SDG 7 nach einem standardisierten Monitoring des erreichten Fortschritts.

Diese Untersuchung zielt auf zwei Ergebnisse ab. Das erste Ziel ist es eine detail-

lierte Analyse des Energiezugangs von Mikrofinanzkunden im S¨udwesten von

Kolumbien zu pr¨asentieren. Dabei liegt der Fokus auf dem Teil der Bev¨olkerung

der Mikrodarlehen besitzt oder schon aufgenommen hat. In Zusammenar-

beit mit dem kolumbianischen Mikrofinanzinstitut (MFI) Contactar, wird die

Studie mit Hilfe des (multi-tiered) Multi-Tier Energy Access Framework (MTF),

welches k¨urzlich von der World Bank, Energy Sector Management Assistance

Program (ESMAP) ver¨offentlicht wurde [Bhatia and Angelou, 2015], durchge-

f¨uhrt. Dieser Ansatz ermittelt mithilfe technischer, ¨okonomischer und sozialer

Attribute die Qualit¨at und Quantit¨at des Energiezugangs. Die Ergebnisse

zeigen die schlechte Qualit¨at der Energieversorgung, verglichen mit den ¨uberdurch-

schnittlich hohen nationalen Zahlen zum landesweiten Energiezugang und ein

Energie stacking Verhalten bei den unterschiedlichen Arten von Brennmaterial

zum Kochen, nach Bezahlbarkeit und Verf¨ugbarkeit durch die Haushalte ermit-

telt. Durch das Aufzeigen der tats¨achlichen Qualit¨at des Energiezugangs, deckt

die Analyse Handlungsbedarf bei der Verl¨asslichkeit, Qualit¨at und Bezahlbarkeit

auf. Außerdem werden Informationen ¨uber die regionale Energieversorgung

bereitgestellt, um die erkannten Engp¨asse besser versorgen zu k¨onnen und es

wird eine detaillierte ¨ubersicht ¨uber den Energiebedarf zum Kochen erstellt.

Um den Nutzen des verwendeten Ansatzes auch f¨ur MFIs zu validieren, werden

ebenfalls die M¨oglichkeiten von etablierten Methoden, wie dem Progress out

of Poverty Index (PPI) [Schreiner, 2004], untersucht, den das MFI aktuell ver-

wendet um die Qualit¨at des Energiezugangs vorher zu sagen (PPI Scorecard f¨ur

Kolumbien). Die Ergebnisse der Untersuchung zeigen, dass der Kolumbianische

PPI weder den aktuellen Grad des Energiezugangs noch die Verf¨ugbarkeit und

Bezahlbarkeit der verwendeten Kochmittel korrekt darstellt. Jedoch weist eine

deutliche Verbindung zwischen dem Konsum und der Wahrscheinlichkeit unter

das Armutslevel zu fallen, eine umgekehrt proportionale Korrelation auf.

Damit die Strategien zur Umsetzung der SDGs in allen Regionen vereinheitlicht

werden und der Multi-Tier Ansatz weiter verbreitet wird, befasst sich der zweite

Teil der Studie mit der Entwicklung eines Toolkit zur Umsetzung von Projekten,

die dazu dienen den Fortschritt von ausgew¨ahlten Attributen bei Haushalten zu

erfassen. Auf Grundlage des ESMAP Tool [Bhatia and Angelou, 2015] und den

Umfrage Ergebnissen, wird MFIs empfohlen als ersten Schritt den Progress out

of Energy Poverty Index (PEPI) (Index zur Bewertung der Energie-Armuts-

Entwicklung) zu verwenden, um die Nachfrage ihrer Kunden nach Elektrizit¨at

und Kochmittel zu identifizieren und zu bedienen sowie deren Fortschritt zu

erfassen. Wird die nahe, langfristige Beziehung, die Infrastruktur der MFI, das

st¨andige aufnehmende Interesse die Nachfrage der Kunden zu bedienen und der

wachsende Markt ber¨ucksichtigt, kann das Tool angepasst werden, so dass es

den MFIs erm¨oglicht in den Markt f¨ur green lending einzusteigen. Außerdem

k¨onnen weitere Organisationen die Energiezugang f¨ur den Base of the Pyramid

(BoP) bereitstellen das Tool nutzen um die Auswirkungen ihrer Produkte, Ser-

vices oder Programme zu erfassen. Der PEPI unterst¨utzt Organisationen bei

(i) der Ermittlung des Energiebedarfs ihrer aktuellen und potentiellen Kunden,

um effektive Programme f¨ur green lending oder Energiezugang zu designen;

(ii) der Messung und Erfassung der Ver¨anderungen durch die Programme beim

Energiezugang von Haushalten; und (iii) der Untersuchung der Auswirkungen

moderner Technologien bei der Verbesserung der Qualit¨at des Energiezugangs

und der H¨ohe der Aufwendungen ¨uber die Zeit des Haushalts f¨ur Elektrizit¨at

und moderne Kochm¨oglichkeiten.

Das Toolkit basiert auf dem angepassten Multi-Tier Framework, um Elektrizit¨at

und Kochmittel bereitzustellen und es bringt (i) eine Auswahl an erweiterten

frameworks mit gr¨oßerer Auswahl an unterschiedlich ausgepr¨agten Merkmalen;

(ii) ein einsatzbereiter Fragebogen f¨ur Energiequellen und Kochkonzepte; und

(iii) ein Index um den Fortschritt beim Energiezugang. Die Modifizierten Pa-

rameter zur Definition des Energiezugang bringen die globalen SDG 7 Ziele

mit der Neudefinition des Sustainable Energy for All (SE4All) Global Tracking

Framework zusammen. Die Merkmale des Energiezugangs werden in globale

Attribute gruppiert und der Fortschritt durch die Variation der Deltas ¨uber

der Zeit untersucht. Damit liefert der Verbund-Index tiefergehende Informa-

tionen ¨uber die Quantit¨at und Qualit¨at des Energiezugangs, welche durch die

globalen Merkmale f¨ur Energiezugang erhalten werden und vermeidet dabei die

Fehlinterpretationen die ein einzelner Index verursachen k¨onnte. Die entwickel-

ten Umfragen sind softwarebasierte Tools zur Datensammlung und Analyse, in

welchen der Zugang ausschließlich anhand von Diensten untersucht wird, welche

in den Haushalten verf¨ugbar sind. Das Umfrage-Tool enth¨alt Energieparameter

zu Elektrizit¨at und Kochmitteln, die auf die untersuchten L¨ander angepasst wer-

den k¨onnen. Kein Verbrauch von Elektrizit¨at wird durch das Tier-Framework

untersucht. Durch den wiederholten Einsatz des Toolkits, ergibt sich der PEPI,

um den Fortschritt aus der Energiearmut ¨uber die verschiedenen Dimensionen

des Energiezugangs, auf Basis des durchschnittlichen Fortschritts der glob-

alen Attribute, zu messen. Durch diesen Ansatz und das hervorheben der

Notwendigkeit eines universellen Energiezugangs, erlaubt das PEPI Toolkit Or-

ganisationen, welche an einem verbesserten Zugang zu Energie arbeiten, den

Energiebedarf ihrer Kunden (oder Regionen) zu ermitteln, Verbesserung auf

der Energieleiter oder ¨ahnlichen Merkmalen zum Energiezugang zu erfassen.

Die Bereitstellung von Tools f¨ur die Mikrofinanzindustrie erlaubt den Akteuren

die inter-disziplin¨are Entwicklung von Energieprogrammen, mit der Rolle des

Mikrofinanzsektors definiert anhand der Errungenschaften der SDGs.

Contents

1. Introduction 1

1.1. Objectives............................... 6

1.1.1. Measuring the Progress out of Energy Poverty . . . . . . 7

1.1.2. The PEPI Toolkit . . . . . . . . . . . . . . . . . . . . . . 9

1.2. ResearchDesign ........................... 10

1.2.1. Hypothesis and Research Questions . . . . . . . . . . . . 10

1.2.2. Methodology ......................... 11

1.2.3. The First Steps in Developing the Toolkit . . . . . . . . . 12

1.3. Overview of the Thesis . . . . . . . . . . . . . . . . . . . . . . . . 13

2. Energy Poverty in Rural Areas 15

2.1. Universal Energy Access for Sustainable Development . . . . . . 16

2.2. Energy Transition Paths . . . . . . . . . . . . . . . . . . . . . . . 17

2.3. Microenergy Systems . . . . . . . . . . . . . . . . . . . . . . . . . 19

2.4. Productive Uses of Energy . . . . . . . . . . . . . . . . . . . . . . 21

2.5. Measuring Energy Poverty . . . . . . . . . . . . . . . . . . . . . . 23

2.6. The Multi-tier Framework Approach . . . . . . . . . . . . . . . . 27

2.6.1. Energy Access at the Household Level . . . . . . . . . . . 28

2.6.2. Energy Access Framework for Productive Engagements . 32

2.6.3. The Energy Access Index . . . . . . . . . . . . . . . . . . 35

3. Tackling Affordability and Access to MES: Green Microfinance 37

3.1. Green Microfinance and Green Lending . . . . . . . . . . . . . . 39

3.2. Clean Energy Technologies Finance Models . . . . . . . . . . . . 42

3.3. Cross-sectoral Cooperation to Tackle Energy Poverty . . . . . . . 43

3.4. Indicators on Green Microfinance . . . . . . . . . . . . . . . . . . 45

3.5. The Progress out of Poverty Index (PPI) . . . . . . . . . . . . . . 51

4. The MFI Contactar 53

4.1. Geopolitical and Economic Contexts . . . . . . . . . . . . . . . . 54

4.2. Portrait of the Microfinance Institution . . . . . . . . . . . . . . 57

4.3. The Diversification of Contactar in Green Microfinance . . . . . 60

4.3.1. A Two-Hand Model Approach . . . . . . . . . . . . . . . 61

4.3.2. Selected Microenergy Systems . . . . . . . . . . . . . . . . 62

4.3.3. The Systematic Approach . . . . . . . . . . . . . . . . . . 65

The PEPI Toolkit for MFIs Contents

5. Energy Usage in Rural Areas: Case Study in Southern Colombia 79

5.1. SampleSelection ........................... 80

5.2. Analysis of Electricity Supply, Services and Consumption . . . . 85

5.2.1. PPI for Assessment of Electricity Services . . . . . . . . . 85

5.2.2. Multi-tier Analysis . . . . . . . . . . . . . . . . . . . . . . 86

5.3. Analysis of Access to Cooking Solutions . . . . . . . . . . . . . . 99

5.4. Multi-tier Ranking vs. PPIData ..................109

5.5. Summary of the Study . . . . . . . . . . . . . . . . . . . . . . . . 112

6. Development of the PEPI Toolkit 115

6.1. ToolkitDesign ............................116

6.1.1. Properties and Limitations of the Multi-tier Framework . 116

6.1.2. The Development Agenda . . . . . . . . . . . . . . . . . . 118

6.1.3. From the ESMAP to the PEPI Framework . . . . . . . . 119

6.1.4. Measuring the Progress . . . . . . . . . . . . . . . . . . . 125

6.1.5. Computation of the Progress Matrix . . . . . . . . . . . . 126

6.1.6. Summary: Properties of the PEPI framework . . . . . . . 128

6.2. PEPI Framework Assessments . . . . . . . . . . . . . . . . . . . . 128

6.2.1. Electricity Access at the Household Level . . . . . . . . . 128

6.2.2. Access to Cooking Facilities . . . . . . . . . . . . . . . . . 131

6.3. A Toolkit for Assessment of Energy Services . . . . . . . . . . . . 137

6.3.1. The PEPI Survey . . . . . . . . . . . . . . . . . . . . . . . 138

6.3.2. Howto ............................140

7. Conclusions 147

7.1. Results and Main Contributions . . . . . . . . . . . . . . . . . . 148

7.1.1. Assessment of Energy Access of Microfinance Clients . . . 148

7.1.2. Development of the PEPI Toolkit . . . . . . . . . . . . . . 150

7.2. Outlook: a Tool for the Microfinance Industry . . . . . . . . . . . 151

Appendices 155

A. Appendix: Contactar Exclusion List 157

B. Appendix: PPI Tool Lookup Tables - Colombia, English version 159

C. Appendix: Maps and Description of Selected Regions 163

List of Tables 166

List of Figures 170

References 175

The PEPI Toolkit for MFIs Acronyms

Abbreviations and Acronyms

AI Access Index

BD Biogas Digesters

BoP Base of the Pyramid

CESPI Correlation Energy Services Proxy Index

CFL Compact Fluorescent Lamp

CL Confidence Level

COP Colombian Peso

COSA Committee on Sustainable Assessment

CSPI Correlation Services Proxy Index

DANE Departamento Administrativo Nacional de Estadstica

DESCO Distributed Energy Service Company

EDI Energy Development Index

e-MFP European Microfinance Platform

EPP Energy Poverty Penalty

ESI Energy Supply Index

ESMAP Energy Sector Management Assistant Program

EU European Union

FI Financial Institution

FPL Fuel Poverty Line

GDSA Gesti´on de Desempe˜no Social y Ambiental

GF Grameen Foundation

GIF Green Inclusive Finance

GPA Green Performance Agenda

GTF Global Tracking Framework

HDI Human Development Index

ICT Information and Communication Technology

IEA International Energy Agency

IHS Integrated Household Survey

ICS Improved Cooking Stoves

IFC International Finance Corporation

ISIC International Standard Industrial Classification of all Economic

Activities

ISO International Organization for Standardization

IWA International Workshop Agreement

LED Light-emitting Diode

LPG Liquified Petroleum Gas

ME Margin of Error

iii

The PEPI Toolkit for MFIs Acronyms

MDG Millennium Development Goal

MECS Multisector Energy Investment Projects

MEPI Multidimensional Energy Poverty Index

MFEPI Microfinance Environmental Performance Index

MES Microenergy System

MEA Microenergy Appliance

MFI Microfinance Institution

MIX The MIX Market Platform

MSME Micro-, Small- and Medium-Enterprises

MTF Multi-tier Framework

NGO Non-Governmental Organization

PAYG Pay-As-You-Go

PCIA Partnership for Clean Indoor Air

PEPI Progress out of Energy Poverty Index

PIFIL Plan de Investigaci´on para el Fortalecimiento Integral de las

Comunidades

PPI Progress out of Poverty Index

PV Photovoltaic

R&D Research and Development

RE Renewable Energy

SCD Solar Crop Driers

SDG Sustainable Development Goal

SE4ALL Sustainable Energy for All

SHS Solar Home System

SME Small and Medium Enterprises

SPI Social Performance Indicators

SPI4 Social Performance Index Tool

SPM Social Performance Management

TEA Total Energy Standard

UN United Nations

UNDP United Nations Development Programme

UNIDO United Nations Industrial Development Organization

USA United States of America

USD United States Dollar

USSP Universal Standards of Social Performance

WB World Bank

WEO World Energy Outlook

Wh Watt hour

The PEPI Toolkit for MFIs Acronyms

WHO World Health Organization

Wp Watt peak

WF Water Filters

WT Water Tanks

1.Introduction

The establishment of the Global Goals of the United Nations (UN) [UNDP,

2016] represents a huge challenge on various scales for both governments and

public authorities. Setting up a development agenda always aims to synchro-

nize initiatives, funds and tools to operationalize programs, which in turn help

meet the targets. At the same time, when exploiting stakeholders’ ability to

push ambitious programs forward, it is imperative to develop functional tools

to track the programs’ progress in achieving the established targets, in order

to correctly guide policies, funding and project follow-ups. Thus, defining a

common language and disseminating best practices among stakeholders are key

issues in achieving the specific goals.

This thesis focuses on the challenges of defining indicators to measure progress

in household-level energy access – e.g., the improvement of electricity and cook-

ing supply – fostered by programs linking microfinance mechanisms with the

provision of modern energy services. To date, a variety of methods to mea-

sure microfinance institutions’ (MFIs) activities focusing on an environmental

agenda fail to assess the energy access quality of microfinance customers. The

primary goal of this dissertation is to propose a measurement and tracking tool

in order to assist the microfinance sector in taking concrete action to achieve

energy access for all.

The Sustainable Development Goals After the UN established the Millen-

nium Development Goals (MDGs) in 2000, which included eight anti-poverty

targets to be met by 2015, there was widespread disappointment that the goals

did not address the challenge of enabling energy access for all. As a result,

the scientific community set about emphasizing the key role of modern clean

energy1for the achievement of the MDGs (see e.g., [AGECC, 2010,Legros et al.,

2009,Ilskog, 2008,MacLean and Siegel, 2007,Modi, 2004,DFID, 2002]). More

specifically, it has been argued that a sustainable, accessible and affordable en-

ergy supply – such as modern electricity for household and community usage,

productive use and cooking solutions – is crucial to enabling development and

1Clean energy refers to renewable energy sources (e.g., solar, biomass, wind, hydropower and

geothermal), liquefied petroleum gas (LPG) as fossil fuels that emit little greenhouse gas, and

traditional fossil and biomass fuels that use technologically advanced processing, practices

and/or products such as energy-efficient cookers [UNCDF, 2012]. Throughout this thesis, the

concepts of modern clean energy and modern energy services will be used interchangeably,

referring to the use of clean energy as defined above.

The PEPI Toolkit for MFIs 1. Introduction

well-being [AGECC, 2010,MacLean and Siegel, 2007,Modi et al., 2006]. Hence,

access to energy services is essential for sustainable development in developing

countries, since these are critical enablers and contributors to a virtuous cycle

of human, economic and social improvements [OECD, 2007].

In 2012, the launch of the Sustainable Energy for All (SE4ALL) program2

– a joint initiative of the Secretary General of the UN and the World Bank

(WB) striving to achieve universal access to modern energy by 2030 – formally

established the need to alleviate energy poverty or energy deprivation, defined as

a broader concept rather than just a lack of energy access. According to [Barnes

et al., 2011], the definition of energy poverty is the point at which people use

the bare minimum of energy, derived from any source, needed to sustain life.

Above this point, energy contributes to welfare and increases economic well-

being, while below this point people lack enough energy to sustain normal

lives.

Figure 1.1.: The Sustainable Development Goals (SDGs)

The SE4ALL program paved the way for channeling mandates to achieve these

objectives and for placing energy access on the list of global priorities. In

September 2015, the UN Member States adopted the Sustainable Development

Goals (SDGs)3(Figure 1.1), including energy access for all (SDG 7) and the

impact of climate changes (SDG 13) as parts of the sustainable development

agenda. However, replacing the MDGs with the SDGs, and thus adding energy

access to the list of Global Goals4, is just the beginning of a greater challenge:

lowering energy deprivation by fostering universal energy inclusion, which, as

opposed to mere energy access, aims to improve the energy service quality for

vulnerable and low-income groups at affordable cost [Groh, 2014].

2See http://www.se4all.org/

3See https://sustainabledevelopment.un.org/post2015.Throughout this thesis, the terms En-

ergy SDG and SDGs will be used interchangeably in referencing to one (SDG 7) or several

of the 17 Global Goals of the 2030 Agenda for Sustainable Development.

4See http://www.undp.org/content/undp/en/home/sdgoverview/post-2015-development-

agenda/goal-7.html

The PEPI Toolkit for MFIs 1. Introduction

Energy Supply for the Base of the Pyramid According to [IFC and WRI,

2007], four billion people constitutes the base of the economic pyramid – BoP,

i.e., people with low annual income population (below $3,000 in local purchasing

power) and living in relative poverty. The BoP represents a considerable share

of the world’s population, and includes over 70% of the population in Africa,

Asia, Eastern Europe, and Latin America and the Caribbean. Since most people

who do not have access to electricity and rely on biomass for cooking live

in the poor parts of emerging and developing countries, i.e., slums and rural

areas, next to food and housing, satisfying the energy needs often represents

the biggest expense [Gradl and Knobloch, 2011]. Indeed, approximately 35%

of rural areas worldwide lack of reliable access to either electricity or clean

cooking facilities [IEA, 2011]. These deficiencies hinder the development of

basic infrastructure, education and health, as well as the productivity and local

value creation. Women and school-aged children from rural areas are affected

in particular, as they have to collect fuelwood, a time-intensive activity that

keeps them away from more beneficial activities [Barnes and Toman, 2006,Saghir,

2005].

Modern energy technologies – referred to in this thesis as microenergy sys-

tems (MES), i.e., decentralized energy systems based on small and locally us-

able energy conversion units using clean energy that enable spatial intercon-

nection between energy demand and energy supply (microenergy appliances,

MEA) [Philipp and Sch¨afer, 2009] – enable increasing profitability and produc-

tivity of micro, small and medium enterprises (MSMEs), small industries, and

agriculture. Hence, affordable access to MES improves people’s quality of life

and contributes to poverty reduction. Following the SDGs’ agenda, one single

principle should guide macro and micro interventions targeting energy access:

enabling access does not mean solely focusing on the source of energy itself,

but also on guaranteeing affordable, reliable and safe energy services that are

essential to the users’ daily well-being. However, to date, there are very few

dissemination methods that reach the population segment that has limited ac-

cess to electricity and improved energy technologies, mainly because of the high

initial capital costs needed for energy poverty alleviation [IEA, 2011,Beck and

Martinot, 2004].

The role of Microfinance in Energy Access Microfinance is generally known

as the provision of financial services for low-income populations, i.e., people

living in poverty who are not considered bankable [Rao and Rao, 2010,Ar-

mend´ariz de Aghion and Morduch, 2005]. Particularly, microfinance mechanisms

are based on the idea that a lack of collaterals, stable employment and verifiable

credit history (making commercial banking inaccessible) can be overcome via

alternative lending techniques (e.g., group lending and loan guarantors), local

information flows and consecutive loans [Ahlin and Neville, 2008]. Hence, the

possibility of splitting up high investment costs into affordable monthly install-

ments makes sustainable financing via microfinance a viable approach, among

others, to overcome the affordability barriers for energy integration. Moreover,

since MFIs have tight local networks and close relationships with their cus-

The PEPI Toolkit for MFIs 1. Introduction

tomers [Kebir, 2009], they also represent a vehicle for the dissemination of MES

striving for sustainability.

The microfinance industry’s increasing engagement in environmental issues has

established the field of green microfinance, combining microfinance services and

products pinpointing environmental responsibility (e.g., [Allet, 2012,Realpe Car-

rillo, 2014,GreenMicrofinance, 2007]). Programs exploiting MFIs’ capabilities for

enhancing energy supply have gained increasing prominence in recent decades

- see Figure 1.2. In several contexts, MFIs have partnered with energy ser-

vice suppliers5to disseminate clean energy technologies (see, e.g., [Srinivasan,

2007,Morris et al., 2007]). MFIs diversify their portfolio by introducing a fi-

nancial product to finance MES and, through these partnership, agree to share

responsibilities with the energy service supplier delivering the technology to the

microfinance client.

Figure 1.2.: Measuring the commitment of the microfinance industry to sustain-

able development. Number of MFIs reporting to the MixMarket

platform (www.mixmarket.org) on their performance within a set

of established Green Performance Indicators. Source: [Pierantozzi

et al., 2015]

In particular, microcredits designed to help low-income populations meet energy

demands using locally based sources through MES aim at alleviating energy

poverty by enabling affordable and sustainable access. Moreover, partnering

MFIs with the energy sector may open up new financial and energy markets,

attract new clients for financial services, alert existing clients to new energy

services, and contribute to poverty alleviation [Morris et al., 2007]. The rationale

is that microfinance and consumer lending make it possible for companies to sell

high-quality (i.e., more expensive) products to customers with low purchasing

power by being paid for them in small installments. As such, it is a tool for

overcoming the price barrier in economically poor areas [Van Elteren, 2007,Hall

et al., 2008,Rippey, 2009].

5Energy service suppliers refers to established companies that offer MES, usually small and

medium enterprises (SMEs). In particular, these include regional, national or local dis-

tributors and/or service providers, and, depending on the technology, it can also include

manufacturers partnering directly with the MFI.

The PEPI Toolkit for MFIs 1. Introduction

Hence, establishing concrete goals and providing sectoral support, microfi-

nance’s potential could be exploited to achieve energy access for all. To this end,

the microfinance sector’s role must be systematically framed, and proper tools

must be designed to support its operationalization and measure its achievements

over time. However, the implementation of MES in rural regions with weak in-

frastructure requires careful design and planning that take economic, social and

environmental dimensions into account [Philipp and Sch¨afer, 2009]. Moreover,

the individual initiatives should be planned in view of the development goals,

and their outcome should be aligned using common metrics.

MFIs Decision-Making Process and Energy Access Monitoring The inter-

sectoral partnerships are at the base of green lending (or green loans) [Re-

alpe Carrillo et al., 2015,Allet and Hudon, 2013,Leva¨ı et al., 2011,Allderdice et al.,

2007], which enable MFIs to disseminate MES, and thereby offer an alternative

to improving energy access and providing viable distribution channels.

At the same time, in order to guarantee a relevant impact of microcredits tai-

lored to meet energy demands, MFIs need to accurately select potential MES to

include in their portfolio, following a “bottom-up” approach, i.e., starting with

an analysis of their clientele. By first identifying their clientele’s energy needs,

uses and the costs involved, MFIs can better configure financial services that

address product sustainability. However, MES choices are often the result of a

demand analysis and top-down approaches (e.g., from funders, donors, devel-

opment cooperations, etc.) [Leva¨ı et al., 2011]. Thus, the possibility of assessing

energy access achievements of selected MES and predicting their relation to

quality improvement is instrumental to the decision-making processes of the

partnership between an MFI and an energy service supplier.

The microfinance sector still has no systematic way of measuring the progress

of MES implemented in regions with unreliable energy infrastructure. This fact

is mainly due to the newness of green lending in the microfinance sector and to

the fact that global metrics for all topics concerning environment management

are primarily established at the organizational level. Furthermore, in the last

decades, there has been no consensus in the energy sector on a conceptual

definition of energy poverty or the appropriate techniques to measure it. This

led to the common use of the traditional methods for energy access assessment:

connected or not connected to the grid, cooking with biomass or not.

Academics and practitioners had made several attempts to come up with a

system for measuring energy progress. In 2013, the SE4All Global Tracking

Framework (GTF) first published a concrete set of indicators using a multi-tier

approach [Global Tracking Framework (GTF), 2013] that assesses the multidimen-

sionality of energy. However, the ability of this approach to efficiently tackle

the main attributes of energy access and assess changes over time undermine its

potential for delivering in-depth assessments of MES potential to tackle afford-

able, reliable and safety energy access. Consequently, statistics on long-term

trends are currently not available for MES that have been financed and imple-

mented in rural areas, particularly in goods-producing sectors like agriculture

or micro-business. Without standardized tools, organizations face challenges in

The PEPI Toolkit for MFIs 1. Introduction

monitoring the impact of MES in households and micro-businesses, a crucial

aspect for policy designers and private implementers.

1.1. Objectives

This research is driven by the desire to support MFIs and stakeholders in the

energy sector throughout their decision-making process and in the interpreta-

tion of their results, and to achieve a better understanding of MES’ potential

to increase energy access in regions with poor infrastructure. Specifically, this

paper is based on the rationale that financing decisions should be based on a

comprehensive analysis, taking into account the MFI clients, the effects of MES

on clients’ energy access level, and the trackability of the level of energy access

achieved by diversifying an MFI’s portfolio. Until now, there has been no tool

supporting organizations in these kind of assessments.

First, this dissertation aims to identify scientific methodologies for defining

the potential of energy access quality obtained through MES at the household

level. As a next step, the findings will be translated into a decision-making

and monitoring tool for assessing the impact of MES. This tool aims eventu-

ally at supporting the portfolio diversification of MFIs and assisting any other

organization involved in improving electricity supply or cooking facilities for

low-income populations.

The methodology has been developed based on the results of field research that

was conducted in southwest Colombia in collaboration with the Colombian MFI

Contactar, with the objective of characterizing the energy access level of the

microfinance clients. The analysis of this case study serves, on the one hand, to

fine-tune the assessment methodology and to further clarify the role of energy

in households and, on the other, to provide a basis for the developed tool to

assess energy access at the micro level.

In detail, the goals of the thesis can be summarized as follows:

I To analyze the household-level energy usage of a selected sample of

microfinance clients using the multi-tier approach

II To assess the attributes describing energy access for electricity supply,

services and consumption and cooking facilities at the household level

III To assess the correlation between poverty metrics and energy poverty

data

IV To develop a systematic and flexible tool – adaptable in different geo-

graphical regions – that financial institutions and energy stakeholders

can use to assess energy access and to identify their clientele’s energy

needs, usages, costs, and expenses

V To develop a methodology to measure progress out of energy poverty

based on the multi-tier approach

VI To validate the developed tool with the collected data

The PEPI Toolkit for MFIs 1. Introduction

1.1.1. Measuring the Progress out of Energy Poverty

Despite the existence of different indices developed to evaluate the performance

of green microfinance (see [Pierantozzi et al., 2015]), the impact of these pro-

grams on the energy access of the microfinance clientele, e.g., in terms of the

alleviation of energy poverty, has not been measured yet. There is also no index

specific to the microfinance industry that measures and tracks the deprivation

of access to modern energy services. Motivated by these needs, this thesis is

based on the application of a multidimensional energy access framework to a

set of microfinance clients in order to develop a Progress out of Energy Poverty

Index (PEPI), which is meant to be a tool for microfinance sector stakeholders

and organizations implementing energy access projects. This index aims at es-

tablishing an effective way of measuring the impact of green lending programs in

order to support MFIs and energy service suppliers in quantifying their ability

to satisfy the different dimensions of energy access of their clientele.

Figure 1.3.: Multi-tier Framework for Electricity Supply

Specifically, this study follows the multi-tier approach fully described and pub-

lished by the global and multi-donor technical assistance trust fund, the En-

ergy Sector Management Assistance Program (ESMAP) of the World Bank

(WB) [Bhatia and Angelou, 2015]. In this approach, diverse aspects of en-

ergy access (such as energy supply, electricity services, energy consumption

and cooking) are analyzed in different frameworks, defining for each of them a

set of attributes (e.g., duration, reliability, quality, etc.) in order to represent

the multidimensionality of energy access. In practice, following empirically pre-

defined criteria for each attribute, each household is then ranked in a specific

tier (from 0 to 5) [Bhatia and Angelou, 2015] to detail the deficiencies in energy

supply performance and to better frame possible interventions (Figure 1.3).

For each framework, the results are condensed into a set of composite indices

that take into account the proportion of the sample households in each tier

and thus describe the energy access of the population. Figure 1.4 depicts an

example of index calculation.

The PEPI Toolkit for MFIs 1. Introduction

Figure 1.4.: An example of an access index (AI) computation according to [Bha-

tia and Angelou, 2015], based on the proportion of households

ranked in tiers 0 to 5

Following the idea of underlying the multi-tier framework (MTF) from the

ESMAP, the PEPI aims to simplify and expand this multidimensional approach

by focusing primarily on household-level electricity access (e.g., supply and

services) and cooking solutions.

For the purposes of this study, the ESMAP MTF for analyzing energy access

has been implemented considering a set of microfinance clients, in collaboration

with the Colombian MFI Corporaci´on Empresa Nari˜no Contactar. Contactar

is a medium-sized MFI (more than 70,000 clients in 2014) operating mainly in

southern Colombia (see Figure 1.5, left). It has a vast track record in the triple

bottom line approach and remarkably prioritizes its environmental impact in

addition to its financial and social objectives.

Lowest Tier

5 4 3 2 1 0

Frequency (%)

Energy Supply

Energy Services

Energy Consumption

Cooking

Figure 1.5.: Left: Regions of Colombia where the case study was carried

out (original map downloaded from: www.your-vector-maps.com)

Right: Summary of the results of the case study in southern Colom-

bia, in terms of tier ranking of households in the four considered

frameworks (energy supply, energy services, energy consumption,

cooking solutions)

By collecting data from the different regions covered by Contactar, the energy

access of its clients has been characterized in detail (in terms of energy supply,

The PEPI Toolkit for MFIs 1. Introduction

electricity services, electricity consumption and cooking facilities, highlighting

the differences between urban and rural areas (see Figure 1.5, right). The case

study has also identified the ways in which energy access lacks most, according

to the multi-tier analysis proposed in [Bhatia and Angelou, 2015] and further

studied and discussed in [Groh et al., 2016].

As a next step, based on the outcome of the survey and the ESMAP MTF,

a multidimensional measure of energy poverty was defined in terms of three

global attributes for electricity supply – reliability,affordability and safety –

and three attributes for cooking facilities – affordability,availability and safety

– aggregating the associated attributes included in the ESMAP MTF approach.

The established framework builds upon the ESMAP conventions [Bhatia and

Angelou, 2015], including a set of modifications echoing the suggestions from

[Groh et al., 2016] and [Bensch, 2014] to fine-tune the deeper ranges of energy

poverty.

Moreover, a hybrid index that takes all attributes into account is proposed in

order to measure the progress out of energy poverty achieved by a household

during a certain period of time. The index has the properties of aligning the

attributes specified in the SDG 7, with the same multidimensional features of

the ESMAP multi-tier metric, and allowing aggregated and disaggregated data

analysis.

1.1.2. The PEPI Toolkit

The PEPI toolkit developed in this thesis comprises specific questionnaires for

measuring access to energy supply and cooking facilities. The toolkit uses the

multidimensional energy access tool and its corresponding tier-ranking matrices

as well as an automated data-processing and analysis tool. Based on easy-to-

use survey sheets, the toolkit (see Figure 1.6) requires only basic training for

its application by an institution in the field.

The toolkit allows energy usage (electricity supply and services and cooking

solutions) to be assessed with a focus on the quality of the services and the costs

of access. By adapting the ESMAP MTF, each household is ranked in three

pillars of global attributes associated with energy access. The energy access

indicators are then normalized and combined, creating a single hybrid index.

Hence, tracking the energy access performance of households in a geographical

area as part of a specific program, the PEPI enables stakeholders to measure

the progress out of energy poverty, assess performance of the intervention and

monitor target achievements.

The ultimate goal of the PEPI is to help MFIs and energy service suppliers

reach the energy poor by providing tools to facilitate products and programs

monitoring in infrastructure-poor areas and assess whether clients’ energy needs

are met or not. Through an extensive and systematic implementation of the

PEPI, MFIs can track multidimensional energy access and progress at the single

household level. Thus, MFIs will be able to make their green lending products

better suited to real customer needs, monitor the performance of their financial

services and products, and track clients’ progress out of energy poverty over

The PEPI Toolkit for MFIs 1. Introduction

Figure 1.6.: PEPI toolkit components

time. At the same time, donors will have a quantitative assessment tool to bet-

ter allocate their funds, while trends of the indices may provide useful insights

into required interventions and specific axes of action.

1.2. Research Design

1.2.1. Hypothesis and Research Questions

Hypothesis 1. Energy access at the household level can be measured by tools

based on multidimensional and multi-tier assessments. Access to electricity

and improved energy services are evaluated based on the capacity of available

appliances.

Research Questions

1.(i) How can the level of energy access in remote rural areas be concretely

and constructively assessed?

1.(ii) How are the different dimensions of energy access related? Does the rel-

evance of the energy access attributes depend on the local context?

Hypothesis 2. The potential of green microfinance and MES to improve the

quality of energy access can be assessed through the multi-tier approach aligned

with the Energy SDG targets.

Research Questions

2.(i) To what extent do MES contribute to the improvement of energy access

quality in energy deprived areas ?

2.(ii) How can the performance of MES be evaluated and included in a holis-

tic energy access assessment of electricity supply and cooking facilities?

The PEPI Toolkit for MFIs 1. Introduction

Which attributes’ characteristics are relevant to assess energy access?

Which global attributes are needed to provide a detailed picture of energy

supply performance in a target area?

2.(iii) How can the energy services enabled by MES be measured and compared

at the household level?

2.(iv) How can the achievements of green lending be measured with regard to

the Energy SDG?

Hypothesis 3. The development of a new assessment methodology will allow

financial institutions and energy service suppliers to systematically evaluate the

progress of energy access provided through MES in a specific geographical area.

Research Questions

3.(i) Which parameters and attributes should be taken into account when as-

sessing the performance of MES?

3.(ii) Which frameworks have to be considered in the evaluation of electricity

access and cooking facilities?

3.(iii) Considering a long-term assessment approach, how should an index be

designed in order to effectively quantify the progress (or the variation, in

general) of the tier-ranking of energy access within the same household?

1.2.2. Methodology

In order to track energy access improvements via green lending programs, the

research methodology is based on a data analysis of a microfinance population

sample.

To this aim, part of the research consists of a case study involving an empirical

investigation with evidence from different data sources. The research database

contains the results of the implementation of the ESMAP multi-tier tool [Bhatia

and Angelou, 2015], juxtaposed with the poverty metric system – the Progress

out of Poverty index (PPI) [Schreiner, 2004] – in order to infer the ability of

poverty assessment metrics to predict energy access performance.

Based on the energy access results of the sample from the MFI Contactar,

conclusions on the measurement algorithms have been drawn. Following these

considerations, a proposal of an improved assessment tool based on an adapted

multi-tier framework has been introduced and described. The tool aims to en-

able financial institutions and other implementers to identify energy needs, and

carry out continuous tracking and monitoring. Through longitudinal research

design, the progress analysis is designed to use panel data, which provides in-

depth information about the processes being studied over time.

By achieving new insights into the impact of microfinance on energy access in

a particular context, this dissertation has also an exploratory nature aiming to

assess current energy access measurement systems. The analysis, as well as the

development of a toolkit for the microfinance and energy sector, is based on the

results of the case study and on a review of recently proposed methodologies.

This dissertation pursues the overall goal of using data collection and analy-

The PEPI Toolkit for MFIs 1. Introduction

sis for decision-making, tracking and monitoring purposes, and is thus based

on a theoretical appraisal and critical assessment, formulating a path for use-

ful and practically adequate measurement methods. The designed metric and

toolkit rely on their potential to best assess the global attributes, and their

corresponding categories associated with energy access.

This doctoral thesis consisted of four main phases. First, a literature review

has been conducted in order to portray the state of the art in the research

linking energy deprivation, energy demand, energy affordability, microfinance,

green lending, and the corresponding measurement and evaluation methodolo-

gies. As a next step, a characterization of the clients of the MFI Contactar

was carried out using the data collected from via the ESMAP MTF survey

application and the institutional available data (disbursed loans, client appli-

cation forms, questionnaires, client feedback). The outcome helped define the

relevant indicators for assessing energy access at the household level. Building

on the results of the second stage, the third stage aimed to develop a system-

atic decision-making tool to help MFIs identify their customers’ energy needs

and expenses, characterize the energy access quality with regard to electricity

supply and cooking facilities, and analyze energy gaps that could be bridged

by market mechanisms. The fourth phase consisted in testing the developed

tool in the existing dataset and comparing those results with the ones obtained

using the ESMAP MTF tool.

1.2.3. The First Steps in Developing the Toolkit

1. Screening of Household-level Energy Access in Selected Regions This

step aimed first to obtain a portrait of the energy access of the microfinance

clientele with regard to the electricity supply, services and consumption pat-

terns. Moreover, an analysis of the cooking facilities is performed. Parallel to

the energy access analysis, the analysis of the PPI of the sample is conducted.

Subsequently, attributes related to the energy supply and services are depicted,

differentiating regional location, and rural and urban results. Simultaneously,

pitfalls of attributes results are identified through the data analysis. The fea-

tured energy supply attributes are clustered for electricity supply, electricity

services, electricity consumption and cooking facilities and tier-index is calcu-

lated for each specific framework.

2. Poverty Data vs. Energy Data Assessment This stage assessed the ca-

pacity of poverty metrics (PPI) to describe the quality of household-level energy

access. To this end, an analysis of the PPI indicators and scoring card used at

the institutional level is conducted. Based on the poverty indicators, electricity

supply, electricity services, electricity consumption and cooking facilities are an-

alyzed. In another stage, an energy access questionnaire with the specific metric

frameworks is developed. The assessment considers three different frameworks

(electricity supply, consumption and cooking facilities). Data collected from

existing clients are analyzed against panel data of poverty indicators, follow-

ing a profile characterization. The energy access is characterized analyzing the

The PEPI Toolkit for MFIs 1. Introduction

compiled data and its correlation with the poverty likelihood (PPI).

3. Definition and Development of the Measurement Tool Indicators, at-

tributes and parameters to be considered within each input data category are

defined. The adaptations of the multi-tier approach are discussed and described

in detail. Specific adapted frameworks for electricity supply and services and

for cooking facilities are developed. Results obtained in the energy access as-

sessment conducted in the first phase will be analyzed and compared to the

previous index derived from the multi-tier approach [Bhatia and Angelou, 2015].

Completing the index toolkit, a scoring mechanism is introduced for measuring

the progress out of energy poverty. Specifically, tier-ranking changes over time

are differentiated and evaluated, referring to the kind of change the household

experienced.

The PEPI survey focuses exclusively on energy access data. Socio-demographical

data are not part of the survey, as it is assumed that the energy access data

can be combined with institutional data available at each of the MFIs. For the

longitudinal analysis of a triple bottom line assessment, panel data of the socio-

demographical data previously collected from the MFI, together with multiple

(min. twice) data collected from the developed toolkit in a specific time period,

are required. At the level of the microfinance clientele, the survey – replica-

ble in any MFI or energy service supplier (i.e., implementer) – will serve to

estimate current energy usages, related costs and energy access tier-ranking.

Particularly, the characterization entails information on the energy attributes

associated with the energy supply.

1.3. Overview of the Thesis

The rest of the thesis is organized as follows: Chapter 2 focuses on energy access

in rural areas, describing the potential of MES, the available measurements of

energy poverty and the ESMAP MTF. In order to complete the research back-

ground, Chapter 3 is dedicated to green microfinance, to the financial models

related to it, and to the available indicators to measure the impact of green

microfinance programs. Chapter 4 focuses on the context of the research, de-

scribing the MFI Contactar, partner of the field research study, describing the

development of its green microfinance initiative and the related geopolitical

and economical contexts. The case study is detailed in Chapter 5, describing

the methodology for the selection of the clients sample and the results of the

ESMAP MTF tool, in terms of electricity supply, consumption and services and

access to cooking solutions. Lastly, the obtained multi-tier ranking is analyzed

with respect to the clients’ PPI data available from Contactar. Chapter 6 is

dedicated to the development of the PEPI toolkit. The chapter describes the

adaptation of the ESMAP framework considering the results of the case study

and enhancing the focuses of the considered attributes in view of the SDGs. Fi-

nally, the conclusions of the research are drawn in Chapter 7, discussing further

work to foster the introduction of the toolkit at the inter-sectoral level.

2.Energy Poverty in Rural Areas

Energy poverty has been defined as the lack of access to modern energy services.

This primarily affects low-income people; constraining their energy consump-

tion, leading to the use of polluting fuels, or resulting in a level of energy con-

sumption that is insufficient to support social and economic development [Bhatia

and Angelou, 2015]. Following the definition from [Pachauri et al., 2012b], mod-

ern energy access includes access to three forms of energy: (i) less polluting

household energy fuels for cooking and heating, which can range from cooking

with improved cooking stoves with traditional fuels, or cooking with non-solid

fuels such as liquid, gaseous, electric or solar-thermal; (ii) electricity for power-

ing appliances and for lighting; and (iii) mechanical power from electricity that

improves productivity.

The concept of access has also been widely discussed, defining it either from

the perspective of the target beneficiaries or the mode of energy supply. Di-

verse attributes of access discussed in literature include affordability, reliability,

quality and adequacy. In particular, access to clean energy has been claimed

to strongly depend on availability and affordability, considering not only the

possibility to acquire energy services, but also the capability of households to

choose between efficient and modern energy services [Brew-Hammond, 2010].

Particularly, the analysis in [AGECC, 2010] suggested to characterize energy

access depending on a given geographic area, i.e., including the availability of

resources, the institutional and technical capacity of the involved stakeholders,

the regulatory and policy environment and the relative cost of technologies.

Hence, by acknowledging the role of these factors, energy access can be fostered

by an optimal combination of different interventions.

Affordable and reliable provision of modern energy services remains a challenge,

impeding economic development, especially in rural areas. Indeed, the lack

of access of modern energy services prevents the development of basic infras-

tructure and improvement of basic living standards [Rao et al., 2009,MacLean

and Siegel, 2007], resulting in a so-called vicious cycle of poverty. In this

context, [Groh, 2014] empirically developed the concept of the energy poverty

penalty (EPP) as a trap that delays rural development at household level. The

penalty describes how energy expenses vary depending on the level of access

to energy services, with significant differences for those experiencing a certain

level of deprivation. The proven causality between energy and development has

driven public policy to establish ambitious goals and call for action to govern-

ments, donors and practitioners in the field.

The PEPI Toolkit for MFIs 2. Energy Poverty

In 2013, 1.2 billion people lacked access to electricity and more than 2.7 billion

people were estimated to rely on biomass for cooking, typically using inefficient

stoves in poorly ventilated spaces [IEA, 2013]. By 2015, the number of people

without the access to electricity had declined to 1.1 billion. However, far less

progress was achieved on access to clean cooking, as 2.9 billion people still

declared to use biomass fuels for cooking and heating [IEA and WB, 2015].

The rural areas were the most affected, with 260 millions of rural households

without access to electricity (87% of the 300 millions of households worldwide),

and, according to [The World Bank, 2008a,AGECC, 2010], it is estimated that

another billion of people only have access to unstable and intermittent electricity

networks. In this context, proper definitions of energy poverty metrics and

assessment methodologies are necessary, in order to support the implementation

of efficient and timely development strategies.

The first part describes different aspects of the relationship between energy

access and development, focusing on the relevance of access to clean energy

technologies, outlining the barriers for technology dissemination (in Section 2.1)

and describing the so-called energy transition path (in Section 2.2). The second

part is dedicated to the role of microenergy appliances (MEAs) and microenergy

systems (MES) in tackling energy poverty (Section 2.3), discussing the relevance

of electricity access in rural areas and in productive activities (Section 2.4).

Finally, the third part reviews selected approaches currently used to measure

the lack of energy access. In particular, a set of energy poverty indicators is

described in Section 2.5, while Section 2.6 focuses on the innovative multi-tier

framework approach (MTF), first introduced in 2013 [Global Tracking Framework

(GTF), 2013] and further discussed in [Bhatia and Angelou, 2015].

2.1. Universal Energy Access for Sustainable

Development

Access to modern energy services, intended as the physical availability of mod-

ern energy carriers and improved end-use devices at affordable prices for all

[Pachauri et al., 2012b], is considered to be crucial for economic growth [Rao

et al., 2009], as well as a starting point for sustainable development [UNDP and

WHO, 2009]. In fact, with the increasing importance of renewable energy and

energy efficiency combined with the failure of national governments to make

significant progress on universal access, the need for alternative solutions has

gained a prominent role in the Post-2015 development agenda.

At the macro level, the correlation between economic development and energy

access has been broadly empirically researched. In particular, [OECD and IEA,

2010] demonstrated the correlation between the Energy Development Index

(EDI) and the Human Development Index (HDI), arguing that the positive

effects of access to modern energy services in development paths include the

impacts on the HDI, on the level of education and on the growth of GDP. How-

ever, the precise cause-effect relationship between energy access and human

welfare remains partially unresolved [Pachauri and Spreng, 2004,Zerriffi, 2007].

The PEPI Toolkit for MFIs 2. Energy Poverty

In fact, while the literature agrees that energy access has a positive impact

on the economic growth, this is not always the case for least-developed coun-

tries [Stern and Cleveland, 2004], due to the low rate at which energy access

increases. Among others, [Goldemberg, 2004] observed a strong positive corre-

lation between energy access and economic growth in early development stages,

but, at the same time, a reversal of this trend after a certain level has been

reached.

Nevertheless, it has been shown that economic progress in the developing world

has yielded an enhancement of energy access for many communities in the

last decades. This has been the case, for instance, of East Asia and Latin

America, where the electricity networks has been extended [Kaygusuz, 2011].

Regions such as South Asia and Sub-Saharan Africa, where about the half of

the population without access to electricity lives, continue to lag far behind the

rest of the respective continents, especially concerning rural electrification [The

World Bank, 2008a].

More specifically, while the urban electrification rate reaches the 90% in devel-

oping countries, the rate of rural electrification reaches only 48.4% in South Asia

and 11.9% in Sub-Saharan Africa, against the yet low average rural electrifica-

tion rate of 58.4% considering all developing countries. In addition, more than

80% of the rural population in developing countries worldwide rely on biomass

for cooking [Barnett, 1990], and are hence exposed to the risk of respiratory and

lung diseases1. Despite this, more efficient and safer fuel solutions for cooking

(such as LPG – liquid petroleum gas) are often unavailable or unaffordable in

rural areas. Hence, the major challenges to be tackled comprise the inefficient

use and production of traditional energy sources, involving relevant economic,

environmental and health hazards, as well as the uneven distribution of electric-

ity access, petroleum, and natural or liquified gas among populations [Barnes

and Floor, 1996].

In fact, the asymmetries in living and in equity conditions, derived from the

lack of energy access (especially electricity), affect income generation and com-

munities development, tending to accentuate already existing social differences.

Hence, the increased poverty, the lack of opportunity for development, and the

uneven distribution of access to energy sources eventually yield considerable

migratory flows to large cities and an increasing disbelief regarding its own

future [Kaygusuz, 2011].

2.2. Energy Transition Paths

The concepts of energy ladder and energy transition refer to the process that

a household undertake in the transition from traditional to modern energy,

depending on the choice of energy sources and energy technologies. The ratio-

nale behind the energy transition is that the potential to acquire better quality

1According to the World Health Organisation (WHO), the indoor smoke inhalation due to

burning biomass causes over 1.6 million premature yearly deaths, whereas half of them are

of children younger than 5 years old [Legros et al., 2009].

The PEPI Toolkit for MFIs 2. Energy Poverty

and more sustainable sources depends monotonically on the household income.

According to this criteria, the level of development of the household can be

associated with a higher energy consumption and with changes in the energy

mix toward higher percentages of modern energy and better service quality. As

previously discussed, rural households in least developing countries, who are

particularly affected by energy poverty, lie on the lowest rung of the energy

ladder, and the concept of energy transition claims that through a more effi-

cient use of resources, modern energy will allow these populations to enter a

sustainable technological path of development.

Figure 2.1.: Energy ladder according to fuel [Kowsari and Zerriffi, 2011]

As an example, Figure 2.1 depicts a metaphorical ladder describing how a house-

hold ascends from using traditional biomass and primitive technologies towards

modern energy sources and more efficient cooking equipments [Kowsari and Zer-

riffi, 2011]. In this case, the attributes appraised through the energy choice

increase entail the efficiency, the cleanliness and the ease of use. In order to

move upwards along the ladder, the ability of acquiring improved technologies

and better fuels is highly related to (increase in) income, as well as to the

availability and to the accessibility of sources and technologies, as observed

by [Barnes and Floor, 1996,Masera et al., 2000,Elias and Victor, 2005] and also

confirmed by econometric evidence [Hosier, 2004].

However, the energy ladder model has been questioned in regards to its over-

simplification on the use and selection of fuel and to its limited view of re-

ality in which these processes take place [Kowsari and Zerriffi, 2011]. Specifi-

cally, [Kowsari and Zerriffi, 2011,van der Kroon et al., 2013] stress the inefficiency

of the model as it relies on a universal hierarchy of fuels as well as household’s

income as the major determinant of fuel choice, failing at identifying further

factors of energy access choice.

The PEPI Toolkit for MFIs 2. Energy Poverty

For instance, parallel technology usage is also observed in rural contexts. Due to

availability, affordability and practicability, households opt for broader options

of energy sources and technologies. This situation is referred as fuel/energy

stacking, i.e., when multiple energy sources (fuels/devices) are simultaneously

used to satisfy household’s energy needs. The main reasons for the fuel stacking

lie in the need of backup fuel options to cope with intermittent energy supply,

to shield from unstable fuel markets and to keep up cultural practices and

preferences, while benefiting from available modern energy sources and tech-

nologies [Pachauri and Spreng, 2004,Elias and Victor, 2005,van der Kroon et al.,

2013] Particularly, the findings from [Masera et al., 2000] describe how stacking

instead of switching to clean fuels is rather the norm in most households. More-

over, although [Masera et al., 2000] observed that an increase in income might

yield a partial or full shift of the energy stacking towards cleaner and more

efficient energy carriers, their empirical research demonstrates that the benefits

from clean fuels or technology adoption are usually smaller with stacking than

those expected from pure switching.

2.3. Microenergy Systems

Among the different sources of energy, electricity is considered to be one of

most important within the path toward economic sustainability, as electricity

is generally the preferred choice for lighting and running appliances [Pachauri

et al., 2013]. Moreover, electricity is often considered representative of the rural

development itself [Kaygusuz, 2011]. However, despite the increasing amount

of global investment in the power sector within the last 15 years, electricity

generators and distribution networks in rural areas still struggle to achieve the

required capacity in order to satisfy the constantly growing power demand.

Moreover, in several regions, traditional and centralized grid-based approaches

are neither physically feasible nor economically affordable for the majority of

unserved households, making the unavailability of electricity one of the most

visible signs of rural-urban differentiation and rural underdevelopment. Indeed,

electrification focusing on connecting rural areas to national or nearby grids are

mostly not considered to be the least-cost option [Saghir, 2005]. According

to [Zerriffi, 2007] the challenge of these initiatives is threefold: first, the high

dispersion and low consumption needs of rural populations result in a lower

return of the initial high costs to supply the energy utility; second, rural house-

holds have a limited ability to pay; third, the lack of required infrastructure and

of constant maintenance often lead to a low quality and reliability of delivered

electricity services.

To face these challenges, [Smith, 2000,Sch¨afer et al., 2011] proposed the con-

cepts of micro-energy appliances (MEAs) and micro-energy systems (MES).

The formers are defined as small and locally usable energy conversion units

that allow a spatial interconnection between energy demand and energy sup-

ply as an energy-converting appliance in the energy sector [Philipp and Sch¨afer,

2009], while MES consist of groups of MEAs and their framing system.

The PEPI Toolkit for MFIs 2. Energy Poverty

Electric Energy

MES Usage

Mini-grid (solar, hydro, wind) Residential, Micro-business

Solar Home System Residential, Micro-Business

Solar Pico PV Lamp Residential, Micro-Business

Solar Water Pump Livestock, Agricultural

Recreational, Residential

Thermic Energy

MES Usage

Solar Water Heater Residential, Micro-business

Livestock (Diary-factory)

Improved Cooking Stove Residential, Micro-business

(Manufactured Solid Fuel Stove)

Improved Baking Oven Residential, Micro-business

Biodigestor Residential, Micro-business

Solar Stove Residential

Solar Crop Dryer Agricultural

Table 2.1.: Examples of Microenergy Systems

Being decentralized energy systems, MES represent a way forward to target par-

ticular energy needs, making use of locally available energy resources and trans-

forming it to direct use at the end-user location [van der Straeten et al., 2014] (see

examples 2.1). Through the increase of the competitiveness of MES in compar-

ison to conventional energy sources (in terms of costs and quality services), its

marketability gains particular attention [Barnes and Floor, 1996,Saghir, 2005].

In fact, the findings from [Zerriffi, 2007] describe how energy delivery models

building on MES have an advantage on sustainability and replicability, com-

pared to centralized electrification projects. This is due to the adaptability to

the specific needs of the MES customers shifting decision-making and imple-

mentation at the local level. Furthermore, MES can provide a technically viable

alternative to conventional energy, especially for the tasks related to agricultural

production and processing (such as land preparation, planting, fertilization, ir-

rigation, harvesting, transport, processing and storage). Moreover, MES can

reduce the labour associated with physical drudgery and animal power, thus

helping to increase the productivity of several agricultural tasks.

However, the dissemination of MES requires the creation of favorable frame-

works [Groh, 2015]. In particular, it is necessary to overcome the prohibitive

initial upfront costs for end-users, the difficulty of defining cost-covered finan-

cial scheme, the need of customized mechanisms for MES distribution, and

the high costs of cleaner energy– due to the subsidies for fossil fuels and non-

renewables [IEA and Photovoltaic Power Systems Program, 2002,Beck and Mar-

tinot, 2004]. In recent years, practitioners and academia have extensively re-

searched preconditions and means for the successful implementation and dis-

semination of MES, as well as the economic and ecological impacts at different

The PEPI Toolkit for MFIs 2. Energy Poverty

levels [van der Straeten et al., 2014].

On the one hand, due to the limited ability to pay for energy services in econom-

ically deprived regions, finding a suitable and sustainable financing mechanism

for capital intensive MES is a major issue. In this context, flexible financing,

i.e., adapted to the end-user, might play an important role in actively foster-

ing energy access at the BoP. On the other hand, in order to be cost-effective,

the selection of sustainable MES should not be biased towards renewable en-

ergy sources, although it should take into account a proper assessment of the

efficiency and operational costs of traditional energy sources. When determin-

ing the most suitable choices for MES in a specific local context, the selection

should be based on local energy needs, source availability and institutional fac-

tors. Consequently, the involvement of local actors in energy access initiatives is

decisive to understand the prevailing local conditions and needs [Agbemabiese,

2009,Zerriffi, 2007].

2.4. Productive Uses of Energy

Besides the importance of energy access at the household level, modern energy

is associated with improvements of living also due to its potential of enhanc-

ing income generation and business opportunities, by creating new economic

activities or by increasing the production outputs of existing ones through a

more efficient use of energy and other resources [Etcheverry, 2003]. In other

words, the link between energy access and economic development is strongly

connected to the increase in productivity, defined as the ratio of value creation

to energy consumption and to the relevance of the considered energy services

for productive uses.

In order to integrate this concept in the context of MES, the work of [Saari,

2006], who defines the concept of business productivity in an operational way,

is of particular relevance. Namely, [Saari, 2006] describes productivity as a part

of economic activity with the main purpose of satisfying human needs. Met by

means of tools, the degree of need satisfaction depends to the success of the

tool in its purpose of use.

From a formal point of view, the definition of productive use of energy (PUE)

has been deeply debated and it has been recently framed under different um-

brellas. Particularly, PUE has been traditionally conceived as any use of energy

that directly helps generate income [Kittelson, 1998]. This view has been later

extended to other productive activities that are able to enhance income and

welfare, taking into account the impact that energy services can have on educa-

tion, health and gender equality [Cabraal et al., 2005,Etcheverry, 2003]. In this

case, the considered productive activities include work for income-generation

and wealth creation, comprising both market production with an exchange

value, and subsistence/home production with actual use value and potential

exchange value clancy-kooijman-2006. Moreover, [Meadows et al., 2003] identi-

fied expenses, collection, production and utilization time, as well as the level of

dependence the business processes on energy inputs as the main factors deter-

The PEPI Toolkit for MFIs 2. Energy Poverty

mining the value of modern energy for micro and small enterprises.

Focusing on rural areas, the major emerging productive uses for renewable en-

ergy include agriculture, powering small industry and commercial services, and

production of electricity for social services such as drinking water, education

and health care facilities [Martinot et al., 2002]. Table 2.2 illustrates how, by

addressing the energy components of agriculture and off-farm activities, the

potential for income generation of rural households and enterprises can be in-

creased, detailing the energy services involved in different productive processes

and the current available options in terms of renewable energy sources.

Energy services Income generating value

to rural HH and SMEs Renewable energy options

Irrigation

Better yields, higher value

crops, greater reliability,

growing during periods

when market prices are

higher

Wind, photovoltaic (PV),

Biomass

Illumination

Reading, many types of

manual production during

evening hours

Wind, PV, Biomass,

Micro-Hydro, Geothermal

Grinding, milling,

husking

Create value-added prod-

uct from raw agricultural

commodity

Wind, PV, Biomass,

Micro-Hydro

Drying, smoking (pre-

serving with process

heat)

Create-value added prod-

uct. Preserve produce to

enable selling to higher-

value markets

Biomass, Solar Heat,

Geothermal

Refrigeration, ice-

making (preserving

with electricity)

Preserve products to en-

able selling to higher-value

markets

Wind, PV, Biomass,

Micro-Hydro, Geothermal

Expelling Produce refined oils from

seeds Biomass, Solar Heat

Transport Reaching markets Biomass (e.g., biodiesel)

TV, radio, computer,

internet, telephone

Education, access to

market news, entertain-

ment, co-ordination with

suppliers and distributors,

weather information

Wind, PV, Biomass,

Micro-Hydro, Geothermal

Battery charging Wide range of services for

enduser

Wind, PV, Biomass,

Micro-Hydro, Geothermal

Table 2.2.: Energy services for productive activities, added-value and renewable

energy options

Several studies have portrayed the role of productive use of energy for develop-

ment of micro- small- and medium-sized enterprises (MSMEs) as a key factor

The PEPI Toolkit for MFIs 2. Energy Poverty

for community development. In fact, the effects of energy access on productive

activities have been claimed to be dependent on the nature of the local com-

munity, on development programs, the business owner’s skills, market access

and demand factors as well as on the access to credit [Barnes, 2007,Peters et al.,

2009].

However, in order to determine to which extent modern energy fosters mi-

croenterprise development, [Meadows et al., 2003] emphasize the importance of

analyzing the heterogeneity of the micro-enterprise sector as it entails a broad

spectrum of income-generating activities with varying needs. For instance, in

previous research on the role of energy in productivity, [Morris et al., 2007]

indicate how productive activities such as irrigating arable land, generating

salable crops, selling cool drinks, charging mobile batteries, or refrigerating fish

have direct advantages from improved energy access. Indeed, businesses are

differently dependent on energy; their energy-intensity and effects of obtaining

access vary accordingly. However, [Rogerson, 1997] stresses that access to energy

might indeed foster the modernization of already established micro-activities,

not necessarily encouraging the growth of new enterprises.

2.5. Measuring Energy Poverty

Energy poverty at the household level, seen as the lack of energy access, has

been traditionally measured by binary metrics, such as having or not having an

electricity connection and cooking with non-solid fuels 2. This characterization

masks major differences in the condition of the delivered energy services, calling

for a modernization of methods for quantifying the energy access.

In fact, the metrics focusing on the availability of grid connections are not able

to capture broader deficiencies in the relevant attributes, such as affordabil-

ity, reliability and quality of service [Global Tracking Framework (GTF), 2015].

Hence, binary metrics are an insufficient measure of energy poverty [AGECC,

2010,Practical Action, 2013,IEA and WB, 2014] when aiming at tracking the

progresses towards the SDGs. Similarly, metrics based on the usage of solid

biomass results in the same lack of information concerning available cooking

solutions, as the households might make use of multiple cooking fuels or cook

with improved cooking stoves to efficiently use fuel, reducing indoor air pollu-

tion and preventing health hazards. Therefore, both academia and practitioners

in the field of energy and development widely accepted that quantifying energy

access goes beyond binary assessments and it needs to entail different attributes,

in order to better capture the quality and quantity of delivered energy [Bazilian

et al., 2010,Pachauri et al., 2012a,Bensch, 2013].

Subsequently, although the definition of “access to modern energy services”

has been largely under discussion (see, e.g., [Pachauri, 2011,Khandker et al.,

2The denomination of non-solid fuels entails liquid fuels (e.g., kerosene, ethanol, other bio-

fuels, etc.), gaseous fuels (natural gas, LPG, and biogas), and electricity or solid fuels, such as

traditional biomass (firewood, charcoal, agricultural residues, and dung), processed biomass

(pellets, briquettes, etc.) and other solid fuels (e.g., coal, lignite, etc.)

The PEPI Toolkit for MFIs 2. Energy Poverty

2012,Bazilian et al., 2010,Bensch, 2013]), targets for 2030 point to the achieve-

ment of “universal access to modern energy services” as dictated by the SE4ALL

decade (2014-2024) of the UN. In order to build upon the results achieved in the

development of metrics, [Groh et al., 2016] point out the need of defining univer-

sal energy access and common assessment tools in order to enable stakeholders

to track endeavors and achievements.

Although a collective consensus on an energy poverty metric has not been

achieved yet, several proposals to measure energy deprivation have been pub-

lished. In contrast to poverty, which is commonly measured based on a relative

poverty measure (poverty line), energy poverty is related to the multidimen-

sional nature of energy access. Hence, the concept faces the challenges of the

absence of a universal definition of energy access and the complexity involved

in achieving accuracy in its measurement [Bhatia and Angelou, 2014].

Over the last decades, several indicators have been developed in order to cap-

ture the different uses of energy, either mirroring other indicators designed for

measuring poverty indices or as an evolution of existing unidimensional en-

ergy access indicators. These measures range from indicators to assess energy

access to indices quantifying the degree of development related to energy or

the deprivation of access to modern energy services. In particular, the inves-

tigations of [Bensch, 2013] categorize existing energy metrics in two groups,

unidimensional and multidimensional indicators, comparing in detail their ad-

vantages and their drawbacks. Similarly, [Nussbaumer et al., 2011] groups the

metrics developed to measure energy poverty and sustainable development in

three broad categories, distinguishing between single indicators, dashboard (a

set of individual indicators) and composite indices. Single indicators are eas-

ier to work with, but they fail in picturing the multidimensional issues, such

as development or poverty. A dashboard, being based on multiple indicators,

improve the comprehensiveness of the aspects to be monitored, although the

quantity of indicators may increase the complexity of analyzing changes over

time, sometimes requiring an aggregation model. Finally, composite indices

are based on a set of sub-indicators, aiming at capturing diverse dimensions of

an issue to be depicted in one indicator. On the one hand, composite indices

maintain the simplicity of single indicators, while still depending on multiple

attributes. On the other hand, the definition of a composite index requires a

reduction process and the final result is then depending on the efficiency of the

aggregation method and to the assumptions of assigned weights (due to the

arbitrariness involved).

The rest of this Section is dedicated to four relevant composite energy metrics

proposed in the last decade, which quantitatively assess energy poverty and

address the multidimensional nature of energy access: the Energy Development

Index (EDI) [IEA, 2004], the Energy Poverty Index (EPI) [Mirza and Szirmai,

2010], the Multidimensional Energy Poverty Index (MEPI) [Nussbaumer et al.,

2011], and the Total Energy Assessment (TEA) [Practical Action, 2012]. These

metrics are also summarized in Table 2.3, in terms of their definition of energy

poverty and of the considered attributes. The need of describing energy access

from a multidimensional perspective has been translated into the proposal of a

The PEPI Toolkit for MFIs 2. Energy Poverty

set of attributes to characterize the overall quality of energy services in devel-

oping countries, setting the basis of an innovative multi-tier framework [Bensch,

2013,Bhatia and Angelou, 2015], which will be described in detail in Section 2.6.

Energy Development Index (EDI) The EDI [IEA, 2004], mirroring the Hu-

man Development Index (HDI) from United Nations Development Programme

(UNDP) [Bensch, 2013], is based on six energy services prescribing a minimum

level of service. It also combines qualitative with quantitative indicators to

analyze household access to fuels, electricity and mechanical power, analyzing

their progression in the use of modern energy services.

Energy Poverty Index (EPI) The EPI, developed by [Mirza and Szirmai,

2010] is a composite index to measure the degree of energy poverty among rural

households (originally tailored for the case of rural Pakistan). In particular, The

EPI measures the inconvenience for the household associated with the use of

the different sources of energy, taking into account its energy shortfall and the

household size.

Multidimensional Energy Poverty Index (MEPI) The MEPI [Nussbaumer

et al., 2011] has been introduced as a modification of the Multidimensional

Poverty Index (MPI). It is composed by a measure of the incidence of energy

poverty and by a quantification of its intensity, focusing on energy services.

Based on micro-data from household surveys, the MEPI enables the estima-

tion of country values from available datasets. However, the arbitrary poverty

cut-off is one of its main deficiencies.

Total Energy Assessment (TEA) The TEA methodology [Cast´an Broto

et al., 2015,Practical Action, 2012] assesses key energy services against minimum

standards, focusing on energy services, on the use of biomass, on electricity

access and on mechanical power. It consists in a survey for energy access at

household level and in three indicators comprising the Energy Supply Index

(ESI).

Additionally, as identified by [Bensch, 2013], a further metric can be constructed

out of the Correlation Services Proxy Index (CSPI) [Rippin, 2011] to create an

energy poverty metric, the Correlation Energy Services Proxy Index (CESPI).

In contrast to the MEPI, the CESPI is more restrictive as deprivation of any

sub-dimension results in classification as energy poor. On the other hand, the

CESPI is more sensitive, as it is able to capture the correlation between energy

poverty indicators. Among other proposals, the approach of [Foster et al., 2000]

quantifies energy poverty by establishing a Fuel Poverty Line (FPL), defined as

the inability by households to meet their energy needs. This metric considers

then energy consumption, energy efficiency and a method to estimate a country

fuel poverty line, providing information on energy consumption and the related

costs to such access.

The PEPI Toolkit for MFIs 2. Energy Poverty

Index Dimensions Energy poverty

definition

EDI

•Per-capita commercial energy con-

sumption

•Share of commercial energy in to-

tal final energy use

•Share of population with access to

electricity

Adequate access if has

access to both, modern

fuels and appliances

EPI

Qualitative and quantitative indicators

•the “energy inconvenience excess”

associated with the energy mix

used

•insufficient energy to meet ba-

sic household needs (energy short-

falls)

30% above the average

value of total energy

inconvenience

MEPI

•Modern cooking fuel and stove us-

age

•Electricity access

•Radio or TV ownership

•Phone ownership

•Fridge ownership

Dual cut-off:

Dimensional cut-offs

for each sub-dimension

& weights

TEA

•Modern cooking fuel and stove us-

age

•Electricity access and usage

•Radio or TV ownership

•Phone ownership

•Fridge ownership

•Energy for enterprises

•Energy for community services

Energy poor if any

dimension is deprived

Table 2.3.: Multidimensional energy poverty metrics

The PEPI Toolkit for MFIs 2. Energy Poverty

2.6. The Multi-tier Framework Approach

Recently, a new multidimensional definition of energy access, has been pro-

posed under the Global Tracking Framework (GTF) initiative, one of the four

initiatives of the SE4ALL Global Knowledge Hub, hosted by the Energy Sector

Management Assistance Program (ESMAP) of the World Bank.

In this approach, energy access is characterized using a multi-tier framework

(MTF), i.e., ranking the different attributes describing energy access in different

tiers, in order to better capture the quantity and quality of electricity supply, as

well as the efficiency, safety and convenience of cooking facilities [Global Tracking

Framework (GTF), 2013]. The concept of a multidimensional measurement of

energy supply was first proposed by AGECC, EnDev and Practical Action

[AGECC, 2010], while the multi-tier ranking according to specific thresholds

was finally brought into light in 2013 as the ESMAP, in consultation with a

diverse group of agencies and programs3, elaborated a new definition of energy

access based not only on energy usage, but taking into account its performance

along a set of specific attributes [Bhatia and Angelou, 2015].

In particular, this innovative methodology, referred as a new “milestone” for

the monitoring of global progress [Bensch, 2013], measures energy access based

on desirable attributes such as: adequateness, availability (when needed), reli-

ability, quality, affordability, legality, convenience, healthy, and safety [Bhatia

and Angelou, 2014]. In each of these dimensions, the energy access performance

is ranked from 0 (the lowest tier) to 5 (the highest tier), depending on specific

thresholds, that must be defined for each attribute. The tier levels reflect a

balance between the diverse spectrum of energy access, attempting to provide

meaningful differentiation between energy access attributes, in order to obtain a

technology-neutral index, which is a key for energy access measurement [Bazilian

et al., 2010].

Moreover, the different areas of energy use are considered, referred as locales of

energy access:

•Energy access at the households level, described by the multi-tier frame-

works for electricity supply, electricity services, electricity consumption,

energy for cooking solutions and energy for space heating.

•Energy for productive use

•Energy for community uses, described by health facilities, educational

facilities, street lighting, government buildings, and public buildings.

For each component of these locales, a separate tier-ranking is calculated, in

order to characterize all the required qualities for energy services. After hav-

ing defined the ranking for each attribute of a particular framework, the final

ranking is obtained using a lower-based rule, i.e., assigning to each individual

3The collaborative effort included organizations such as EnDev, Lighting Africa, Practical

Action, The Global Alliance for Clean Cookstoves, UNDP, UNIDO, World Bank and WHO,

with previous experiences in energy poverty indexes development (such as Practical Action,

The Global Alliance for Clean Cookstoves), or which rigorously track their impact through

internal metrics (such as EnDev, Lighting Africa).

The PEPI Toolkit for MFIs 2. Energy Poverty

(household or enterprise) the lowest tier among all attributes.

Considering a population sample, the multitier ranking allows to compute dif-

ferent access indices, based on a weighted average of the overall performance.

At this stage, the different frameworks can be combined in composite indices.

The household locale (which will be considered for the case study described in

Chapter 5) and the locale for productive use (which is the most relevant in the

context of microbusiness development) will be shortly described in Sections 2.6.1

and Section 2.6.2, while Section 2.6.3 details the approach proposed in [Bhatia

and Angelou, 2015] for the composite index calculation.

2.6.1. Energy Access at the Household Level

Electricity supply In order to overcome the limits of binary measurements of

electricity access, the multi-tier framework for electricity supply considers the

key attributes that constrain its usefulness (such as limited quality, affordability,

the presence of illegal connection and the risk of accidents). In detail, household

electricity is described through the follow attributes: (i) capacity, (ii) duration

(including daily supply and evening supply), (iii) reliability, (iv) quality, (v)

affordability, (vi) legality, and (vii) health and safety (see Table 2.4).

Electricity Services and Electricity Consumption Besides electricity supply,

a separate multi-tier framework can be defined for access to electricity services,

in order to track how an improvement in electricity supply reflects in increased

and improved access to services. In particular, the matrix measuring access to

household electricity is based on the type of appliances used in the household

(see Table 2.5, right). Hence, different tier ratings across access to electricity

supply and access to electricity services aim at reflecting the case when appli-

ances are available, but supply is poor, or the case when appliances (or high

consumption) are unaffordable, despite adequate supply.

A further multi-tier framework is defined for electricity consumption, closely

aligned with the one for electricity services (see Table 2.5, left). In this case,

the tier thresholds are based on annual and daily consumption levels, without

focusing on the diversity of appliances actually used by the household. More-

over, the potential use of energy efficiency appliances not necessarily reflected

in the estimated thresholds [Bhatia and Angelou, 2015].

Electricity Supply

Capacity Duration Reliability Quality Affordability Legality Health/Safety

(Power, Daily capac-

ity Service) (Hours) (No. of disrup-

tions) binary binary binary binary

gradual gradual gradual

Tier 5

Min 3000 kWh and

8219 Wh or very high

power appliances

Min 22 hours/day, min

4 hours/night

Max 3 disrup-

tions per week,

max 2h long

Good quality

of energy

supply

Cost of consumed

365kWh less than

5% of income

Legal en-

ergy supply

No accident or

risk

Tier 4

Min 1250 kWh and

3425 Wh or high

power appliances

16-22 hours/day, min 4

hours/night

Max 14 disrup-

tions per week

Tier 3

Min 365 kWh and

1000 Wh or medium

power appliances

8-16 hours/day

More than 14

disruptions per

week

Poor qual-

ity, damaged

appliances

Illegal en-

ergy supply

Accidents or risk

feeling

Tier 2

Min 73 kWh and 200

Wh or low power ap-

pliances

4h-8h/day, min

2h/night

Cost of consumed

365kWh more than

5% of income

Electric lighting,

air circulation (if

needed), TV & phone

charging

Tier 1

Min 4.5 kWh and 12

Wh or very low power

appliances

Less than 4 hours/day,

min 1h/night

Lighting 1000 lmhrs/

day & phone charging

(Min 12 Wh)

Tier 0 No electricity Less than 4 hours/day,

less than 1h/night

Table 2.4.: Multi-tier Matrix: Thresholds of attributes and tier ranking standards for the electricity supply framework

The PEPI Toolkit for MFIs 2. Energy Poverty

Electricity Consumption Electricity Services

Tier 5 More than 3000 kWh / Year Very high-power services (air condi-

tioning, electric water heater)

Tier 4 More than 1250 kWh / Year High-power services (microwave,

hair dryer, toaster, iron)

Tier 3 More than 365 kWh / Year

Medium-power services (fridge,

freezer, washing machine, mixer,

rice cooker, water pump)

Tier 2 More than 73 kWh / Year Low-power services (TV, PC,

printer, ventilator)

Tier 1 More than 4.5 kWh / Year Very-low power services (light bulbs,

phone charger, radio)

Tier 0 Less than 4.5 kWh / Year None of the above

Table 2.5.: Thresholds of attributes and tier ranking standards for Electricity

Consumption and Electricity Services Frameworks

Energy for Cooking Solutions In order to measure the access to energy for

cooking, the multitier framework considers the following attributes: (i) health

(based on indoor air pollution), (ii) convenience (based on fuel collection time

and stove preparation time), (iii) safety, (iv) affordability (based on costs on

cookstove and fuel), (v) efficiency, (vi) quality, and (vii) fuel availability.

This framework has been defined consistently with a rating system proposed by

the International Workshop Agreement on Cookstoves (IWA)4for measuring

cookstove performance [Bhatia and Angelou, 2015]. The framework matrix is

detailed in Table 2.6. Note that three of the seven attributes composing this

framework (health, safety and efficiency) require measurement by a competent

agency in order to correctly evaluate the energy access of the household.

Energy for Space Heating At the household level , energy for space heating

(where needed) can be availed through a range of solutions, including electric

heating, fuel-based centralized district heating, fuel-based standalone heating,

and direct solar heating. In this case, the multitier framework is based on (i)

capacity, (ii) duration, (iii) quality, (iv) convenience (fuel collection time), (v)

affordability, (vi) reliability, (vii) health (air quality) and (viii) safety.

4The IWA was organized in 2012 by the The Partnership for Clean Indoor Air (PCIA) and the

International Organization for Standardization (ISO) organized the IWA in February 2012.

The rating system agreed by the participants proposed to evaluate cookstove models in tiers

of performance in different areas: fuel usage, emissions, indoor emissions and safety.

Cooking Facilities

To be measured by a local competent agency

Health Safety Efficiency Convenience

Availability

of primary

fuel

Affordability Quality

IWA Safety

Tiers, accidents

or perceived

risks

Stove preparation

(SP) time & fuel

acquisition and

preparation (FAP)

time gradual

gradual binary binary

Tier 5 PM2.5<10 mg/m3,

CO<7 mg/m3

IWA Tier 4, no

accidents or per-

ceived risks

Tier 4

SP time <0.5

min/meal, FAP time

<2 min/meal

Primary fuel is

readily available

all year

Stove and fuel

cost <5% of HH

income

No major effect

of primary fuel

quality

Tier 4 PM2.5<35 mg/m3,

CO<7 mg/m3

SP time <1.5

min/meal, FAP time

<5 min/meal

Primary fuel is

readily available

for at least 9

months/year

(80%)

Tier 3

PM2.5<100

mg/m3, CO<20

mg/m3

IWA Tier 3 Tier 3

SP time <3

min/meal, FAP

time <10 min/meal

Available

less than 9

months/year

Stove and fuel

cost >5% of HH

income

Heat rate vari-

ates

Tier 2

PM2.5<250

mg/m3, CO<50

mg/m3

IWA Tier 2, ac-

cidents or per-

ceived risks

Tier 2

SP time <7

min/meal, FAP

time <15 min/meal

Tier 1

PM2.5<250

mg/m3, CO<50

mg/m3

IWA Tier 1 Tier 1

Tier 0 higher emissions IWA Tier 0 Tier 0

Table 2.6.: Multi-tier Matrix: Thresholds of attributes and tier ranking standards for cooking solutions framework

The PEPI Toolkit for MFIs 2. Energy Poverty

2.6.2. Energy Access Framework for Productive Engagements

The complexity of defining a common metric for energy access for productive

use is due to the wide diversity of productive activities, to their varying scales

of operations and degrees of mechanization.

In this case, the multi-tier methodology is built upon the concepts that produc-

tive uses of energy refer to those that allow to increase income or productivity,

i.e., value-adding activities, and that these uses involve diverse sources of en-

ergy. Therefore, in order to measure the level of access of energy, the different

applications are grouped into broader categories as lighting, information and

communication, motive power, space heating, product heating and water heat-

ing [Bhatia and Angelou, 2015]. For each of these categories, the multi-tier

framework (detailed in Tables 2.7 and 2.8), is built on nine attributes that

determine the usefulness of the supply for each application needed for the pro-

ductive activity.

In the course of a household energy survey, the assessment consists in the fol-

lowing steps:

•Identification of earning members

•Identification of relevant energy appliances used based on their significant

impact on productivity, sales, cost or quality

•Identification of primary energy source for each application

•Assessment for the key attributes of energy supply and lower-based overall

tier ranking

Attribute Tier 0 Tier 1 Tier 2 Tier 3 Tier 4 Tier 5

Capacity

Electricity (Power) Min 3W Min 50 W Min 200 W Min 800 W Min 2kW

Electricity (Daily

supply) Min 12 Wh Min 200 Wh Min 1.0 kWh Min 3.4 kWh Min 8.2 kWh

Electricity (Typical

source)

Solar

lanterns

Solar Home

Systems

Generator or mini-

grid Generator or grid Grid

Non-electric

Available non-electric

energy partially

meets requirements

Available non-electric

energy largely meets

requirements

Available non-electric

energy fully meets all

requirements

Both No application is absent solely due to energy supply constraints

Duration of daily

supply

Electricity Min 2 hrs Min 4 hrs Min 50% of working

hours

Most of working

hours (Min 75%)

Almost all of working

hours (Min 95%)

Non-electric

Available non-electric

energy partially

meets requirements

Available non-electric

energy largely meets

requirements

Available non-electric

energy fully meets all

requirements

Both No application is absent solely due to energy supply constraints

Reliability

No reliability issues

that have severe im-

pact

No reliability issues

or little impact

Quality No quality issues that

have severe impact

No quality issues or

little impact

Table 2.7.: Thresholds of attributes and tier ranking standards for Productive Uses of Energy Frameworks - Part 1 (Note: W: watts;

Wh: watt-hours)

Attribute Tier 0 Tier 1 Tier 2 Tier 3 Tier 4 Tier 5

Affordability

Variable cost of en-

ergy is less than two

times the grid tariff

Variable cost of en-

ergy is less than grid

tariff

Legality Energy bill is paid to authorized body

Convenience

Time and effort in se-

curing and preparing

energy does not cause

severe impact

No convenience issues

or little impact

Health (indoor

air quality from

use of fuels)

PM2.5 (mg/m3)

— CO (mg/m3) [To be specified by competent agency (WHO)] <35(WHO, IT-1) —

<7(WHO Guideline)

<10 (WHO Guide-

line) — <7(WHO

Guideline)

OR Use of fuels

(BLEENS) non-BLEENS for heating (open/with smoke extraction) Use of BLEENS or equivalent (if any)

Safety

Energy supply solu-

tions have not caused

any accident over

the past year that

required professional

medical assistance

Energy supply have

not caused any ac-

cident over the past

year

Table 2.8.: Thresholds of attributes and tier ranking standards for Productive Uses of Energy Frameworks - Part 2 (Note: BLEENS

consists of: biogas, LPG, ethanol, electricity, natural gas and solar; CO-carbon monoxide; kW: kilowatts; kWh: kilowatt-

hours; LPG: liquid petroleum gas; PM: particulate matter)

The PEPI Toolkit for MFIs 2. Energy Poverty

2.6.3. The Energy Access Index

The approach described in [Bhatia and Angelou, 2015] also introduces a com-

posite Access Index (AI), with the aim of measuring the energy access level

of an entire population (i.e., a region or a country). According to [Bhatia and

Angelou, 2015], the index is computed as an arithmetic mean via

AI =

k=0

VkPk,(2.1)

where kis the tier number, Pkis the proportion of households in tier kand Vk

is a measure of the degree of access of the people in tier k. In particular, the

choice Vk= 20kis used in [Bhatia and Angelou, 2015], which yields an Access

Index ranging from 0 to 100. In this case, the AI can be interpreted as the

overall access percentage in a particular sector for the considered population.

Figure 2.2 sketches two particular example of the AI computation, in which two

different multi-tier rankings produce the same AI.

As observed in [Bhatia and Angelou, 2015], there is no particular reason for

taking Vk= 20k. As well, different approaches for computing the AI (such as

a geometric mean instead of an arithmetic mean) could be more appropriate in

different circumstances.

Figure 2.2.: Two examples of AI computation according to (2.1), producing

the same composite index. In the first situation, the population

is mostly concentrated in tier 0, while, in the second example, the

sample is more distributed among tiers

One of the motivations behind the introduction of the AI is the possibility

of measuring the improvement in energy access of a population over time, by

multiple and consecutive implementations of the energy survey. On the one

hand, by reducing the multi-tier rankings (one for each framework) to a set of

composite indices allows the consideration of different samples of the population

for each evaluation, as only the average performance is taken into account.

On the other hand, the tiers depend on predefined thresholds, while several

attributes are binary (i.e., consisting only of tier 0 or 5). Hence, it is difficult

to estimate to which extent the composite index, which is solely defined by the

lowest tier among all attributes, might be biased by these thresholds.

3.Tackling Affordability and Access to

MES: Green Microfinance

The lack of energy access in rural areas can be attributed to a variety of factors,

including economic meltdowns, market performance and legal and regulatory

arenas [Beck and Martinot, 2004,The World Bank, 2008a]. Among these, the fi-

nancial barriers, which result in both a limited ability to pay and in the unavail-

ability of high up-front investments, have a considerable impact. As discussed

in Chapter 2, decentralized energy solutions, such as MES, might provide more

affordable options due to their adaptability to the particular context (e.g., solar

home systems, biogas digesters, solar water heaters, grain-mills, or improved

cooking stoves) [Barnes and Floor, 1996,Smith, 2000] and by enabling a gradual

increase in the size of investments. However, the initial costs still remain a

major issue in several cases.

The issue of affordability of energy services, as observed by [Winkler et al., 2011],

does not affect only the energy access, but also the energy use. Particularly, in

the case of electricity, [Lucas et al., 2003] maintain that the populations deprived

of energy access, besides needing to overcome the initial connection costs (or the

costs for the acquisitions of power source equipments), often lack the means to

acquire efficient and appropriate MES. Hence, financial barriers are considered

to be a cause of the so-called vicious cycle of poverty, a concept that describes

the correlation between micro-enterprising activities (income-generating activi-

ties which sustain livelihoods) and access to modern energy services. According

to [Lucas et al., 2003], such a cycle is broken only by efficiently tackling effi-

ciently the energy poverty, i.e., by combining improved energy services with

additional income generation.

Microfinance is an approach to deliver financial services to a segment of the pop-

ulations otherwise excluded from commercial banking. It is based on the princi-

ple that the BoP, with access to improved financial services, has the capacity to

generate higher income by increasing the output of economic activities. In order

to overcome the informational and institutional barriers usually associated with

commercial banking [Armend´ariz de Aghion and Morduch, 2005,Banerjee et al.,

2010], microfinance makes use of social pressure and characteristic lending tech-

niques (small amounts, consecutive group lending, community engagement as

loan guarantors and social default penalties, among others).

In recent years, microfinance has been seen as a viable strategy for tackling

financial barriers to energy access, such as initial investment costs. In fact,

The PEPI Toolkit for MFIs 3. Green Microfinance

by supplying financial services to the non-bankable (by commercial banking),

microfinance can overcome the liquidity constraints of households and micro-

entrepreneurs, enabling the financing and affordability of MES over time1.

Several financial institutions worldwide have developed strategies devoted to

improving energy access for underserved populations via microfinance mech-

anisms. These approaches often consist of establishing partnerships with en-

ergy product service suppliers and adapting existing financial services with new

credit products, so-called green loans, which finance the purchase of renewable

energy or energy efficient technologies [Wenner, 2002,Hall et al., 2008,Allet,

2012]. These intersectoral partnerships demonstrate the fact that by enabling

the acquisition of modern technologies and designing tailored microloans, MFIs

can help to reduce poverty [Srinivasan, 2007,Rao et al., 2009], attract new cus-

tomers, and improve access to MES for existing customers, thus enlarging the

market of energy product service suppliers [Morris and Kirubi, 2009].

However, despite these developments, the microfinance movement to address

energy access has advanced at a slow pace, and the successful scaling-up of

projects combining energy access and microfinance schemes is still the subject

of open discussions (see e.g., [Groh and Taylor, 2015]). In fact, while energy

service suppliers are often unwilling to operate in remote rural areas without

a guarantee that financing is available for their customers, MFIs are unwill-

ing to issue loans for MES without a guaranteed partnership with an energy

service supplier with high-quality products [Morris and Kirubi, 2009]. On the

other hand, the scale at which MFIs operate is insufficient to alleviate energy

poverty on a larger scale [OECD and IEA, 2010], as, without additional subsi-

dies, microfinance mechanisms encounter difficulties adjusting loan instalments

to prior expenditures for kerosene, candles, diesel or battery charge [Martinot

et al., 2001,Morris and Kirubi, 2009]. In this respect, through scenarios simu-

lations, [Pachauri et al., 2012b] analyzed the best strategies in order to achieve

universal energy access by 2030. The results showed that supporting policies

that provide a combination of subsidies and microfinance, thus increasing af-

fordability of running costs, are likely to be most successful and cost-effective

in achieving the final targets.

This chapter focuses on the linkage between financial inclusion and energy ac-

cess, on the debated role of microfinance in scaling up access to MES and on

available metrics to measure the environmental performance of microfinance.

Firstly, Section 3.1 introduces the role of microfinance in overcoming financial

1This research focuses on the role and the potential of microfinance mechanisms for the financ-

ing of MES, in order to explore green microfinance’s role in assisting those MFIs involved

in green lending programs. It must also be mentioned that the use of mobile banking and

mobile cash services in retail and wholesale shops have enabled new opportunities for credit

access and increased the affordability of MES. In fact, by eliminating the need for in-person

transactions, mobile payments have helped to reduce barriers to trade (see, e.g., the case

of M-Pesa in Kenia and bKash in Banglasdesh, both as mobile banking and cash transfer

services used for selling products and buying from their distributors) and such technical and

financial innovations on end-user financing have enabled distributed energy services com-

panies (DESCOs) to rapidly expand their market, under schemes such as Pay-As-You-Go

(PAYG).

The PEPI Toolkit for MFIs 3. Green Microfinance

barriers for energy access, as well as the mechanisms used to facilitate energy

access, distribution and delivery or installation. Next, Section 3.2 describes the

financial models behind the green lending programs, while Section 3.3 explores

approaches that can lead to inter-sectoral collaborations that serve as models

of partnerships or ventures, in order to enhance energy access in a specific geo-

graphical region. The last part of the chapter is dedicated to relevant indicators

used in the field of microfinance: Section 3.4 discusses the assessment of the

impact of green microfinance, while Section 3.5 focuses on the Progress out of

Poverty Index (PPI), which is used by MFIs to estimate the level of poverty of

their clients.

3.1. Green Microfinance and Green Lending

Microfinance delivers financial services to low-income populations who other-

wise lacking access to commercial banking services, by exploiting peer pressure

and innovative lending techniques. In particular, in order to be successful,

the business model of MFIs relies on a solid distribution network and on con-

stantly maintaining a close relationship with clients. In the last decade, since

the launch of the International Year of Microcredit 2005 by the United Nations

(UN) and the honouring of the Grameen Bank and its founder Muhammad

Yunus (awarded the Peace Nobel Prize in 2006), the expectations regarding the

impact of microfinance in development processes have steadily increased.

The popularity of microfinance, together with microfinance success stories world-

wide, lead to an explosion of interest in the microfinance sector in developing

countries, for both non- and for-profit institutions [Ghosh, 2013]. However, the

impact of microfinance has been severely questioned as there remains a lack

of empirical evidence to demonstrate the benefits. One of the main criticisms

claims is that microfinance intensifies mechanisms that ensures the short-term

profitability of MFIs while increasing the vulnerability of the poorest and only

benefiting the better-off [Banerjee et al., 2010,Coleman, 2006]. Other authors

have argued that microfinance resulted in a growing bubble of expectation de-

spite of the institutional and sectoral meltdowns, disappointing those who hoped

it would alleviate poverty [Hulme, 2000,Maren et al., 2011,Ghosh, 2013].

According to [Groh, 2014], in order to positively affect the quality of life of low-

income clients, one must take into account the symbiotic relationship between

financial and energy inclusion, which is based on their mutual benefits and on

a bidirectionally causality. Access to finance can lead to energy inclusion in

terms of affordability and better financial means, i.e., people who have access

to financial services are able to finance their basic energy needs and either pay

for grid-supplied electricity or purchase a distributed energy generation system.

On the other hand, by financing the purchase of a distributed energy genera-

tion system through small monthly installments to retailers or intermediaries,

those at the BoP can then use the energy generated to increase their produc-

tive capacities and repay over the course of two to three years [MicroEnergy

International, 2014].

The PEPI Toolkit for MFIs 3. Green Microfinance

Due to their infrastructure (i.e., multiple offices, established communication

channels, local loan officers networks) and to their customer relationship man-

agement, MFIs have been acknowledged as vehicles which help enable energy

access, by allowing access to MES via their microlending mechanisms [Devine

et al., 2010,UNDP, 2000,Mohiuddin, 2006,Srinivasan, 2007,Rao et al., 2009]. The

label green microfinance refers to the ensemble of microfinance services address-

ing the triple bottom line, i.e., fostering an impact at the economic, social and

environmental levels. More specifically, green microfinance entails a variety of

internal or external activities that MFIs undertake with the common goal of

fostering green businesses and contributing to environmental preservation [Re-

alpe Carrillo, 2014]. [Hall et al., 2008] listed the following as main motivations

to “go green”: scale risk management, regulation procedures, competition pres-

sure, ethical considerations and access to funding [Van Elteren, 2007], as the

main key drivers, while [Allet, 2011] also indicated risk mitigation interests,

ethical responsibilities, donor pressure or business opportunities, directed by

both ethical and instrumental arguments.

Among the activities comprised in the green microfinance spectrum, green lend-

ing aims at facilitating access to clean energy technologies- renewable energy or

energy efficient appliances- by designing customized credit products for the tar-

get population in order to finance energy systems2[Realpe Carrillo, 2014,Pieran-

tozzi et al., 2015].

If the loans for clean energy technologies are appropriately designed to closely

match installments with existing expenditures on fuels or income flows [Mor-

ris et al., 2007], energy access lending programs can result in attractive self-

repaying credit schemes, as a consequence of automatically generated sav-

ings [Leva¨ı et al., 2011]. Interested MFIs in diversifying their portfolio design

loans for MES [Leva¨ı et al., 2011], facilitating necessary after-sale services and

expanding their coverage by developing new business models [Allderdice et al.,

2007,Kebir et al., 2013]. However, as green loans are built upon the linkage be-

tween the microfinance and energy sectors, MFIs must possess the willingness

and the capabilities to channel capital into loans for MES, as well as a high

capacity to assume the largest risk [Rao et al., 2009].

Hence, pursuing these opportunities requires not only strategic decisions from

the MFIs’ management followed by identification of adequate products and care-

ful program design, piloting, and roll-out [Leva¨ı et al., 2011], but also extensive

and unconditional support from the energy product service suppliers [Morris

et al., 2007]. These considerations are in line with the results of [d’Almeida

and Roberts, 2014]. Based on an analysis of 17 Latin American countries, the

authors studied the hypothesis that high demand for MES and low barriers to

market entry promote entry into the green microfinance market. Among others,

2[Shuite and Forcella, 2015] framed green lending within Green Inclusive Finance (GIF), which

refers to financial services that support economic growth in a clean, resilient and sustainable

manner, and focus on the BoP including micro, small, and medium-sized enterprises in

low-income countries or such subsets of population within other developing countries [IFC,

2013]. As such, green lending can be seen as one of many instruments that address climate

change, under the umbrella of GIF comprising the multidimensional purpose of economic

development, social inclusion and environmental sustainability.

The PEPI Toolkit for MFIs 3. Green Microfinance

they show a low correlation between the fact that an MFI decided to offer green

loans and the high demand of MES from the population due to low electrifica-

tion rates or high costs of substitutes (e.g., high electricity prices, high diesel

prices, etc.). Their findings specifically revealed that government interventions

are not necessarily required in order to encourage firms to diversify their port-

folio with green loans, provided that there is a strong business environment for

microfinance with low barriers to enter.

Figure 3.1.: Number of MFIs reporting on Green Performance Indicators (see

[Pierantozzi et al., 2015])

The portfolio diversification of MFIs, in order to integrate green lending prod-

ucts, can be seen as an opportunity to extend their market and to better reach

vulnerable communities. However, whether driven by environmental conscious-

ness or business opportunities, green initiatives still appear to lack long-term

sustainability. Moreover, the scaling-up of green lending programs still remains

a challenge, and profitable business models have not been validated on a large

scale yet [Realpe Carrillo, 2014]. Indeed, the fact that the number of MFIs of-

fering green lending -despite increase- is still relatively small invites a reflection

(see Figure 3.1).

The analysis from [Shuite and Forcella, 2015] claims that the main arguments

restraining financial institutions to integrate green loans into their portfolio

includes high investment costs, the low environmental awareness, and the over-

coming of the issues related to the establishment of complex partnerships. Ac-

cordingly, despite of the fact that green microfinance contributes to overcome

the barriers of access to credit, green loans and microcredits, especially indi-

vidual loans, seldom target the poorest of the poor, to whom MES are still

not affordable [Groh and Taylor, 2015]. For instance, [Beck and Martinot, 2004]

stress that notwithstanding the availability of microfinance services aiming at

improving MES access, loan conditions and duration are still significant barriers

for this segment of the population. Moreover, it has been claimed that loans

are not always sufficient to overcome the barrier of the initial cost of MES in

The PEPI Toolkit for MFIs 3. Green Microfinance

remote areas. The studies presented by [Beck and Martinot, 2004,The World

Bank, 2008b], among others, argue that, in order to guarantee fair competition

between renewable and conventional energy sources, proper legal and regulatory

framework are needed, such as limiting subsidies for fossil fuels and introducing

additional grants for supporting clean energies.

3.2. Clean Energy Technologies Finance Models

Initiating a green lending program requires strategic decisions from the MFIs’

management and operational capacities. As a next step, MFIs and energy prod-

uct service suppliers negotiate their role (responsibilities and ownership) based

on different delivery models, that vary worldwide with respect to engagement

strategies, product offerings, service delivery, and specific business models. In

this context, [Parkerson, 2005] differentiates three different credit sale forms

•the lease purchase model,

•the dealer credit model (one-hand model), and

•the end-user credit model (two-hand model).

While in the first two approaches only one organization is responsible for the

production, delivery, financing and after-sales services of the energy systems,

in the two-hand model (see also Figure 3.2) the financial institution establishes

a cross-sectoral partnership with energy product service suppliers, so that the

responsibilities between the two parts are clearly shared, and the business re-

lationship stresses on utilizing each actor’s expertise concerning the end-users.

Specifically, financial institutions provide financing and technology suppliers

install the system, train and offer after sale maintenance3.

A further benefit of two-hand models is that MFIs are rather attracted to part-

ner with energy companies [Morris and Kirubi, 2009], instead of incurring heavy

organizational changes or assuming the entire supply chain of the technologies

by themselves [MEI and PF, 2010,Kebir and Heipertz, 2010]. However, crucial

prerequisites for an appropriate supply chain design are both reliable energy

product service suppliers and a high level of commitment of the MFIs’ manage-

ment and of their operational forces [Morris et al., 2007]. Indeed, diversifying an

MFIs’ portfolio by introducing green loans represents an opportunity for both

partners to extend their markets jointly, thereby reaching vulnerable communi-

ties in need of both access to finance and/or energy [Groh, 2014]. Specifically,

the assumption behind the two-hand model is that, through a systematic ap-

proach, MFIs are able to build up commercial relationships with energy product

service suppliers and thus to (i) diversify the MFI’s portfolio, (ii) foster local

3An exemplary one-hand model approach takes place in Bangladesh at Grameen Shakti.

Founded in 1996, Grameen Shakti is a worldwide leader in energy lending that dissemi-

nates solar home systems (SHS) to energy deprived populations. The organization has been

able to convert the challenges of energy supply into business cases along the supply chain.

Besides SHS, Grameen Shakti also offers financing for improved cooking stoves and biogas

digesters. As of September 2016, the organization has financed 1,692,194 SHS, 949,984 ICS

(since 2006) and 32,668 biogas gasifiers (since 2005), leading the MF green sector in MES

financing. For further information see www.gshakti.org.

The PEPI Toolkit for MFIs 3. Green Microfinance

industries by increasing the market outreach and facilitating know-how and

technology transfer and (iii) satisfy the energy needs of the underserved clients.

Figure 3.2.: Two-Hand Model Scheme [Realpe Carrillo et al., 2015]

Among the green microfinance initiatives, green lending still appears to lack

long-term sustainability, and therefore the scaling up of a two-hand model ap-

proach remains a challenge for both parties. The major obstacles are the access

to better technical assistance, the development of efficient supply chains on a

large scale, and the lack of profitable business models [Realpe Carrillo, 2014].

More in detail, the high investment costs in the learning process and the limited

access to knowledge and to technical expertise are two main obstacles for the

MFIs as they pursue large-scale commercialization. Concerning local energy

companies (mostly SMEs) involved in two-hand green lending models, the main

challenges reside in the design of the supply chain and in the adaptation of their

products and services in order to be able to expand their market together with

the partner MFIs. Furthermore, efficiently reaching remote clients becomes a

challenge for both MFIs and energy product service suppliers.

3.3. Cross-sectoral Cooperation to Tackle Energy

Poverty

The joint effort of the microfinance and energy sector towards improving the

access to energy services may contribute substantially to the achievement of

the SDGs. However, in order to achieve results on a significant scale with

the implemented electrification projects, serious investments are required from

multilateral and bilateral donor community, as budgetary resources of MFIs

and energy product service suppliers are constrained.

Regarding this issue, two approaches building on cross-sector cooperation have

been proposed by [De Gouvello and Durix, 2008]. They describe the experience

of a multi-sector committee in Senegal which, given earlier attempts from the

energy sector to contribute to poverty reduction, made key policy decisions in

order to support a renewable energy program joining the financial and the en-

ergy sectors. In view of this study, [De Gouvello and Durix, 2008] describe two

The PEPI Toolkit for MFIs 3. Green Microfinance

operational methodologies to foster cooperation between sectors with comple-

mentary steps from a systematic and a pragmatic perspective. Their motivation

consisted in proposing an approach to rural electrification that ensures a pos-

itive impact on communities, targeting the direct impact on livelihoods and

revenue generation beyond the provision of connections and kilowatt hours.

Rather than waiting for spontaneous positive effects of electrification projects

to trickle-down in rural areas, the alternative proposed by [De Gouvello and

Durix, 2008] focuses on productive uses of electricity. They specifically identi-

fied the necessity of working across sectors and the importance of a political

will for implementation. By reaching out to other sectors to help generate their

output, the authors call from political will to implementation.

The systematic approach (Figure 3.3, left) defines the necessary steps prior

to cross-sectors energy program, such as the financing of modern technologies

through microcredit, based on the impact of the selected energy technologies

on the community. In particular, it consists in the identification of the produc-

tive activities and of the energy usage of the target population, followed by a

categorization of productive activities carried out in a specific rural area. Next,

an analysis of the contribution of electricity to the productivity of the consid-

ered economic activities shall be conducted, followed by an identification of the

required technologies or equipment and hence an assessment of basic economic

viability.

The pragmatic approach (Figure 3.3, right) entails primarily the identification

and promotion of productive uses of electricity. It consists of an analysis of rural

development priorities, the identification of sector programs and areas of cooper-

ation, and the analysis and implementation of the technical design and costing.

The main aim of this approach is to establish an effective stakeholder coopera-

tion within renewable energies concessions. Within this approach, [De Gouvello

and Durix, 2008] introduce the multi-sector energy investment projects (MECs),

as an opportunistic and practical method to speed up the delivery of positive

impact of renewable energy.

Figure 3.3.: The systematic (left) and the pragmatic approach (right) for cross-

sectoral cooperation according to [De Gouvello and Durix, 2008]

The PEPI Toolkit for MFIs 3. Green Microfinance

3.4. Indicators on Green Microfinance

Stakeholders involved in implementing green microfinance have been interested

in the development of analytical tools for MFIs aimed at fostering the inte-

gration of environmental objectives within their strategic plans. Consequently,

categories and terminology have been harmonized and green practitioners have

coordinated their effort to agree on metrics and standards definitions of the

concept of “green”. Specifically, several indexes in the area of environment and

microfinance have fostered the potential and ability of MFIs to commit to envi-

ronmental goals. This section briefly describes the Microfinance Environmental

Performance Index, proposed by [Allet, 2011], the Green Performance Agenda,

developed by the companies Enclude and Hivos, the MIX set of environmen-

tal indicators [Pierantozzi et al., 2015], and the Green Index, developed by the

European Microfinance Platform (e-MFP) Action Group Microfinance & Envi-

ronment [Allet, 2014]. The main characteristics of these methodologies are also

summarized in Table 3.1.

Microfinance Environmental Performance Index (MFEPI) The MFEPI4aims

at measuring the environmental performance of MFIs, hence assessing the en-

gagement of MFIs in preserving the environment. The MFEPI has been a

pioneer tool in investigating the impact of microfinance on the triple bottom

line, i.e., tackling social, economic and environmental goals. Given the increas-

ing interest of MFIs in implementing environmental management programs,

the tool developed by [Allet, 2011] aimed at facilitating the microfinance sec-

tor to measure environmental performance, focusing in actions and processes

rather than in outputs. The tool is based on management performance indi-

cators, assessing the following five dimensions: environmental policy, ecological

footprint, environmental risk assessment, green microcredit, and environmental

non-financial services. The indicators within each axis are depicted in Table

3.2.

4In the original publication [Allet, 2011], the Microfinance Environmental Performance Index

has been shortened using the acronym MEPI. However, for the sake of clarity, in this section

this index is abbreviated as MFEPI, in order to distinguish it from the MEPI introduced in

Section 2.5.

The PEPI Toolkit for MFIs 3. Green Microfinance

Index Outputs Axis of actions

MFEPI

Assessment of MFI’s

environmental

performance

I Environmental policy

II Ecological footprint

III Environmental risk assessment;

IV Green microcredit

V Environmental non-financial

services

GPA

Self-assessment of MFI

today’s and tomorrow’s

performance

List of activities and tools

MIX

Env.

Indi-

cators

Binary metrics (MFI

performs or not specific

environmental activities)

I Environmental policy

II Environmental risk assessment

III Environmental risk outstanding

loans assessment

IV Green loans

V Environmental micro-insurance

Green

Index

Assessment of MFI’s

environmental

performance

I Environmental strategy

II Environmental risk manage-

ment

III Green opportunities

Table 3.1.: Green Microfinance Indicators

The PEPI Toolkit for MFIs 3. Green Microfinance

Environmental Policy

Mission Environmental protection mentioned in the official vision, mis-

sion, or values (1 pt.)

Environmental Pol-

icy Formal policy on environmental responsibility (1 pt.)

Manager A person appointed to manage environmental issues (1 pt.)

Incentives Incentive system to encourage employees to take into account

specific environmental objectives (1 pt.)

Ecological Footprint

Carbon Audit Previous realization of a carbon audit (1 pt.)

Footprint Objec-

tives

Specific objectives to reduce ecological footprint (e.g., reduction

in energy consumption, carbon emissions, waste, etc.) (1 pt.)

Staff Awareness

Toolkits to raise employees’ awareness of good practices in

paper, water, and energy consumption, transportation, waste

management, etc. (1 pt.)

Reporting Inclusion of environmental performance indicators in annual

report (paper, water, and energy consumption, etc.) (1 pt.)

Environmental Risks Assessment

Exclusion List Use of an environmental exclusion list (1 pt.)

Screening Tools Use of specific toolkits to evaluate the environmental risks of

clients’ activities (1 pt.)

Staff Training Training module to teach loan officers how to evaluate the en-

vironmental risks of their clients’ activities (1 pt.)

Monitoring Infor-

mation System

(MIS)

Inclusion of indicators into MIS to track the environmental per-

formance of clients (1 pt.)

Green Microcredit

RE&EE Loans Provision of credits to promote access to renewable energy or

energy efficient technologies (RE&EE) (2 pts.)

Green EEN IGAs

Loans

Provision of loans with reduced interest rates to promote the

development of environmentally-friendly activities (2 pts.)

Environmental Non-Financial Services

Client Chart Environmental chart to be signed by clients (1 pt.)

Client Awareness Programs to raise clients’ awareness on environmental risks (1

pt.)

Promotion Action Organization of actions to promote environmentally-friendly

microenterprises (1 pt.)

Client Training Training and other services to support clients who want to de-

velop environmentally-friendly activities (1 pt.)

Table 3.2.: Microfinance Environmental Performance Index (MFEPI) [Allet,

2011]. The score of each evaluation axis (right column) is indicated

in brackets.

The PEPI Toolkit for MFIs 3. Green Microfinance

Green Performance Agenda (GPA) With the purpose of supporting MFIs

with a practical guide that addresses their environmental agenda, Enclude and

Hivos designed the GPA, consisting of a self-assessment tool which appraises

the MFI’s perspective of their environmental performance ‘today’ and ‘tomor-

row’5. The application of the GPA provides a list of tools to plan and improve

their performance suggesting the desirable impact of conducting such activities

without quantifying in detail the effects.

MIX Indicators The largest available set of green indicators (and, more broadly,

of social performance indicators) are the one proposed within the microfinance

platform MIX Market6[Pierantozzi et al., 2015], where data have been collected

since 2009. The qualitative indicators on environmental performance, available

to the public, are solely based on binary metrics, reflecting whether the MFI

does or does not conduct a particular ”green” activity. Within the set of the

MIX indicators linked to environmentally friendly products and/or practices

(see Table 3.3), only two describe whether an institution offers green loans en-

tailing the financing of clean energy technologies e.g., (i) Products related to

renewable energy (e.g., solar panels, biogas digesters) and (ii) Products related

to energy efficiency (e.g., insulation, improved cooking stoves, etc.).

(1) The institution includes clauses in loan contracts that require clients

to improve environmental practices/mitigate environmental risks.

(2) The institution offers specific loans linked to environmentally

friendly products and/or practices. The environmentally friendly credit

product offering are environmentally friendly practices or products re-

lated to environmentally friendly practices (e.g., organic farming, re-

cycling, waste management, agroforestry or silvopasture, clean water,

etc.).

(3) The institution offers specific loans linked to environmentally

friendly products and/or practices:

- Products related to renewable energy (e.g., solar panels, biogas di-

gesters, etc,)

- Products related to energy efficiency (e.g., insulation, improved cook

stoves, etc.)

- None of the above

Table 3.3.: MIX Environmental Indicators: Environmental Policies and

Initiatives.

An analysis of these indicators [Shuite and Forcella, 2015,Pierantozzi et al., 2015]

showed that the number of MFIs offering products related to renewable energy

and to energy efficiency is increasing (see Figure 3.4). However, a drawback of

this classification is that environmentally friendly credit products not strictly

included in the renewable energy or energy efficiency categories or practices are

5See www.gpa4mf.blogspot.com

6See http://reports.mixmarket.org/crossmarket

The PEPI Toolkit for MFIs 3. Green Microfinance

classified as “none of the above”.

Figure 3.4.: Number of MFIs offering financing to products related to renewable

energy (RE) and to energy efficiency (EE) according to yearly MIX

reports in 2013 (N:1335) and 2014 (N:1466)

Green Index Presented to the microfinance sector by the Microfinance and En-

vironment Action Group from the European Microfinance Platform (e-MFP),

the Green Index provides a full picture of the environmental engagement of an

MFI, depending on (i) its formal environmental strategy, (ii) its environmental

risk management, and (iii) the offer of green opportunities, including in the

latter, among others, the provision of financial services to purchase environ-

mentally friendly technologies (see Table 3.4).

The Green Index (Table 3.4) includes the 6 dimensions of the Universal Stan-

dards and offers it as an optional module7. Without aiming at measuring

impact, the index intends to measure processes providing an overview of the

means of the MFI to reach its environmental objectives [Allet, 2014].

The research of [Pierantozzi et al., 2015] provides a picture of available qualita-

tive and quantitative indicators. It also provides an analysis of the easiness and

usefulness of tracking environmental indicators, describing the perceptions of

MFIs involved in green activities. Results show that MFIs consider tracking of

green loans easier and more useful method than tracking other green activities.

However, authors recommend to differentiate between the types of green loan

disbursed. Categories include renewable energy, energy efficiency, agroforestry,

waste management, clean water, sanitation and ecotourism. Yet, indicators

do not differentiate between financed technologies or basic need (e.g., energy,

water, sanitation, etc.).

Since 2014, the Green Index has been included in the Social Performance Indica-

tors (SPI) of CERISE8, a French association focused on disseminating knowl-

7As of November 2015, 31 MFIs completed the Green Index. 24 institutions within accompa-

nied self-assessments, while 7 of them as a self-assessments [e-MFP, 2015].

8See www.cerise-microfinance.org

The PEPI Toolkit for MFIs 3. Green Microfinance

Numeration Green Indicator

Green a The institution addresses environmental issues through a

formalized strategy.

Green a11 The institution defines its environmental strategy.

Green a12 The institution implements its environmental strategy.

Green b The institution manages its environmental risks.

Green b11 The institution implements actions to reduce its internal ecological

footprint.

Green b12 The institution monitors its internal environmental risks.

Green b21 The institution evaluates the level of environmental risk of its

clients.

Green b22 The institution includes the level of environmental risk as a factor

in the loan approval.

Green b23 The institution monitors the external environmental risks.

Green b24 The institution raises clients’ awareness on environmental risks.

Green c The institution fosters green opportunities.

Green c11 The institution provides specific green loan products.

Green c12 The institution provides other green financial products.

Green c13 The institution provides green non-financial services.

Table 3.4.: Green Index - Set of Indicators (summarized)

edge and tools for ethical finance, and in the Social Rating of the company

MicroFinanza Rating9, demonstrating further efforts of stakeholders to system-

atically assess the environmental performance of MFIs.

In particular, CERISE, founded by five French organizations specialized in mi-

crofinance, involves MFIs, networks, technical assistance providers, investors,

and donors worldwide. Its working areas include impact finance and social

performance, governance, rural and agricultural finance. So far, by November

2015, the results have showed CERISE that many MFIs do not see the rele-

vance of the index and that practices are lagging. While beginners recognized

the tool has contributed to their understanding of the concept of green, more

experienced MFIs (in any of the green MFI axis) showed to be more structured

and more mature in managing their environmental performance.

On the other hand, MicroFinanza Rating integrated the Green Index early 2015

and it has back-tested the environmental performance results according to the

Green Index structure. Results of the implementation confirmed the limited

number of financial institutions involved in green activities, and that only a

share of the double bottom line financial institutions have a triple bottom line

impact [e-MFP, 2015]. Specifically, in regards to green lending, MicroFinanza

Rating found that the demand among the target population and the increasing

availability of technical providers in a country constitute a high potential for

the financial institution to offer green financial products. Further findings from

those MFIs which completed the Green Index, demonstrated that MFIs are

9See http://www.microfinanzarating.com

The PEPI Toolkit for MFIs 3. Green Microfinance

considering the feasibility of introducing and scaling-up loan products for eco-

housing, SHS and water purification [e-MFP, 2015].

3.5. Microfinance Metrics to Assess Poverty: The

Progress out of Poverty Index (PPI)

For the institutions serving the BoP, the capability of measuring the level of

poverty of their clients is of fundamental relevance, in order to quantify the

effectivity of their programs or projects in targeting the poor.

To this aim, the Progress out of Poverty Index (PPI) [Schreiner, 2004] is a

worldwide used statistical based tool which assesses the poverty level. In par-

ticular, the PPI consists of 10 questions with specific scoring systems, which

can be adjusted to each country. As a result, the PPI score is associated with

the likelihood of a household being below the poverty line (based on a daily in-

come threshold which depends on the country). In order to provide a complete

benchmark for reporting the contextual reality of the households, especially in

relation to overall poverty levels in the country, the PPI manages to portray

the most relevant indicators directly associated with poverty levels, instead of

using monetary indicators (total income, net income from crops, total house-

hold assets, etc.). Using the net income as a proxy for poverty may provide

an incomplete, although valuable, understanding of the socioeconomic status

of the beneficiaries [COSA, 2015]. The hypothesis of the tool methodology is

that other non-crop factors (or monetary income) affect their poverty, there-

fore, a broader tool is particularly useful. Measuring the PPI over time allows

to efficiently track the households progress out of poverty along the organiza-

tions’ activities in a geographical area. A test conducted by the Committee on

Sustainable Assessment (COSA) debates the accuracy of the PPI. Specifically,

since the tool is created using national consumption data, the likelihood that

the tool efficiently describes the level of poverty of different populations within

a country may fail. For instance, considering populations that are by their

nature agricultural producers may not be, in general, nationally representative.

Within the microfinance sector, the Grameen Foundation (GF) has promoted

the use of the PPI for more than a decade. As an example, the PPI scorecard

developed for Colombia (and currently used by the MFI Contactar, see Chapter

4) is shown in Figure 3.5. The survey is based on the 2009 Integrated Household

Survey (IHS) conducted by DANE (the Colombian National Administrative De-

partment of Statistics, for its acronym in Spanish), updated in December 2012,

accommodating the new poverty lines created by the Colombian government,

which are still currently used [Grameen-Foundation, 2012].

The PEPI Toolkit for MFIs 3. Green Microfinance



This PPI was updated in November 2012 based on data from 2009. For more information about the PPI, please visit

www.progressoutofpoverty.org



PPI® Scorecard for Colombia

Entity Name ID Date (DD/MM/YY)

Member:

Joined:

Field agent:

Today:

Service point:

Household size:

Indicator Value Points Score

1. How many household members are

18-years-old or younger?

A. Four or more 0

B. Three 5

C. Two 11

D. One 17

E. None 23

2. What is the highest educational

level reached by the female

head/spouse?

A. None, or pre-school 0

B. Primary or middle school 3

C. High school 6

D. No female head/spouse 8

E. Post-secondary or college (1 to 4 years) 9

F. Post-secondary or college (5 years or more) 17

3. How many household members spent most of the past week working? A. None 0

B. One 9

C. Two or more 14

4. In their main line of work, how many household members work as

wage or salary employees for a private firm or the government?

A. None 0

B. One 4

C. Two or more 11

5. What is the residence’s

rate class for

electricity?

A. No class or zero (no connection, pirated connection, or

generator), one, or two 0

B. Three 4

C. Four, five, or six 9

6. What fuel or energy

source does the

household usually

cook with?

A. Firewood, wood, charcoal, coal, electricity, gasoline,

petroleum, kerosene, alcohol, or waste material 0

B. LPG from a cylinder or tank 2

C. Natural gas from a public network 3

D. Does not cook 6

7. Does the household have a working clothes washing machine? A. No 0

B. Yes 4

8. Does the household have a working refrigerator or freezer? A. No 0

B. Yes 3

9. Does the household have a working DVD? A. No 0

B. Yes 4

10. Does the household have a motorcycle and/or a

car for its own use?

A. None 0

B. Motorcycle only 3

C. Car (regardless of motorcycle) 9

By Mark Schreiner of Microfinance Rish Management L.L.C., developer of the PPI. Score:

Figure 3.5.: PPI Scorecard for Colombia, originally developed in November

2012 by Mark Schreiner of Microfinance Risk Management, L.L.C..

Source: http://www.progressoutofpoverty.org/country/colombia.

4.The MFI Contactar

The field research presented in this thesis (see Chapter 5) has been carried out in

collaboration with the Corporaci´on Empresa Nari˜no Contactar1, a Colombian

medium-sized MFI covering four departments in south-western Colombia (in

order of outreach size: Nari˜no, Huila, Putumayo and Tolima, see Figure 4.1 for

details), established in 1995 as an NGO.

Figure 4.1.: Regions of Colombia covered by Contactar (original map down-

loaded from: www.your-vector-maps.com)

By offering a variety of products adapted to the needs to their clients, Contactar

addresses the triple bottom line. In particular, the institution offers inclusive

financing, based on individual and group credits for basic financial needs, as well

as customized financial products for agriculture and micro-insurances, targeting

1www.contactar-pasto.org

The PEPI Toolkit for MFIs 4. The MFI Contactar

primarily the rural population dedicated to agriculture and cattle. Moreover,

as an institution with a vast track record in diversifying its portfolio, Contactar

also offers specialized non-financial products, seeking to develop the private

business initiatives and entrepreneurial skills of their clients. In particular, since

2011 the MFI has acquired valuable experience in green lending, developing

financial products for the acquisition of clean energy technologies.

In 2014, the German consulting company MicroEnergy International GmbH

(MEI)2was hired by the Citi Foundation to support Contactar in developing

its green strategy. This assignment established a strong partnership between

Contactar and MEI. In this context, given the interest of the MFI in identi-

fying the energy needs of the clients, as well as in establishing a systematic

methodology for the energy access assessment, the proposal to conduct field

research parallel to the consulting services provided by MEI was very welcomed

within the institution. Moreover, Contactar has a department dedicated to all

activities related to the environment, called the the Social and Environmental

Performance Management Department (GDSA)3. The goal of the GDSA is to

transform Contactar’s social and environmental mission and vision into real-

ity4. This department, also in charge of all research activities, has a strong

interest in the systematic collection of client data, thus the proposed research

aligned with their current strategy. At the same time, the fact that the MFI

serves clients living in very diverse climatic and economic conditions, and the

possibility of accessing the demographic information from the MFI database,

made Contactar an ideal partner for implementing and analyzing the detailed

multi-tier assessment survey [SE4ALL and WB, 2014].

The objective of this chapter is to introduce the partner MFI and the local con-

text in which the field research was framed. In particular, Section 4.1 presents

the regional Colombian context, in terms of the microfinance market, energy

access and green microfinance, while in Section 4.2 the profile of Contactar will

be briefly described, portraying its clients and its regions of work. Finally, Sec-

tion 4.3 is dedicated to the green lending program implemented at Contactar,

describing the energy technologies offered and the methodology used by the

institution to select them.

4.1. Geopolitical and Economic Contexts

Colombia is a country rich in natural resources, with a high biodiversity and

an ample ecosystem in need of preservation. In this context, clean energy

and sustainable development play an important role in keeping up the pace of

economic growth. With 1’446,000 inhabitants without access to electricity and

modern fuels (in 2015) [The World Bank, 2016], there is a huge potential for the

2www.microenergy-international.com

3For its acronym in Spanish: Gesti´on de Desempe˜no Social y Ambiental

4By raising financial and environmental awareness among clients and internal staff, the GDSA

team provides the MFI with a competitive advantage and enhances the institution’s commit-

ment to improve the living standards of its clientele. The GDSA is also in charge of managing

communication with suppliers and customer satisfaction follow up.

The PEPI Toolkit for MFIs 4. The MFI Contactar

dissemination of clean energy technologies and of related financial products.

Concerning the macro picture of the microfinance sector, Colombia is the third

largest country in Latin America per number of microfinance clients [Serrano,

2009]. With more than 2,8 million clients and 43 MFIs (reporting to the MIX

Market platform in 2016), the sector has been steadily increasing in the last

decade with the support of the governmental program Banca de las Oportu-

nidades5. Since 2011, two national financial networks (Asomicrofinanzas and

Asobancaria) have contributed to the collection, exchange and management of

information within the sector at a national level. This initiative led to the in-

troduction of the topic “green microfinance” in the country. In particular, the

awareness-raising campaign undertaken by these organizations has encouraged

MFIs to commit to environmental protection by integrating an environmental

axis within their market strategies.

Green Microfinance As an emerging market with, as of recently, very little

(anecdotical) practice6, the green microfinance sector has a strong potential for

growth and replication. Particularly, 2012 marked the signing of the Green

Protocol [CNG and Asobancaria, 2012], an official commitment from Colombian

financial institutions in conjunction with the Ministry of Environment and Sus-

tainable Development to emphasize the importance of green microfinance prod-

ucts and to incorporate them into their portfolio7.

However, the expansion of the green microfinance market still faces challenges

which can only be tackled over a longer time, by building experience, raising

awareness among the key stakeholders and sharing the positive results of imple-

menting innovative strategies. As observed by [Guti´errez and Reddy, 2015], the

most important barriers towards access to finance for rural populations include:

•the limited participation of private financial institutions lending in rural

areas;

•the existence of public support programs for rural agricultural credit with

perverse incentives;

•the limited range of assets used as collateral;

•the limited credit history and financial education of the rural population;

and

•the lack of agricultural insurance to support risk management and facili-

tate access to credit.

A further aspect, in the specific context of Colombia, is that most of the small

and medium enterprises (SMEs) are rural and their profiles are considered to

be risky, making the access to microcredits through the usual financial system

more difficult. Moreover, the legal framework for green microfinance in the

5See www.microfinancegateway.org/es/pais/colombia and www.bancadelasoportunidades.gov.co

6According to Climatescope 2012, only 3 financial institutions were offering green products:

the MFIs Activos y Finanzas SA,Fundesmag (i.e., 2 out of 29 MFIs) and the national

development bank Banc´oldex. Clean energy credit providers included: Bancolombia, BBVA

Colombia and Banco de Bogot´a [MIF and Bloomberg, 2012]

7See the Green Protocol (official document) [CNG and Asobancaria, 2012]

The PEPI Toolkit for MFIs 4. The MFI Contactar

country is still in its incipient phase, leading to a lack of clarity on legal aspects

of terms and conditions. Lastly, due to the high variability of geographical and

socio-economic characteristics of the different Colombian regions, there is a lack

of standardization on product development, which increases the complexity of

scaling up pioneer green microfinance initiatives.

Energy Access Colombia ranks seventh in the national electrification rate

within Latin America, with 97% coverage [IEA, 2015] (99,7% urban and 87,9%

rural electrification rate in 2012), slightly above the average of 95% of the

neighboring countries8. However, the 2% of population uncovered by the grid,

still represents about 1.2 million people, the seventh largest population in Latin

America. In particular, according to the latest national census from 2005,

Contactar focus regions (e.g., Nari˜no, Huila, Putumayo and Tolima) suffer from

limited access to infrastructure as shown in Table 4.1 and Figure 4.2.

Households Off-grid

Off-grid, with-

out water sup-

ply nor sanita-

tion

Without natural

gas

Huila 232,621 18,547 (8%) 11,629 (5%) 129,815 (56%)

Nari˜no 321,112 43,383 (14%) 37,880 (11.8%) 321,112 (100%)

Tolima 328,280 24,632 (8%) 17,849 (5%) 209,934 (64%)

Put. 61,032 20,346 (33%) 18,972 (31%) 60,531 (99%)

Total 943,045 106,908 (11%) 86,330 (9%) 721,392 (76%)

Colombia 9’742,928 623,141 (6%) 510,794 (5%) 5’797,241 (59%)

Table 4.1.: Access to infrastructure in the working regions of Contactar. Re-

gions with the lowest access performance (in percentage) are marked

in red

Hence, regardless of the positive trend of coverage shown in national statis-

tics [IEA, 2015], Colombia still faces several challenges in the energy sector.

Specifically, the Council of the Americas Energy Action Group [Viscidi, 2010]

listed:

•the erratic pace of infrastructure development in comparison to invest-

ments in renewables energy and the strong dependence on oil and gas

production;

•the expansion of drilling into new areas with environmental and social

consequences;

•the lack of security, which puts at risk economic, political and social

stability.

8Electricity access in 2013 - regional aggregates in Latin America: 22 million people without

access to electricity; with 98% urban and 84% rural electrification rates

The PEPI Toolkit for MFIs 4. The MFI Contactar

Nar Hui Tol Put COL

Households (%)

100 Offgrid

Offgrid, w/o water

W/o natural gas

Figure 4.2.: Infrastructure access in working regions of Contactar

4.2. Portrait of the Microfinance Institution

Contactar was founded in 1995 in Pasto (Department of Nari˜no, Colombia).

Registered as a non-governmental organization (NGO), Contactar is currently

the 10th largest MFI in Colombia (out of 43 MFIs currently reporting to the

MIX Market platform)9with 81,068 clients (67,498 rural and 13,570 urban).

Contactar offices cover four Colombian departments in the southwest region of

the country: in order of outreach, Nari˜no, Huila, Putumayo and Tolima (see

also Figure 4.1 and Appendix C for more details). Table 4.2 and Figure 4.3

depict the distribution of clients in the different departments, together with the

total department population. The table 4.2 includes also, for each department,

the total population and the year when Contactar opened the first office on the

regional territory.

Clients Active

since Population

Nari˜no 41,022

(50.6%) 1995 1’744,000

Huila 18,428

(25.4%) 2009 1’174,000

Tolima 8,931

(11.0%) 2015 1’410,000

Put. 10,504

(13.0%) 2010 345,000

Table 4.2.: Distributions of Contactar clients across the served regions

9Source: www.mix-market.org (data recovered on July 30th, 2016).

The PEPI Toolkit for MFIs 4. The MFI Contactar

Nar Hui Tol Put

Clients

5000

15000

25000

35000

Figure 4.3.: Distributions of Contactar clients across the served regions

In order to provide a more detailed picture of the MFI, Figure 4.4 describes

the statistics of the clientele. Most of the customers belong to the age group

30 to 50 years (60%), with only primary education (60%) and only about 5%

have completed a university degree. Moreover, the vast majority (about 95%)

declare themselves to be self-employed, i.e., without a formal work contract.

As shown in Figure 4.5, the microcredits offered by Contactar are mostly indi-

vidual (95%, compared to 5% for group loans), and more than 60% of loans are

disbursed for business purposes, while about 20% are related to personal use or

dedicated to improve housing conditions.

Age

<30 30-40 41-50 51-60 >61

Clients (%)

Education level

None Prim Secn Tech Univ

Clients (%)

Type of occupation

Independent Employee

Clients (%)

Figure 4.4.: Brief portrait of Contactar clients according to age (left), educa-

tion(center) and type of occupation (right)

In order to evaluate the impact of its programs, Contactar utilizes the Uni-

versal Standards of Social Performance10 into practice, monitoring the level of

poverty of its clients over time through the Progress out of Poverty Index (PPI)

developed by [Schreiner, 2004] (see Section 3.5).

The international organization Grameen Foundation (GF)11, with the support

10The Universal Standards of Social Performance consist in a comprehensive manual of

best practices created for microfinance stakeholders as a resource to help financial ser-

vice providers achieve their social goals. For more information: http://sptf.info/universal-

standards-for-spm/start-here

11The Grameen Foundation, pioneer developer of the PPI, seeks to continuously enhance its

tool by providing support to the organizations applying the tool and disseminating the best

practices for its implementation.

The PEPI Toolkit for MFIs 4. The MFI Contactar

Type of credit

Individual Groupal

Credit(%)

Purpose

Cons. Educ. Pers. Busin House

Credist (%)

70 All Credits

Individual

Groupal

Figure 4.5.: Brief portrait of Contactar credits. Left: Type of Credit. Right:

Purposes (Consume, Education, Personal Use, Business, Housing)

of the Cisco Foundation, introduced and implemented the PPI at Contactar. To

this end, in 2012 Contactar loan officers and employees were trained on the PPI

and on the interpretation of the results for strategic and operational decision-

making processes within the organization. Since then, the MFI has consistently

applied the tool to its clients. A portrait of the PPI scores of Contactar clients

is provided in Figure 4.6.

PPI Score (2014)

0 20 40 60 80 100

Number of clients

3000

6000

9000

12000

15000

Figure 4.6.: Histogram of the PPI results among Contactar clients (data col-

lected in 2014)

According to the analysis conducted by the Cisco Foundation in 2014 [Cisco

Foundation, 2014], Contactar has a rather diverse spectrum of clients. The re-

sults showed that about 10% of the clients live in extreme poverty (i.e., PPI

score less than 30), while 58% are vulnerable to falling below the poverty line

(PPI score below 45) or are living in impoverished conditions. On the con-

trary, only 32% of the clients have a reduced likelihood of falling into poverty12.

Particularly, the Cisco Foundation suggested that the wide range of categories

might require a correct combination of social and economic strategies to effec-

tively target the clients13.

12Further details on the interpretation of PPI scores are provided in Appendix B.

13Note that Contactar does not register the PPI information nor due diligence data of rejected

The PEPI Toolkit for MFIs 4. The MFI Contactar

4.3. The Diversification of Contactar in Green

Microfinance: ConSuPlaneta

Since its establishment, Contactar has developed a range of products and train-

ing programs according to the needs of its clientele, innovated approaches and

services, and prioritized its environmental impact, in addition to its financial

and social objectives. Some examples of the financial products developed for

tackling specific basic needs include ConSuVivienda (dedicated to housing),

ConSuEducaci´on (education), ConSuTransporte (transport). In 2011, Contac-

tar identified cooking and drying crops as a major energy related issue for its

rural clients, developing a Green Finance Program with its own means and

initiative.

The portfolio diversification of Contactar started the same year, with a launch-

ing of a small pilot project that fosters access to clean energy technologies in the

department of Nari˜no. Through a strategic association with a local energy ser-

vice supplier, the pilot began with a donation of fixed improved cooking stoves

(ICS) for 35 rural families. In 2012, combining microcredits with donations for

ICS and biogas digesters (BD), the pilot project reached 81 rural families. One

year later, through partnerships with small local suppliers in the region, and

following a two-hand model approach, Contactar created its first green financial

product ConSuPlaneta (see also Section 4.3.1) and started offering financing of

locally distributed clean energy technologies, formally launching its Green Fi-

nance Program. In 2013, the program was based solely on credit. By then,

the financed technologies included, besides ICS and BD, also solar crop dryers

(SCD), water tanks (WT) and water filters (WF). A more detailed description

of these MES and of the selection criteria is provided in Section 4.3. As of

April 2015, the MFI has disbursed more than 800 green loans of the diverse set

of technologies, targeting mostly the rural population, and satisfying specific

energy needs for cooking, drying crops and water supply.

Currently (July 2016), the average loan period for Contactar’s financial prod-

ucts is 17 months and the average monthly interest rate is 3.2051% (max.

3.5042% min. 2.0831%). For the ConSuPlaneta credits, a commission fee is

charged in order to cover marketing costs, and the collateral was established at

its minimum. The credit conditions of the financial product do not differentiate

among technologies (see Table 4.3) except for the water filters, which are sold

directly over the counter given the reduced price.

Since the launch of the green microcredit ConSuPlaneta, Contactar has demon-

strated wide management capacities in designing customized microcredits to

finance MES [Matas et al., 2015]. Specifically, the MFI has explored the differ-

ent market segments in association with potential energy partners, in order to

determine the best energy solutions for its customers, an essential requirement

for the success of lending programs designed for financing MES [Leva¨ı et al.,

loan applications. Furthermore, there is no consecutive data collection or systematized

procedures for one-time clients, i.e., those who did not apply for a consecutive loan. Hence,

an analysis of causalities or relationships between poverty measures and loan effects is not

yet possible.

The PEPI Toolkit for MFIs 4. The MFI Contactar

Water Solar Crop Improved Cooking Biogas Water

Tank Dryer Stove Digester Filter

Loan size (USD) 750 675 415 325 35

Period (Max.) 12 12 12 12 Cash

Interest rate 27% 27% 27% 27% –

Commission Fee 4.5% 4.5% 4.5% 4.5% –

Collateral 0.7% 0.7% 0.7% 0.7% –

Table 4.3.: Portfolio of energy technologies offered by Contactar through the

program ConSuPlaneta

2011]. By scaling up its Green Finance Program, Contactar seeks to reach the

economically challenged population of Nari˜no, Huila, and Putumayo14, as well

as at creating the socio-economic and environmental impact needed to attract

other Colombian financial institutions towards developing their green strate-

gies [Matas et al., 2015].

4.3.1. A Two-Hand Model Approach

In the case of Contactar, the delivery methodologies, i.e., the set of systems

and procedures that an institution develops in order to deliver its services [Wa-

terfield, 2001], were based on a two-hand model. Namely, the MFI collaborated

with MES suppliers following a bottom-up approach, whose steps are described

in Figure 4.7.

In the case of Contactar, the delivery methodologies, i.e., the set of systems

and procedures that an institution develops in order to deliver its services [Wa-

terfield, 2001], were based on a two-hand model. Namely, the MFI collaborated

with MES suppliers following a bottom-up approach, whose steps are described

in Figure 4.7.

According to the bottom-up approach [Realpe Carrillo et al., 2015], the finan-

cial product ConSuPlaneta targets clients with specific energy needs: (i) de-

crease consumption of firewood for cooking by replacing firewood with gas or

improving stove efficiency; (ii) improve the quality of agricultural crop dry-

ing processes; and (iii) improve the quality of the water supply. As observed

in [Waterfield, 2001] the proper design of financial products and their delivery

methodologies is fundamental in order to ensure their efficiency and sustain-

ability. Hence, [Waterfield, 2001] advocates that credit institutions need to

incorporate financial products and delivery methodologies which are both ap-

propriate and sustainable, i.e., appropriate for the needs of the target group and

sustainable in order to guarantee the continuity and long-term sustainability of

the provided services.

14The green loan is not yet offered in Tolima, as the MFI expanded the supply of financial

services to this department only in late 2015

The PEPI Toolkit for MFIs 4. The MFI Contactar

Figure 4.7.: Bottom-up approach of a two-hand model according to initiator

[Realpe Carrillo et al., 2015]

The proper systematic approach described by [Realpe Carrillo et al., 2015] sug-

gests the implementation of this scheme in order to guarantee a balanced part-

nership and a controlled supply chain design. In fact, in a two-hand model,

both parties commit to work jointly and to select the technology based on the

energy needs of the MFI clients and on their institutional profiles.

4.3.2. Selected Microenergy Systems

The clean energy technologies to be offered within ConSuPlaneta corresponded

to those with the highest potential of increasing the income and/or the saving

generation of domestic and economic activities [MicroEnergy International, 2014].

These MES were selected following a systematic approach [De Gouvello and

Durix, 2008] (see Section 4.3.3). Particularly, in order to select the most suitable

technologies, the main opportunities have been identified in the increase of labor

productivity, in the improvement of product yield and quality (especially for

coffee dryers), as well as in the development of new economic activities for the

rural population (current and potential clients) [Matas et al., 2015]. Within

the green strategy developed by MEI for Contactar, a market analysis was

conducted in order to assess the market viability of the existing and potential

MES to integrate in ConSuPlaneta portfolio. This study consisted in desk

research, a client database analysis, and a survey carried out among the loan

officers.

The technologies offered by Contactar since 2012 are briefly described below

(see also Figures 4.8 and 4.9).

The PEPI Toolkit for MFIs 4. The MFI Contactar

Figure 4.8.: Left: Solar Crop Dryer (for coffee beans). Right: Biogas Digester

(for farmers with at least 10 pigs)

Figure 4.9.: Left:Water Tank. Right: Residential Water Filter

The PEPI Toolkit for MFIs 4. The MFI Contactar

Solar Crop Dryer (SCD) A SCD consists in a structure made out of guadua

also known as American Bamboo, an organic material available in the area

that is commonly used for housing. The system has a size of 5 meters (length)

by 4 meters (width). The exterior of the product is made of plastic and the

installation lasts normally 3 working days. [SantaMar´ıa, 2011] has validated the

technology in Peru in a region geographically similar to Nari˜no. The comparison

between the traditional drying methods and the ones used with the Solar Crop

Dryer underlines the advantage of the increase in productivity (working-time

spent in drying) and in quality (see also Table 4.16).

Biogas Digester (BD) A BD is a system measuring 1 meter (width) by 5

to 10 meters (length). The system converts animal residues into a mixture of

gases. The biogas is composed of 60% methane gas (CH4) and 40% carbon

dioxide (CO2), and this enables the mixture to be used for cooking15, among

other uses.

Thus, the BD offers cooking with methane gas and bio-fertilizer. The require-

ments for a functional biogas digester consist of daily animal waste and water

(1 kg dung requires 3 water buckets). The installation lasts about 3 days.

Improved Cooking Stove (ICS) The fixed ICS consists of a grill for cooking, a

boiler, an oven, a chimney and the respective gates for the firewood, ventilation

and disposal of ashes. Its dimensions are 105 cm (length) by 73 cm (width);

its installation lasts 4 hours. For the mobile ICS, measuring 110 cm by 85 cm

by 68 cm (height), the installation lasts about a day. Both ICS models reduce

firewood consumption, enable 4 spaces for cooking, an oven for baking and

channel smoke out of the kitchen through the chimney.

Water Tank Seeking to cope with the adversities of climate change, the water

tank enables rural families to reserve water for dry seasons. The water tank has

a height of 2,20 m and a diameter of 3,30 m. Water is preserved from rainwater,

close brooks or community tanks. The water tank capacity is conceived for

an estimated consumption of 180 liters per day (5,400 liters per month) per

household. In order to feed the water tank, the household roof must cover an

area of at least 7 square meters.

Water Filters The Eco-Filtro functions as water-maker for households. It

can filter up to 2 liters per hour and has a storage capacity of 35 liters. Its

dimensions are: 39 cm diameter and 59 cm height and it weighs 8 kg without

water. The water filters are sold for cash and they are relatively affordable

compared to the average loan amount of 2’172,277 COP16.

15Gas combustion is possible only if the concentration of methane gas is above 50%.

16Approximately 650 e, based on the exchange rate of October 15th, 2016.

The PEPI Toolkit for MFIs 4. The MFI Contactar

4.3.3. The Systematic Approach

The implementation of the Green Finance Program and of the green microcredit

ConSuPlaneta resulted from a set of steps, which can be framed under the the

so-called systematic approach introduced by [De Gouvello and Durix, 2008]:

1 identification and categorization of productive activities of the targeted

rural areas;

2 identification of areas of improvement in productive processes;

3 contribution of energy access through identified equipment;

4 economical assessment of new production process;

5 promotional campaigns for consumer targets.

Besides the application of the systematic approach, the following factors were

also decisive in designing the business model of the Green Finance Program

within the MFI:

•solid skills to develop an effective business model: in particular, the ability

to demonstrate the positive impact on economic efficiency, on energy and

monetary savings, and on improved education and quality of life;

•well-defined supply chain’s involvement of all actors on the supply chain,

whilst improving the operational efficiency of each one;

•standardization of product development: a well-structured supplier and

commercialization base, avoiding the concentration of suppliers and/or

clients;

•clear-cut credit characteristics: transparent and clear tenors, dates, amounts,

operational processes, payments and grace periods.

The following sections describe in detail the steps undertaken within the sys-

tematic approach. Notice that, in this context, the approach is not restricted

to electricity usage, as the authors initially aimed sought, but it also concerns

mechanical and thermal energy usage, as the selected MES enable further uses

of energy.

1. Identification and Categorization of Productive Activities of the Targeted

Rural Areas Following the steps proposed by [De Gouvello and Durix, 2008],

Contactar has first identified the type of productive activities taking place in

its working area. Complying with the principles of classification by industry

Activities (ISIC) of the United Nations, Contactar’s internal codification applies

the ISIC Colombian codification provided by the Colombian National Admin-

istrative Statistics Department (DANE, for its acronym in Spanish). Table 4.4

shows the ISIC codification used for the energy needs assessment, while Tables

4.5 and 4.6 show the sector-system matrix of clients by June 2014 [Matas et al.,

2015].

The PEPI Toolkit for MFIs 4. The MFI Contactar

ISIC

Code Productive activity

A Agriculture, hunting, forestry and fishing

BMining and quarrying

CManufacturing

DElectricity, gas, steam and air conditioning

EWater distribution; disposal and treatment of wastewater,

waste management and environmental remediation activities

FConstruction

GWholesale and retail trade; repair of motor vehicles and mo-

torcycles

HTransportation and storage

IAccommodation and food services

JInformation and communication

KFinancial and insurance activities

LReal estate activities

MProfessional, scientific and technical activities

NAdministrative activities and support services

OPublic administration and defense; compulsory social secu-

rity schemes

PEducation

QActivities of human health care and social assistance

RArts, entertainment and recreation

SOther service activities

Activities of individual households as employers; undiffer-

entiated activities of individual households as producers of

goods and services for its own usage

UActivities of extraterritorial bodies and organizations

Table 4.4.: DANE Codification of productive activities used in Contactar

database for loan disbursements

The PEPI Toolkit for MFIs 4. The MFI Contactar

ISIC

Code Productive Activity Number

of clients

– Others 2,629

AAgriculture, hunting, forestry and fishing 43,755

A011 Transitory agricultural crops 8,436

A012 Permanent agricultural crops 20,877

A013 Plant propagation (nurseries activities, except forest nurseries)

A014 Livestock 14,049

A015 Mixed farming (crop and livestock) 39

A016 Support activities to agriculture and livestock, and post harvest

activities

A017 Ordinary hunting and trapping and activities of related services

A021 Silviculture and other forestry activities 14

A022 Wood removal

A023 Collection of forest products other than timber

A024 Support services to forestry

A031 Pesca 5

A032 Aquaculture 335

BMining and quarrying 40

CManufacturing 3,623

C10 Manufacture of food products 676

C11 Manufacture of beverages 19

C12 Manufacture of snuff 0

C13 Manufacture of textiles 494

C14 Manufacture of clothing 471

C15

Tanning and re-tanning of leather; shoemaking; manufacture of

suitcases, handbags and similar articles and manufacturing of sad-

dlery and harness; dressing and dying of fur

449

C16 Wood processing and manufacture of products of wood and cork,

except furniture; manufacture of articles of straw and plaiting 178

C17 Manufacture of paper, cardboard and paper products and card-

board 7

C18 Activities and production of printing copies from originals 41

C19 Coking, manufacture of oil refining and fuels blending activities 1

C20 Manufacture of chemicals and chemical products 28

C21 Manufacture of pharmaceuticals, medicinal chemicals and botan-

ical products pharmaceutical use

C22 Manufacture of rubber and plastic 14

C23 Manufacture of other non-metallic mineral products 51

C24 Manufacture of basic metal products 23

Table 4.5.: Productive activities (ISIC coditification A to C24) among Contac-

tar clients – June 2014

The PEPI Toolkit for MFIs 4. The MFI Contactar

ISIC

Code Productive Activity Number

of clients

C25 Manufacture of fabricated metal products, except machinery and

equipment 146

C26 Manufacture of computer, electronic and optical products 1

C27 Manufacture of appliances and electrical equipment

C28 Manufacture of machinery and equipment n.c.p. 11

C29 Manufacture of motor vehicles, trailers and semitrailers 13

C30 Manufacture of other transport equipment 0

C31 Manufacture of furniture, mattresses and box springs 361

C32 Other manufacturing 475

C33 Installation, maintenance and repair of specialized machinery and

equipment 164

DElectricity, gas, steam and air conditioning

EWater distribution; disposal and treatment of wastewater, waste

management and environmental remediation activities 74

FConstruction 1,467

GWholesale and retail; repair of motor vehicles and motorcycles 11,151

HTransportation and storage 1,708

IAccommodation and food services 2,235

JInformation and communication 274

KFinancial and insurance activities

M Professional, scientific and technical activities 95

LReal estate activities 814

NAdministrative activities and support services 228

OPublic administration and defense; compulsory social security

schemes

P Education 22

QActivities of human health care and social assistance 62

RArts, entertainment and recreation 295

SOther service activities

S94 Activities of memberships 47

S95 Maintenance and repair of computers and personal and household

goods 42

S96 Other activities of personal services 1,563

Activities of individual households as employers; undifferentiated

activities of individual households as producers of goods and ser-

vices for its own usage

UActivities of extraterritorial bodies and organizations

Total 70,159

Table 4.6.: Productive activities (ISIC coditification C25 to U) among Contac-

tar clients – June 2014

The PEPI Toolkit for MFIs 4. The MFI Contactar

2. Identification of Areas of Improvement in Productive Processes From

Table 4.5–4.6, it can be observed that, out of the 71,059 clients of Contactar,

the vast majority are engaged in:

•permanent and transitory agricultural crops (27.8% and 11.2% respec-

tively);

•livestock (18.7%) and wholesale and retail;

•repair of motor vehicles and motorcycles (14.9%).

The agriculture in all four regions of work (Nari˜no, Huila, Putumayo and

Tolima) is mostly dominated by permanent agricultural crops rather than tran-

sitory ones. The livestock industry is characterized by its concentration in

raising cattle throughout the four departments, especially in Huila. A more

detailed client division within the agricultural categories is presented in the

sector-system matrices per region in Tables 4.7 and 4.8, while Table 4.9 details

the livestock activities.

Nari˜no Huila Put. Tolima

A011 - Transitory agriculture crops 14,388 6,906 12 13,800

A0111

Cereals (except rice), 27 Barley 20 0 0 0

pulses, oilseeds 48 Beans 978 1,725 0 2,824

59 Corn 5,540 2,088 7 4,054

77-78 Quinoa 18 0 0 0

89 Wheat 395 0 0 0

5030 Cocoa 29 477 0 1,044

A0112

Rice 48 435 1 709

A0113

Vegetables, 71 Potato 3,559 42 0 440

roots and tubers 91 Yuca 602 1,365 4 2,307

A0115

Fibre 5080 Fique 302 1 0 1

Table 4.7.: Transitory agriculture activities in Contactar working regions ac-

cording to the 2005 National Census

For the main economic activities (transitory and permanent agriculture, live-

stock), the production processes are mostly dependent on fossil fuels (diesel

and firewood) or on manual (mechanical) technologies. Table 4.10 depicts the

production processes involved and the current machinery used depending on

the different economic activities.

3. Contribution of Energy Access Through Identified Equipment By green-

ing the financial services and scaling up the access offered to its clients, Contac-

tar enables both low income households and micro-, small- and medium-sized

businesses (MSMEs) an increase in welfare, productivity and competitiveness.

The PEPI Toolkit for MFIs 4. The MFI Contactar

Nari˜no Huila Put. Tolima

A012 - Permanent agriculture crops 6,278 14,998 16 18,264

A0122

5159 Banana 962 1,416 11 2,877

A0123

5031 Coffee 3,214 9,320 0 9,384

A0124

Sugar 5035-5037 Cane 431 1,103 1 1,452

A0125

Cut flower 2-4 Achira 179 30 0 0

A0128

Aromatic and medic-

inal spices and plants 38-39 Coriander 58 13 0 28

73 Parsley 4 0 0 0

119 Thyme 1 7 0 0

124 Peppermint 3 0 0 1

125 Mint 1 0 0 1

127-129 Lemongrass 42 54 0 150

131 Chamomile 1 0 0 0

132 Oregano 3 0 0 0

5000 Achiote 2 1 0 2

5006 Chili 7 3 0

5015-5016 Anise 7 0 0

5061 Coca 2 2 0 0

5165 Romero 1 0 0 0

5167-5168 Ruda 3 0 0 0

Table 4.8.: Permanent agriculture activities in Contactar working regions ac-

cording to the 2005 National Census

Livestock Nari˜no Huila Put. Tolima

A0141 Cattle and buffalo 45,800 171,755 1,071 310,637

A0142 Horses, other equines 7,785 25,198 101 36,090

A0143 Sheep and goats 2,632 6,926 27 17,348

A0144 Pig 29,408 146 38,476 60,513

A0145 Breeding poultry 792,397 1’063,416 535 1’195,590

A0149 Other animals n.c.p. 344,483 39,019 - 35,776

TOTAL 1’222,505 1’328,205 1,880 1’633,917

Table 4.9.: Livestock inventory according to 2005 National Census

Moreover, this creates new jobs within the institution and among the providers,

and it disseminates affordable access to a sustainable energy supply and green

products and services, thus ultimately improving the quality of life for vulner-

The PEPI Toolkit for MFIs 4. The MFI Contactar

Activity Production process Machinery used

Transitory and Security, Safety, Lighting, Electric systems for,

permanent Irrigation, Crop drying, Water pump, Crop Dryers,

agriculture Tools cleaning Water heating

Livestock Security, Safety, Tools

cleaning, Irrigation

Fridges, Lighting Systems,

Electric systems for water

heating, Water pumps

Table 4.10.: Production processes and machinery in Contactar’s working

regions

able smallholders.

Considering the goal of satisfying energy needs in the regions of work, Contactar

identified a range of modern technologies that is particularly effective for both

household and productive uses [Matas et al., 2015]. The outcome is summarized

in Tables 4.11 and 4.12. These shows, for each identified energy need, the dif-

ferent affected technologies (both for household and productive use). Moreover,

Table 4.13 shows the market potential identified among Contactar’s client pool,

according to the energy needs based on the economic activities with which the

clients are engaged.

Energy Need Technology

Usage: Household

(HH), Productive

Use (PU)

Lighting LED Lighting HH & PU

Fluorescent lamps HH, PU

Magnetic induction lighting PU

Automatic photovoltaic lamps HH, PU

Refrigeration Efficient industrial refrigerator PU

Cold rooms for storage and freezing PU

Efficient household refrigerator HH, PU

Thermal Waterproofing roofs HH, PU

insulation Thermal insulation doors HH, PU

Solar control films, window tinting HH, PU

Conditioning floors HH

Energy Efficient Efficient stoves HH, PU

Cooking Improved cooking ovens HH, PU

Gas production Biogas digester for gas generation HH, PU

Agricultural Solar crop dryer for fruits or

products drying agricultural products PU

Table 4.11.: Potential technologies for specific energy usages I - adapted from

[Matas et al., 2015]

The PEPI Toolkit for MFIs 4. The MFI Contactar

Energy Usage Technology

Type of usage:

Household (HH),

Productive Use

(PU)

Cooking Electrical stove HH, PU

Solar stove HH, PU

Improved cooking stove HH, PU

Biodigester HH, PU

Electrical Sys-

tems PV (Peak and Mini) HH, PU

DC Photovoltaic system (Pico-PV and

Mini Systems)

Energy efficiency

in industrial pro-

cesses

AC voltage controller - Residential / In-

dustrial HH, PU

Demand control (Servo Controls) PU

High efficiency electric motors PU

Automation and Remote Monitoring PU

Variable speed drives PU

Electronic ballasts PU

Movement sensors PU

Centrifugal pumps PU

Solar Water Pumps PU

Pump-Pressurized and constant pressure

supply kit PU

Air compressors PU

Brick-making oven PU

Washing and dry-

ing clothes Washing machines HH

Clothes dryers HH

Water treatment UV water purifier HH, PU

Water purifier reverse osmosis equipment HH, PU

Commercial water softeners PU

Greensand filters HH, PU

Water reservation Water treatment system (septic tanks) HH, PU

Water heating Thermo-syphon solar water heaters HH, PU

Water heater for pools (polyethylene ab-

sorbers)

Water heater high-performance direct fire

Water pumping Solar water pumps PU

Table 4.12.: Potential technologies for specific energy usages II - adapted from

[Matas et al., 2015]

The PEPI Toolkit for MFIs 4. The MFI Contactar

For the market research of potential technologies, the authors built upon the

assumptions that each group of clients could benefit from the identified relevant

technologies for its respective productive activity. According to the database

of productive activities, the authors added the number of clients for which the

considered technologies could be of relevance. The results are depicted in Table

4.13. Considering the offer of energy service suppliers within regions of interest

and the relevance of the technologies within the existing market, Contactar

opted in early 2011 to initiate the Green Finance Program offering BD and,

a year later, including ICS, in order to first tackle the efficiency of cooking

solutions.

Energy Usage Activity Market Po-

tential

Lighting Residential and Productive

activity 120,247

Refrigeration

Commercial agriculture and

for subsistence, Cattle, Retail

trade

92,272

Thermal insulation Residential and Productive

activity 1,793

Energy efficiency for

food preparation Restaurants, Food selling 24,285

Washing and drying

clothes Residential 18,905

Water treatment Residential 25,351

Water reservation Residential 58,266

Electrical Systems

Commercial agriculture

and for subsistence, Cattle,

Wholesale business

107,857

Water heating Cattle 29,804

Biogas digestor Cattle 14,049

Product drying Commercial agriculture and

for subsistence 4,299

Water pumping Commercial agriculture and

for subsistence, Cattle 43,736

Energy efficiency in in-

dustrial processes Industrial manufacturing 12,361

Table 4.13.: Market potential of energy technologies among the clientele of

Contactar (first column) and corresponding energy usages (second

column)

It shall be noticed that lighting, refrigeration and electrical systems have a

relatively high market potential. In fact, by estimating a multiple use of ap-

pliances per household and MSME, the market potential is even higher than

the number of Contactar’s clients. However, despite this predicted market size,

The PEPI Toolkit for MFIs 4. The MFI Contactar

the systematic approach also identified the presence of obstacles concerning the

maturity of the offer in the selected regions, and thus focused instead on more

viable technologies (e.g., biogas digesters, improved cooking stoves and solar

crop dryers).

4. Economical Assessment of New Production Processes The economic

viability of the production processes involved a different assessment procedure

for each technology. Furthermore, a necessary condition for the implementation

of the assessment included the design of a supply chain in coordination with

each technology supplier.

The following table illustrates the added value of the selected technologies.

Specifically, Table 4.14 depicts the contribution of energy access to the com-

modity value chain, due to the improvement in cooking practices, lighting fuels

and drying techniques, while Table 4.15 summarizes the cost reductions ob-

tained by using BD, ICS and SCD in productive activities.

MES Process supported Contribution to the commodity

value chain

ICS Firewood burning Fuel saving (energy efficiency)

Digestion of waste to re-

fine biogas for cooking

and production of bio-

fertilizer

Improvement of energy efficiency

SCD Crop, fruit drying

Decrease of time spent in same pro-

cess; increase of quality of dried

product; decrease of waste product

Table 4.14.: Illustration of an example of the contribution of energy access to

the commodity value chain

The PEPI Toolkit for MFIs 4. The MFI Contactar

MES Activity Former associated cost Costs reduction

Daily firewood

consumption and

LPG consump-

tion

Firewood cost: free (de-

forestation practices)

Time consumption:

dung collection; water

mixing; daily-weekly

maintenance

ICS

Daily firewood

consumption /

LPG consump-

tion

Firewood cost: free (de-

forestation practices)

Time consumption:

firewood collection

time, preparation time,

cooking time

SCD Time spent in

drying processes

Daily expenses of farm-

ers to spent in drying

process

Reduction of days dedi-

cated to drying

Product lost after

quality check-up

% Lost coffee pro-

duction due to moist,

garbage and contami-

nated beans

Full usage of amount

dried

Table 4.15.: Cost reductions obtained from the use of MES in productive

activities

Table 4.16 shows a comparison in further detail of crop drying time with and

without a SCD, while Tables 4.17 shows the results of a water boiling test with

ICS. In both tables the advantages of the proposed technologies over the tradi-

tional methods (for drying and cooking, respectively) can be easily appreciated.

Finally, Table 4.18 summarizes the main characteristics of a BD.

Without SCD With SCD

Productivity

Duration (time spent

for crop drying) Up to 12 days Less than 7 days

Time spent per day

for human activities in-

volved in drying process

3 working days 1 working day

Quality

Humidity Humidity absorbed None reabsorption

Waste 10% 0

Exportable quality 70% 75%

Table 4.16.: Comparison of traditional drying processes (without SCD) and

with a SCD, for a crop quantity of 90 Quintals (9000 kgs) of coffee

beans

The PEPI Toolkit for MFIs 4. The MFI Contactar

3-stone fire ICS ICS

(traditional) (mobile) (fixed)

Boiling time - Cold start 21 38 43

Boiling time - Hot start (min) 49 29 21

Fuel consumption - Cold start (g/L) 382 129 225

Fuel consumption - Hot start (g/L) 513 124 113

Energy consumption - Cold start (kJ/L) 6722 3511 4047

Energy consumption - Hot start (kJ/L) 9841 2168 2043

Combustion speed - Cold start (g/min) 61 43 24

Combustion speed - Hot start (g/min) 35 43 27

Thermal efficiency - Cold start (%) 6 11.2 11

Thermal efficiency - Hot start (%) 5 17 20.2

Table 4.17.: Comparison of water boiling test results with different cooking

stoves

Small BD Medium

BD Large BD

Size (Plastic calibre 8, diameter 1,25

m) 5m 7m 10m

Production capacity (biogas liters) 600l 900l 1300l

Required materials (TB: Tubular

Plastic; CP: Ceiling Plastic)

TP 14m -

CP 8m

TP 18m -

CP 10m

TP 24m -

CP 13m

Animals required (Average weather

/ Temperature conditions 18◦C to

23◦C)

10 14 25

Total dung per day (proportion 1:3)

(40kg/day) 1000 1400 2000

Water consumption 3000

(120l/day)

4400

(180l/day)

6400

(255kg/day)

Materials price 380,000

COP

460,000

COP

580,000

COP

Installation price 300,000

COP

760,000

COP

980,000

COP

Guaranty time 5 - 7 years for materials

Table 4.18.: Characteristics of biogas digesters according to size

The PEPI Toolkit for MFIs 4. The MFI Contactar

5. Promotional Campaigns for Consumer Targets Contactar directly coor-

dinated the design and implementation of promotional campaigns for the Green

Finance Program that were tailored to each type of end-user. The materials

developed for this objective include flyers, training manuals for loan officers,

posters, in-house trainings, and supplier trainings at installation processes as

well as radio spots. In particular, the promotional campaigns involved the

following stakeholders:

•Contactar: communication and marketing strategy development for Green

Program promotion;

•Radio stations: promotion and communication of Green Program at re-

gional levels (Ipiales, Cumbal, Sandon´a, Gualmat´an, T´uquerres, La Uni´on);

•Technology suppliers: promotion of Green Program at local levels and

trainings for loan officers on technologies;

•MicroEnergy International: support for Contactar and promotion of Green

Program at an international level;

•Citi Foundation: promotion and dissemination of Green Program through

network events.

As of October 2016, Contactar has managed to establish its program in the

departments of Nari˜no, Huila and Putumayo and it continuously coordinates

with suppliers to further shape the respective supply chains of clean energy

technologies.

5.Energy Usage in Rural Areas: Case

Study in Southern Colombia1

Part of this research is framed under Contactar’s motivation and willingness to

better identify energy needs among its clients and to track the improvements in

energy access through the financed technologies. To this aim, specific metrics

that are able to assess the level of provision of energy, as well as to quantify the

achievements of MES in supplying quality energy services, should be considered.

In fact, the currently available national statistics concerning energy supply fail

to describe specific quality attributes of the electrification or cooking fuels in

specific regions. Hence, a detailed analysis of such access might provide Con-

tactar valuable information for their Green Finance Program. Motivated by

this need, the research included a case study dedicated to the assessment of

energy access in the regions where Contactar operates. The study consisted in

applying the MTF assessment tool [Bhatia and Angelou, 2014,Bhatia and An-

gelou, 2015] to a sample of the MFI’s clients, comparing also the results of the

panel PPI data available from the institution.

The energy access assessment study considered two aspects of energy access at

the household level: electricity (supply, services and consumption) and cooking

solutions. The main motivation to conduct this twofold assessment lied in the

interest of the MFI to investigate energy uses at household level. Specifically,

for Contactar it was important to describe relevant aspects of energy assess in

the context of the MES currently offered in its Green Finance Program (Section

4.3), which include improved cooking stoves and biogas digesters. Moreover,

an additional motivation to include the assessment of cooking solutions was

to explore the possibility of testing their viability leaving out the attributes

requiring laboratory tests (safety, efficiency and convenience).

This chapter summarizes the results of this case study, which was conducted

from April to June 2015, based on the version of the multi-tier questionnaire

available by March 2015. Firstly, in Section 5.1 the criteria for the selection of

the sample clientele are described in detail, while, Sections 5.2 and 5.3 present

the detailed multi-tier ranking results, focusing on electricity supply (and ser-

vices) and cooking solutions, respectively. Finally, in Section 5.4 the multi-tier

energy access results have been analyzed relating them with the PPI scores.

1Part of this chapter is currently prepared for submission to the Elsevier journal Energy for

Sustainable Development.

The PEPI Toolkit for MFIs 5. Case Study in Southern Colombia

5.1. Sample Selection

Sample Size The selection of the sample size is a crucial aspect of a survey

study. Increasing the sample size often yields a more statistically significant

results (in the sense of a lower probability of errors and an improved ability to

detect less probable events).

In practice, the size of the sample is related to the precision of the estimate in

terms of margin of error and confidence level. The margin of error describes

the amount of sampling error in the results, expressing the likelihood that the

survey result, conducted among the sample subjects, is representative of the

description of the whole population. The confidence level is related to the level

of significance of the results and it describes the probability that the estimated

population parameters fall within a certain range of values (for instance, within

the prescribed margin of error).

Formally, margin of error (ME) and confidence level (CL) can be related to the

sample size nbased on on the probability distribution assumed for the target

population. Assuming a normal distribution yields

ME = z(CL)rσ2

n,(5.1)

where σ2stands for the variance of the distribution and where the value z(CL),

also called z-value, refers to the distance from the population mean µ, so that

the interval [µ−z(CL), µ +z(CL)] covers a ratio of the population equal to CL

(e.g., z= 1.96 for a 95% confidence level, see also Figure 5.1 – in this case, 95%

of the area under the normal distribution lies within 1.96 standard deviations

of the mean–.)

Figure 5.1.: Sketch of the meaning of z-value for a normal distribution

For dichotomies (questions allowing only two answers, e.g., positive and neg-

ative) it holds σ2=p(1 −p), where prefers to the probability of positive (or

negative) answers. Hence, the above formula reduces to

ME = z(CL)rp(1 −p)

n.(5.2)

The PEPI Toolkit for MFIs 5. Case Study in Southern Colombia

If no a priori information is available, p= 0.5 is usually taken, giving equal

probability to both events. This choice leads to

ME = z(CL)r0.25

n⇒n= 0.25z(CL)2

ME2.(5.3)

In order to implement the Energy Survey (Household Questionnaire) [SE4ALL

and WB, 2014], the size of the sample was selected based on a 95% confidence

level (i.e., z(0.95) = 1.96) and a margin of error of 5% (ME = 0.05) for di-

chotomies, resulting in a sample size

n= 0.25z(0.95)2

0.052= 384.(5.4)

Notice that the size of the sample for a given confidence level and margin of error

depends, in general, also on the total size of the target population. Particularly,

for small populations the required confidence can be achieved also with a smaller

sample. In the case of dichotomies, the sample size for a finite population N,

and prescribed confidence level and margin of error, can be estimated by

n(N)=0.25z(CL)2

ME21 + 0.25z(CL)2

ME2N−1

.(5.5)

The sample size nNgiven by (5.5) is always smaller than the size nprovided

by (5.3). For relatively large N(larger than about 150,000 individuals), nN

eventually coincides with n. Specifically, in the case of Contactar, considering

a client pool of 77,1502, the required sample size (confidence level 95%, margin

of error 5%) would be, according to Equation 5.5, of 382 households.

Selection of Clients Based on Geographical Areas Considering the four re-

gions in southern Colombia where Contactar operates (Nari˜no, Huila, Putu-

mayo and Tolima (see 4)), the first selection criteria was based on covering

both rural and urban clientele3reflecting the proportion of clients in each of

the regions (see Table 5.1). To this aim, different climate zones were iden-

tified, which determined substantial cultural differences in terms of economic

activities, diet and socially accepted living standards. Therefore, the routes

to be followed by the interviewers were selected by considering diverse areas

within each department, covering the diverse topography and climatic condi-

tions. Particularly, the highest altitude registered was 3,486 m (a.s.l.), while

the minimum was 321 m (a.s.l.), with characteristic mean temperatures varying

between 11◦C and 38 ◦C (see Table 5.2 and Figure 5.2 for details – (maps gen-

2MixMarket profile database for 2015.

Source: https://www.themix.org/mixmarket/profiles/contactar

3According to the DANE – National Administrative Department of Statistics – urban area is

characterized by several blocks, delimited by roads or avenues, and it is provided with basic

services such as aqueduct, sewage, electricity grid, hospitals or schools. On the contrary, rural

municipalities are characterized by disperse distribution of households, without official street

names, and mainly without access to public services or facilities. See also www.dane.com

The PEPI Toolkit for MFIs 5. Case Study in Southern Colombia

erated with the free application GpsPrune [GpsPrune, 2015])–). However, due

to security issues and given the current national conflict in Colombia, several

villages where Contactar is present were excluded from the research.

Total sample Nari˜no Huila Put. Tolima

Urban 69.5 % 43.8 % 84.2 % 100 % 58.2 %

Rural 30.5 % 56.2 % 15.8 % 0 % 41.8 %

Table 5.1.: Summary of the number of selected clients in urban and rural areas

Subjects Mean Altitude (min–max) Mean Temperature (min–max)

Nari˜no 50 % 2383 m (1276 – 3486) 21.6 ◦C (21 – 36)

Huila 29.7 % 1347 m (450 – 1929) 24.7 ◦C (18 – 38)

Putumayo 16.9 % 1852 m (321 – 2933) 24.2 ◦C (13 – 38)

Tolima 3.4 % 321 m 31.6 ◦C (27 – 35)

Total 100 % 2032 m (321 – 3486) 23.3 ◦C (11 – 38)

Table 5.2.: Summary of the number of selected clients in the four considered

regions

Figure 5.2.: GPS coordinates of the interviewed clients. Left: political map;

right: physical map.

Selection of Clients Based on Availability of PPI Data The second selection

criteria reflected the historical relation of the clients with the MFI, selecting the

clients who had filled the PPI questionnaire applied in 2012 and in 2014 (38,449

and 82,758 loans applications respectively). This means that selected clients

had a longer track record with the institution. However, in terms of energy

access, a picture of their energy situation should not differ from the one of the

clients who had filled out the PPI questionnaire only once (one-time client)

or who have applied only recently for a financial service at the institution for

the first time. Notice that the clients involved in the green lending program

ConSuPlaneta were not selected for this study. This was due to the fact that

The PEPI Toolkit for MFIs 5. Case Study in Southern Colombia

their energy supply situation had been already affected by the acquired energy

technologies and to the unavailability of data concerning their conditions prior

to the financing.

As introduced in Section 3.5, Contactar collects the PPI data of all clients

applying for any financial service at the institution since 2011. As for December

2015, almost a third of the clients filled out the PPI questionnaire twice (32,567

entries), i.e., they had applied and successfully received more than one loan

at the institution. Note, however, that since Contactar extended its branches

in Tolima only since 2013, PPI data were not available for the clients in this

department (3.12% of the total sample). The average scores of the considered

sample are shown in Table 5.3, according to the region and area of the clients.

Specifically, this characterization shows that the PPI score is mostly uniform

for the different regions despite of the diverse climatic conditions and cultures,

while it tends to be higher (47.9 versus 39.7, on average) for urban clients

compared to the rural ones.

Total sample Nari˜no Huila Putumayo

2012 2014 2012 2014 2012 2014 2012 2014

Urban 47.9 50.1 47.0 48.0 47.7 48.8 49.5 54.6

Rural 39.7 40.9 40.9 42.2 37.0 37.8 – –

Table 5.3.: Summary of mean PPI scores (2012 and 2014) of the selected clients,

according to their area (urban, peri-urban and rural) and region

Based on the above described criteria, the selection was randomized using the

free software Research Randomizer [Urbaniak and Plous, 2013] for each branch

cluster. Prior to the full survey implementation, a test-survey was conducted

among 12 individuals in the Pasto peri-urban areas, where the headquarters

of Contactar are located. The test served to adjust the terminology of the

questionnaire, to clarify skipping patterns and to include other existing income

sources.

Survey Implementation The survey was based on the multi-tier framework

questionnaire relative to the household locale4. The questionnaire has been

translated into Spanish by the Author5(see also Figures 5.3-5.4). The survey

included 1276 variables entries with an average of 30 questions per module.

The modules considered in this dissertation and their respective amount of

questions and variables are described in Table 5.4, not including the questions

for the identification of household, the data registering interview details (such

as location or time) and the Household Roster first module.

4English Version 5 from 13th of December of 2014. At the time of survey implementation

(March 2015), the tool included only 11 modules (A-K) [SE4ALL and WB, 2014].

5The Spanish version is available on nataliarealpecarrillo.weebly.com. This translation was

not created by The World Bank or the International Energy Agency and should not be

considered an official translation of either organization. Neither the World Bank nor the

International Energy Agency shall be liable for any content or error in this translation.

The PEPI Toolkit for MFIs 5. Case Study in Southern Colombia

For its application, the original survey [SE4ALL and WB, 2014] layout was trans-

formed into practical sheets so that the interviewers were able to conduct it us-

ing their notebooks. Therefore, data was directly uploaded into the data sheets

ready for data analysis For each household, the survey took between 80 and

120 minutes, depending on the household size, i.e., on the number of family

members living in the house.

With the collaboration and coordination of the general managers of Contactar

in each office, the interviewers coordinated visits with the clients upon arrival

(via mobile phone)6.

Modules No. of Questions

B: Supply of Electricity 34

C: Use of Electricity Services 6

D: Sources of Lighting Used within Household 5

G: Use of Cooking Stoves 42

K: Income 6

Table 5.4.: Modules, questions and variables considered in the implemented sur-

vey, following [SE4ALL and WB, 2014]

Figure 5.3.: Excerpt of the translated survey tool (electricity supply module)

used for the field study

6In some of the villages, prior to the research, the national government had distributed flyers

suggesting not to answer any type of survey or interview conducted by strangers, due to

recent extortion cases. Hence, in a number of cases a coordination with the local office, prior

to the interview, was needed to ensure trust from clientele.

The PEPI Toolkit for MFIs 5. Case Study in Southern Colombia

Figure 5.4.: Excerpt of the translated survey tool (cooking module) used for the

field study

5.2. Analysis of Electricity Supply, Services and

Consumption

5.2.1. PPI for Assessment of Electricity Services

As mentioned in Section 5.1, the pool of clients from which the sample was

chosen had as requisite to have received a PPI score twice. In order to further

motivate the need of specific indicators and of better measurement methodology

for energy access, this section presents a summary of the conclusions drawn from

the PPI questions related to electricity services.

The data reported in Table 5.5 shows that at least 91.7% of the interviewed

clients (in 2014) have access to electricity (i.e., they own at least a washing

machine, a fridge or a DVD player) in terms of enabled electricity services.

The analysis revealed also a notable difference between urban and rural areas

(97.2 % vs. 84.2 %), as well as a general improvement between 2012 and 2014

(+2% and +2.5% for the urban and rural area, respectively). However, no in-

formation was provided on the electricity supply capacity (power system), nor

on the electricity consumption nor on the quality of the available electricity.

These observations demonstrate how the PPI metric fails in correctly depict-

ing the evolution of energy access, as the indicators on appliances ownership

show an increase in all three categories (washing machine, fridge and DVD).

Hence, based on the information provided by the PPI, conclusions may only

be extracted in regards to the increasing level of power at the household (e.g.,

washing machines ownership increased). The ownership of fridges, on the other

hand, is not a reliable indicator as some areas covered by the MFI are above

3,000 m.a.s.l., where populations exclude fridge acquisition from their basic

The PEPI Toolkit for MFIs 5. Case Study in Southern Colombia

needs.

Furthermore, due to the formulation of the PPI questionnaire, off-grid house-

holds are classified in the same category (i.e., they receive the same scoring)

as the households classified in the lower ’rate-class’ (0, 1 and 2). For instance,

according to the PPI questionnaire, in 2012, 95.6% of the population lacked of

access to electricity, or had an illegal connection or pertained to socio economi-

cal class of 0, 1 or 2; in 2014, the size of this group increased by 1.6%. However,

no distinction can be made on the proportion of off-grid households. Moreover,

considering that the rural households are considered to belong to first or second

rate class, thus the indicator is not able to assess grid access in these areas.

Total Urban Rural

2012 2014 2012 2014 2012 2014

Q.5 No connec-

tion, classes 1–2 95.6 % 97.2 % 93.3 % 95.2 % 99.3 % 100 %

Q.7 WM 38.1 % 43.8 % 54.3 % 61.1 % 15.8 % 19.7 %

Q.8 Fridge 77.3 % 82.9 % 88.6 % 93.4 % 61.8 % 68.4 %

Q.9 DVD 69.6 % 74.9 % 73.3 % 79.6 % 64.5 % 68.4 %

At least one ap-

pliance 89.5 % 91.7 % 95.2 % 97.2 % 81.6 % 84.2 %

Table 5.5.: Information on electricity services (ownership of washing machine,

fridge/freezer and DVD) extracted from the PPI tool

The questions of the PPI survey related to cooking solutions (Table 5.6) revealed

that 95% of clients in the urban area (2014) cook with gas (either from a tank

or connected to a network), while only 50% in the rural area (2014) have access

to these facilities. Since the categories given to the cooking solution question

combined a variety of fuels, it is not possible to distinguish the usage of modern

fuels (gas or electricity) nor the efficient use of biomass (through improved

cooking stoves). Moreover, solid fuels such as firewood, wood, charcoal, mineral

charcoal, kerosene, alcohol or waste are combined with non-solid fuels such as

electricity and gas. Therefore, the results impede a more detailed analysis of

the quality of access to cooking fuels. Furthermore, considering the usage of

multiple fuels at household level, the indicator is constrained to the information

of the primary stove; it does not reveal the variety of fuels used at the household

neither the proportion of households that use more than one stove.

5.2.2. Multi-tier Analysis

This section is dedicated to the evaluation of electricity supply, electricity con-

sumption and access to electricity services of the sample, according to the

ESMAP MTF ranking approach with the tier thresholds established by [Bhatia

and Angelou, 2015] (see also Tables 2.4 and Table 2.5).

Aiming at capturing the largest dataset to provide detailed insights and at

identifying the most relevant indicators for the tier-ranking, the comprehensive

The PEPI Toolkit for MFIs 5. Case Study in Southern Colombia

Total Urban Rural

2012 2014 2012 2014 2012 2014

Gas (network) 14.4 % 18.5 % 24.3 % 30.1 % 0.7 % 2.0 %

LPG (tank) 54.7 % 58.4 % 65.7 % 65.4 % 39.5 % 48.7 %

Other 30.4 % 23.1 % 9.0 % 4.3 % 59.9 % 49.3 %

Do not cook 0.6 % 0 % 1.0% 0 % 0 % 0 %

Table 5.6.: Information on cooking facilities extracted from the PPI tool

“Other” include: firewood, wood, charcoal, mineral charcoal, elec-

tricity, gas, kerosene, alcohol or waste

framework was implemented among the three levels proposed by [Bhatia and

Angelou, 2015, pag 17]7.

After the data collection, the sample was ranked according to the tier-ranking

proposed in [Bhatia and Angelou, 2015]. In what follows, the results for each

attribute are described in detail, while the whole multi-tier matrix is summa-

rized in Table 5.8 (percentage of households in each tier), Table 5.9 (number of

households in each tier) and Figure 5.7. Moreover, the resulting access indices

are reported in Table 5.10 and Figure 5.22.

Electricity Supply

Capacity The results of the survey show that the totality of the sample is

connected to the grid, however, no information concerning the capacity of the

connection was available8. Hence, the household peak capacity has been esti-

mated considering the appliances used in the household, according to the values

reported in Table 5.7.

As shown in Table 5.8 and Figure 5.5, peak capacities are concentrated between

Tier 4 and Tier 5 (above 1,250 kWh and 3,425 Wh of daily supply, respectively)

in the urban area (53% and 29%, respectively), while the vast majority of rural

households fall in Tier 3 and Tier 4 (between 365 kW and 2,999 kW power, or

between 1,000 Wh and 8,218 Wh of daily supply) (36% and 54%).

7Given the amount of data required for the described approach, three levels of the framework

were proposed addressing the complexity of the MTF: a comprehensive, a simplified and

a minimalistic framework, varying in the level of data aimed to be collected under each

framework.

8The electricity utility only provided electricity consumption prices. Capacity was manually

estimated following average usages of owned appliances.

The PEPI Toolkit for MFIs 5. Case Study in Southern Colombia

Appliance Power (W)

Light Bulb 100

CFL 30

FL 28

LED 10

TV 150

PC 150

Electric Stove 1000

Fridge 400

Freezer 400

Ventilator 100

Washing Machine 500

Microwave 600

Air Conditioning 1000

Table 5.7.: Appliance power used to estimate the household capacity. Sources:

Daftlogic and ABS Alaskan.

Estimated Capacity (kWh)

0 1000 2000 3000 4000 5000 6000

Frequency (n. of clients)

80 Total

Urban

Rural

Figure 5.5.: Histogram of capacity data (considering total sample, urban and

rural areas)

Duration The best performing duration of electricity supply was found in

Huila with almost 95% of the sample assigned to Tier 4 and Tier 5, compared to

the lowest, Nari˜no with only 70% in the same tiers, and 15% with less than four

hours a day of electricity supply. Specifically, 10% of the sample had less than

four hours of electricity during the day due to several unpredictable outages,

which last more than one day in rural areas. In Putumayo, no information on

longer interruptions than two hours is found, ranking its electricity supply as

the most stable from the sample.

The PEPI Toolkit for MFIs 5. Case Study in Southern Colombia

As for the seasons, the months of May and June are considered the ones in

which the largest amount of electricity supply interruptions occur. In average,

the sample suffers of 1.99 outages per week lasting around 17.5 hrs. Particularly,

the duration of outages last from five minutes up to two days in the worst cases.

Indeed, 18.5% of the interviewed complain that unpredictable interruptions are

the most relevant issue concerning their electricity supply, and the 20% rank it

as the second barrier.

Reliability In both urban and rural areas, the connection is reported to be

unstable and affected by unpredictable interruptions. Specifically, Nari˜no and

Tolima appear to be the most affected regions. Note that urban households

dispose of a more deficient connection, with more than 84.3% suffering of more

than two outages per day against more than 67.5% of the rural households.

Quality Concerning the quality of the supplied energy, rural households are

the most affected (27.3% of the rural clients, against 14% of the urban clients,

reported damaged appliances due to outages). Particularly, the department

of Nari˜no is the one with the highest amount of cases of affected households.

Moreover, from those affected by voltage fluctuations, 57.4% confirm that the

presence of voltage fluctuation is not related to a a specific time of the year. The

remaining households affirm that fluctuations would mostly happen during the

first semester. Similar to the reliability of electricity supply, 19.5% of the sample

identifies the low voltage and fluctuations problem to be the main barrier the

households faced in regards to their electricity supply. Only 3% ranks it as a

second barrier.

Affordability Analyzing electricity expenses related to an annual consumption

of 365 kWh, in relation to the clients income, it is observed that, according to

the metrics proposed by [Bhatia and Angelou, 2015], a very small proportion of

households in urban and rural areas has excessive expenditures (more than 5%

of income) in electricity (4.8% and 8.7% respectively).

In order to provide a more detailed view of the affordability results, Figure 5.6

shows the percentage of clients (on the y-axis) whose energy expenses stays

below a given thresholds (x-axis) between 0% and 15% of the monthly income.

From these curves, one can see that the rural population declare in general

to have higher expenses (with respect to the monthly income). However, by

considering only the 5% threshold, it masks these differences, as for both regions

more than 90% of the sample is assigned in the highest tier.

Likewise, without considering the constraint on electricity consumption pack-

age, 22.9% and 21.7% of urban and rural households spend more than 5% of

their income in electricity supply. Moreover, in Huila only 9.4% of households

spend more than 5% of their income in electricity, while in Nari˜no less than 7%

of the clients have high energy expenses. Note, however, that this department

appears also to be the region characterized by the worse quality and reliability.

Notably, despite the fact that the majority of the sample was ranked in the

The PEPI Toolkit for MFIs 5. Case Study in Southern Colombia

Percentage of monthly income

0 0.05 0.1 0.15

Frequency

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

Total

Urban

Rural

5% Threshold

Figure 5.6.: Proportion of clients (y-axis) whose energy expenses does not ex-

ceed a certain percentage of the monthly income (x-axis)

highest tiers, one of the most relevant concerns of the interviewed population

was the high cost of the electricity supply. Specifically, the 67.5% of the clients

claimed it to be one of the two major issues the household faces, indicating that

the 5% threshold might not necessarily reflects the perception and the priori-

ties of the population regarding the distribution of their income and respective

expenses.

Legality Except for two cases (rural households in Huila), all clients declared

to be legally connected to the grid. However, the two subjects without le-

gal connections pay significant amounts to neighbouring households for their

connection and are satisfied with their current situation.

Health and Safety None of the sample have had any accident in the past year,

though a minimal proportion (2.6% of households in Huila) feels a safety risk

(e.g., due to the outages) for the need of using complementary energy sources.

In fact, 80.7% use candles or fuel-run lamps as substitute energy source for

illumination during the outages, leading to a perceptible increase of health-

hazard in case of any unfortunate accident.

The PEPI Toolkit for MFIs 5. Case Study in Southern Colombia

Tier

5 4 3 2 1 0

Frequency (%)

Capacity

Reliability

Duration

Quality

Affordability

Legality

Health-Safety

Tier

5 4 3 2 1 0

Frequency (%)

100

Capacity

Reliability

Duration

Quality

Affordability

Legality

Health-Safety

Tier

5 4 3 2 1 0

Frequency (%)

Capacity

Reliability

Duration

Quality

Affordability

Legality

Health-Safety

Figure 5.7.: Tier ranking for the different attributes (electricity supply), con-

sidering the total sample (top), the urban area (bottom-left) and

the rural area (bottom-right)

The PEPI Toolkit for MFIs 5. Case Study in Southern Colombia

Area Departments

Total Urban Rural Nari˜no Huila Put. Tolima

Capacity

Tier 5 20.8% 29.1% 8.4% 16.7% 17.5% 37.9% 25.0%

Tier 4 53.9% 53.5% 54.6% 55.7% 57.9% 40.9% 58.3%

Tier 3 25.0% 17.4% 36.4% 27.1% 24.6% 21.2% 16.7%

Tier 2 – – – – – – –

Tier 1 0.3% – 0.6% – – – –

Tier 0 – – – – – – –

Reliability (short unpredictable interruptions)

Tier 5 65.6% 72.6% 55.2% 51.0% 81.6% 80.3% 66.7%

Tier 4 34.4% 27.4% 44.8% 49.0% 18.4% 19.7% 33.3%

Duration (including day and night electricity supply duration)

Tier 5 68.0% 75.2% 57.1% 54.2% 84.2% 78.8% 75.0%

Tier 4 11.2% 7.8% 16.2% 14.6% 10.5% – 25.0%

Tier 3 5.2% 4.3% 6.5% 7.3% 0.9 7.6% –

Tier 2 5.5% 4.3% 7.1% 7.8% 1.8 6.1% –

Tier 1 3.6% 4.3% 2.6% 4.7% – 7.6% –

Tier 0 6.5% 3.9% 10.4% 11.5% 2.6% –

Quality (appliances not damaged due to outages)

Tier 5 80.7% 86.1% 72.7% 69.8% 93.0% 87.9% 100%

Tier 0 19.3% 13.9% 27.3% 30.2% 7.0% 12.1% –

Affordability (electricity expenses below 5% of HH income)

Tier 5 93.6% 95.2% 91.3% 93.8% 90.6% 96.6% 100%

Tier 0 6.4% 4.8% 8.7% 6.2% 9.4% 3.4% –

Legality (legal connection)

Tier 5 99.2% 100% 98.1% 100% 97.4% 100% 100%

Tier 0 0.8% – 1.9% – 2.6% – –

Health and Safety (absence of accidents or risks)

Tier 5 99.2% 100% 98.0% 100% 97.4% 100% 100%

Tier 0 0.8% – 2.0% – 2.6% – –

Table 5.8.: Results (in percetages) of the tier-ranking for electricity supply

The PEPI Toolkit for MFIs 5. Case Study in Southern Colombia

Area Departments

Total Urban Rural Nari˜no Huila Put. Tolima

Capacity

Tier 5 80 67 13 32 20 25 3

Tier 4 207 123 84 107 66 27 7

Tier 3 96 44 56 52 28 14 2

Tier 2 – – – – – – –

Tier 1 1 – 1 – – – –

Tier 0 – – – – – – –

Reliability (short unpredictable interruptions)

Tier 5 252 167 85 98 93 53 8

Tier 4 132 63 69 94 21 13 4

Duration (including day and night electricity supply duration)

Tier 5 261 173 88 104 96 52 9

Tier 4 43 18 25 28 12 52 3

Tier 3 20 10 10 14 1 – 9

Tier 2 21 10 15 1 2 3

Tier 1 14 10 4 9 – 4 –

Tier 0 25 19 16 22 3 – –

Quality (appliances not damaged due to outages)

Tier 5 310 198 112 134 106 58 12

Tier 0 74 32 42 58 8 8 –

Affordability (electricity expenses below 5% of HH income)

Tier 5 323 197 126 180 77 56 10

Tier 0 22 10 12 8 2 –

Legality (legal connection)

Tier 5 381 230 151 192 111 66 12

Tier 0 3 – 3 – 3 – –

Health and Safety (absence of accidents or risks)

Tier 5 381 230 151 192 111 66 12

Tier 0 3 – 3 – 3 – –

Table 5.9.: Results (number of households) of the tier-ranking for electricity

supply

The PEPI Toolkit for MFIs 5. Case Study in Southern Colombia

Electricity Consumption

In regards to the consumption of electricity supply, urban households have

a similar average consumption patterns as rural houses, with, however, higher

consumption peaks (Figure 5.8). Nevertheless, only few rural households (1.4%)

achieved the highest tier (16.9% in the urban areas). In detail, monthly con-

sumption estimation showed that an average yearly consumption of 1,500 kWh

across the regions, with a clear difference between urban (1,860 kWh) and rural

(1,070kWh) areas.

Energy consumption (kWh/year)

0 1000 2000 3000 4000 5000 6000

Frequency (n. of clients)

70 Total

Urban

Rural

Figure 5.8.: Histogram of consumption data (considering total sample, urban

and rural areas)

Energy Services

Consistently to what has been reflected in the capacity metrics, most of the

clients are ranked in Tier 3 and Tier 4 according to the usage of medium and

high-power appliances. More than half of rural households use low or medium

power appliances (ranking in Tier 2 and Tier 3), and this situation resulted

most frequent in Huila and Tolima.

Tier Rankings

The comparison of tier-rankings (see Table 5.10) drawn from electricity supply

(composed of multiple attributes), electricity services and electricity consump-

tion revealed a notable difference of characterization of the electricity access

in urban and rural households (see subsections 5.2.2, 5.2.2 and 5.2.2. Using

the lowest-tier assignment rule, according to the supply ranking, about 30% of

households have been ranked in the lowest tier (Tier 0), and more than 35% of

the households have been ranked in the lowest tiers (Tiers 0 to 2). Moreover,

the analysis of electricity supply revealed differences between urban and rural

The PEPI Toolkit for MFIs 5. Case Study in Southern Colombia

areas. The access index, calculated using equation 2.1 (see also Section 2.6.3),

was slightly above 50 (52.4) for the total sample. However, urban households

achieved an index of 61.7, while rural households only of 38.7.

On the other hand, in the case of electricity consumption and electricity ser-

vices, the vast majority of clients (more than 90%) achieved a better tier (3–5),

yielding multi-tier indices above 70 in all cases, except for the case of electricity

consumption in rural areas (index of 64.4).

Total Urban Rural

Sup Serv Cons Sup Serv Cons Sup Serv Cons

Tier 5 12.5% 20.8% 10.5% 19.1% 23.9% 16.9% 2.6% 16.2% 1.4%

Tier 4 35.4% 47.1% 38.6% 40.9% 50.4% 44.9% 27.3% 42.2% 29.0%

Tier 3 15.6% 28.9% 44.9% 13.5% 23.9% 34.8% 18.8% 36.4% 60.1%

Tier 2 4.2% 2.1% 5.8% 3.0% 0.9% 3.4% 5.8% 3.9% 9.4%

Tier 1 2.6% 1.0% – 2.6% 0.9% – 2.6% 1.3% –

Tier 0 29.7% – – 20.9% – – 42.9% – –

Index 52.4 76.8 70.6 61.7 79.2 75.0 38.5 73.6 64.4

Table 5.10.: Index and tier assignment according to Electricity Supply (Sup),

Electricity Services (Serv) and Electricity Consumption (Cons)

Frameworks

Sensitivity Analysis

In order to complete the results of the multi-tier approach, based on appling the

lowest tier ranking among all attributes, this section is dedicated to a sensitivity

study aiming to analyze the dependence of the final result from single attributes.

Table 5.11 shows the frequency of attributes ranked as the “lowest” within the

sample. While reliability,legality and health and safety almost never determined

the tier-ranking of the household energy access, capacity was in about half of

the cases the lowest-ranked attribute among all. Furthermore, quality and

affordability also appeared to be relevant, since these attributes are ranked

lowest in about 20% of the cases.

Furthermore, Figures 5.9 and 5.10, together with Tables 5.12, 5.13 and 5.14,

summarize the results obtained leaving out one attribute and recomputing the

lowest-based tier ranking.

From these pictures, one can conclude that the framework is mainly sensitive to

capacity (consistently with Table 5.11), and to quality. Specifically, removing

the capacity yielded a shift in the highest tiers, moving households from Tiers 3

and 4 to Tier 5 (in both urban and rural areas). The quality affected in a greater

extent the lowest tiers. Without this attribute, about 20% of the households

in Tier 0 have been shifted to Tier 3 and Tier 4. Considering only the rural

households, the framework resulted sensitive also to duration and affordability,

as excluding one of these attributes leaded a shifts from Tier 0 to Tier 4.

These conclusions are also visible in Figure 5.11, which shows the sensitivity of

The PEPI Toolkit for MFIs 5. Case Study in Southern Colombia

Area

Total Urban Rural

Capacity 49.5% 51.3% 46.8%

Reliability 1.6% 2.6% –

Duration 11.5% 9.6% 14.3%

Quality 19.0% 13.9% 26.6%

Affordability 17.4% 22.6% 9.7%

Legality 0.3 – 0.6

Health and Safety 0.8% – 1.9%

Table 5.11.: Frequency of attributes ranked as the lowest tier for each household

(energy supply)

Without:

MTF –Cap –Rel –Dur –Qua –Aff –Leg –HS

Tier 5 12.5% 50.7% 13.5% 12.8% 14.1% 12.8% 12.5% 12.5%

Tier 4 35.4% 9.9% 34.3% 43.5% 43.0% 39.1% 35.4% 35.4%

Tier 3 15.6% 3.1% 15.6% 17.7% 21.1% 16.4% 15.9% 15.9%

Tier 2 4.2% 4.2% 4.2% – 5.5% 4.2% 4.2% 4.2%

Tier 1 2.6% 2.3% 2.6% 0.3% 3.6% 2.9% 2.6% 2.6%

Tier 0 29.7% 29.7% 29.7% 25.8% 12.8% 24.7% 29.4% 29.4%

Index 52.4 62.7 52.6 58.2 64.0 56.1 52.5 52.5

Table 5.12.: Results of sensitivity study of the tier ranking for electricity supply

(total sample): lowest tier ranking leaving out single attributes,

compared with the full MTF ranking

Without:

MTF –Cap –Rel –Dur –Qua –Aff –Leg –HS

Tier 5 19.1% 59.1% 20.9% 19.1% 21.7% 19.6% 19.1% 19.1%

Tier 4 40.9% 10.4% 39.1% 47.4% 47.4% 43.5% 40.9% 40.9%

Tier 3 13.5% 3.9% 13.5% 15.2% 14.8% 13.9% 13.5% 13.5%

Tier 2 3.0% 3.0% 3.0% – 4.3% 3.0% 3.0% 3.0%

Tier 1 2.6% 2.6% 2.6% – 3.9% 3.0% 2.6% 2.6%

Tier 0 20.9% 20.9% 20.9% 18.3% 7.8% 17.0% 20.9% 20.9%

Index 61.7 71.6 62.0 66.2 71.0 64.5 61.7 61.7

Table 5.13.: Results of sensitivity study of the tier ranking for electricity sup-

ply (urban area): lowest tier ranking leaving out single attributes,

compared with the full MTF ranking

the index when leaving out a single attribute. Leaving out capacity and quality

yielded an increase in the composite index of about 15 points, while leaving out

affordability or duration increased the index mostly in the rural area (about 10

points).

The PEPI Toolkit for MFIs 5. Case Study in Southern Colombia

Without:

MTF –Cap –Rel –Dur –Qua –Aff –Leg –HS

Tier 5 2.6% 38.3% 2.6% 3.2% 2.6% 2.6% 2.6% 2.6%

Tier 4 27.3% 9.1% 27.3% 37.7% 36.4% 32.4% 27.3% 27.3%

Tier 3 18.8% 1.9% 18.8% 12.5% 30.6% 20.1% 19.5% 19.5%

Tier 2 5.8% 5.8% 5.8% – 7.1% 5.8% 5.8% 5.8%

Tier 1 2.6% 1.9% 2.6% 0.6% 3.2% 2.6% 2.6% 2.6%

Tier 0 42.9% 42.9% 42.9% 37.0% 20.1% 36.4% 42.2% 42.2%

Index 38.6 49.5 38.6 46.4 53.5 43.5 40.0 40.0

Table 5.14.: Results of sensitivity study of the tier ranking for electricity sup-

ply (rural area): lowest tier ranking leaving out single attributes,

compared with the full MTF ranking

Tier

5 4 3 2 1 0

Frequency (%)

50 w/o Capacity

w/o Reliability

w/o Duration

w/o Quality

w/o Affordability

w/o Legality

w/o Health-Safety

MTF

Figure 5.9.: Results of sensitivity study of the tier ranking for electricity sup-

ply: histogram of lowest tier ranking leaving out single attributes,

compared with the full MTF ranking (in pink)

The PEPI Toolkit for MFIs 5. Case Study in Southern Colombia

Tier

5 4 3 2 1 0

Frequency (%)

Tier

5 4 3 2 1 0

Frequency (%)

Figure 5.10.: Results of sensitivity study (in urban (left) and rural (right) areas)

of the tier ranking for electricity supply: lowest tier ranking leav-

ing out single attributes, compared with the full ESMAP MTF

ranking (in pink). Legend as in Figure 5.9.

Area

Total Urban Rural

Index

70 w/o Capacity

w/o Reliability

w/o Duration

w/o Quality

w/o Affordability

w/o Legality

w/o Health-Safety

MTF

Figure 5.11.: Results of sensitivity study of the tier ranking for electricity sup-

ply: composite multi-tier index computed leaving out single at-

tributes, compared with the full MTF index

The PEPI Toolkit for MFIs 5. Case Study in Southern Colombia

5.3. Analysis of Access to Cooking Solutions

Regarding the access to cooking solutions, the considered regions are charac-

terized by an ample availability of options (summarized in Table 5.15). Partic-

ularly, households are mainly provided by LPG (close to 70%, both in urban

and rural areas), while access to natural gas is only found in urban areas, and

mostly in Huila and Putumayo. Besides gas, the use of solid biomass cooking

fuel for the primary stove (10% of the total sample) can be encountered in the

rural areas of Nari˜no and Huila.

Furthermore, almost half of the rural households (48%) own a second stove. In

these cases, the majority of secondary solutions (75%) consists of a hand-made

stove, using firewood as cooking fuel.

Area Departments

Total Urban Rural Nari˜no Huila Put. Tolima

Primary Cooking Solution

LPG 67.7% 66.5% 69.5% 91.1% 20.2% 90.9 16.7%

Natural gas 20.3% 31.7% 3.2% – 54.4% 9.1 83.3%

Firewood 9.9% 0.4% 24.0% 4.7% 25.4% – –

Electrical 0.3% – 0.6% 5.2% – – –

Owns a 2nd

stove 26.3% 11.7% 48.1% 43.2% 14.9% 1.5% –

Type of secondary solution (if any)

Firewood 75.2% 85.2% 71.6% 88.0% 17.6% – –

LPG 19.8% 3.7% 25.4% 7.2% 82.4% – –

Liquid 3.0% 7.4% 1.4% 3.6% – – –

Electrical 2.0% 3.7% 1.4% 1.2% – 100% –

Table 5.15.: Summary of cooking solutions

The following part presents and discusses the tier ranking for each attribute

characterizing access to cooking solutions, according to the MTF and following

the thresholds established by [Bhatia and Angelou, 2015] (see Table 2.6). The

results are also summarized in Table 5.16, Table 5.17 and Figure 5.14.

The PEPI Toolkit for MFIs 5. Case Study in Southern Colombia

Area Departments

Total Urban Rural Nari˜no Huila Put. Tolima

Convenience (time spent in stove and fuel acquisition and preparation)

Tier 5 57.2% 76.5% 27.7% 40% 70.3% 77.3% 91.7%

Tier 4 16.0% 15.0% 17.6% 21.6% 5.4% 21.2% –

Tier 3 6.4% 5.8% 7.4% 12.4% – 1.5% –

Tier 2 2.7% 1.8% 4.1% 5.4% – – –

Tier 0 17.6% 0.9% 43.2% 20.5% 24.3% – 8.3%

Availability (fuel availability throughout the year)

Tier 5 85.4% 82.2% 90.9% 94.8% 81.6% 63.6% 100%

Tier 4 9.4% 13.0% 3.9% 5.2% 3.5% 33.3% –

Tier 0 4.9% 4.8% 5.2% – 14.9% 3.1% –

Affordability (fuel and stove expenses below 5%)

Tier 5 70.2% 78.0% 58.2% 65.4% 70.0% 81.8% 83.3%

Tier 0 29.8% 22.0% 41.8% 34.6% 30.0% 18.2% 16.7%

Quality (absence of heat variation of fuel)

Tier 5 96.8% 97.8% 95.3 % 93.5% 100% 100% 100%

Tier 0 3.2% 2.2% 4.7% 6.5% – – –

Safety

Tier 5 99.7% 99.6% 100 % 99.5% 100% 100% 100%

Tier 0 0.3% 0.4% –% 0.5% – – –

Table 5.16.: Results (in percentage) of the tier-ranking for cooking facilities for

the different attributes

100

The PEPI Toolkit for MFIs 5. Case Study in Southern Colombia

Area Departments

Total Urban Rural Nari˜no Huila Put. Tolima

Convenience (time spent in stove and fuel acquisition and preparation)

Tier 5 214 173 41 74 78 51 11

Tier 4 60 34 26 40 78 14 –

Tier 3 24 13 11 23 6 1 –

Tier 2 10 4 6 10 – – –

Tier 0 66 2 64 38 27 – 1

Availability (fuel availability throughout the year)

Tier 5 329 189 140 182 93 42 12

Tier 4 36 30 6 10 4 22 –

Tier 0 19 11 8 – 17 2 –

Affordability (fuel and stove expenses below 5%)

Tier 5 262 177 85 121 77 54 10

Tier 0 111 50 61 64 33 12 2

Quality (absence of heat variation of fuel)

Tier 5 365 222 143 173 114 66 12

Tier 0 12 5 7 12 – – –

Safety (absence of past accidents)

Tier 5 374 225 149 182 114 66 12

Tier 0 1 1 – 1 – – –

Table 5.17.: Results (number of households) of the tier-ranking for cooking

facilities

101

The PEPI Toolkit for MFIs 5. Case Study in Southern Colombia

Convenience Concerning the convenience (an attribute based on the time

required for preparing/collecting the fuel and the time needed for cooking),

more than 75% of the urban households were assigned to the highest tier, and

more than 90% achieved Tier 4 or Tier 5. This result is consistent with the

fact that the majority of urban clients has access to gas (LPG or network) for

cooking. On the contrary, only about one fourth (28%) of rural households

achieved the highest tier ranking, although, as shown in Table 5.15, more than

70% use gas for cooking. The low assignment is due to the time spent in

acquisition/collection of LPG in rural areas. Moreover, the majority of rural

households (55%) is ranked in Tiers 1 to 3, further highlighting the difference

between fuel access in urban and rural areas.

Figure 5.12 provides a more detailed picture of cooking time, fuel preparation

time and collection time. It reveals that, while cooking time might be mainly

associated with the usage of gas (implying very short time), both in urban and

rural areas, the fuel preparation and collection times affects in a greater extent

the rural households (whose times are roughly uniformly distributed between

1 and 8 hours per week) than the urban ones (as the majority needs less than

1h/week).

Cooking time (min)

1 10 30 60

Frequency (n. of clients)

100

150

200

250

300 Total

Urban

Rural

Fuel preparation and collection time (h per week)

1 2 4 8 15 30 60 100

Frequency (n. of clients)

100

120

140 Total

Urban

Rural

Figure 5.12.: Histogram of convenience data considering cooking time (left) and

fuel preparation and collection time (right)

Availability On the one hand, most of the households declared to have fuel

available for at least 10 months/year (except from the region of Putumayo with

over 35% of affected households and almost 15% of households in Nari˜no). Rural

households seem to be better placed in terms of availability rather than urban

households. On the other hand, it is worth noticing that this might be due to

the fact that households that are not able to use gas (i.e., not connected to a

network), mostly in the rural area, need to adjust their cooking solution to the

availability of fuel and of budget.

Affordability The affordability of cooking solutions (considering fuel costs per

year, plus the price of the stove adjusted to the usage time) reveals that only

22% of urban households spent more than 5% of their income in cooking fuel,

102

The PEPI Toolkit for MFIs 5. Case Study in Southern Colombia

while almost the 42% of rural households exceeded this threshold of their ex-

penses (with a peak of 35% in Nari˜no, where for the interviewed clients belong

mainly to rural area). Compared with the expenses in electricity supply, house-

holds’ expenses to satisfy their energy needs are dominated by their cooking

needs.

Figure 5.13 shows the percentage of clients (on the y-axis) whose expenses for

cooking fuel is below a given threshold (x-axis) between 0% and 50% of the

monthly income. As for the fuel supply, the rural population seems to be more

affected by affordability issues than the urban one (i.e., having, in average,

higher expenses). However, unlike in the case of affordability of electricity

supply, concerning cooking fuel setting the threshold between lowest and highest

tier at 5% allows to differentiate between rural and urban population.

Percentage of monthly income

0 0.1 0.2 0.3 0.4 0.5

Frequency

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

Total

Urban

Rural

5% Threshold

Figure 5.13.: Percentage of clients (y-axis) whose expenses for cooking fuel do

not exceed a certain percentage of the monthly income (x-axis)

Quality Most of the households claimed not to be affected by the quality of

the cooking fuel. Particularly, only the 3.12% that reported quality issues, and

all these used LPG as cooking fuel.

Safety Only two cases of accidents have been reported in Nari˜no, and the

remaining households did not consider that their health or security are at risk

due to their cooking solution.

103

The PEPI Toolkit for MFIs 5. Case Study in Southern Colombia

Tier

5 4 3 2 1 0

Frequency (%)

Convenience

Safety

Availability

Affordability

Quality

Tier

5 4 3 2 1 5

Frequency (%)

Convenience

Safety

Availability

Affordability

Quality

Tier

5 4 3 2 1 0

Frequency (%)

100

Convenience

Safety

Availability

Affordability

Quality

Figure 5.14.: Tier ranking for the different attributes (cooking solutions), con-

sidering the total sample (top), the urban area (bottom-left) and

the rural area (bottom-right)

Tier Rankings

The tier ranking based on the lowest tier among all attributes is detailed in

Table 5.18. Two main conclusions can be drawn from these results. Firstly, the

majority of the population is ranked either in the highest or in the lowest tier

(about 38% in tier 5 and about 44% in tier 0). This behavior depends on the

level of threshold used for the attributes. Secondly, there are large differences

between the urban and the rural areas. More than 50% of urban households

are ranked in tier 5 and less than 30% in tier 0. On the contrary, less than 20%

of rural households reach the highest tier, while about 68% of the households

remain in tier 0.

Sensitivity Analysis

This section describes the result of a sensitivity study, in order to analyze the

dependence of the final result from single attributes.

104

The PEPI Toolkit for MFIs 5. Case Study in Southern Colombia

Total Urban Rural

Tier 5 37.8% 50.4% 18.8%

Tier 4 13.3% 17.4% 7.1%

Tier 3 3.9% 3.5% 4.6%

Tier 2 1.3% 0.9% 1.9%

Tier 1 –––

Tier 0 43.7% 27.8% 67.5%

Index 51.2 66.8 28.1

Table 5.18.: Results of the (lowest-based) tier-ranking and of the composite

index for cooking facilities

Table 5.19 shows the frequency of attributes which were ranked as the “lowest”

for each household. Unlike the case of electricity supply, the assessment of access

to cooking facilities reveals a dominant role of convenience and affordability,

while the remaining attributes rather determined in a lesser extent the tier

assignment of the household. Specifically, since the affordability attribute is

binary (in the sense that a household can be either in Tier 0 or in Tier 5),

the threshold set for this attribute is responsible of placing more than 20% of

households in the lowest tier.

Area

Total Urban Rural

Convenience 67.4% 65.6% 70.1%

Safety 0.8% 0.4% 1.3%

Affordability 21.6% 21.3% 22.1%

Availabilty 8.3% 11.3% 3.9%

Quality 1.8% 1.3% 2.6%

Table 5.19.: Frequency of attributes ranked as the lowest tier for each household

(cooking solutions)

Similar conclusion concerning the tier ranking sensitivity with respect to these

two attributes can be drawn from Tables 5.20, 5.21 and 5.22, which show the

lowest-based tier ranking obtained leaving out one attribute at a time. The

outcome of this sensitivity study is also summarized in Figures 5.15 and 5.16.

Leaving out the affordability leaded to a shift of about 20% of households from

Tier 0 to Tier 4 and Tier 5, with a greater effect in the urban area. Also the

effect of removing the convenience attribute was different for urban and rural

areas. In the former, it produced a slight shift of households from Tier 4 to

Tier 5, while, in the latter, households moved mainly from Tier 0 to Tier 5.

The above conclusions are also reflected in the composite access index, shown

in Figure 5.17 for the different cases.

105

The PEPI Toolkit for MFIs 5. Case Study in Southern Colombia

Without:

MTF –Conv –Saf –Ava –Aff –Qua

Tier 5 37.8% 57.8% 37.8% 41.7% 49.7% 38.3%

Tier 4 13.3% 7.0% 13.3% 10.4% 17.7% 14.6%

Tier 3 3.9% – 3.9% 3.9% 6.0% 3.9%

Tier 2 1.3% – 1.6% 1.3% 2.3% 1.3%

Tier 1 –% – – – – –

Tier 0 43.7% 35.2% 27.4% 40.9% 24.2% 41.9%

Index 51.2 63.4 51.3 52.9 68.4 52.8

Table 5.20.: Results of sensitivity study of the tier ranking for cooking solutions

(total sample): lowest tier ranking leaving out single attributes,

compared with the full MTF ranking

Without:

MTF –Conv –Saf –Ava –Aff –Qua

Tier 5 50.4% 61.7% 50.4% 58.7% 63.9% 50.9%

Tier 4 17.4% 10.9% 17.4% 11.7% 21.3% 18.3%

Tier 3 3.5% – 3.5% 3.5% 5.2% 3.5%

Tier 2 0.9% – 1.3% 0.9% 1.3% 0.9%

Tier 1 – – – – – –

Tier 0 27.8% 27.4% 27.4% 23.9% 8.3% 26.5%

Index 66.8 70.4 67.0 70.5 84.6 69.0

Table 5.21.: Results of sensitivity study of the tier ranking for cooking solutions

(urban area): lowest tier ranking leaving out single attributes, com-

pared with the full MTF ranking

Without:

MTF –Conv –Saf –Ava –Aff –Qua

Tier 5 18.8% 51.9% 18.8% 18.9% 28.6% 19.5%

Tier 4 7.1% 1.3% 7.1% 8.4% 12.3% 9.1%

Tier 3 4.6% – 4.6% 4.6% 7.1% 4.6%

Tier 2 1.9% – 1.9% 1.9% 3.9% 1.9%

Tier 1 – – – – – –

Tier 0 67.5% 46.8% 67.5% 66.2% 48.1% 64.9%

Index 28.0 52.9 28.0 29.1 44.3 30.3

Table 5.22.: Results of sensitivity study of the tier ranking for cooking solu-

tions (rural area): lowest tier ranking leaving out single attributes,

compared with the full MTF ranking

106

The PEPI Toolkit for MFIs 5. Case Study in Southern Colombia

Tier

5 4 3 2 1 0

Frequency (%)

w/o Convenience

w/o Safety

w/o Availability

w/o Affordability

w/o Quality

MTF

Figure 5.15.: Results of sensitivity study of the tier ranking for cooking solu-

tions: lowest tier ranking leaving out single attributes, compared

with the full MTF ranking

Tier

5 4 3 2 1 0

Frequency (%)

Tier

5 4 3 2 1 0

Frequency (%)

Figure 5.16.: Results of sensitivity study (in urban and rural areas) of the tier

ranking for electricity supply: lowest tier ranking leaving out sin-

gle attributes, compared with the full MTF ranking. Legend as

in Figure 5.15.

107

The PEPI Toolkit for MFIs 5. Case Study in Southern Colombia

Area

Total Urban Rural

Index

80 w/o Convenience

w/o Safety

w/o Availability

w/o Affordability

w/o Quality

MTF

Figure 5.17.: Results of sensitivity study of the tier ranking for cooking so-

lutions: composite multi-tier index computed leaving out single

attributes, compared with the full MTF index

108

The PEPI Toolkit for MFIs 5. Case Study in Southern Colombia

5.4. Multi-tier Ranking vs. PPI Data

As a following step, the performance for the different attributes and the tier

ranking have been combined with the PPI data of the clients9.

Electricity Supply The major differences across the sample have been observed

for the attributes capacity,reliability and duration. The tier assignments asso-

ciated to the attribute capacity are correlated to the PPI, as the average PPI

scores increases per tier (e.g. 41.4 for Tier 3, 45.9 for Tier 4, 53.5 for Tier 5

(urban: 45.6, 49, 55, and rural: 38.6, 41.6, 46.2, respectively). As mentioned

before, since one third of the PPI questionnaire depends on electricity services,

the correlation is reflected in to the capacity, through the use of appliances. On

the contrary, the reliability behaves differently; while an increase in the PPI

scores also reflected an increase in reliability in urban households (46.7 aver-

age of PPI in Tier 4, 55.6 in Tier 5), in rural areas the PPI scores were more

uniformly distributed across Tier 4 and Tier 5 (without any household in Tier

0). The correlation in urban area suggested that in households that are better

off also have better electricity connection. Concerning the attribute duration,

the average PPI did not correlate with the tier ranking, excluding the highest

score for urban households in Tier 5. This suggests that a decrease in likeli-

hood to fall under the poverty is related to the availability of longer electricity

connection. For the remaining attributes, the household are mostly ranked in

the highest tier, thus it was not possible to analyze their correlations with the

PPI.

Considering the tier assignment according to the electricity supply framework,

a relation between attributes and PPI could not be seen. In fact, average PPI

scores per tier oscillated (between 41 and 55), as shown in Figure 5.18. However,

households assigned in Tier 5 have by far the highest PPI scores. Similar PPI

correlation behavior is also observed in urban and rural areas (not shown).

Electricity Consumption and Services The tier rankings of the electricity

consumption and services frameworks appeared to be strongly related to the

PPI score of the households. Specifically, in both cases a global increase of

average PPI scores with the tier assignment was observed (see Figure 5.19–5.20).

Moreover, a linear correlation analysis also revealed that the overall electricity

expenses, not only consumption, increased with the PPI average scores.

Cooking Solutions Concerning the tier ranking for the cooking solutions, an

increase in PPI was observed only between households in Tier 0 (average PPI

9Besides the PPI score, the monthly income (from the survey) and the loan of the clients

(from the database of Contactar) were also available. However, these three variables result

positively correlated, yielding also similar results with respect to the energy access attributed.

Between the loan amount and the PPI score, the correlation resulted of 0.20, while between

the loan amount and the monthly income of the household the correlation was of 0.15. Hence,

in the following analysis, only the results using the latest PPI score (2014) to characterize

the financial data will be reported.

109

The PEPI Toolkit for MFIs 5. Case Study in Southern Colombia

●● ●

012345

20 40 60 80

Energy Supply Lowest Tier

PPI Score (2014)

Figure 5.18.: Mean values and standard variations of PPI score vs. tier ranking

for energy supply

●

1 2 3 4 5

20 30 40 50 60 70 80

Energy Services Tier

PPI Score (2014)

Figure 5.19.: Mean values and standard variations of PPI score vs. tier ranking

for electricity consumption

slightly above 40) and the remaining tiers (average PPI score between 46.6 and

51), see Figure 5.21.

The detailed analysis of the correlation of PPI scores and tier rankings was

limited to the attributes affordability and convenience, since, for the remaining

attributes, the vast majority of households were assigned in Tier 5.

Namely, the average PPI scores increase with the tier assignment in affordability

(e.g. Tier 3: 43.1, Tier 5: 47.8). This is valid both for urban and rural areas.

However, in the case of convenience the average PPI resulted less related to the

final tier ranking, suggesting that the fuel collection/preparation time might

depend on other aspects (e.g., on the area), rather than on the poverty level

(see also Figure 5.12).

Finally, Table 5.23 reports the Spearman correlation coefficients of the different

attributes of electricity supply against the PPI score (2014). Results showed

that only the capacity attribute is significantly correlated to the likelihood of

falling below the poverty line. From an analogous analysis, in the case of cooking

110

The PEPI Toolkit for MFIs 5. Case Study in Southern Colombia

●

2345

20 30 40 50 60 70 80

Energy Consumption Tier

PPI Score (2014)

Figure 5.20.: Mean values and standard variations of PPI score vs. tier ranking

for energy services

●

●●

0 2 3 4 5

20 40 60 80

Cooking Lowest Tier

PPI Score (2014)

Figure 5.21.: Mean values and standard variations of PPI score vs. tier ranking

for cooking solutions

Spearman correlation vs. PPI

Energy Supply (lowest tier) 0.19

Capacity 0.34

Duration 0.15

Reliability 0.13

Quality 0.01

Affordability 0.02

Legality 0.11

Safety 0.04

Table 5.23.: Spearman correlation coefficients of tier ranking (energy supply)

and PPI score

solutions (Table 5.24), it was observed that affordability and convenience are

the only attributes correlated to the PPI score. This is in line with the graphs

111

The PEPI Toolkit for MFIs 5. Case Study in Southern Colombia

in Figure 5.14, which show that the results for these attributes were better in

the urban area than in the rural one.

Spearman correlation vs. PPI

Cooking solutions (lowest tier) 0.20

Convenience 0.28

Safety -0.08

Affordability 0.18

Quality -0.12

Availability 0.02

Table 5.24.: Spearman correlation coefficients of tier ranking (cooking solutions)

and PPI score

5.5. Summary of the Study

As a conclusion, Figure 5.22 provides the results of the application of the dif-

ferent multi-tier frameworks, comparing electricity supply, electricity services,

electricity consumption and cooking solutions.

The outcome of the case study can be summarized as follows:

•The analysis of electricity supply showed that the majority of house-

holds lies in high tiers (4 or 5) for most of the attributes. However, the

capacity attribute is dominant as in most cases it determines a lower tier

ranking.

•Moreover, the quality of electricity supply in rural areas (with 20% of

households reporting quality problems) requires intervention on the grid

service in order to avoid voltage problems

•Concerning the access to cooking solutions, the analysis resulted in

a higher number of households ranked in Tier 0, in comparison to the

electricity supply assessment. In most cases, this low ranking was due to

poor performance in affordability of the cooking solution.

•Moreover, rural households were also affected by convenience issues (time

spent for cooking fuel acquisition and preparation), with more than 40% of

the households in Tier 0 revealing a lack of suitable technologies according

to fuel availability.

•The access indices showed large differences between urban and rural areas.

Particularly, rural households achieved an access index of 38.5 concerning

electricity supply (compared to 61.7 of the urban area) and of an index

of 28.1 to cooking solutions (compared to 67.8 for the urban area).

•The rankings for electricity supply and cooking solutions showed only a

low correlation with the PPI ranking (correlation coefficient of 0.20 in

both cases).

112

The PEPI Toolkit for MFIs 5. Case Study in Southern Colombia

Lowest Tier

5 4 3 2 1 0

Frequency (%)

Energy Supply

Energy Services

Energy Consumption

Cooking

Lowest Tier

5 4 3 2 1 0

Frequency (%)

Energy Supply

Energy Services

Energy Consumption

Cooking

Lowest Tier

5 4 3 2 1 0

Frequency (%)

60 Energy Supply

Energy Services

Energy Consumption

Cooking

Figure 5.22.: Tier ranking (based on the lowest tier among all attributes) for

the different frameworks, considering the total sample (top), the

urban area (bottom-left) and the rural area (bottom-right)

113

6.Development of the PEPI Toolkit1

Microfinance institutions, whose primary social mission is the financial inclu-

sion, are rarely aware of the energy needs of their customers. Moreover, since the

core of their business lies in providing financial services, so far only a small num-

ber of institutions, although steadily increasing [Pierantozzi et al., 2015,Shuite

and Forcella, 2015], have exploited their potential to also offer financing for

modern energy services to foster energy inclusion [Realpe Carrillo et al., 2015].

These considerations suggest that MFIs, as well as other stakeholders, could

benefit from better and more detailed information management tools towards

efficiently addressing their triple bottom line (combined social, financial and

environmental goals).

According to this view, the objectives of this research is to develop the Progress

out of Energy Poverty Index (PEPI), a toolkit targeted to financial institutions

and energy service suppliers, in order to support them in improving energy

access of the base of the pyramid. Aimed at providing tools to be used initially

by the microfinance industry, the presented toolkit seeks to enable MFIs to

assess the energy access of their clients, and, at the same time, focuses on

practicability for data collection and analysis.

Targeting the assessment of energy access, the PEPI is based on the methodol-

ogy of the MTF (described in Section 2.6 and used for the case study discussed

in Chapter 5), aligning the MTF attributes with the SDG targets. In particu-

lar, this is done by focusing on the energy services accessed by the population,

considering a different set of attributes and tier thresholds, and, finally, defining

specific metrics to measure progress on global attributes of energy access over

time at the household level. From a practical point of view, the main qualities

of the designed tool and progress metric are that it allows a comprehensive

analysis of energy access at the household level, based on a simple and struc-

tured survey implementation tool, which supports flexible thresholds of energy

attributes, in order to easily adapt the framework to different contexts and to

efficiently evaluate the different attributes of energy access. Namely, in order

to facilitate data collection within the financial institutions, the PEPI toolkit

provides a set of surveys (based on precompiled Excel sheets), to be used by

organizations together with their standard data collection processes.

The proposed tool is described in detail in this Chapter. Firstly, Section 6.1

1Part of this chapter was published as [Realpe Carrillo, Natalia, ’Development of the Progress

out of Energy Poverty Index (PEPI) Toolkit’, Technical Report of MicroEnergy International

GmbH, 15.06.17].

115

The PEPI Toolkit for MFIs 6. Development of the PEPI

discusses in detail the design of the PEPI building upon the experience of the

MTF, addressing its limitations, proposing some modifications according to

the Energy SDG, and introducing the progress measure related to the adapted

framework. Next, the practical aspects of the tool are described in detail Section

6.3. Finally, in Section 6.2 the proposed frameworks are applied to the micro-

finance clients sample of Contactar, comparing the results with the outcome of

the MTF (presented in Chapter 5).

6.1. Toolkit Design

6.1.1. Properties and Limitations of the Multi-tier Framework

As presented in Chapter 2, in recent years the discussion on the relevant at-

tributes to describe and assess energy access has considered several overviews.

In the search of energy access assessment tools, [Pachauri et al., 2012a] claimed

that a fair metric of energy access should be able to capture more than the

sole physical access to energy. In particular, they argued that, even among

populations with physical access to electricity and modern fuels, the lack of

affordability and reliable supplies would considerably limit the extent to which

a transition to using these could occur. Similarly, [Nussbaumer et al., 2011]

highlighted the relevance of these attributes, claiming the importance of mea-

suring the extent of provision of energy services. However, the index developed

in [Nussbaumer et al., 2011] (MEPI, see also Section 2.5) was defined based

on the MPI and focused mainly on selected uses of energy. Therefore, limited

attention was still devoted to other specific attributes of the provided energy.

The importance of quality, reliability and affordability of the delivered energy,

besides monitoring the mere expansion of physical infrastructure or the fuel

usage, has been also acknowledged by [Kittelson, 1998,Rehman et al., 2012,Groh

et al., 2016], in order to effectively measure access to modern energy. Moreover,

in regards to cooking facilities, [Modi et al., 2006] also claimed that one of

the main objectives of better energy access is to improve the affordability, the

availability and the safety of cooking fuels and practices.

Focusing on the role of clean energy technologies to address energy access, the

developments of end user finance as well as the flexibility of payment should

be considered when defining the affordability of electricity supply and cooking

solutions. In fact, whether framed within green microfinance programs or as a

consequence of mobile banking mechanisms, the flexibility of payment influences

directly the affordability of end users for the acquisition and the usage of the

technology.

The ESMAP MTF [Bhatia and Angelou, 2015] (described in more detail in Sec-

tion 2.6) represents a first important step to improve traditional measurement

approaches. By focusing on the different dimensions of energy, rather than only

on the binary assessment of access, the main novelty of MTF lies in the assess-

ment of energy access from the perspective of multiple attributes for all required

energy applications across households, productive enterprises and community

116

The PEPI Toolkit for MFIs 6. Development of the PEPI

institutions.

However, in a preliminary step, its limitations should be cross-examined. As

observed by [Groh et al., 2016], as there is no consensus on the amount of energy

needed to meet human needs, the tiered-spectrum from the MTF highlights a

preliminary point of debate. On the one hand, what is the optimal level of

energy measured at the useful level from an end-user perspective that facilitates

human development? On the other hand, how can these levels be translated into

different tiers to be achieved on average? Given that needs vary significantly

among countries and regions, social customs, weather and other specific region-

and society-specific factors, a set of minimum basic energy needs is not yet

accepted [Pachauri et al., 2004].

Upon the implementation of this methodology in Bangladesh (considering about

200 households), [Groh et al., 2016] critically assessed the MTF, stressing the

relevance of using appropriate metrics to effectively track universal energy ac-

cess objectives, and suggesting possible ways forward and improvements to the

multi-tier approach. Their analysis reveals a clear trade-off between captur-

ing the multidimensionality of energy access and defining an easy-to-use global

framework. Despite the adoption of certain suggestions in the latest version

(see [Bhatia and Angelou, 2015]), the assessment is still focused on the sup-

ply rather than addressing the “on-demand” side. Specifically, focusing on

the meaningfulness of existing measurements to end users, [Groh et al., 2016]

highlight the importance of “putting the services at the core of the metric”

when defining an energy poverty measure. That aside, rather limited atten-

tion has been given to the quality, reliability and affordability of energy ser-

vices. Furthermore, by defining the tiers based on specific combinations of

attributes, [Groh et al., 2016] observed that the approach is prone to errors

due to its complexity and the limitations of fixed thresholds and dichotomous

attributes. In particular, as quality deficiencies might vary along a specific at-

tribute, providing further scales of deprivation would increase the transparency

of the energy access assessment. For instance, the goal of the tier ranking is

to measure the ability of the energy supply to cater to specific energy applica-

tions. Notwithstanding, the tier ranking for electricity services and electricity

consumption have separated frameworks of electricity supply [Bhatia and An-

gelou, 2015]. Therefore, their inter-relationship is not directly reflected, but

the tiers are assigned independently of the performed quality of the electricity

supply and the electricity services provided.

One of the outcomes of the MTF is the set of energy access indices (for the

different frameworks), defined for a set of households (e.g., for a geographical

area), which can be calculated as a weighted average of the tier ranking results

in the different frameworks (e.g., households, productive uses and community

facilities) [Bhatia and Angelou, 2015]. While these composite indices allow for

the combination of multiple data in order to identify specific deficiencies of the

energy supply, as also discussed by [Bhatia and Angelou, 2015], whether the

proposed index provides an effective measure for assessing the impact of energy

access programs is still an open question. Namely, as the weights are associated

with the proportion of households ranked in the given tiers, the final indices

117

The PEPI Toolkit for MFIs 6. Development of the PEPI

might be biased by the concentration of the population in a specific (low) tier

rather than providing a global picture of relevant attributes. Moreover, since

the final tier-ranking, on which the indices are based on, depends only on the

lowest tier from the assessed attributes, the performances might result to be

very sensitive to a particular attribute while not tracking the improvements

in the overall picture, thus yielding to an incorrect interpretation. A further

limitation of the MTF access index consists in the fact that, by definition, it

reflects the overall access of a whole population (the total sample considered

for the survey implementation), and it does not allow to track changes at the

household, micro-enterprise or installation level according to the tier-ranking of

each attribute.

In conclusion, aiming at a multi-tier measurement of energy access able to sup-

port governments in setting and monitoring targets, both the condensed metric,

obtained by applying the lowest tier among the attributes, and the composite

index, depending on the average performance of the population, might not show

where improvements have been achieved (e.g., in the case that the lowest tier re-

mained constant, but some of the other attributes improved), possibly yielding

biased conclusions and misleading decision-making processes.

6.1.2. The Development Agenda

The SDG 7, which dictates “ensure affordable and reliable energy access by

2030”, recognizes not only the crucial role of energy access for development

but also the importance of the quality of delivered energy. Specifically, this

development goal is subdivided in three sub-targets:

•SDG 7.1: Ensure universal access to affordable, reliable and modern energy

services

•SDG 7.2: Increase substantially the share of renewable energy in the global

energy mix

•SDG 7.3: Double the global rate of improvement in energy efficiency

In order to achieve a significant change, it is important to be able to track the

progresses towards the goal of this SDG, by setting clear milestones, defining

specific indicators and aligning definitions and methodologies in the involved

sectors.

However, several methods for measuring energy access significantly underesti-

mate the scale of the challenge [Global Tracking Framework (GTF), 2015], and

the need for a robust set of measurement tools to set common goals is claimed

by different stakeholders, from academics to donors and practitioners.

A first effort to adopt the MTF to the SDG targets was conducted by [Stevens

et al., 2015]. The authors adopted the original indicators and attributes defined

by the MTF, proposing a methodology to assess progress towards the SDG 7

sub-targets:

•percentage of population with access to electricity of at least MTF Tier

3 (Target 7.1);

118

The PEPI Toolkit for MFIs 6. Development of the PEPI

•percentage of population with access to clean and efficient cooking fuels

and technology of at least MTF Tier 4 (Target 7.1);

•renewable energy share in the total energy final energy consumption (Tar-

get 7.2).

Furthermore, [Stevens et al., 2015] suggested that an indicator for the sub-

target 7.1 should at least include a measure of the safety of energy access, in

order to avoid potential conflicts with the overall Energy SDG or with other

climate and health-related SDGs and targets. The importance of the attribute

safety, especially when considering the assessment of energy access in rural areas

through MES, was also acknowledged by [Groh et al., 2016], who, however,

observed that the safety of energy access within the MTF is rather vaguely

defined.

6.1.3. From the ESMAP to the PEPI Framework

Since its pioneer publication [Global Tracking Framework (GTF), 2013], the MTF

has achieved a broad acknowledgement, yielding several improvements, which

recognize energy access enhancing processes. As such, given the level of detail

of the approach, future metrics aiming at tracking energy access should support

its adoption and strive the harmonization of all decisions on the choices of sub-

indicators and modifications [Bensch, 2014].

Considering the multidimensional nature of energy poverty, despite the trade-

offs, implies that assessment tools must achieve a balance between methodologi-

cal sophistication and theoretical accuracy, as well as between applicability and

transparency [Bazilian et al., 2010]. Additionally, from an energy poverty metric,

it is expected that the tools will combine political attractiveness and usefulness

for policy design, as well as match practical data availability [Nussbaumer et al.,

2011]. The scope of the PEPI is, therefore, to provide a framework for rating

energy access quality against tiers of performance for the series of attributes,

building upon the MTF and considering the indicators proposed from [Stevens

et al., 2015] together with the observations and recommendations from [Groh

et al., 2016].

To this end, the PEPI frameworks for electricity supply considers the following

MTF attributes: capacity (based on power use), quality (based on efficiency

of the supplied energy), duration, reliability, affordability, legality, health and

safety. Besides which, seeking to enable stakeholders to assess the household’s

access to cooking facilities without the need of laboratory values (i.e., in the ab-

sence of the appropriate tools for smoke measurement), the PEPI framework for

cooking solutions only takes into account the following previous attributes from

the MTF: convenience,availability,affordability and safety, including a sepa-

rate health attribute. However, in order to better align the metric to the Energy

SDG target, the methodology groups these attributes in wider categories, that

will be called global attributes: The framework is constructed measuring specif-

ically the indicators assigned for each sub-target and global measurements of

three groups of the MTF attributes.

119

The PEPI Toolkit for MFIs 6. Development of the PEPI

Figure 6.1.: Sketch of the PEPI Frameworks and Attributes

•Reliability,Affordability, and Safety, for electricity supply and ser-

vices

•Availability,Safety and Affordability, for cooking facilities.

Figure 6.1 sketches the considered frameworks, while the subdivisions for energy

supply and cooking are detailed in Tables 6.1 and 6.2, respectively, showing

the selected attributes aggregated into the three main global attributes and

indicating the corresponding information to be collected in each matrix.

120

The PEPI Toolkit for MFIs 6. Development of the PEPI

Electricity Services & Supply

(Global)

Attributes Affordable Reliable Safe

SDG explicit goal

Suggestion

[Stevens et al.,

2015]

MTF attributes

[Bensch, 2014]

•Capacity

•Affordability

•Legality

•Duration

•Reliability

•Quality

•Health

•Safety

Corresponding

indicators

•Services

•Electricity ex-

penses

•Payment, fre-

quency and in-

come

•Electricity

hours

•Unpredictable

interruptions

•Voltage fluctu-

ations

•Electricity

source

•Accidents

•Risks

Table 6.1.: PEPI Methodology: aggregation of the MTF attributes (electric-

ity supply and electricity services) to adapt the framework to the

Energy SDG

Cooking facilities

(Global)

Attributes Affordable Available Safe

SDG explicit goal

Suggestion

[Stevens et al.,

2015]

Indicators

•Initial invest-

ment

•Monthly fuel

expenses

•Maintenance

costs

•Fuel Availabil-

ity

•Fuel quality

•Fuel and stove

convenience

•Health

•Safety

Table 6.2.: PEPI Methodology: aggregation of the MTF attributes (cooking

solutions) to adapt the framework to the Energy SDG

121

The PEPI Toolkit for MFIs 6. Development of the PEPI

As envisioned, an unbiased picture of the state of play in regard to energy

access for analysts and policy makers [Bensch, 2013], can be provided by a

hybrid approach from the dashboard of the independent indicators integrated

in the PEPI following the MTF. Hence, by defining the measured dimensions

according to the SDG dictated goal, policy design can address targets in each

axis and track actions and achievements.

Moreover, a series of modifications of the original frameworks (see Tables 2.4

and 2.6) have been added to the PEPI matrices in order to provide a deeper

visualization of the quality of energy access, incorporating several suggestions

from [Groh et al., 2016]. Electricity services have been kept in the same tier-

ranking (based on the “usual” power demanded by the electrical appliance, see

Table 6.3 and [Bhatia and Angelou, 2015]). However, the services are assessed

within the same electricity supply framework, instead of defining a separate

framework.

Output services ESMAP Service classification

Tier 5

Cooling/heating spaces; very

high-power mechanical loads;

electric cooking

Very high-power services

Tier 4 Heating; high-power mechan-

ical loads;

High-power services (microwave, hair

dryer, toaster, iron)

Tier 3 Refrigeration; Medium-power

mechanical loads

Medium-power services (fridge, freezer,

washing machine, mixer, rice cooker,

water pump)

Tier 2 Entertainment; Information;

Cooling

Low-power services (TV, PC, printer,

ventilator)

Tier 1 Illumination; Communica-

tion

Very-low power services (light bulbs,

phone charger, radio)

Tier 0 None None of the above

Table 6.3.: Thresholds of tier-ranking standards for electricity services consid-

ered in the PEPI frameworks

Similarly, the PEPI does not contain a separate electricity consumption frame-

work. The motivation behind this choice is to constrain the assessment of energy

access within the quality of services that the household is able to use (and how

the household makes use of them). In fact, according to [Groh et al., 2016], the

inputs to measure the capacity attribute undermine efficiency goals, since the

real service output is not measured. This is the case, for instance, in efficient

lighting: while a higher consumption might lead the household to higher tiers,

luminosity is rather lower than in traditional lighting.

At the level of single attributes, within the PEPI the capacity is measured as

a function of the household usage of electricity services, while the availability

of energy is no longer a dichotomy (either Tier 0 or Tier 5), but is described

by multiple tiers, in order to provide a more accurate picture for the eval-

uation of off-grid solar applications, as suggested by [Groh et al., 2016]. A

122

Electricity Supply and Services

SDG

Attributes

Associated

Attributes Tier 0 Tier 1 Tier 2 Tier 3 Tier 4 Tier 5

Reliable Reliability Outages Max no. of outages per

week 28 21 14 3

Duration of

outages

Max. duration of outages

(min) 299 239 179 119

Duration Day Min. duration of power

supply (hours) 4 4 8 16 23

Night Min. duration of power

supply (hours) 1 2 3 4 4

Quality Damages Damaged appliances or

risk of damages

TRUE -

risk FALSE FALSE

Affordable Affordability % Max. % of income for elec-

tricity consumption 10% 5% 2%

Capacity Appliances Service output of cate-

gorised appliances

Task

lighting,

Phone

charging

Lighting,

TV,

(Fan)

any

medium-

power

app

any

high-

power

app

any very

high-

power

app

Legality Connection Illegal connection

TRUE

- pay-

ing for

services

FALSE FALSE

Safe Health Affected

health

Health hazards due to en-

ergy source FALSE

Safety Accidents

or risks

Accidents or risks of in-

juries due to power supply Accidents High risk Low risk No acci-

dents

No acci-

dents

Table 6.4.: PEPI Multi-tier Matrix: Thresholds of attributes and tier-ranking standards for Electricity Supply and Services at house-

hold level

Cooking Facilities

SDG

Attributes

Associated

Attributes Tier 0 Tier 1 Tier 2 Tier 3 Tier 4 Tier 5

Available Convenience

Stove

preparation

time

Max. minutes to prepare

stove for a meal 7 3 1,5 0,5

Fuel acqui-

sition and

preparation

time

Max. hours per week to

acquire and prepare fuel 15 10 5 2

Availability

of primary

fuel

% of fuel availability

throughout the year 50% 80% 100%

Quality

Quality of

primary

fuel

Variation of fuel quality

for cooking No No

Affordable Affordability Fuel and

stove costs

Max. % of income spent

for stove and fuel 20% 5% 5%

Safe Health Affected

health

Health hazards due to fuel

inhalation FALSE FALSE

Ventilation Absence of chimney - ex-

traction of smoke FALSE FALSE

Safety

Cooking

area (for

traditional

fuels)

Cooking place same as of

sleeping area FALSE FALSE

Accidents

or risks

Accidents or risks of in-

juries due to cooking fuel

or stove

FALSE FALSE FALSE

Table 6.5.: PEPI Multi-tier Matrix: Thresholds of attributes and tier-ranking standards for Cooking Facilities at household level

The PEPI Toolkit for MFIs 6. Development of the PEPI

further modification, with respect to the MTF, is that the attributes reliabil-

ity and legality have been expanded to include further indicators, in order to

enable better tracking of the improvements of interventions. Furthermore, the

attributes health and safety are evaluated using different indicators, although

they are eventually kept under the same umbrella of the safety global attribute.

In particular, the global attribute considers health hazards, pains and diseases

that might be caused by the electricity power supply or fuel.

In the case of the attributes for describing access to cooking solutions, the

indicators describing the health and safety attributes have been expanded to

take into account the level of ventilation, the conditions of the cooking area and

the level of hazards the household is exposed to due to the cooking fuel or to

the stove.

6.1.4. Measuring the Progress

The tier rankings corresponding to the global attributes provide a first one-

shot assessment of the energy access level. As a further step, the PEPI aims at

becoming a methodology for long-term evaluation of energy access progresses

at the household level. To this end, the goal of the PEPI is to be integrated

within the standard survey tools of MFIs used to characterize their clients, thus

allowing for multiple data-collections of the same households.

Namely, in order to quantify the (multidimensional) improvement of energy

access at the household level, instead of calculating a composite index based on

a one-shot evaluation (as is the case for the MFT access indices), we propose

a methodology for defining the Progress out of Energy Poverty Index (PEPI),

which consists in a composite metric for the variation of tier-ranking (positive

or negative) over time.

Therefore, it is important to separately assess the attributes related to different

SDG targets, in order to better track the improvements over time along the

different dimensions. Specifically, a separate progress indicator is defined for

the (SDG) global attributes in each framework (electricity supply and services,

cooking solutions). The final hybrid index is then calculated as the arithmetic

mean of the values of the indicators.

In order to formally introduce the Index, the three global attributes for elec-

tricity supply and services will be denoted with Si, i = R,A,S (reliability,

affordability, safety, respectively) while Ci, i = Af,Av,S stand for the three

global attributes describing cooking solutions (affordability, availability, safety,

respectively). Moreover, for each global attribute, let us denote with Si,j or

Ci,j the tiers of specific dimensions (with j= 1,2 or j= 1,2,3 depending on

the sub-attribute). For instance, SA,3will denote the tier of the legality of

electricity supply (third dimension of the affordability global attribute).

Next, it is assumed to have available two different data collections at two dif-

ferent times, denoted by S1

i,j,C1

i,j and S2

i,j,C2

i,j.

The progress is quantified assigning weights ∆(Si,j) and ∆(Ci,j) depending on

the (positive or negative) tier variation within each specific category of the

125

The PEPI Toolkit for MFIs 6. Development of the PEPI

three global attributes. These weights vary from -1 (the worst case, i.e., when a

household moved back from Tier 4 or 5 to Tier 1 or 0) to 1 (best case, denoting

a considerable improvement, e.g., from Tier 0 or 1 to 4 or 5). The values of

the weights are detailed in Table 6.6 (see also Section 6.1.5 for more details),

depending on the tier variation from the first to the second evaluation.

First evaluation Second evaluation

Tier 0 Tier 1 Tier 2 Tier 3 Tier 4 Tier 5

Tier 0 00.2 0.6 0.8 1 1

Tier 1 0 0 0.3 0.8 1 1

Tier 2 -0.2 -0.2 0.2 0.6 1 1

Tier 3 -0.6 -0.6 -0.4 0.6 0.9 1

Tier 4 -1 -1 -0.6 -0.2 0.8 1

Tier 5 -1 -1 -0.8 -0.4 -0.2 1

Table 6.6.: Matrix of values for measuring the progress in terms of tier variation

Finally, for each framework (electricity supply and services, cooking facilities)

the two-dimensional PEPI is obtained by averaging, for each global attribute,

the progresses of the different categories. Particularly,

PEPIsupply =







∆(SR,1) + ∆(SR,2) + ∆(SR,3)

∆(SA,1) + ∆(SA,2) + ∆(SA,3)

∆(SS,1) + ∆(SS,2)







(6.1)

and

PEPIcooking =







∆(CAv,1) + ∆(CAv,2) + ∆(CAv,3)

∆(CAf,1) + ∆(CAf,2) + ∆(CAf,3)

∆(CS,1) + ∆(CS,2)







.(6.2)

6.1.5. Computation of the Progress Matrix

The coefficients displayed in Table 6.6, measuring the progress from t1to t2,

have been computed according to the following considerations.

As in the previous Section, let us denote with t1and t2the tier rankings obtained

in two different evaluations for a particular attribute.

•If t1< t2, it is assumed that the (positive) value of the coefficient depends

on the difference t1−t2, taking the value 0 if t1=t2and approaching 1

if t1= 5 (maximum tier). Moreover, increasing the value of t1(i.e., if the

starting tier is higher), the function measuring progress must approach

value 1 at a faster rate. In detail, the following function is considered (if

126

The PEPI Toolkit for MFIs 6. Development of the PEPI

t2> t1)

f+(t1, t2)=1−e−1+t2

5(t2−t1)2

.(6.3)

The profiles of f(t1, t2), for different t1are depicted in Figure 6.2.

•If t1=t2(no progress), the progress coefficient still takes a positive value

if the initial tier is higher than 2. This is needed in order to avoid biasing

the final results when households remain in high tiers (e.g., 3, 4 or 5)

in both evaluations, thus giving a positive value based on the ability of a

household to maintain its energy access conditions. Specifically, remaining

in Tier 0 or in Tier 1 will be evaluated with zero, while, for t1>1 the

following function is used:

f0(t1) = 1 −e−(t1−1)2

5.(6.4)

As observed above, (6.4) can be interpreted as the value, in terms of

progress, of remaining in tier t1.

•In the case of a negative variation (t1> t2) the coefficient is computed as

the difference between the values of f0(t2) and f0(t1).

0 1 2 3 4 5

0.2

0.4

0.6

0.8

t1 = 0

t1 = 1

t1 = 2

t1 = 3

t1 = 4

Figure 6.2.: Profile of the f(t1, t2) (6.3) for different values of t1(initial tier),

depending on t2(final tier)

In summary, Table 6.6 has been obtained via the following rules:

f(t1, t2) =











1−e−1+t2

5(t2−t1)2

t1< t2

1−e−(t1−1)2

5t1=t2, t1>1

0t1=t2= 0,1

e−(t1−1)2

5−e−(t2−1)2

5t1> t2, t2>0

e−(t1−1)2

5−1t1> t2, t2≤1

(6.5)

rounding the outcome to the first decimal digit.

127

The PEPI Toolkit for MFIs 6. Development of the PEPI

6.1.6. Summary: Properties of the PEPI framework

The main properties of the proposed framework and progress measure tool can

be summarized as follows:

•The PEPI allows the improvements across different attributes to be fol-

lowed and for the variations between the fulfilment of attributes to be

better characterized. This is the result of considering the new classifica-

tions in two different frameworks and of monitoring the progress via the

indices (6.1) and (6.2) at the household level (instead of considering only

the lowest tier among all attributes).

•The measure of the improvement depends less on the particular definitions

(thresholds) of the different tiers, than the static tier-ranking based on

the lowest tier. The attribute variations are renormalized by the tier-

ranking and by the progress matrix, making the PEPI less sensitive to

modifications of the thresholds underlying the tier definitions.

•By considering different thresholds (e.g., varying thresholds at regional

level, depending on climatic conditions or cultural traditions), the PEPI

can be used to compare and monitor progress in different areas.

•Therefore, the dashboard of indicators provides analysts with critical per-

spectives on where improvements should be made, measuring whether

policies and interventions have achieved their goals and recovered those

pitfalls of quality access at household level.

6.2. PEPI Framework Assessments

The methodology of the PEPI described in Section 6.1 has been applied to the

sample of microfinance clients considered in Chapter 5, evaluating the global

attributes for electricity supply and services and cooking solutions. The calcula-

tion has also included the proposed modifications of the frameworks (reported

in blue in Tables 6.4 and 6.5) and, for each global attribute, the lowest tier

among the individual indicators entailing each group has been selected. The

following sections are dedicated to the results for the individual attributes of

the PEPI frameworks. In particular, Tables 6.7–6.9 and Figure 6.5 report the

detailed tier ranking of the sample.

6.2.1. Electricity Access at the Household Level

The attribute duration is kept with its original thresholds providing the same

results. By expanding the categories describing the remaining attributes, re-

sults demonstrate a broader allocation of tier assignment. This is the case of

reliability and quality; in the former, by extending the range of number of out-

ages per week and their duration, the sample was further classified in inferior

tiers as tier 3. In the latter, households, whose electricity supply faced voltage

fluctuations throughout the year were allocated to tier 3 and those with risk and

damages to tier 2. Furthermore, according to the affordability of the electricity

128

The PEPI Toolkit for MFIs 6. Development of the PEPI

supply, while previously most of the population was found to be in tier 4 or

5, the graph 5.13, depicted in chapter 5, shows the distribution of the afford-

ability, suggesting an increasing variability below and above the 5% threshold.

The same can be seen in the assignments of the tier-ranking varying from tier

5 to tier 2. In regards to the capacity of the electricity supply it is estimated

based on the electricity services the household makes use of, rather than on its

consumption or daily capacity. The sample is more condensed in tiers 4 and 3.

Considering that the whole population was connected to the grid, the energy

source did not cause any health hazard and the safety attribute kept its values

from the MTF [Bhatia and Angelou, 2015] tier ranking results.

The rankings for the single attributes are provided in Table 6.7, while Figure 6.3

shows the lowest-based ranking, comparing the results with the energy supply

ranking of the MTF (see Chapter 5). The lowest results are obtained for the af-

fordability, with the majority of households ranked in tiers 3 and 5. Comparing

with Table 6.7, the low rankings is shown to depend mainly on capacity (lack

of services). The reliability is mainly a concern for rural households (majority

of households below tier 3), with 10% of households reporting a short duration

of the supply and about 30% affected by unpredictable interruption and quality

issues. Finally, almost all households achieved tier 5 in regards to the attribute

safety.

Further conclusions can be drawn comparing the results with the ESMAP MTF

ranking (see Figure 6.3). While the ESMAP MTF lowest-based ranking identi-

fied about 30% of households lacking sufficient energy supply (40% in the rural

area), it is worth noticing that considering the attributes separately yields a

clearer picture of the missing services, in particular, in terms of low reliability

in the rural areas.

Sensitivity Study and Relation with PPI

As in the previous chapter, the last part of the framework assessment focuses on

a sensitivity study, computing the tier ranking and leaving out single attributes.

Notice that safety has not been considered as almost all households achieved

the highest ranking in both sub-attributes (see Table 6.7). The results, shown

in Figure 6.4, confirm that the ranking is most sensitive to capacity. Namely,

leaving out this attribute yields a shift of households from tiers 3 and 4 to 5,

in both urban and rural areas. With respect to the global attribute of Reliabil-

ity, the sub-attributes reliability and quality appear to be the most important.

Particularly, leaving out quality increases of about 10% the households in tier

Finally, Table 6.8 provides the Spearman correlation coefficients for the three

tier rankings (Affordability, Reliability, Safety) with respect to the PPI data,

showing that the PEPI attributes are not strongly correlated with the PPI score.

Specifically, while capacity shows the highest correlation (0.2, in line with the

results obtained for the ESMAP MTF), the attribute quality, which resulted to

be very important in the reliability ranking, seems to be uncorrelated with the

poverty index.

129

The PEPI Toolkit for MFIs 6. Development of the PEPI

Area Departments

Total Urban Rural Nari˜no Huila Putumayo Tolima

Reliable

Reliability(short unpredictable interruptions)

Tier 5 65.6% 72.6% 55.2% 51.0% 81.6% 80.3% 66.7%

Tier 4 6.5% 7.3% 5.2% 6.3% 2.6% 15.2% –

Tier 3 5.7% 5.7% 5.8% 7.8% 2.6% 4.6% 8.3%

Tier 1 22.1% 14.3% 33.8% 34.9% 13.2% – 25%

Duration(including day and night electricity supply duration)

Tier 5 68.0% 75.2% 57.1% 54.2% 84.2% 78.8% 75.0%

Tier 4 11.2% 7.8% 16.2% 14.6% 10.5% – 25.0%

Tier 3 5.2% 4.3% 6.5% 7.3% 0.9 7.6% –

Tier 2 5.5% 4.3% 7.1% 7.8% 1.8 6.1% –

Tier 1 3.6% 4.3% 2.6% 4.7% – 7.6% –

Tier 0 6.5% 3.9% 10.4% 11.5% 2.6% –

Quality(damages or risks due to voltage fluctuations)

Tier 5 70.3% 76.5% 61.0% 59.4% 83.3% 81.8% 58.3%

Tier 3 10.4% 9.6% 11.7% 10.4% 9.6% 6.1% –

Tier 2 19.3% 13.9% 27.3% 30.2% 7.0% 12.1% 41.7%

Affordable

Affordability(electricity expenses below defined thresholds of HH income)

Tier 5 82.0% 86.9% 74.6% 78.1% 81.2% 93.1% 100%

Tier 4 11.6% 8.2% 16.7% 15.6% 9.4% 3.5% –

Tier 3 4.6% 2.9% 7.2% 4.7% 8.2% – –

Tier 2 1.7% 1.9 % 1.5% 1.6% 1.2% 3.5% –

Capacity(enabled services)

Tier 5 19.8% 23.9% 13.6% 28.1% 1.7% 30.3% –

Tier 4 42.7% 46.5% 37.0% 30.7% 64.9% 36.4% 58.3%

Tier 3 33.6% 27.4% 42.8% 36.5% 28.9% 33.3% 33.3%

Tier 2 2.1% 1.3% 3.2% 2.1% 2.6% – 8.3%

Tier 1 1.3% 0.8% 1.9% 2.6% – – –

Tier 0 0.5% – 1.2% – 1.7% – –

Legality(legal connection and costs)

Tier 5 99.2

%100 % 98.1

%100% 97.4% 100% 100%

Tier 3 0.8% – 1.9% – 2.6% – –

Safe

Health(absence of diseases or health hazards)

Tier 5 100% 100% 100% 100% 100% 100% 100%

Safety(absence of accidents or risks)

Tier 5 99.2% 100% 98.0% 100% 97.4% 100% 100%

Tier 3 0.8% – 2.0% – 2.6% – –

Table 6.7.: PEPI Framework: Results of the tier ranking for electricity supply

130

The PEPI Toolkit for MFIs 6. Development of the PEPI

Tier

543210

Frequency (%)

E Supply (GTF)

Reliability

Affordability

Safety

Tier

543210

Frequency (%)

100

E Supply (GTF)

Reliability

Affordability

Safety

Tier

543210

Frequency (%)

E Supply (GTF)

Reliability

Affordability

Safety

Figure 6.3.: Comparison of the results of the ESMAP MTF tier ranking and of

the PEPI Frameworks tier ranking for the total sample (top) and

for urban (bottom left) and rural (bottom right) areas

6.2.2. Access to Cooking Facilities

Through the PEPI frameworks, the original attributes convenience and qual-

ity were kept with the original thresholds provided by the ESMAP MTF. The

attributes availability and affordability were expanded in further thresholds for

tier 3. The results displayed in Table 6.9 show minimal portions of the sample

assigned in the lowest tiers, though still providing a clearer view of the condi-

tions of fuel access and affordability of the households. Considering that the

framework has excluded those attributes to be assessed by a third party (lab-

oratory), the attribute safety has been modified including an indicator of the

cooking place and its ventilation. Both of these can provide a picture of the

health hazards that might attempt to affect the household when using tradi-

tional stoves, in addition to the risks or accidents the household has had related

to the cooking fuel or stove. As observed in 6.9, around 20% of the households

have poor ventilation, though are better off in terms of safety; marginal portions

cook inside the sleeping area and almost none have had any accident.

131

The PEPI Toolkit for MFIs 6. Development of the PEPI

Tier

543210

Frequency (%)

80 w/o Affordability

w/o Capacity

w/o Legality

Affordability (PEPI)

Tier

543210

Frequency (%)

60 w/o Reliability

w/o Duration

w/o Quality

Reliability (PEPI)

Tier

543210

Frequency (%)

80 w/o Affordability

w/o Capacity

w/o Legality

Affordability (PEPI)

Tier

543210

Frequency (%)

w/o Reliability

w/o Duration

w/o Quality

Reliability (PEPI)

Tier

543210

Frequency (%)

70 w/o Affordability

w/o Capacity

w/o Legality

Affordability (PEPI)

Tier

543210

Frequency (%)

50 w/o Reliability

w/o Duration

w/o Quality

Reliability (PEPI)

Figure 6.4.: Sensivity study for the PEPI energy supply framework: tier rank-

ing for affordability (left) and reliability (right), leaving out single

attributes. Top row: total sample; middle row: urban area; bottom

row: rural area.

The rankings for the single attributes are provided in Table 6.9, while Figure 6.5

shows the lowest-based ranking, comparing the results with the energy supply

ranking of the ESMAP MTF (see Chapter 5). The main issues are related to

the availability of fuel, affecting 50% of rural households. A further notable

aspect is that about 20% of urban households appears to be affected by safety

issues (lack of ventilation, see Table 6.9).

Comparing the results of both household’s energy access assessments, the ESMAP

MTF lowest-based ranking divided mainly the households in two categories

(80% of households are split between tier 0 and 5), while the PEPI framework

is more able to identify the different issues, also considering the differences be-

132

The PEPI Toolkit for MFIs 6. Development of the PEPI

Spearman correlation vs. PPI

Affordability (lowest tier) 0.18

Affordability 0.07

Capacity 0.21

Legality 0.11

Reliability (lowest tier) 0.14

Reliability 0.18

Duration 0.15

Quality 0.01

Safety (lowest tier) 0.04

Safety 0.04

Health 0.04

Table 6.8.: Spearman correlation coefficients of energy supply tier ranking (for

the three main attributes and the corresponding sub-attributes) and

PPI score

tween rural and urban areas. Figure 6.5 compares the results with the ranking

obtained using the ESMAP MTF, the depiction of the attributes assessment of

the total sample, and in urban and rural areas it shows the range and diversity

of information unobserved under a single metric.

Sensitivity Study and Relation with PPI

As a further step a sensitivity study is performed, computing the tier ranking

leaving out single attributes in the category availability. The global attribute

safety has not been considered as almost all households achieved the highest

ranking in the sub-attribute safety and the ranking mainly depends on the

health attribute (see Table 6.9). while the attribute affordability has not been

considered as it does not have sub-attributes. The results show that the con-

venience is the attribute which determines at most the lowest tier. When the

attribute is not considered, about 25% of households in the total sample (and

about 50% of rural households) are shifted to the highest tier.

Finally, Table 6.10 provides the Spearman correlation coefficients for the three

tier ranking (availability, affordability, safety) with respect to the PPI data,

showing that the PEPI global attributes are not strongly correlated with the

PPI score. Particularly, only convenience shows a large correlation with PPI

(0.28, in line with the results obtained for the ESMAP MTF).

133

The PEPI Toolkit for MFIs 6. Development of the PEPI

Area Departments

Total Urban Rural Nari˜no Huila Putumayo Tolima

Available

Convenience(time spent in stove and fuel acquisition and preparation)

Tier 5 57.2% 76.5% 27.7% 40.0% 70.3% 77.3% 91.7%

Tier 4 16.0% 15.0% 17.6% 21.6% 5.4% 21.2% –

Tier 3 6.4% 5.8% 7.4% 12.4% – 1.5% –

Tier 2 2.7% 1.8% 4.1% 5.4% 24.3% – 8.3%

Tier 1 17.6% 0.9% 43.2% 20.5% – – –

Availability(fuel availability throughout the year)

Tier 5 85.7% 82.2% 90.9% 94.8% 81.6% 63.6% 100%

Tier 4 9.4% 13.0% 3.9% 5.2% 3.5% 33.3% –

Tier 3 3.6% 3.5% 3.9% – 10.5% 3.0% –

Tier 2 1.3% 1.3% 1.3% – 4.4% – –

Quality(absence of heat variation of fuel)

Tier 5 96.8% 97.8% 95.3% 93.5% 100% 100% 100%

Tier 3 3.2% 2.2% 4.7% 6.5% – – –

Affordable

Affordability(fuel and stove expenses below defined thresholds of HH income)

Tier 5 70.2% 78.0% 58.2% 65.4% 70.0% 81.8% 83.3%

Tier 3 24.7% 18.9% 33.6% 28.67% 25.4% 13.6% 16.7%

Tier 2 5.1% 3.1% 8.2% 5.9% 4.6 4.6% –

Safe

Safety(cooking place outside of sleeping area and no accidents)

Tier 5 97.3% 96.0% 99.3% 99.0% 97.4% 93.9% 91.7%

Tier 3 0.3% 0.4% – 0.5% – – –

Tier 2 2.4% 3.6% 0.7% 0.5% 2.6% 6.1% 8.3%

Health(sufficient ventilation)

Tier 5 58.0% 48.5% 72.5% 83.2% 28.3% 43.9% 25.0%

Tier 4 23.9% 29.5% 15.4% 16.8% 31.9% 27.3% 41.7%

Tier 3 18.1% 22.0% 12.1% – 39.8% 28.8% 33.3%

Table 6.9.: PEPI Framework: Results of the tier ranking for cooking facilities

134

The PEPI Toolkit for MFIs 6. Development of the PEPI

Tier

543210

Frequency (%)

Cooking (GTF)

Availability

Affordability

Safety

Tier

543210

Frequency (%)

70 Cooking (GTF)

Availability

Affordability

Safety

Tier

543210

Frequency (%)

Cooking (GTF)

Availability

Affordability

Safety

Figure 6.5.: Comparison of the results of the ESMAP MTF tier ranking (cook-

ing solutions) and of the PEPI Frameworks tier ranking for the

total sample (top), the urban area (bottom left) and the rural area

(bottom right)

Spearman correlation vs. PPI

Availability (lowest tier) 0.18

Convenience 0.28

Availability -0.12

Quality 0.02

Affordability (lowest tier) 0.14

Safety (lowest tier) -0.1

Safety -0.04

Health -0.08

Table 6.10.: Spearman correlation coefficients of cooking solution tier ranking

(for the three main global attributes and the corresponding sub-

attributes) and PPI score

135

The PEPI Toolkit for MFIs 6. Development of the PEPI

Tier

543210

Frequency (%)

80 w/o Convenience

w/o Availability

w/o Quality

Availability (PEPI)

Tier

543210

Frequency (%)

w/o Convenience

w/o Availability

w/o Quality

Availability (PEPI)

Tier

543210

Frequency (%)

80 w/o Convenience

w/o Availability

w/o Quality

Availability (PEPI)

Figure 6.6.: Sensivity study for the availability leaving out single sub-attributes.

Top: total sample; Bottom-left: urban area; Bottom-right: rural

area

136

The PEPI Toolkit for MFIs 6. Development of the PEPI

6.3. A Toolkit for Assessment of Energy Services

Within the goals of the PEPI, the provision of practical tools to the microfinance

industry aims at enabling MFIs to identify the electricity and cooking needs

of their clients and at allowing them tracking the effects of their green lending

programs by measuring the progress out of energy poverty.

Figure 6.7.: Components of the PEPI Toolkit

To this purpose, the PEPI frameworks and the progress measure described in

Sections 6.1.3 and 6.1.4, respectively, are embedded in a PEPI toolkit, contain-

ing a precompiled survey, designed to facilitate the field data collection over

time, as well as data management and analysis.

Specifically, by avoiding text documents with narrative surveys and the use of

multiple, separate files and tools for data collection, data cleaning and analysis,

the toolkit enables the user (i.e., the surveyor) to apply the tool directly on

the field. Moreover, the gathered data is automatically organized in a single

database, in order to simplify the final format and to unify the data analysis

and the interpretation of the results. Algorithms for tier ranking calculation

are integrated in the data analysis R file, fed by the database collected in the

field.

As ultimate goals, this toolkit aims at contributing to the MFIs effort in in-

novating in products and services and striving the access to clean energy tech-

nologies through microfinance services by providing a measure of microfinance

performance in contributing to the SDGs. Furthermore, the toolkit can serve

as a model for develop further extension to identify, measure and monitor the

progress and performance of other basic-needs related microloans, such as for

loans sustainable housing, education, health, water, and transportation.

137

The PEPI Toolkit for MFIs 6. Development of the PEPI

The PEPI Toolkit developed in the framework of this research can be down-

loaded upon request from nataliarealpecarrillo.weebly.com.

6.3.1. The PEPI Survey

Keeping energy applications at the core of the approach in order to ensure

its relevance to users [Nussbaumer et al., 2011,Bhatia and Angelou, 2014,Groh

et al., 2016], the PEPI toolkit entails a comprehensive survey, contained in

multiple Microsoft Excel R

sheets, for electricity and cooking facilities access

assessment. The focus of the electricity supply is based on the enabled services

the electricity supply for the household and its quality. Meanwhile, the access

to cooking facilities builds on the types of cookstoves the household uses.

The survey is based on the ESMAP MTF [Bhatia and Angelou, 2015] modified

as described in Section 6.1 and on the corresponding tier-ranking, as previously

described in Tables 6.4 and 6.5. Moreover, the tool enables the modification

of the thresholds defining the tier-ranking. and the comparison of the results

with the ones obtained using the original ESMAP MTF thresholds. The survey

tools allow to store data in a fill-in table, which, combined with a periodical

use of the tool, allows to easily compute the evolution of energy access based

on the designed PEPI (see also Table 6.6).

Figure 6.8.: PEPI Survey Sketch and Components

The toolkit builds on MFIs ability to periodically collect large amount of so-

cial, financial and demographical data for assessment and monitoring of their

client. Hence, the toolkit survey focuses exclusively on capturing information

related to electricity and cooking solutions at household level, excluding data

which can be extracted from the database of the institution (social and financial

data). Whether institutions decide for implementing the tool with its complete

clientele or a selected sample will depend on their capacities and resources. In

138

The PEPI Toolkit for MFIs 6. Development of the PEPI

either cases, the toolkit can be viewed as a marginal extension in the amount

of data to be collected, which, however, will allow MFIs to have a broader

overview of the scale of energy access of the target population and make data

available for policy design and internal strategic objectives. On the other hand,

correlations and data analysis combining both socio-economical and electricity

and/or cooking data cannot be integrated in the toolkit as they require the

institutional databases. However, organizations interested in such analysis can

extract the PEPI database and combine it with their own database for these

purposes.

Improved Data Collection and Management

The PEPI survey simplifies data collection, management and analysis under

different aspects. The properties of the toolkit can be summarized as follows.

•Labelled questions: Each question of the survey is associated to an

attribute and a global attribute. This ensures that all the questions of

the survey are relevant for the final assessment and enhances transparency

on the attributes assessed per question.

•Questions and associated variables: The variables derived from each

question are already defined and labelled. Hence, the surveyor/organization

does not need to create an additional file or sheet for data collection nor

label the specific variables. The fill-in tables have hidden the variables

column between the questions and the fill-in spaces for the answers. The

column displaying the variables is only used for the subsequent statistical

and econometrical analysis.

•Multiple choice answers: Each question has a predefined list of possi-

ble answers, i.e., there are no open questions. All the answers are given in

numbers or a numerical value; whether categories are titled with numbers

or the variable is integer with a specific range.

•Flexible thresholds: One of the main differences with the ESMAP

MTF methodology [Bhatia and Angelou, 2015]) consists in the possibil-

ity of modifying the tier thresholds. Specifically, taking into account the

on-going debate on attributes [Groh et al., 2016], their cut-offs and their

relevance [Stevens et al., 2015], the PEPI toolkit integrates a dashboard

enabling modifications according to the established limits for each tier

assignment. These frameworks can be updated according to the develop-

ments of the sector and to the geographical region of interest.

•Association of answers with tier ranking: Following the thresholds,

depicted in the corresponding PEPI framework matrices, the tier assign-

ment can be automatically estimated.

•Fill-in table design: The toolkit survey provides the questions in rows,

while subjects data is collected in columns. Thus, the fill-in table vali-

dates answers according the established ranges of each questions as quality

control. Moreover, observing the individual answers in different columns

facilitates the comparison between subjects, as well as helps to continu-

139

The PEPI Toolkit for MFIs 6. Development of the PEPI

ously validate the data collection.

•Efficient survey implementation: Thanks to the filtering of questions

related to the power sources and the cooking devices used by the house-

hold, the questionnaire avoids skipping patterns, thus the survey can be

efficiently implemented in 5 to 20 minutes. Moreover, the absence of skip-

ping patterns reduces errors due to misunderstanding of the questionnaire

by the implementer. As an example, households without electricity (using

kerosene, candles, or similar light sources) only answer a maximum of 13

questions, while grid-connected households answer up to 40 questions.

•Formatted database At last, the fill-in table can be easily transposed

(unhiding the column with the variables) and exported (e.g., as CVS or

text file), in order to be read into an external program for data analysis

(e.g., R,Stata,SPSS). Hence, the toolkit can be used within an automatic

workflow to deliver the final results.

6.3.2. How to

The following part describes in detail the steps for the implementor to take

in regards to the contents and objectives of the different components of the

Excel-based toolkit.

I. Required Sample Size

As a preliminary step, the toolkit contains a precompiled sheet for determining

the most convenient sample size for the data collection. Particularly, the user

has to provide the desired margin of error and the confidence level, and the tool

computes the sample size based on the selected statistical significance and on

the size of the total population (see also Section 5.1 for more details).

II. Tiers Thresholds Matrices

The tier ranking depends on several thresholds, which must be set for each at-

tribute. The established thresholds incorporate characteristics of MES perfor-

mances in order to better assign them. Frameworks from ESMAP MTF [Bhatia

and Angelou, 2015] for electricity supply, electricity services, electricity con-

sumption and cooking facilities are also described indicating the thresholds

within each attribute. The integration of these frameworks aim at showing

the user the differences between the ESMAP MTF and the PEPI framework

matrices displayed in the next sheet. In the PEPI toolkit, the ESMAP MTF

thresholds are used as default values, with slight modification motivated by the

recommendations of [Groh et al., 2016] previously described (see Section 6.1).

However, in the dedicated sheet, the threshold for each attribute can be set

independently and adjusted to the local conditions.

140

The PEPI Toolkit for MFIs 6. Development of the PEPI

III. PEPI Matrices

Within the toolkit provided in the Excel file, the PEPI framework matrices

for the assessment of access to electricity and to cooking facilities access are

depicted. As the user can observe, the PEPI is based on the attributes of

the ESMAP MTF frameworks, with the above mentioned modifications of the

thresholds and considering only a single framework for measuring electricity

access. All modifications of the thresholds with respect to the ESMAP MTF

published in [Bhatia and Angelou, 2015] are marked in blue (see Tables 6.4 and

6.5).

IV. Questionnaire Electricity Services and Supply

In order to assess the quality of access to electricity, the designed questionnaire

filters at first place the source of power at the household, for which the enabled

services and the quality of performance of energy supply are assessed. All ques-

tions are related to specific attributes measuring the quality of energy access,

associated also with the described global attributes (see excerpt in Figure 6.9).

Thus, the methodology allows the analysis of energy services costs depending on

the energy supply source and the related costs of the (multiple) power source(s)

of the household.

V. Questionnaire Cooking Facilities

Mirroring the assessment of electricity supply and services, access to cooking

facilities is assessed through the lens of the cooking stoves the household use

for the preparation of its meals. Excluding the attributes to be measured at

laboratory and considering only the proposed in the PEPI frameworks, all ques-

tions are as well related to the corresponding attributes and global attributes

(see excerpt in Figure 6.10).Through this methodology, multiple usage of cook-

ing stoves and their corresponding fuels can be easily analyzed as well as their

respective quality of performance.

VI. – VII. Fill-in Surveys (Electricity Supply and Services and Cooking

Facilities)

The fill-in sheets are aimed for the ease of data collection in the field. By choos-

ing first the power supply or cooking stove used at the household, respectively,

the interviewer only captures the relevant information for the household as-

sessment avoiding instructions for skipping patterns. All variables are numeric,

whether categorical or continuous. Excerpts of the surveys are displayed in

Figures 6.11 and 6.12, for electricity access and cooking solutions assessment,

respectively. While the questionnaire shows the relationship between each ques-

tion, attributes and global attributes, the fill-in sheets serve the purpose to

directly respond to the specific survey.

141

Figure 6.9.: Excerpt of PEPI Questionnaire indicating the respective attributes to be assessed - Access to Electricity Supply and

Electricity Services

Figure 6.10.: Excerpt of PEPI Questionnaire indicating the respective attributes to be assessed - Access to Cooking Facilities

Figure 6.11.: Excerpt of PEPI Survey tool to assess access to electricity supply and electricity services

Figure 6.12.: Excerpt of PEPI Survey tool to assess access to cooking facilities

7.Conclusions

The SDGs incorporated in 2015 identify “energy access for all” as part of the

global goals to be adopted by all nations, recognizing that universal energy

access is an imperative goal to enable global development.

Besides developing technical and financial solutions aimed at improving the

access to energy, the ability to measure energy poverty rates and to determine

the quantity and quality of energy access of the population with insufficient and

unsatisfactory access is necessary . Governments, development communities,

NGOs, private companies and financial institutions require this information in

order to identify policies and potential projects tailored to the energy needs of

populations, to track the progress and evolution of energy access over time and

along programs implementation, and to identify the effects of interventions and

other initiatives that are intended to alleviate energy deprivation.

Traditional energy access assessments, specifically for electricity and cooking

facilities, are often reduced to the availability of a physical grid connection,

or on the dependence of biomass for cooking. Thus, the true value of energy

of energy access and variations across its many attributes are not discussed.

By assessing the quality of the connection, the role of decentralized electricity

access approaches and the efficient use of fuels for cooking, among other fac-

tors, innovative metric systems research has led to more detailed definitions,

acknowledging the multidimensional nature of energy access and the need of

multiple frameworks to capture the various attributes thereof.

The growing consensus on the definition of energy poverty and the demand for

practical but comprehensive measurement tools across the diverse attributes

of energy access reveals the potential of decentralized MES for fulfilling en-

ergy needs, transforming global energy discourse. Nevertheless, the potential

for the dissemination of MES has to date strongly depended on the potential

for providing customized financial approaches to tackle affordability challenges.

Among a range of possibilities, green microfinance strategies enable MFIs to dis-

tribute MES by tackling affordability and sustainability challenges through so

called green lending programs. Through these programs, following a two-hand

model approach, financial institutions and energy service suppliers establish

partnerships in order to develop financial products to enhance access to specific

technologies and meet latent energy needs.

The benefit of these intersectorial partnerships, between the financial and the

energy sectors, is twofold. If, on the one hand, MFIs can play a relevant role

147

The PEPI Toolkit for MFIs 7. Conclusions

in the achievements of the SDG 7 by improving access to renewable and effi-

cient energy technologies. Furthermore, noting that MFIs are condemned to

constantly innovate in order to remain attractive to clients, green loans are a

promising avenue to consolidate or expand market share. Green lending can

be also seen as a model for other “directed” microloans, whether targeted to

housing, education, health, transport or water (as in the case of the partner

MFI Contactar) among other basic needs.

In accordance with these considerations, the focus of this dissertation is the

assessment of household energy access and needs in the context of green mi-

crofinance, with the ultimate goal of supporting MFIs with dedicated decision-

making tools. To this end, the first part of the thesis consisted of an introduc-

tion concerning energy poverty, discussing its definition and relevant metrics,

and concerning the role of microfinance in tackling energy access (via MES dis-

semination through green lending programs). The outcomes of a case study in

Southern Colombia where then presented, introducing the profile of the Colom-

bian MFI Contactar and the integration of its green lending program; describing

the detailed results of this program through application of the ESMAP multi-

tier frameworks for electricity and cooking on a sample of microfinance clients

(households). Besides the analysis of energy access in this particular context,

the case study aimed at testing the potential and limitations of the ESMAP

MTF approach, and at assessing the ability of poverty metrics to describe elec-

tricity supply and cooking fuels of the organization’s clients. Furthermore, it

provided the basis for a novel tool for measuring the provision and quality of

electricity supply and services and cooking facilities among microfinance clients.

Finally, this research aimed at introducing the Progress out of Energy Poverty

Index (PEPI) toolkit, initially targeted to the green microfinance industry. This

toolkit is designed to be used for measuring the energy access at household level

in regards to electricity supply and services as well as cooking solutions, in order

to facilitate green lending programs design, implementation and tracking.

7.1. Results and Main Contributions

7.1.1. Assessment of Energy Access of Microfinance Clients

The field research performed in collaboration with Contactar provided a de-

tailed picture of electricity and improved cooking facility access of the clients,

portrayed by the ESMAP MTF frameworks [Bhatia and Angelou, 2015]. At the

same time, the implementation of the case study provided relevant experience

concerning the requirements of an assessment tool in order to be successfully

used by the microfinance industry, in terms of practical usage and in terms of

the relevance of the different attributes to assess.

From the point of view of the electricity supply, services and consumption, the

case study revealed that the majority of households are in high tiers (i.e., well

performing) for most of the attributes. However, among all attributes, the

capacity of electricity services appeared to be the one that in most cases deter-

148

The PEPI Toolkit for MFIs 7. Conclusions

mined a lower overall tier ranking. Moreover, although the physical access to

electricity was almost completely legal, several rural households reported quality

problems (related to voltage problems and/or to unexpected service interrup-

tions), hence demonstrating that electricity access assessment should consider

multiple dimensions, rather than the mere existence of a grid connection.

Concerning the access to cooking solutions, the analysis based on the ESMAP

MTF resulted in a higher number of households ranked in Tier 0 due to the

limited affordability of the cooking solution. In this case, this is strongly de-

termined by the thresholds set in the ESMAP MTF, for which a household is

ranked in Tier 0 whenever the fuel costs overshoot 5% of the monthly income,

hence not taking into account the possibility of financing via, for example, mi-

crocredits or mobile banking. Moreover, rural households are also affected by

issues related to the convenience of the cooking solution (time spent for cook-

ing fuel acquisition and preparation), revealing the lack of suitable technologies

according to fuel availability.

Overall, the obtained access indices showed large differences between the urban

and the rural areas, quantifying to which extent the lack of infrastructure might

affect the access to energy (especially in the case of cooking solutions).

The study cross-analyzed the energy access assessment and the poverty metric

PPI panel data available for each client from the database of the MFI. The

results revealed only a relatively low correlation between the tier-ranking of a

household in terms of energy access and the probability that it lies below the

poverty line. Specifically, the electricity consumption, the capacity of services

and the convenience of cooking solution resulted to be the variables with the

most correlation with PPI data.

Beyond the detailed analysis of energy access of the microfinance clients, the

successful collaboration with an MFI has to be considered one of the major out-

comes of the case study. Particularly, the preparation of the case study required

a close collaboration with the Social and Environmental Performance Manage-

ment Department of Contactar, which benefitted the quality of this research.

Thanks to the deep level of knowledge, commitment and to the valuable prac-

tical experience of Contactar employees, the field research methodology, time

plan and logistics were smoothly and efficiently organized. At the same time,

Contactar considered the study as an important opportunity to innovate and

to stand ahead in their green initiative. These aspects further motivated the

institution in developing tools to monitor its achievements.

Moreover, since the data collection was planned and carried out in collabo-

ration with recent economics graduates of the local university (Universidad de

Nari˜no), the case study also represented an important opportunity to strengthen

the link between practitioners and academia. Contactar has been invited by the

PIFIL1group of the Universidad de Nari˜no and by the Universidad Pontificia

Bolivariana from Medell´ın to present the results of the PEPI assessment, and

has applied to the Citi Award for Microentrepreneurs and to the Master-Card

1Plan de Investigaci´on para el Fortalecimiento Integral de las Comunidades (Research Plan

for the Strengthening of Communities)

149

The PEPI Toolkit for MFIs 7. Conclusions

Client Centric Awards with its green initiative and tools development.

7.1.2. Development of the PEPI Toolkit

The main motivation for the development of the PEPI toolkit lies in the po-

tential of green microfinance to improve electricity and cooking solutions access

and in the fact that, despite the existing debate concerning the long-term sus-

tainability and successful upscaling of projects combining energy access and

microfinance schemes, there is an increasing number of MFIs disbursing green

loans. As the concept of green inclusive microfinance gains increasing attention,

two clear needs arise, including; the development of different metrics to track

performance of MFIs in the broad spectrum of environmental management and

a clear methodology to assess the performance of green loans in enhancing en-

ergy access. Henceforth, the PEPI toolkit was conceived and designed to fill

this methodological gap.

Within this research, the development of the toolkit consisted of two main steps.

On the one hand, the existing energy poverty metrics and green microfinance

indicators have been reviewed, highlighting their main properties. On the other

hand, the relevant practical aspects of the field study in collaboration with

Contactar have been taken into account, aiming at developing an efficient toolkit

to be used by MFIs to collect client data.

Among the available methodologies for the assessment of energy access, the

ESMAP MTF provides by far the most complete set of attributes to characterize

the different dimensions of energy. Acknowledging the qualities of the MTF, the

PEPI has been built from this multi-tier approach, taking into account a set of

modifications motivated by recent applications of the MTF in different contexts

(see, e.g., [Groh et al., 2016]) and aligning the considered attributes with the

Energy SDG targets. Specifically, the PEPI framework retains the attributes

of the original ESMAP MTF, grouping them in global attributes (reliability,

affordability and safety for assessing electricity supply and services; availability,

as well as safety and affordability for assessing cooking facilities) reflecting the

sub-targets of the SDG 7.

An important aspect of the PEPI is that the toolkit builds on the ability of

MFIs to periodically collect large amount of detaileddata at the household

level for monitoring their client. Particularly, aiming at being integrated as

a toolkit within the standard data collection of MFIs, the PEPI toolkit pro-

poses a methodology to quantify the progress out of energy poverty taking into

account the value of quantity and quality assessments (tier-ranking) of electric-

ity and cooking access.

From a practical point of view, the design of the PEPI toolkit is based on a cost-

effective approach in regards to its methodology: the ready to fill-out surveys

are contained in Microsoft Excel R

sheets, for electricity supply and services

(condensed) and cooking facilities access assessment, and the questions can be

easily filtered in order to focus on relevant aspects. Moreover, the outcomes can

be easily exported in different formats to be directly analyzed using professional

statistical programs.

150

The PEPI Toolkit for MFIs 7. Conclusions

The surveys take into account the attributes of the ESMAP MTF [Bhatia and

Angelou, 2015], whose thresholds have been modified taking into account the

aforementioned considerations on the frameworks including several features to

increase its flexibility, its interpretation and to favor its dissemination. Firstly,

each question is clearly associated to an attribute, ensuring that all the ques-

tions of the survey are relevant for the final assessment; enhancing transparency

on the attributes assessed per question and without skipping patterns. At the

same time, for each question the relevant variables for data analysis are la-

beled. Additionally, the user has the possibility to hide the variables column

in the fill-in tables between the questions, hence using the variables only for

the subsequent statistical and econometrical analysis. Secondly, taking into ac-

count the on-going debate on the role of attribute cut-offs and nature (binary

or graded) [Groh et al., 2016,Stevens et al., 2015], the PEPI survey toolkit in-

tegrates a dashboard enabling modifications according to the established limits

for each tier assignment, computing the resulting ranking accordingly. This

feature allows the user to update the frameworks according, for instance, to

the developments of a particular sector and/or to the geographical region of

interest. Finally, the fill-in table of the PEPI survey can be easily transposed

(unhiding the column with the variables) and exported (e.g., as CVS or text

file), in order to be imported into an external program for data analysis (e.g.,

R,Stata,SPSS), allowing to use the PEPI toolkit within an automatic workflow

to deliver the final results.

7.2. Outlook: a Tool for the Microfinance Industry

The SDG 7 targets must be linked with robust and well chosen indicators.

The adoption of multi-tier approaches makes it possible to reveal the realities

behind the traditional binary measures, i.e. the on-grid and off-grid population

and of those cooking with solid and non-solid fuels. To date, unfortunately,

national indicators and available statistics to measure and monitor the different

dimensions of energy access are extremely scarce, particularly for the least

developed countries and regions where the issue is the most pressing [Pachauri

et al., 2012a]. However, empirical evidence calls for further research on the role

of microfinance on improving energy access at household and MSMEs level in

order to; estimate its potential, to assess its impact, and to fulfil the data needs

of microfinance stakeholders. Thus it also assists global programs on green

lending.

The long-term goal of the PEPI toolkit is to support the microfinance indus-

try in the standardised self-monitoring and self-assessment of its achievements,

by providing a tailored easy to use toolkit and to integrate within existing

clients’ surveys. Moreover, it supports the management and the automatic post-

processing of the results. The latter can be both integrated within the processes

of the organization or outsourced to deliver a report of results. Through the

implementation of the PEPI in the microfinance sector, an industry specialized

in serving remote populations, the collected data will help to understand power

source and fuels usage and needs, and provide an extremely valuable picture of

151

The PEPI Toolkit for MFIs 7. Conclusions

the actions to be undertaken. This overview serves to develop a framework for

sectoral interventions and policy design towards realistic environmental goals.

Based on the proposal and on the results of this research, further steps towards

these goals can be listed as:

•Implementation of PEPI in the microfinance industry

Through the publication of the results of the case study and of the PEPI

toolkit, interested MFIs are invited to track the impact of their green

lending programs within their institutions. The toolkit provides MFIs

with a comprehensive approach to detail the electricity supply and services

and cooking facilities of the clientele at any stage of experience in green

lending, identifying, among others, further needs, uses and costs of their

clients, as well as payment capabilities.

•Assessment of energy needs for productive activities

Energy access, by improving living standards, has an effective and long-

lasting impact on livelihood and income generation. However, besides the

provision of energy services at the household level, productive activities

that yield improved income generation should also be assessed. In fact,

productive activities can have a positive impact on the economic and so-

cial benefits of energy access, while increasing the economic sustainability

of energy access projects as the ability to pay of end-users increases with

their newly generated profits [Etcheverry, 2003,Kapadia, 2004]. As de-

scribed in Chapter 4 such assessment is of utmost importance from the

point of view of MFIs when developing strategic plans for green microfi-

nance programs. Moreover, while, most of the research has been focused

on rural electrification, the role of thermal or mechanical energy has been

explored in a lesser extent. A further development of the PEPI toolkit

will be aimed at including an additional framework with a focus on pro-

ductive uses of energy, taking into account different sources and different

uses of energy.

•Support the testing of the PEPI toolkit

Future research will be dedicated to developing a set of automatic analysis

and reporting templates, in order to process more efficiently the collected

data and support the continuous implementation of the survey among

existing and new microfinance clients. Firstly, these tools will be impor-

tant to test the potential and the limitations of the proposed metric and

framework, both in terms of their capability of assessing energy needs and

in terms of practical implementation. At the same time, MFIs willing to

share information and learned lessons will be encouraged to publish their

results and comments, making them available for academia and interested

stakeholders.

•Capacity building and empowerment process for MFIs

Given the complexity of such initiatives, typically, intensive technical as-

sistance is required from development institutions to assist MFIs in their

approach to the topic of energy [von Wolff and Phalpher, 2014]. For im-

plementation and results monitoring, technical and financial assistance

152

The PEPI Toolkit for MFIs 7. Conclusions

should be provided to help guide MFIs in implementation (e.g. stan-

dard sample-sizes/regions/clusters) and support the planning of multi-

stakeholder interventions aiming at improving specific attributes for en-

ergy access and track its success. Moreover, long-term testing and pilot

cases will reveal additional challenges and add-ons to integrate in the

toolkit. Once the institution estimates the effort, time constraints and

data uploading capabilities, a systematization of the tool within their

management information system should follow.

153

Appendices

155

A.Contactar Exclusion List

According to the policy of Contactar, the following activities are banned to

receive any kind of financing:

•Illegal, sale and trafficking of narcotics, trafficking crops

•Production or activities involving forced or child labor

•Trade of wildlife

•Fishing on the marine environment with nets larger than 2.5 km long

•The destruction of critical habitat or any forestry project which is not

carried out a plan for sustainable development

•Illegal logging, for manufacture and sale of charcoal

•Production, use or trade of hazardous materials such as fibers, asbestos

and products containing PCBs

•Production, use or trade of pharmaceuticals, pesticides, herbicides, chem-

icals, substances that deplete the ozone layer and other hazardous sub-

stances that affect the atmosphere and others

•Cross-border trade of wastes and residues

•Production or trade in arms and ammunition

•Brick manufacturing

•No legalized gambling

•Raffles, sales and purchases, lenders

•Illegal activities, brothels and / or disreputable establishments

157

B.PPI Tool - Colombia

Figure B.1.: Colombia- PPI Scorecard Tool Lookup Table (Part 1). Source:

http://www.progressoutofpoverty.org/country/colombia from Mi-

crofinance Risk Management, L.L.C, based on Colombia’s 2009

Encuesta Integrada de Hogares

159

The PEPI Toolkit for MFIs Appendix

Figure B.2.: Colombia- PPI Scorecard Tool Lookup Table (Part 2). Source:

http://www.progressoutofpoverty.org/country/colombia from Mi-

crofinance Risk Management, L.L.C, based on Colombia’s 2009

Encuesta Integrada de Hogares

Figure B.3.: Colombia- PPI Scorecard Tool Lookup Table (Part 3). Source:

http://www.progressoutofpoverty.org/country/colombia from Mi-

crofinance Risk Management, L.L.C, based on Colombia’s 2009

Encuesta Integrada de Hogares

160

The PEPI Toolkit for MFIs Appendix

Figure B.4.: Colombia- PPI Scorecard Tool Lookup Table (part 4) Source:

http://www.progressoutofpoverty.org/country/colombia from Mi-

crofinance Risk Management, L.L.C, based on Colombia’s 2009

Encuesta Integrada de Hogares

161

C.Maps and Description of Selected

Regions

Tolima

The department of Tolima is the most northern among the four departments

considered in this study. It is located in the Andean regions (mid-west Colom-

bia). It has a total surface of 23,562 km2(1.74% of the total surface of Colombia)

and a total of 1’410,000 inhabitants, divided in 47 municipalities.

Figure C.1.: Maps of the departments of Tolima (left) and Huila (right). Source:

www.vmapas.com

Huila

The department of Huila is locate in south-west of Colombia, within the Andean

region. It has a surface of 19,890 km2, with a population of about 1’174,000

inhabitants divided in 37 municipalities.

163

The PEPI Toolkit for MFIs Appendix

Nari˜no

The department of Nari˜no (where the head office of Contactar is located) defines

the south-east border between Colombia and Ecuador, and its geography is

divided in the plain Pacific region, characterized by elevate temperature and

abundant rains, in the Andean region, with high mountains (up to 4.700m) and

low temperatures and in the Amazonas region, mainly covered by forests. The

total surface of the department of Nari˜no is of 33,268 km2, with about 1’744,000

inhabitants, divided in 64 municipalities.

Figure C.2.: Map of the department of Nari˜no. Source: www.vmapas.com

Putumayo

The department of Putumayo, located in the south-eastern part of Colombia,

defines the border between Colombia, Ecuador and Peru. With a surface of

24,885 km2, Putumayo has a population of 345,000 inhabitants, divided in 13

municipalities.

164

The PEPI Toolkit for MFIs Appendix

Figure C.3.: Map of the department of Putumayo. Source: www.vmapas.com

165

List of Tables

2.1. Examples of Microenergy Systems . . . . . . . . . . . . . . . . . 20

2.2. Energy services for productive activities, added-value and renew-

ableenergyoptions.......................... 22

2.3. Multidimensional energy poverty metrics . . . . . . . . . . . . . . 26

2.4. Multi-tier Matrix: electricity supply . . . . . . . . . . . . . . . . 29

2.5. Multi-tier Matrices: Electricity Consumption and Services . . . . 30

2.6. Multi-tier Matrix: Cooking Solutions . . . . . . . . . . . . . . . . 31

2.7. Multi-tier Matrices: Productive Uses . . . . . . . . . . . . . . . . 33

2.8. Multi-tier Matrices: Productive Uses . . . . . . . . . . . . . . . . 34

3.1. Green Microfinance Indicators . . . . . . . . . . . . . . . . . . . . 46

3.2. Microfinance Environmental Performance Index (MFEPI) [Allet,

2011]. The score of each evaluation axis (right column) is indi-

catedinbrackets. .......................... 47

3.3. MIX Environmental Indicators: Environmental Policies and Ini-

tiatives. ................................ 48

3.4. Green Index - Set of Indicators (summarized) . . . . . . . . . . . 50

4.1. Access to infrastructure in the working regions of Contactar . . . 56

4.2. Distributions of Contactar clients across the served regions . . . 57

4.3. Portfolio of energy technologies offered by Contactar through the

program ConSuPlaneta ....................... 61

4.4. DANE Codification of productive activities used in Contactar

database for loan disbursements . . . . . . . . . . . . . . . . . . . 66

4.5. Productive activities (ISIC coditification A to C24) among Con-

tactar clients – June 2014 . . . . . . . . . . . . . . . . . . . . . . 67

4.6. Productive activities (ISIC coditification C25 to U) among Con-

tactar clients – June 2014 . . . . . . . . . . . . . . . . . . . . . . 68

4.7. Transitory agriculture activities in Contactar working regions ac-

cording to the 2005 National Census . . . . . . . . . . . . . . . . 69

4.8. Permanent agriculture activities in Contactar working regions

according to the 2005 National Census . . . . . . . . . . . . . . . 70

4.9. Livestock inventory according to 2005 National Census . . . . . . 70

4.10. Production processes and machinery in Contactar’s working regions 71

4.11. Potential technologies for specific energy usages (part 1) . . . . . 71

4.12. Potential technologies for specific energy usages (part 2) . . . . . 72

167

The PEPI Toolkit for MFIs List of Tables

4.13. Market potential of energy technologies among the clientele of

Contactar (first column) and corresponding energy usages (sec-

ondcolumn).............................. 73

4.14. Illustration of an example of the contribution of energy access to

the commodity value chain . . . . . . . . . . . . . . . . . . . . . 74

4.15. Cost reductions obtained from the use of MES in productive

activities................................ 75

4.16. Comparison of traditional drying processes (without SCD) and

dryingwithaSCD.......................... 75

4.17. Comparison of water boiling test results with different cooking

stoves ................................. 76

4.18. Characteristics of biogas digesters according to size . . . . . . . . 76

5.1. Summary of the number of selected clients in urban and rural

areas.................................. 82

5.2. Summary of the number of selected clients in the four considered

regions................................. 82

5.3. Summary of mean PPI scores (2012 and 2014) of the selected

clients ................................. 83

5.4. Modules, questions and variables considered in the implemented

survey................................. 84

5.5. Information on electricity services extracted from the PPI tool . 86

5.6. Information on cooking facilities extracted from the PPI tool . . 87

5.7. Appliance power used to estimate the household capacity. Sources:

Daftlogic and ABS Alaskan. . . . . . . . . . . . . . . . . . . . . . 88

5.8. Results (in percetages) of the tier-ranking for electricity supply . 92

5.9. Results (number of households) of the tier-ranking for electricity

supply................................. 93

5.10. Tier Assignment: Electricity Supply, Services and Consumption . 95

5.11. Frequency of attributes ranked as the lowest tier for each house-

hold(energysupply)......................... 96

5.12. Tier Assignment for Electricity Supply: Sensitivity Study (total

sample) ................................ 96

5.13. Tier Assignment for Electricity Supply: Sensitivity Study (urban

area).................................. 96

5.14. Tier Assignment for Electricity Supply: Sensitivity Study (rural

area).................................. 97

5.15. Summary of cooking solutions . . . . . . . . . . . . . . . . . . . . 99

5.16. Results (in percentage) of the tier-ranking for cooking facilities

for the different attributes . . . . . . . . . . . . . . . . . . . . . . 100

5.17. Results (number of households) of the tier-ranking for cooking

facilities................................101

5.18. Results of the (lowest-based) tier-ranking and of the composite

index for cooking facilities . . . . . . . . . . . . . . . . . . . . . . 105

5.19. Frequency of attributes ranked as the lowest tier for each house-

hold (cooking solutions) . . . . . . . . . . . . . . . . . . . . . . . 105

5.20. Tier Assignment for Cooking Solutions: Sensitivity Study . . . . 106

168

The PEPI Toolkit for MFIs List of Tables

5.21. Tier Assignment for Cooking Solutions: Sensitivity Study (urban

area)..................................106

5.22. Tier Assignment for Cooking Solutions: Sensitivity Study (rural

area)..................................106

5.23. Spearman correlation coefficients of tier ranking (energy supply)

andPPIscore.............................111

5.24. Spearman correlation coefficients of tier ranking (cooking solu-

tions)andPPIscore .........................112

6.1. PEPI Methodology: aggregation of the MTF attributes (elec-

tricity supply and electricity services) to adapt the framework to

theEnergySDG ...........................121

6.2. PEPI Methodology: aggregation of the MTF attributes (cooking

solutions) to adapt the framework to the Energy SDG . . . . . . 121

6.3. Thresholds of tier-ranking standards for electricity services con-

sidered in the PEPI frameworks . . . . . . . . . . . . . . . . . . . 122

6.4. PEPI Multi-tier Matrix: electricity supply and services . . . . . . 123

6.5. PEPI Multi-tier Matrix: cooking facilities . . . . . . . . . . . . . 124

6.6. Matrix of values for measuring the progress in terms of tier variation126

6.7. PEPI Framework: Results of the tier ranking for electricity supply130

6.8. Spearman correlation: energy supply tier ranking vs. PPI score . 133

6.9. PEPI Framework: Results of the tier ranking for cooking facilities134

6.10. Spearman correlation: cooking solution tier ranking vs. PPI score135

169

List of Figures

1.1. The Sustainable Development Goals (SDGs) . . . . . . . . . . . . 2

1.2. Number of MFIs reporting to the MixMarket platform on their

performance within a set of established Green Performance In-

dicators ................................ 4

1.3. Multi-tier Framework for Electricity Supply . . . . . . . . . . . . 7

1.4. An example of an access index (AI) computation according to

[Bhatia and Angelou, 2015], based on the proportion of households

rankedintiers0to5......................... 8

1.5. Left: Regions of Colombia where the case study was carried out.

Right: Summary of the results of the case study. . . . . . . . . . 8

1.6. PEPI toolkit components . . . . . . . . . . . . . . . . . . . . . . 10

2.1. Energy ladder according to fuel [Kowsari and Zerriffi, 2011] . . . . 18

2.2. Two examples of Access Index computation according to (2.1),

producing the same composite index . . . . . . . . . . . . . . . . 35

3.1. Number of MFIs reporting on Green Performance Indicators (see

[Pierantozzi et al., 2015])....................... 41

3.2. Two-Hand Model Scheme [Realpe Carrillo et al., 2015]....... 43

3.3. The systematic (left) and the pragmatic approach (right) for

cross-sectoral cooperation according to [De Gouvello and Durix,

2008].................................. 44

3.4. Number of MFIs offering financing to products related to re-

newable energy (RE) and to energy efficiency (EE) according to

yearly MIX reports in 2013 (N:1335) and 2014 (N:1466) . . . . . 49

3.5. PPI Scorecard for Colombia (November 2012) . . . . . . . . . . . 52

4.1. Regions of Colombia covered by Contactar (original map down-

loaded from: www.your-vector-maps.com)............. 53

4.2. Infrastructure access in working regions of Contactar . . . . . . . 57

4.3. Distributions of Contactar clients across the served regions . . . 58

4.4. Brief portrait of Contactar clients according to age (left), educa-

tion(center) and type of occupation (right) . . . . . . . . . . . . 58

4.5. Brief portrait of Contactar credits. Left: Type of Credit. Right:

Purposes (Consume, Education, Personal Use, Business, Housing) 59

4.6. Histogram of the PPI results among Contactar clients (data col-

lectedin2014) ............................ 59

171

The PEPI Toolkit for MFIs List of Figures

4.7. Bottom-up approach of a two-hand model according to initiator

[Realpe Carrillo et al., 2015] ..................... 62

4.8. Left: Solar Crop Dryer (for coffee beans). Right: Biogas Digester

(for farmers with at least 10 pigs) . . . . . . . . . . . . . . . . . . 63

4.9. Left:Water Tank. Right: Residential Water Filter . . . . . . . . . 63

5.1. Sketch of the meaning of z-value for a normal distribution . . . . 80

5.2. GPS coordinates of the interviewed clients. Left: political map;

right:physicalmap. ......................... 82

5.3. Excerpt of the translated survey tool (electricity supply module)

used for the field study . . . . . . . . . . . . . . . . . . . . . . . . 84

5.4. Excerpt of the translated survey tool (cooking module) used for

thefieldstudy ............................ 85

5.5. Histogram of capacity data (considering total sample, urban and

ruralareas) .............................. 88

5.6. Proportion of clients (y-axis) whose energy expenses does not

exceed a certain percentage of the monthly income (x-axis) . . . 90

5.7. Tier ranking for the different attributes (electricity supply), con-

sidering the total sample (top), the urban area (bottom-left) and

the rural area (bottom-right) . . . . . . . . . . . . . . . . . . . . 91

5.8. Histogram of consumption data (considering total sample, urban

andruralareas)............................ 94

5.9. Results of sensitivity study of the tier ranking for electricity sup-

ply: histogram of lowest tier ranking leaving out single attributes,

compared with the full MTF ranking (in pink) . . . . . . . . . . 97

5.10. Results of sensitivity study (in urban (left) and rural (right) ar-

eas) of the tier ranking for electricity supply: lowest tier ranking

leaving out single attributes, compared with the full ESMAP

MTF ranking (in pink). Legend as in Figure 5.9. . . . . . . . . . 98

5.11. Results of sensitivity study of the tier ranking for electricity sup-

ply: composite multi-tier index computed leaving out single at-

tributes, compared with the full MTF index . . . . . . . . . . . . 98

5.12. Histogram of convenience data considering cooking time (left)

and fuel preparation and collection time (right) . . . . . . . . . . 102

5.13. Percentage of clients (y-axis) whose expenses for cooking fuel do

not exceed a certain percentage of the monthly income (x-axis) . 103

5.14. Tier ranking for the different attributes (cooking solutions), con-

sidering the total sample (top), the urban area (bottom-left) and

the rural area (bottom-right) . . . . . . . . . . . . . . . . . . . . 104

5.15. Results of sensitivity study of the tier ranking for cooking solu-

tions: lowest tier ranking leaving out single attributes, compared

with the full MTF ranking . . . . . . . . . . . . . . . . . . . . . . 107

5.16. Results of sensitivity study (in urban and rural areas) of the tier

ranking for electricity supply: lowest tier ranking leaving out

single attributes, compared with the full MTF ranking. Legend

asinFigure5.15............................107

172

The PEPI Toolkit for MFIs List of Figures

5.17. Results of sensitivity study of the tier ranking for cooking so-

lutions: composite multi-tier index computed leaving out single

attributes, compared with the full MTF index . . . . . . . . . . . 108

5.18. Mean values and standard variations of PPI score vs. tier ranking

forenergysupply...........................110

5.19. Mean values and standard variations of PPI score vs. tier ranking

for electricity consumption . . . . . . . . . . . . . . . . . . . . . . 110

5.20. Mean values and standard variations of PPI score vs. tier ranking

forenergyservices ..........................111

5.21. Mean values and standard variations of PPI score vs. tier ranking

for cooking solutions . . . . . . . . . . . . . . . . . . . . . . . . . 111

5.22. Tier ranking (based on the lowest tier among all attributes) for

the different frameworks, considering the total sample (top), the

urban area (bottom-left) and the rural area (bottom-right) . . . 113

6.1. Sketch of the PEPI Frameworks and Attributes . . . . . . . . . . 120

6.2. Profile of the f(t1, t2) (6.3) for different values of t1(initial tier),

depending on t2(finaltier) .....................127

6.3. Comparison of the results of the ESMAP MTF tier ranking and

of the PEPI Frameworks tier ranking for the total sample (top)

and for urban (bottom left) and rural (bottom right) areas . . . 131

6.4. Sensivity study for the PEPI energy supply framework: tier rank-

ing for affordability (left) and reliability (right), leaving out sin-

gle attributes. Top row: total sample; middle row: urban area;

bottom row: rural area. . . . . . . . . . . . . . . . . . . . . . . . 132

6.5. Comparison of the results of the ESMAP MTF tier ranking

(cooking solutions) and of the PEPI Frameworks tier ranking

for the total sample (top), the urban area (bottom left) and the

rural area (bottom right) . . . . . . . . . . . . . . . . . . . . . . 135

6.6. Sensivity study for the availability leaving out single sub-attributes.

Top: total sample; Bottom-left: urban area; Bottom-right: rural

area ..................................136

6.7. Components of the PEPI Toolkit . . . . . . . . . . . . . . . . . . 137

6.8. PEPI Survey Sketch and Components . . . . . . . . . . . . . . . 138

6.9. Excerpt of PEPI Questionnaire indicating the respective attributes

to be assessed - Access to Electricity Supply and Electricity Ser-

vices..................................142

6.10. Excerpt of PEPI Questionnaire indicating the respective attributes

to be assessed - Access to Cooking Facilities . . . . . . . . . . . . 143

6.11. Excerpt of PEPI Survey tool to assess access to electricity supply

and electricity services . . . . . . . . . . . . . . . . . . . . . . . . 144

6.12. Excerpt of PEPI Survey tool to assess access to cooking facilities 145

B.1. Colombia- PPI Scorecard Tool Lookup Table (Part 1) . . . . . . 159

B.2. Colombia- PPI Scorecard Tool Lookup Table (Part 2) . . . . . . 160

B.3. Colombia- PPI Scorecard Tool Lookup Table (Part 3) . . . . . . 160

B.4. Colombia- PPI Scorecard Tool Lookup Table (Part 4) . . . . . . 161

173

The PEPI Toolkit for MFIs List of Figures

C.1. Maps of the departments of Tolima (left) and Huila (right).

Source: www.vmapas.com ......................163

C.2. Map of the department of Nari˜no. Source: www.vmapas.com . . 164

C.3. Map of the department of Putumayo. Source: www.vmapas.com 165

174

Bibliography

[Agbemabiese, 2009] Agbemabiese, L. (2009). A Framework for Sustainable Energy

Development Beyond the Grid: Meeting the Needs of Rural and Remote Populations.

Bulleting of Science, Technology and Society.

[AGECC, 2010] AGECC (2010). Energy for a Sustainable Future: Report and Rec-

ommendations. The Secretary Generals Advisory Group on Energy and Climate

Change (AGECC).

[Ahlin and Neville, 2008] Ahlin, C. and Neville, J. (2008). Can Micro-Credit Bring

Development? Journal of Development Economics, 86.

[Allderdice et al., 2007] Allderdice, A., Wienicki, J., and Morris, E. (2007). Using Mi-

crofinance to Expand Access to Energy Services. Summary of Findings. Technical

report, The SEEP Network, Washington DC.

[Allet, 2011] Allet, M. (2011). Measuring the Environmental Performance of Microfi-

nance. CEB Working paper, 11/045.

[Allet, 2012] Allet, M. (2012). Microfinance and Environment: Why Do Microfinance

Institutions Go Green? CEB Working paper, 12/015.

[Allet, 2014] Allet, M. (2014). The Green Index, an Innovative Tool to Assess the Envi-

ronmental Performance of MFIs. Technical report, European Microfinance Platform.

Prepared in collaboration with the e-MFP Microfinance and Environment Action

Group.

[Allet and Hudon, 2013] Allet, M. and Hudon, M. (2013). Green Microfinance. Charac-

teristics of Microfinance Institutions Involved in Environmental Management. Jour-

nal of Business Ethics, 126(3):395–414.

[Armend´ariz de Aghion and Morduch, 2005] Armend´ariz de Aghion, B. and Morduch, J.

(2005). The Economics of Microfinance. The MIT Press, Cambridge, London.

[Banerjee et al., 2010] Banerjee, A., Duflo, E., Glennester, R., and Kinnon, C. (2010).

The Miracle of Microfinance? Evidence from a Randomised Evaluation. Technical

report, BREAD Working Paper no.278.

[Barnes, 2007] Barnes, D. (2007). The Challenge of Rural Electrification: Strategies for

Developing Countries. Resources for the Future (Rff) Press.

[Barnes et al., 2011] Barnes, D., Khandker, S. R., and Samad, H. A. (2011). Energy

Poverty in Rural Bangladesh. Energy Policy, 39:894–904.

[Barnes and Toman, 2006] Barnes, D. and Toman, M. (2006). Energy, Equity, and

Economical Development, chapter Economic Development and Environmental Sus-

tainability. Oxford University Press. Lopez. R. and Toman, M.

[Barnes and Floor, 1996] Barnes, D. F. and Floor, W. (1996). Rural Energy in Devel-

oping Countries: A Challenge for Economic Development. Annual Review of Energy,

21(1):497530.

[Barnett, 1990] Barnett, A. (1990). The Diffusion of Energy Technology in The Rural

175

The PEPI Toolkit for MFIs Bibliography

Areas of Developing Countries: A Synthesis of Recent Experience. World Develop-

ment, 18:539–553.

[Bazilian et al., 2010] Bazilian, M., Nussbaumer, P., Cabraal, A., Centurelli, R.,

Detchon, R., Gielen, D., Rogner, H.-H., Howells, M., McMahon, H., V., M., and

Nakicenovic, N. (2010). Measuring Energy Access: Supporting a Global Target.

Technical report, The Earth Institute, Columbia University.

[Beck and Martinot, 2004] Beck, F. and Martinot, E. (2004). Renewable Energy Poli-

cies and Barriers. In Cleveland, C. J., editor, Encyclopedia of Energy, pages 365–383.

New York: Elsevier.

[Bensch, 2013] Bensch, G. (2013). Inside the Metrics: An Empirical Comparison of

Energy Poverty Indices for Sub-Saharan Countries. Technical report, Ruhr Economic

Papers, No. 464.

[Bensch, 2014] Bensch, G. (2014). Tracking the Energy Poor – Empirical Insights on

Energy Poverty Measurement Approaches. Technical report, International Associa-

tion for Energy Economics. First Quarter 2014.

[Bhatia and Angelou, 2014] Bhatia, M. and Angelou, N. (2014). Capturing the Multi-

Dimensionality of Energy Access. World Bank. Energy Sector Management Assis-

tance Program (ESMAP).

[Bhatia and Angelou, 2015] Bhatia, M. and Angelou, N. (2015). Beyond Connections:

Energy Access Redefined. Technical report, Energy Sector Management Assistance

Program (ESMAP) Technical Report 008/15, Washington. USA: The World Bank

Group.

[Brew-Hammond, 2010] Brew-Hammond, A. (2010). Energy Access in Africa: Chal-

lenges Ahead. Energy Policy, 38(5):2291–2301.

[Cabraal et al., 2005] Cabraal, A., Barnes, D., and Agarwal, S. . (2005). Productive

Uses of Energy for Rural Development. Annual Review of Environment and Re-

sources, 30(1):117–144.

[Cast´an Broto et al., 2015] Cast´an Broto, V., Stevens, L., and Salazar, D. (2015). En-

ergy Access and Urban Poverty. Poor People’s Energy Briefing, 4.

[Cisco Foundation, 2014] Cisco Foundation (2014). Mejora en la toma de decisiones en

Instituciones que sirven a los pobres a trav´es de un mejor uso del PPI. Technical

report, Cisco Foundation and Grameen Foundation.

[CNG and Asobancaria, 2012] CNG and Asobancaria (2012). Green Protocol. Technical

report, Colombian National Government (CNG) & Asobancaria.

[Coleman, 2006] Coleman, B. E. (2006). Microfinance in northeast Thailand: who

benefits and how much? World Development, 34(9):1612–38.

[COSA, 2015] COSA (2015). Testing the Progress out of Poverty Index. Technical

report, Committee on Sustainable Assessment (COSA) for Ford Foundation.

[d’Almeida and Roberts, 2014] d’Almeida, S. and Roberts, M. (2014). Power to the Peo-

ple: What’s Driving the Supply of Green Microfinance? Technical report, DPIBE.

[De Gouvello and Durix, 2008] De Gouvello, C. and Durix, L. (2008). Maximizing the

Productive Uses of Electricity to Increase the Impact of Rural Electrification Pro-

grams. Technical report, Energy Sector Management Assistance Program (ESMAP),

The World Bank, Washington DC.

[Devine et al., 2010] Devine, G., Sheldon, T., and Smith, S. (2010). Making the Con-

nection: Partnerships in Sustainable Energy and Development Finance. Technical

report, Yale School of Management.

[DFID, 2002] DFID (2002). Energy for the Poor – Underpinning the Millennium De-

176

The PEPI Toolkit for MFIs Bibliography

velopment Goals. UK Department for International Development (DFID).

[e-MFP, 2015] e-MFP (2015). Green Training - Environmental Management and MFIs.

e-MFP Action Group Microfinance and Environment.

[Elias and Victor, 2005] Elias, R. and Victor, D. (2005). Energy Transitions in Devel-

oping Countries: A Review of Concepts and Literature. Working Paper 40, Stanford

University, Program on Energy and Sustainable Development.

[Etcheverry, 2003] Etcheverry, J. (2003). Renewable Energy for Productive Uses: Strat-

egy to Enhance Environmental Protection and the Quality of Rural Life. Technical

report, Department of Geography and Institute for Environmental Studies.

[Foster et al., 2000] Foster, V., Tre, J.-P., and Wodon, Q. (2000). Energy Prices, En-

ergy Efficiency, and Fuel Poverty.

[Ghosh, 2013] Ghosh, J. (2013). Microfinance and the Challenge of Financial Inclusion

for Development. Cambridge Journal of Economics.

[Global Tracking Framework (GTF), 2013] Global Tracking Framework (GTF) (2013).

Progress Toward Sustainable Energy. Technical report, International Bank for Re-

construction and Development/The World Bank & the International Energy Agency.

[Global Tracking Framework (GTF), 2015] Global Tracking Framework (GTF) (2015).

Progress Toward Sustainable Energy. Technical report, International Bank for Re-

construction and Development/The World Bank & the International Energy Agency.

[Goldemberg, 2004] Goldemberg, J. (2004). The Case for Renewable Energies. In Re-

newables 2004.

[GpsPrune, 2015] GpsPrune (2015). Computer software (Version 18.2) available at

http://activityworkshop.net/software/gpsprune.

[Gradl and Knobloch, 2011] Gradl, C. and Knobloch, C. (2011). Energize the BoP!

Energy Business Model for Low Income Markets. Technical report, Endeva.

[Grameen-Foundation, 2012] Grameen-Foundation (2012). PPI Glance - Colombia.

Grameen Foundation.

[GreenMicrofinance, 2007] GreenMicrofinance (2007). Microfinance and the Environ-

ment: Setting the Research and Policy Agenda. Roundtable, May 5–6, 2006.

Philadelphia: GreenMicrofinance–LLC.

[Groh, 2014] Groh, S. (2014). The Role of Energy in Development Processes – The

Energy Poverty Penalty: Case Study of Arequipa (Peru). Energy for Sustainable

Development, 18:83–99.

[Groh, 2015] Groh, S. (2015). The Role of Access to Electricity in Development Pro-

cesses. PhD thesis, Aalborg University.

[Groh et al., 2016] Groh, S., Pachauri, S., and Narasimha, R. (2016). What are We

Measuring? An Empirical Analysis of Household Electricity Access Metrics in Rural

Bangladesh. Energy for Sustainable Development, 30:21–31.

[Groh and Taylor, 2015] Groh, S. and Taylor, H. (2015). The Role of Microfinance

in Energy Access: Changing Roles, Changing Paradigms, and Future Potential.

Enterprise Development and Microfinance.

[Guti´errez and Reddy, 2015] Guti´errez, E. and Reddy, R. (2015). Expanding Opportu-

nities for Rural Finance in Colombia. Report. Finance and Markets Practice, Latin

America and the Caribbean Region. AUS10747, World Bank.

[Hall et al., 2008] Hall, J., Collins, L., Israel, E., and Wenner, M. (2008). The

Missing Bottom Line: Microfinance and the Environment. In Philadelphia:

GreenMicrofinance-LLC.

[Hosier, 2004] Hosier, R. (2004). Energy Ladder in Developing Nations. Enciclopedia

177

The PEPI Toolkit for MFIs Bibliography

of Energy, 2.

[Hulme, 2000] Hulme, D. (2000). Is Microdebt Good for Poor People? A Note on the

Dark Side of Microfinance. Small Enterprise Development, 11(1):26–29.

[IEA, 2004] IEA (2004). World Energy Outlook 2004. Technical report, International

Energy Agency (IEA).

[IEA, 2011] IEA (2011). World Energy Outlook 2011 - Special Report - Energy for

All Financing Access for the Poor. Technical report, International Energy Agency

(IEA).

[IEA, 2013] IEA (2013). World Energy Outlook 2013. Technical report, International

Energy Agency (IEA).

[IEA, 2015] IEA (2015). World Energy Outlook 2015. Technical report, International

Energy Agency (IEA).

[IEA and WB, 2014] IEA and WB (2014). Sustainable Energy for All 2013-2014: Global

Tracking Framework Report. Technical report, International Energy Agency (IEA)

and The World Bank (WB).

[IEA and WB, 2015] IEA and WB (2015). Sustainable Energy for All 2015 – Progress

Toward Sustainable Energy. Technical report, International Energy Agency (IEA)

and The World Bank (WB).

[IEA and Photovoltaic Power Systems Program, 2002] IEA and Photovoltaic Power Sys-

tems Program (2002). Implementing Agreement on Photovoltaic Power Systems.

Technical report, International Energy Agency (IEA) and Photovoltaic Power Sys-

tems Program.

[IFC, 2013] IFC (2013). Mobilising Public and Private Funds for Inclusive Green

Growth Investment in Developing Countries. Technical report, International Fi-

nance Corporation (IFC).

[IFC and WRI, 2007] IFC and WRI (2007). The Next 4 Billion. Market Size and Busi-

ness Strategy at the Base of the Pyramid. Technical report, International Financial

Corporation (IFC) and World Resource Institute (WRI).

[Ilskog, 2008] Ilskog, E. (2008). Indicators for Assessment of Rural Electrification An

Approach for the Comparison of Apples and Pears. Energy Policy, 36(7):2665–2673.

[Kapadia, 2004] Kapadia, K. (2004). Productive Uses of Renewable Energy. A Review

of Four Bank-GEF Projects. DRAFT Productive Use of Renewables.

[Kaygusuz, 2011] Kaygusuz, K. (2011). Energy Services and Energy Poverty for Sus-

tainable Rural Development. Renewable and Sustainable Energy Reviews, 15:936–

947.

[Kebir, 2009] Kebir, N. (2009). Development of a Certification Process for Microenergy

Systems. In International Conference on Engineering Design.

[Kebir and Heipertz, 2010] Kebir, N. and Heipertz, J. (2010). Financing Energy. The

Role of Microfinance Institutions. Technical report, UMM Workshop Report E-MFP

Action Group. Bergamo: PlaNet Finance.

[Kebir et al., 2013] Kebir, N., Spiegel, N., Schrecker, T., Groh, S., Scott, C., and

Aliaga Ferrufino, G. (2013). Exploring Energy SME Financing in Emerging and

Developing Countries. Technical report, California: Sustainable Business Institute.

[Khandker et al., 2012] Khandker, S. R., Barnes, D. F., and Samad, H. A. (2012). Are

the Energy Poor also Income Poor? Evidence from India. Energy Policy, 47:1–12.

[Kittelson, 1998] Kittelson, D. (1998). Productive Uses of Electricity: Country Expe-

riences. In Village Power’ 98. (Conference Paper).

[Kowsari and Zerriffi, 2011] Kowsari, R. and Zerriffi, H. (2011). Three Dimensional

178

The PEPI Toolkit for MFIs Bibliography

Energy Profile. Energy Po, 39(12):7505–7517.

[Legros et al., 2009] Legros, G., Havet, I., Bruce, N., and Bonjour, S. (2009). The

Energy Access Situation in Developing Countries: A Review Focusing on the Least

Developed Countries and Sub-Saharan Africa. Technical report, United Nations

Development Programme (UNDP) and World Health Organization.

[Leva¨ı et al., 2011] Leva¨ı, D., Rippey, P., Rhyne, E., and Allderdice, A. (2011). Mi-

crofinance and Energy Poverty: Findings from the Energy Links Project. Technical

report, Final Report to AFD (now FHI 360) and USAID under the FIELD Project.

Center for Financial Inclusion. Publication No. 13.

[Lucas et al., 2003] Lucas, H., Barnett, A., Standing, H., Yuelai, L., and Jolly, S.

(2003). Energy, Poverty and Gender: A Review of the Evidence and Case Stud-

ies in Rural China. Technical report, Report for the World Bank by The Institute

of Development Studies at the University of Sussex, UK.

[MacLean and Siegel, 2007] MacLean, J. and Siegel, J. M. (2007). Financing Mech-

anisms and Public/Private Risk Sharing Instruments for Financing Small Scale

Renewable Energy Equipment and Projects. (Research Paper). Technical report,

United Nations Environment Programme (UNEP), Global Environment Facility

(GEF).

[Maren et al., 2011] Maren, D., Palmer-Jones, R., Copestake, J. G., Hooper, L., Loke,

Y., and N., R. (2011). What is the Evidence of the Impact of Microfinance on the

Well-being of Poor People? Technical report, EPPI-Centre, Social Science Research

Unit, Institute of Education, University of London.

[Martinot et al., 2001] Martinot, E., Cabraal, A., and Marthu, S. (2001). World

Bank/GEF Solar Home System Projects: Experiences and Lessons Learned 1993-

2000. Renewable and Sustainable Energy Reviews, 5(1):39–57.

[Martinot et al., 2002] Martinot, E., Chaurey, A., Lew, D., Moreira, J., and Wa-

mukonya, N. (2002). Renewable Energy Markets in Developing Countries. Annual

Review of Energy, 27(1):309–348.

[Masera et al., 2000] Masera, O., Saatkamp, B., and Kammen, D. (2000). From Linear

Fuel Switching to Multiple Cooking Strategies: A Critique and Alternative to the

Energy Ladder Model. World Development, 28(12):2083–2103.

[Matas et al., 2015] Matas, B., Ramirez, J., and Kahlen, L. (2015). Green Microfinance

in Contactar. Market Study: Pre-selection of Green Technologies of Relevance. Tech-

nical report, MicroEnergy International.

[Meadows et al., 2003] Meadows, K., Riley, C., Rao, G., and Harris, P. (2003). Mod-

ern Energy: IImpact on Micro-enterprises. A Literature Review into the Linkages

between Modern Energy and Micro-Enterprise. Technical report, UK Department

for International Development, London.

[MEI and PF, 2010] MEI and PF (2010). Fact Sheet: The Potential of Linking Micro-

finance and Energy Supply. Technical report, MicroEnergy International (MEI) and

PlaNet Finance (PF).

[MicroEnergy International, 2014] MicroEnergy International (2014). Green Profile -

Contactar. Internal Report.

[MIF and Bloomberg, 2012] MIF and Bloomberg (2012). Climatescope 2012. Technical

report, Multilateral Investment Fund (MIF) and Bloomberg New Energy Finance.

[Mirza and Szirmai, 2010] Mirza, B. and Szirmai, A. (2010). Towards a New Measure-

ment of Energy Poverty: A Cross-Community Analysis of Rural Pakistan. Technical

report, Maastricht Economic and Social Research and Training Centre on Innovation

and Technology.

179

The PEPI Toolkit for MFIs Bibliography

[Modi, 2004] Modi, V. (2004). Energy Services for the Poor (Commissioned paper

for the Millennium Project Task Force 1). Technical report, Earth Institute and

Department of Mechanical Engineering, Columbia University.

[Modi et al., 2006] Modi, V., McDade, S., Lallement, D., and Saghir, J. (2006). Energy

Services for the Millennium Development Goals. Technical report, New York: Energy

Sector Management Assistance Programme (ESMAP), United Nations Development

Programme, UN Millennium Project, and World Bank.

[Mohiuddin, 2006] Mohiuddin, S. (2006). Expanding the Role of Microfinance in Pro-

moting Renewable Energy Access in Developing Countries. The Georgetown Public

Policy Review, 11(1).

[Morris and Kirubi, 2009] Morris, E. and Kirubi, G. (2009). Bringing Small Scale Fi-

nance to the Poor for Modern Energy Services: What is The Role Of Government?

Experiences from Burkina Faso, Kenya, Nepal and Tanzania. Technical report,

United Nations Development Programme.

[Morris et al., 2007] Morris, E., Winiecki, J., Chowdhard, S., and Cortiglia, D. (2007).

Using Microfinance to Expand Access to Energy Services: Summary of Findings.

Small Enterprise Education and Promotion Network (SEEP) Network.

[Nussbaumer et al., 2011] Nussbaumer, P., Bazilian, M., Modi, V., and Yumkella, K.

(2011). Measuring Energy Poverty: Focusing on What Matters. Oxford Poverty

& Human Development Initiative (OPHI) Oxford. Technical report, Department

of International Development Queen Elizabeth House (QEH), University of Oxford,

OPHI Working Paper Nr. 42.

[OECD, 2007] OECD (2007). Energy for Sustainable Development. Technical report,

The Organisation for Economic Cooperation and Development (OECD). Contri-

bution to the United Nations Commission on Sustainable Development 15. Paris:

OECD.

[OECD and IEA, 2010] OECD and IEA (2010). Energy Poverty: How to make Modern

Energy Access Universal? Energy Poverty: Special early excerpt of the World En-

ergy Outlook for the UN General Assembly on the Millennium Development Goals?

Technical report, The Organisation for Economic Cooperation and Development

(OECD) and International Energy Agency (IEA).

[Pachauri, 2011] Pachauri, S. (2011). Reaching an International Consensus on Defining

Modern Energy Access. Current Opinion in Environmental Sustainability, 3(4):235–

240.

[Pachauri et al., 2012a] Pachauri, S., Brew-Hammond, A., Barnes, D. F., Bouille,

D. H., Gitonga, S., Modi, V., Prasad, G., Rath, A., and Zerriffi, H. (2012a). Global

Energy Assessment – Toward a Sustainable Future, chapter Energy Access for De-

velopment (Chapter 19), pages 1401–1458. Cambridge University Press, Cambridge,

UK and New York, NY, USA and the International Institute for Applied Systems

Analysis, Laxenburg, Austria.

[Pachauri et al., 2004] Pachauri, S., Mueller, A., Kemmler, A., and Spreng, D.

(2004). On Measuring Energy Poverty in IIndia Households. World Development,

32(12):2083–2104.

[Pachauri et al., 2012b] Pachauri, S., Rao, D., Nagai, Y., and Riahi, K. (2012b). Access

to Modern Energy Assesment and Outlook for Developing and Emerging Regions.

Technical report, International Institute for Applied Systems Analysis (IIASA),

United Nations Industrial Development Organization (UNIDO), Global Environ-

ment Facility (GEF).

[Pachauri and Spreng, 2004] Pachauri, S. and Spreng, D. (2004). Energy Use and En-

ergy Access in Relation to Poverty. Economic and Political Weekly, 39(3):271–278.

180

The PEPI Toolkit for MFIs Bibliography

[Pachauri et al., 2013] Pachauri, S., van Ruijven, B., Nagai, Y., Riahi, K., Van Vuuren,

D., Brew-Hammond, A., and Nakicenovic, N. (2013). Pathways to achieve universal

household access to modern energy by 2030. Environment Research Letters, 8:024015

(7pp).

[Parkerson, 2005] Parkerson, D. (2005). Assessing the Potential for Microfinance Insti-

tutions to Finance Solar Photovoltaic Systems in the Dominican Republic. Technical

report, SEEP Network. Working Paper.

[Peters et al., 2009] Peters, J., Harsdorff, M., and Ziegler, F. (2009). Rural Electrifi-

cation: Accelerating Impact with Complementary Services. Energy for Sustainable

Development, 13(1):38–42.

[Philipp and Sch¨afer, 2009] Philipp, D. and Sch¨afer, M. (2009). Interdisziplin¨arer

Forschungsansatz zur Analyse von Mikroenergie-Systeme. Discussion Paper.

Graduiertenkolleg Mikroenergie-Systeme.

[Pierantozzi et al., 2015] Pierantozzi, A., Allet, M., Forcella, D., Huybrechts, F.,

Mauro, A., Realpe Carrillo, N., Spiaggari, L., and Shuite, G. (2015). Assessing

Green Microfinance Qualitative and Quantitative Indicators for Measuring Environ-

mental Performance. Technical report, Mix Market.

[Practical Action, 2012] Practical Action (2012). Poor People’s Energy Outlook 2012:

Energy for Earning a Living. Practical Action Publishing, Rugby, UK.

[Practical Action, 2013] Practical Action (2013). Poor Peoples Energy Outlook 2013:

Energy for community services. Practical Action Publishing, Rugby, UK.

[Rao and Rao, 2010] Rao, D. and Rao, K. (2010). Micro Credit and Economic Devel-

opment, chapter The Role of Micro Credit in Poverty Alleviation, pages 193–207.

New Delhi: Regal Publications.

[Rao et al., 2009] Rao, P., Miller, J., Wang, Y., and Byrne, J. (2009). Energy-

Microfinance Intervention for Below Poverty Line Households in India. Energy Pol-

icy, 37(5):1694–1712.

[Realpe Carrillo, 2014] Realpe Carrillo, N. (2014). A Practitioner’s Outlook on the

Debate Why Green Microfinance, and if so, How? In 10th University Meets Micro-

finance Workshop. Frankfurt School of Finance & Management. July, 2013, pages

43–38. UMM Thematic Paper. Understanding the Challenges: New Insights from

Practice & Research on Mobile Banking, Remittances and Green Microfinance.

[Realpe Carrillo et al., 2015] Realpe Carrillo, N., Kahlen, L., and Dumitrescu, R.

(2015). Implementing a Green Microfinance – The Case of Contactar. In Inter-

national Conference Micro Perspectives for Decentralized Energy Supply. Bangalore,

India.

[Rehman et al., 2012] Rehman, I. H., Kar, A., Banerjee, M., Kumar, P., Shardul, M.,

Mohanty, J., and Hossain, I. (2012). Understanding the Political Economy and

Key Drivers of Energy Access in Addressing National Energy Access Priorities and

Policies. Energy Policy, 47:27–37.

[Rippey, 2009] Rippey, P. (2009). Microfinance and Climate Change: Threats and

Opportunities. Technical report, CGAP Focus Note, No. 53.

[Rippin, 2011] Rippin, N. (2011). A Response to the Weaknesses of the Multidimen-

sional Poverty Index (MPI): The Correlation Sensitive Poverty Index (CSPI). Tech-

nical report, Deutsches Institut f¨ur Entwicklungspolitik, Briefing Paper 19/2011.

[Rogerson, 1997] Rogerson, C. (1997). Rural Electrification and the SMME Economy

in South Africa. Technical report, University of Cape Town.

[Saari, 2006] Saari, S. (2006). Productivity Theory and Measurement in Business. In

European Productiviy Conference, Finland.

181

The PEPI Toolkit for MFIs Bibliography

[Saghir, 2005] Saghir (2005). Energy and Poverty: Myths, Links, and Policy Issues.

Energy Working Notes 4, The World Bank, Energy and Mining Sector Board, Wash-

ington D.C.

[SantaMar´ıa, 2011] SantaMar´ıa, H. S. (2011). Sistema T´ecnico Mejorado de Secado

de Caf´e a Nivel Familiar. Secador Solar vs. Sistema Tradicional. Technical report,

Instituto Nacional de Innovaci´on Agraria (INIA) - Fondo Nacional de Capacitaci´on

Laboral y Promoci´on del Empleo (FONDOEMPLEO).

[Sch¨afer et al., 2011] Sch¨afer, M., Kebir, N., and Neumann, K. (2011). Research Needs

for Meeting the Challenge of Decentralized Energy Supply in Developing Countries.

Energy for Sustainable Development, 15(3):24–32.

[Schreiner, 2004] Schreiner, M. (2004). A Simple Scorecard for Nepal. Microfinance

Risk Management. L.L.C., microfinance.com.

[SE4ALL and WB, 2014] SE4ALL and WB (2014). Energy Survey. Household Ques-

tionnaire.

[Serrano, 2009] Serrano, J. (2009). Microfinanzas e Instituciones Microfinancieras en

Colombia. Technical report, Comisi´on Econ´omica para Am´erica Latina y el Caribe

(CEPAL).

[Shuite and Forcella, 2015] Shuite, G. and Forcella, D. (2015). Green Inclusive Finance.

Status, Trends and Opportunities! Technical report, NpM, Enclude, Hivos.

[Smith, 2000] Smith, J. (2000). Solar-Based Rural Electrification and Microenterprise

Development in Latin America: A Gender Analysis. Technical report, National

Renewable Energy Laboratory (NREL), Golden, Colorado.

[Srinivasan, 2007] Srinivasan, S. (2007). Microfinance for Renewable Energy: Financing

the Former Poor. World Review of Entrepreneurship, Management and Sustainable

Development, 3(1):79–89.

[Stern and Cleveland, 2004] Stern, D. and Cleveland, C. (2004). Energy and Economic

Growth. Renlasser working papers on economics, Rensselaer Polytechnic Institute.

[Stevens et al., 2015] Stevens, L., Wykes, S., and Singer, S. (2015). Measuring what

matters in the Energy SDG. infoHub (Practical Action).

[The World Bank, 2008a] The World Bank (2008a). Designing Sustainable Off-Grid

Rural Electrification Projects: Principles and Practices. Technical report, Energy

and Mining Sector Board, The World Bank Group (Washington DC).

[The World Bank, 2008b] The World Bank (2008b). REToolkit: A Resource for Re-

newable Energy Development. Technical report, The World Bank, Washington D.C.

[The World Bank, 2016] The World Bank (2016). World Development Indicators. Sta-

tistical Report.

[UNCDF, 2012] UNCDF (2012). Clean Start Detailed Business Plan. Technical report,

United Nations Capital Development Fund (UNCDF).

[UNDP, 2000] UNDP (2000). World Energy Assessment: Energy and the Challenge of

Sustainability. Technical report, United Nations Development Programme (UNDP).

[UNDP, 2016] UNDP (2016). UNDP Support to the Implementation of the 2030

Agenda for Sustainable Development. United Nations Development Programme

(UNDP) Policy and Programme Brief.

[UNDP and WHO, 2009] UNDP and WHO (2009). The Energy Access Situation in

Developing Countries: A Review Focusing on the Least Developed Countries and

Sub-Saharan Africa. Technical report, United Nations Development Programme

(UNDP) and World Health Organization (WHO), New York.

[Urbaniak and Plous, 2013] Urbaniak, G. C. and Plous, S. (2013). Research Randomizer

182

The PEPI Toolkit for MFIs Bibliography

[Computer software] – URL: http://www.randomizer.org/. (Version 4.0).

[van der Kroon et al., 2013] van der Kroon, B., Brouwer, R., and van Beukering, P.

(2013). The Energy Ladder: Theoretical Myth or Empirical Truth? Results from a

Meta-Analysis. Renewable and Sustainable Energy Reviews, 20:504–513.

[van der Straeten et al., 2014] van der Straeten, J., Friederici, K., and Groh, S. (2014).

Taking a Micro-Perspective on the Global Challenge of Climate Change: The

Microenergy Systems Research Focus at the Technische Universit¨at Berlin. In

Leal Filho, W., editor, International Perspectives on Climate Change, Climate

Change Management, pages 135–149. Springer International Publishing.

[Van Elteren, 2007] Van Elteren, A. (2007). Environmental and Social Risk Manage-

ment and Added Value at MFIs and MFI funds the FMO Approach. Technical

report, The Hague: Netherlands Development Finance Company (FMO).

[Viscidi, 2010] Viscidi, L. (2010). Colombia’s Energy Renaissance. Technical report,

Americas Society and Council of the Americas Energy Action Group.

[von Wolff and Phalpher, 2014] von Wolff, S. and Phalpher, K. (2014). Green Finance.

Successes and Challenges - A Landscape Overview. Technical report, Finance in

Motion.

[Waterfield, 2001] Waterfield, C. (2001). Designing Microfinance Loan Products. In-

stitutional paper, MFI Solutions, LLC, Lancaster, PA USA.

[Wenner, 2002] Wenner, M. (2002). Microenterprise Growth and Environmental Pro-

tection. Microenterprise Development Review, 4(2):1–8.

[Winkler et al., 2011] Winkler, H., Sim˜oes, A. F., L´ebre La Rovere, E., Alam, M.,

Rhaman, A., and Mwakasonda, S. (2011). Access and Affordability of Electricity in

Developing Countries. World Development, 39(6):1037–1050.

[Zerriffi, 2007] Zerriffi, H. (2007). Making Small Work: Business Model for Electrifying

the World. Technical report, Stanford University.

183