i
A compa ison o he pe o mance o G een bond
unds, bond Mu ual unds and bond ETFs
Ana Ri a Rosado Rod igues
Disse a ion p esen ed as pa ial equi emen o ob aining
he Mas e ’s deg ee in S a is ics and In o ma ion
Managemen
i
NOVA In o ma ion Managemen School
Ins i u o Supe io de Es a ís ica e Ges ão de In o mação
Uni e sidade No a de Lisboa
A COMPARISON OF THE PERFORMANCE OF GREEN BOND FUNDS,
BOND MUTUAL FUNDS AND BOND ETFS
by
Ana Ri a Rosado Rod igues
Disse a ion p esen ed as pa ial equi emen o ob aining he Mas e ’s deg ee in S a is ics and
In o ma ion Managemen , wi h a specializa ion in Risk Analysis and Managemen .
Ad iso : P o . Luís Albe o Fe ei a de Oli ei a, PhD
Feb ua y 2023
ii
ACKNOWLEDGEMENTS
Reaching he end o his disse a ion was no easy, so I would like o sha e my g a i ude.
To my amily, o hei uncondi ional suppo , conce n and o gi ing me he oppo uni y o s udy and
deli e his inal wo k.
To my boy iend, o all his ad ice, encou agemen , pa ience and mo i a ion. My g ea es suppo
h oughou his p ocess.
To my eam a CGD, especially my di ec o Isabel and my colleagues Ped o and Hen ique, o
unde s anding my posi ion in he las mon h and o p o iding me all he suppo I needed.
To P o esso Jo ge B a o, wi hou whose help I would no ha e ound he mos app op ia e opic and
ad iso o his disse a ion.
Finally, o P o esso Luís Oli ei a, o accep ing he challenge o being my ad iso , o all he a ailabili y
he has shown and all he wisdom he has sha ed.
Thank you all.
iii
ABSTRACT
This disse a ion measu es and compa es he pe o mance o g een bond unds, bond mu ual unds
and bond ETFs, all domiciled in he Eu opean egion, om Janua y 2005 o Decembe 2019. This pe iod
has been di ided in o h ee subpe iods in o de o analyse he pe o mance be o e, du ing and a e
he inancial c isis o 2007-2008. The sample consis s o mon hly da a o a o al o 3,484 unds and
hei pe o mance was assessed by using adi ional isk-adjus ed measu es, namely Sha pe a io,
T eyno a io, and Jensen’s Alpha. The main indings show ha , on a e age, bond mu ual unds
ou pe o med bond ETFs and g een bond unds in all he s udied subpe iods. Fu he mo e, when
analysing each und ca ego y sepa a ely, all und ca ego ies ha e pe o med bes du ing he c isis
pe iod, which can be conside ed a ly- o-sa e y e en , whe e he p ices o sa e asse s end o ise.
Rega ding he pe o mance o g een bond unds, al hough hey ha e ou pe o med hei pee s a
some poin s in ime, he e is no clea e idence o suppo his. Howe e , some in es o s may p e e
o in es in his ype o asse due o he g een bond unds' en i onmen al con ibu ion.
KEYWORDS
Risk-adjus ed pe o mance; G een bond und; Bond mu ual und; Bond ETF; ESG; Financial c isis
Sus ainable De elopmen Goals (SGD):
i
INDEX
1. In oduc ion .................................................................................................................. 1
2. Li e a u e e iew .......................................................................................................... 3
2.1. Sus ainable In es ing ............................................................................................. 3
2.2. Pe o mance Measu emen .................................................................................. 5
3. Da a and Me hodology ................................................................................................. 8
3.1. Me hodology ......................................................................................................... 8
3.1.1. Sha pe Ra io ................................................................................................... 9
3.1.2. Pa ame e s Es ima ion: The Single-Index Model ......................................... 10
3.1.3. T eyno Ra io ................................................................................................ 11
3.1.4. Jensen’s Alpha .............................................................................................. 11
3.2. Da a ..................................................................................................................... 13
4. Resul s and discussion ................................................................................................ 18
4.1. Sha pe Ra io ........................................................................................................ 18
4.2. T eyno Ra io ....................................................................................................... 20
4.3. Jensen’s Alpha ..................................................................................................... 22
5. Conclusions and Fu he De elopmen s .................................................................... 25
Bibliog aphy..................................................................................................................... 27
Appendices ...................................................................................................................... 30

LIST OF FIGURES
Figu e 2.1 - Examples o issues o ESG ac o s ........................................................................... 4
Figu e 3.1 - To al numbe o unds by yea .............................................................................. 14
Figu e 3.2 - To al numbe o new unds pe yea .................................................................... 15
LIST OF TABLES
Table 3.1 - Desc ip i e s a is ics o mon hly e u ns o bond mu ual unds, bond ETFs, g een
bond unds and benchma k index ................................................................................... 16
Table 4.1 - Sha pe a io esul s ................................................................................................ 18
Table 4.2 - Two-sample T- es o he a e age Sha pe a ios (by subpe iod) ......................... 19
Table 4.3 - Two-sample T- es o he a e age Sha pe a ios (by und ca ego y) ................... 20
Table 4.4 - Be a esul s ............................................................................................................. 20
Table 4.5 - T eyno a io esul s ............................................................................................... 21
Table 4.6 - Two-sample T- es o he a e age T eyno a ios (by subpe iod) ........................ 22
Table 4.7 - Two-sample T- es o he a e age T eyno a ios (by und ca ego y) .................. 22
Table 4.8 - Jensen’s alpha esul s ............................................................................................. 23
Table 4.9 - Two-sample T- es o he a e age Jensen's alphas (by subpe iod) ...................... 24
Table 4.10 - Two-sample T- es o he a e age Jensen's alphas (by und ca ego y).............. 24
i
LIST OF ABBREVIATIONS AND ACRONYMS
ARCH Au o eg essi e Condi ional He e oscedas ici y
CAPM Capi al Asse P icing Model
EGARCH Exponen ial Gene alized Au o eg essi e Condi ional He e oscedas ic
ESG En i onmen al, Social and Go e nance
ETF/ETFs Exchange-T aded Fund / Exchange-T aded Funds
GARCH Gene alized Au o eg essi e Condi ional He e oscedas ici y
LAP Loss-A e se Pe o mance
NAV Ne Asse Value
OLS O dina y Leas Squa es
SRI Socially Responsible In es men
1
1. INTRODUCTION
Fo some ime now we a e acing a clima e change ha esul s om wo cen u ies o accumula ed
unsus ainable de elopmen . Global emissions o g eenhouse gases a e ising. Acco ding o Wo ld
Me eo ological O ganiza ion (2021), concen a ions o g eenhouse gases in he a mosphe e oday a e
149 pe cen highe han p e-indus ial le els and annual a es o inc ease ha e ne e been so high.
O e he pas decade, people ha e shown signs o awa eness in se e al a eas o he inancial sys em
h ough capi al mobiliza ion o g een ac i i ies. G een and sus ainable inance can help imp o ing
en i onmen al sus ainabili y, educing ca bon emissions and de eloping a s ong clima e
in as uc u e (Aglia di & Chechulin, 2020). The wo ld is inc easingly conce ned abou his opic and
he manage s and in es o s hemsel es wan o p ese e he en i onmen and he wo ld ha we li e
in. In o de o do ha , some companies a e issuing dep in he o m o g een bonds.
One o he bigges d i e s o in es o awa eness o isk managemen and sus ainabili y was he
inancial c isis o 2007-2008. When he US housing ma ke collapsed, i igge ed a sub-p ime c isis
ha led o he insol ency o some banks and o he inancial ins i u ions. This quickly sp ead o he es
o he inancial sys em and had a global impac , wi h banks and go e nmen s ha ing o ake eme gency
measu es o s abilise he inancial sys em and a oid a o al collapse o he global economy
(Sama bakhsh & Shah, 2021).
The c isis ein o ced he need o a mo e esponsible and sus ainable app oach, esul ing in a g ea e
ocus on en i onmen al, social and go e nance (ESG) issues. Ne e heless, he inancial c isis o 2007-
2008 can be seen as a u ning poin o he g een bond ma ke , as i gained g ea e isibili y (Sampei,
2018).
As he g een bond ma ke has de eloped, so has he ma ke o g een bond unds. As an asse ha
can be in es ed in a di e si ied way, managed by p o essionals and wi h lowe isk, i has gained
popula i y especially among p i a e in es o s. Howe e , i is use ul o unde s and whe he hese g een
bond unds o e he same o be e e u ns han hei non-g een pee s.
The e ha e been se e al academic s udies on he pe o mance o mu ual unds and ETFs, al hough no
so many on bond unds in pa icula , and e en ewe ha includes g een bond unds. Gi en he e e -
inc easing en i onmen al conce ns a ound he wo ld, as well as he impo an poli ical decisions and
economic changes o he las decade, g een in es ing is a e y in e es ing subjec o s udy. As a as
he au ho is awa e, he e is also a isible lack o s udies on he Eu opean ma ke , as many s udies a e
ca ied ou on he US ma ke .
So, o add ess hese gaps, he main pu pose o his disse a ion is o conduc a s udy on he compa ison
o he pe o mance o unds ha in es on g een bonds, bond mu ual unds and Exchange-T aded
Funds (ETFs) ha only in es on bonds, all Eu opean domiciled.
In o de o achie e his goal, i is necessa y o analyse he isk-adjus ed e u ns o g een bond unds,
bond mu ual unds and bond ETFs. In addi ion, he pe o mance o hese h ee ca ego ies o unds
o e he chosen pe iod, which includes he inancial c isis o 2007-2008, will also be analysed. To his
end, he models used in his hesis include he Sha pe Ra io, he T eyno Ra io, as well as he Jensen's
Alpha.
2
This disse a ion is s uc u ed as ollows. In chap e 2, a heo e ical app oach is made by p esen ing
some concep s and models ega ding sus ainable in es ing and pe o mance measu emen , h ough
a li e a u e e iew. Fu he , in chap e 3, is p esen ed he me hodology and he da a used o achie e
he s udy objec i e. The esul s a e p esen ed and discussed in chap e 4 and he espec i e
conclusions a e s a ed in chap e 5, along wi h he limi a ions iden i ied du ing he cou se o his s udy
and he sugges ions o u u e wo ks.
9
𝑃𝑍𝐶=𝐹
1+𝑟
12
(3)
Whe e:
𝑃𝑍𝐶 – Ze o-coupon bond p ice a mon h 𝑡
𝐹 – Face alue o he bond
𝑟 – Risk- ee a e o he asse
Mo eo e , as he compu ed und e u ns a e exp essed using con inuous compounding, hei ola ili y
can be de ined as he his o ical s anda d de ia ion o he mon hly e u ns p o ided by und. The
ola ili y o he und 𝑖 is gi en by:
𝜎𝑖= √𝜎𝑖2= √1
𝑛−1∑(𝑅𝑖,𝑡−𝑅𝑖
)2
𝑛
𝑡=1
(4)
Whe e:
𝜎𝑖2 – Va iance o he und 𝑖 expec ed e u n
𝑅𝑖,𝑡 – Re u n o he und 𝑖 a mon h 𝑡
𝑅𝑖
 – His o ical a e age o e u n, 1𝑛𝑅𝑖,𝑡
𝑛 – Numbe o mon hs
An al e na i e o his calcula ion would be o model ola ili ies using he au o eg essi e condi ional
he e oscedas ici y (ARCH) model (Engle, 1982), he gene alized au o eg essi e condi ional
he e oscedas ici y (GARCH) model (Bolle sle , 1986) o he exponen ial gene alized au o eg essi e
condi ional he e oscedas ic (EGARCH) model (Nelson, 1991). These models a emp o cap u e
a ia ions in ola ili y, ecognising ha i is no cons an h ough ime. Ne e heless, in his
disse a ion, he ola ili y is es ima ed using he s anda d app oach, as i is simple and mo e
s aigh o wa d.
The ollowing subsec ions desc ibe he isk-adjus ed measu es used o assess po olio pe o mance.
The nex e e enced excess e u n o a und o e he isk- ee a e o he mon h 𝑡 is calcula ed using
he las a ailable isk- ee a e in ha mon h. Meaning ha , o example, o he Ma ch 2006 excess
e u n calcula ion, he las a ailable a e o making a isk- ee in es men was Feb ua y 2006.
3.1.1. Sha pe Ra io
The Sha pe a io (Sha pe, 1966) measu es he excess e u n o a und ela i e o he isk- ee a e o
a gi en pe iod, conside ing he ola ili y o he und’s e u ns. I is gi en by he ollowing equa ion:

10
𝑆ℎ𝑎𝑟𝑝𝑒 𝑅𝑎𝑡𝑖𝑜= 𝑅𝑖,𝑡−𝑅𝑓,𝑡
𝜎𝑖
(5)
Whe e:
𝑅𝑖,𝑡 – Re u n o he und 𝑖 a mon h 𝑡
𝑅𝑓,𝑡 – Re u n o he isk- ee asse a he beginning o mon h
𝑡
𝜎𝑖 – S anda d de ia ion o he und 𝑖 e u ns
As al eady e e ed, he s anda d de ia ion is a measu e o he ola ili y o he e u ns, and i is
calcula ed by equa ion (4).
As a as he in e p e a ion o he Sha pe a io is conce ned, i is qui e s aigh o wa d. The highe he
a io, he be e he pe o mance and he e o e he highe he e u n ea ned o each uni o isk aken
( ola ili y).
3.1.2. Pa ame e s Es ima ion: The Single-Index Model
Based on Ma kowi z (1952), Sha pe (1963) in oduced he single-index model ha desc ibe he
ela ionship be ween he e u ns o an indi idual asse and he o e all ma ke . This model is p esen ed
as:
(𝑅𝑖,𝑡−𝑅𝑓,𝑡)=𝛼𝑖+𝛽𝑖×(𝑅𝑚,𝑡−𝑅𝑓,𝑡)
(6)
Whe e:
𝑅𝑖,𝑡 – Re u n o he und 𝑖 a mon h 𝑡
𝑅𝑓,𝑡 – Re u n o he isk- ee asse a he beginning o mon h 𝑡
𝑅𝑚,𝑡 – Re u n o he ma ke a mon h 𝑡
𝛼𝑖 – Alpha o he und 𝑖 (Jensen’s alpha)
𝛽𝑖 – Be a o he und 𝑖
The a iables 𝛼𝑖 and 𝛽𝑖 a e es ima ed using an O dina y Leas Squa es (OLS) eg ession
1
.
Be a Es ima ion
Be a (𝛽𝑖) is he sys ema ic o ma ke isk coe icien , i.e., i measu es he ola ili y o e u ns ela i e
o he o e all ma ke . Sys ema ic isk ep esen s he non-di e si iable isk o he isk ha emains e en
a e ex ensi e di e si ica ion. This isk a ec s he ma ke as a whole a he han a speci ic indi idual
1
O dina y Leas Squa es is a commonly used s a is ical me hod o es ima ing he pa ame e s o a linea
eg ession model. When ce ain assump ions a e me , OLS p o ides he bes linea unbiased es ima e o he
eg ession coe icien s.
11
asse . (Bodie e al., 2014) As such, i is gene ally conside ed o be he le el o isk ha in es o s a e
compensa ed o aking on.
I he be a is nega i e, i means ha he po olio is nega i ely co ela ed wi h he ma ke , i.e., as
ma ke isk inc eases, he po olio’s e u ns dec ease, o ice- e sa. A be a o 1 means ha he
ola ili y o he po olio is pe ec ly co ela ed wi h he ma ke , sugges ing ha he mo emen o he
po olio e lec s he mo emen o he ma ke . The e o e, a po olio wi h a be a g ea e han 1
indica es a highe le el o isk compa ed o he ma ke a e age isk and a be a below 1 indica es ha
he po olio is less isky han he ma ke a e age isk.
Alpha Es ima ion
When in es ing in a und, in es o s o en conside how he manage has con ibu ed o he und's
pe o mance. Looking a alpha (𝛼𝑖) is one way o measu e whe he a und manage has added alue
beyond simply in es ing in he index.
3.1.3. T eyno Ra io
Ano he ela i e isk-adjus ed measu e is he T eyno a io (T eyno , 1965). The pu pose o his
measu e is o de e mine whe he an in es o has been adequa ely compensa ed o he isk aken by
being abo e he ma ke . I is calcula ed using he ollowing equa ion:
𝑇𝑟𝑒𝑦𝑛𝑜𝑟 𝑅𝑎𝑡𝑖𝑜= 𝑅𝑖,𝑡−𝑅𝑓,𝑡
𝛽𝑖
(7)
Whe e:
𝑅𝑖,𝑡 – Re u n o he und 𝑖 a mon h 𝑡
𝑅𝑓,𝑡 – Re u n o he isk- ee asse a he beginning o mon h 𝑡
𝛽𝑖 – Be a o he und 𝑖
Compa ing equa ion (7) wi h equa ion (5), i is possible o see ha while he Sha pe a io uses s anda d
de ia ion as a measu e o ola ili y, he T eyno a io uses he be a coe icien . The be a used o
calcula e equa ion (7) is ob ained om he eg ession desc ibed in subsec ion 3.1.2.
In e ms o in e p e ing he T eyno a io, he highe he a io, he be e he pe o mance and
he e o e he highe he e u n ea ned o each uni o sys ema ic isk aken. A a io less han ze o
shows ha he in es men has unde pe o med he ma ke , while a a io g ea e han ze o shows
ou pe o mance.
3.1.4. Jensen’s Alpha
The Jensen’s alpha (Jensen, 1968) is an absolu e measu e de ined as he expec ed e u n o he
in es o ’s po olio ela i e o he ma ke e u ns gi en i s le el o sys ema ic isk. I is o mula ed as
ollows:
12
𝐽𝑒𝑛𝑠𝑒𝑛′𝑠 𝑎𝑙𝑝ℎ𝑎=𝑅𝑖,𝑡−[𝑅𝑓,𝑡+𝛽𝑖×(𝑅𝑚,𝑡−𝑅𝑓,𝑡)]
(8)
Whe e:
𝑅𝑖,𝑡 – Re u n o he und 𝑖 a mon h 𝑡
𝑅𝑓,𝑡 – Re u n o he isk- ee asse a he beginning o mon h 𝑡
𝑅𝑚,𝑡 – Re u n o he ma ke a mon h 𝑡
𝛽𝑖 – Be a o he und 𝑖
In e ms o in e p e ing he Jensen’s alpha, his measu e assesses he con ibu ion o he manage 's
decisions o he pe o mance o he po olio. The highe he alpha, he be e he manage 's
pe o mance. Mo eo e , i Jensen’s alpha is nega i e means ha he und manage has
unde pe o med he ma ke , whe eas i i is posi i e indica es ha he und manage has
ou pe o med he ma ke .
13
3.2. DATA
Fo he analysis o he pe o mance di e ences be ween g een bond unds, bond mu ual unds and
bond ETFs, da a was ob ained om he Bloombe g pla o m. A und sc eening ool was used o
gene a e a lis o unds based on a se o sea ch c i e ia. The e o e, some common il e s we e applied.
Only unds wi h an ac i e ma ke s a us and a p ima y sha e class o 'Yes' we e included, o a oid
possible bias om he a ie y o in es men classes ha ha e he same a ibu es and a e managed
by he same und manage s. Mo eo e , as he main in en ion o his s udy is bond unds, only ixed
income was selec ed in he und asse class ocus. By Bloombe g’s de ini ion, a und classi ied as ixed
income means ha a leas 80% is in es ed in ixed income secu i ies.
Apa om using common c i e ia, se e al speci ic il e s we e applied o dis inguish each ca ego y o
und. Fo mu ual unds, he und ype selec ed was open-ended and o ETFs, as he name sugges s,
he ype o und was “ETF”. Fo g een unds, since hey a e bo h mu ual unds and ETFs, he gene al
a ibu e "En i onmen ally F iendly" was used in addi ion o he p e ious il e s. In o de o a oid
o e lapping da a in he sample, his las a ibu e was excluded om he i s wo ca ego ies o unds.
Fu he mo e, o he pu poses o his disse a ion, we e conside ed unds domiciled in he Eu opean
egion and wi h a ime ho izon om Janua y 2005 o Decembe 2019. This pe iod was chosen in o de
o assess he po en ial impac o he 2007-2008 global inancial c isis on he pe o mance o he abo e-
men ioned unds. Funds wi h no da a o wi h less han six mon hs o da a a ailable on Bloombe g we e
excluded om he sample. Funds ha disappea ed o we e me ged/in eg a ed in o o he unds we e
also no conside ed. Thus, he sample is no comple ely ee om su i o ships bias because i only
conside s he unds wi h unin e up ed end-o - he-mon h NAV be ween Janua y 2005 and Decembe
2019 o de i e mon hly a es o e u n.
Gi en ha he p ices we e quo ed in di e en cu encies (e.g., US Dolla , B i ish Pound, e c.) and o
a oid dealing wi h mul i-cu ency da a, he p ices we e all collec ed in Eu o. The NAVs do no include
managemen ees, he e o e he analysis and he esul s will be p esen ed om an in es o 's
pe spec i e.
The applica ion o he be o e men ioned me hodology also equi es a benchma k and a isk- ee a e.
The Bloombe g Eu o Agg ega e Bond To al Re u n Index was chosen as he benchma k and "p oxy" o
he Eu opean ixed income ma ke . This index measu es he ma ke o in es men g ade, ixed a e
bonds denomina ed in Eu o, including T easu ies, go e nmen , co po a e and secu i ised issues, and
is he e o e aligned wi h he po olios unde analysis. Al hough speci ic indices exis o benchma king
g een bonds, i was decided o use a single ma ke benchma k o acili a e in e p e a ion and
compa ison o he pe o mance o he h ee po olios. The Bloombe g Eu o Agg ega e Bond To al
Re u n Index, launched in June 1998 and main ained by Bloombe g, is also widely used by in es o s
and und manage s as a benchma k o he pe o mance o eu o-denomina ed ixed income secu i ies.
I p o ides a comp ehensi e iew o he eu ozone bond ma ke and is an impo an ool o assessing
he pe o mance o in es men po olios.
Fo he isk- ee a e, conside a ion was gi en o using one-mon h Ge many Go e nmen Bond,
howe e , he one-mon h Eu ibo was chosen as i is a gene ally accep ed ma ke e e ence and has a
high liquidi y p o ided by he eu ozone in e bank ma ke . Eu ibo a es a e based on he in e es a es
a which Eu opean c edi ins i u ions bo ow unds om each o he . This ma ke a e e e en ial
14
e lec s highes daily ading olume o business and he ma ke in e enien s deno e a bes quali y
c edi a ing, high e hical s anda ds, and an excellen epu a ion. The le el o Eu ibo a es is p ima ily
de e mined by he law o supply and demand, bu he e a e also ex e nal ac o s ha can in luence
his le el, such as in la ion and economic g ow h (Eu opean Money Ma ke s Ins i u e, 2023). Bo h he
benchma k index and he isk- ee a e ha e been ex ac ed om he Bloombe g pla o m.
O e all, he inal sample comp ises 3,100 bond mu ual unds, 328 bond ETFs and 56 g een bond unds,
which ansla es in a o al o 3,484 unds. Figu e 3.1 shows he dis ibu ion o o al numbe o bond
unds in each yea o he sample pe iod (2005-2019).
Figu e 3.1 - To al numbe o unds by yea
The o e all inc ease in he o al numbe o unds obse ed is ema kable, wi h an inc ease o a ound
262% since he i s yea . As he e we e 963 bond unds in 2005 and 3,484 bond unds in 2019. The
la ges inc ease was in bond ETFs, wi h an a e age annual g ow h o 131% o e he pe iod, ollowed
by g een bond unds, which ha e epo ed an a e age annual g ow h o 20% h oughou he obse ed
pe iod. Finally, he smalles , bu s ill no able, g ow h was seen in bond mu ual unds, which inc eased
by 17% pe yea , on a e age.
A he end o he obse ed pe iod, bond mu ual unds ep esen 89% o he whole sample, being he
bigges g oup o unds conside ed. Bond ETFs accoun o 9.4% o he o al unds conside ed, making
hem he second la ges g oup, while g een bond unds made up he emaining 1.6% o he sample,
ep esen ing he smalles numbe o obse ed unds.
In addi ion, a he han jus looking a he o al numbe o obse a ions, i migh also be in e es ing o
look a how many new bond unds we e added o he sample each yea . The Figu e 3.2 displays he
numbe o new bond mu ual unds, bond ETFs and g een bond unds launched each yea .

15
Figu e 3.2 - To al numbe o new unds pe yea
Since 2005, he e ha e been pe iods o ups and downs in he numbe o new bond unds launched
each yea . Fo ins ance, 2008 has he smalles inc ease in he numbe o bond und incep ions du ing
he sample pe iod, accoun ing o only 95 new bond unds, while he pe iod om 2016 o 2018
epo ed he la ges inc ease in he numbe o bond unds incep ed each yea , om 216 o 288 new
bond unds.
Rega ding g een bond unds, Figu e 3.2 also shows ha he e we e no new unds in 2007, so he e was
no g ow h ha yea . In 2009, he e was a s ong g ow h in his ca ego y o und, wi h six new unds
launched. In he ollowing yea s, he g ow h was less e iden , bu since 2017 un il 2019, he g ow h
a e has been highe again, wi h six new unds c ea ed in each yea .
In o de o be e analyse he di e ences in pe o mance o each ca ego y o und o e he pe iod,
he sample was di ided in o h ee subpe iods. A e he in e na ional inancial c isis in he Uni ed
S a es and G ea B i ain in mid-2007, we obse ed he p oli e a ion o i s e ec s on o he Eu opean
ma ke s. Wi h he s a o he so e eign deb c isis, inancial ma ke s expe ienced a pe iod wi h a high
le el o unce ain y ha , gi en he ola ili y le els, e lec s dis inc in es men condi ions han in
p e ious pe iods.
The e o e, he i s subpe iod co esponds o he p e-c isis pe iod, om Janua y 2005 o July 2007; he
second subpe iod co esponds o he US c edi c isis and he subsequen Eu opean so e eign deb
16
c isis, om Augus 2007 o Decembe 2012; and he hi d subpe iod, co esponds o he pos -c isis
pe iod, a e Decembe 2012 un il Decembe 2019.
Using equa ions (1) and (2) om sec ion 3.1., he mon hly e u ns ha e been calcula ed, espec i ely,
o each und and o he benchma k index. Table 3.1 gi es a desc ip i e summa y o each ca ego y
o und in he h ee men ioned subpe iods. Fo he pu pose o he calcula ions, in each subsample,
unds wi h less han six mon hs o da a we e no included.
Table 3.1 - Desc ip i e s a is ics o mon hly e u ns o bond mu ual unds, bond ETFs, g een bond
unds and benchma k index
P e
Du ing
Pos
A e age
Re u ns (%)
Mu ual
0.039
0.315
0.157
ETF
-0.184
0.350
0.188
G een
-0.017
0.342
0.015
Benchma k
0.013
0.551
0.353
1s Qua ile (%)
Mu ual
-0.200
0.113
0.008
ETF
-0.333
0.156
0.035
G een
-0.071
0.187
-0.320
Benchma k
-0.835
-1.370
-0.733
Median (%)
Mu ual
0.031
0.261
0.134
ETF
-0.180
0.363
0.183
G een
0.019
0.302
0.001
Benchma k
0.295
0.186
0.161
3 d Qua ile (%)
Mu ual
0.189
0.510
0.281
ETF
0.033
0.526
0.291
G een
0.072
0.506
0.162
Benchma k
0.886
1.819
1.304
Minimum (%)
Mu ual
-2.085
-10.755
-8.007
ETF
-0.726
-2.228
-1.448
G een
-0.309
0.054
-0.394
Benchma k
-2.486
-4.834
-3.185
Maximum (%)
Mu ual
2.981
3.387
9.472
ETF
0.125
1.910
1.312
G een
0.221
0.913
0.931
Benchma k
3.102
6.475
6.823
S anda d
De ia ion (%)
Mu ual
1.354
2.373
1.875
ETF
1.280
2.013
1.556
G een
1.583
2.202
1.545
Benchma k
1.298
2.507
1.646
In his able “P e” ep esen s he p e-c isis pe iod (Jan2005-Jul2007), “Du ing” ep esen s he c isis pe iod (Aug2007-
Dec2012) and “Pos ” ep esen s he pos -c isis pe iod (Jan2013-Dec2019).
Conside ing Table 3.1, in he p e-c isis pe iod bond mu ual unds had he highes a e age mon hly
e u n (0.039%), in he c isis and pos -c isis pe iods he highes a e age mon hly e u ns we e gi en
by bond ETFs (0.350% and 0.188% espec i ely). Al hough hese we e he highes mon hly a e age
17
e u ns, only in he p e-c isis pe iod did bond mu ual unds ou pe o m he benchma k (Bloombe g
Eu o Agg ega e Bond To al Re u n Index), while in he o he wo subpe iods none o he bond unds
ou pe o med he benchma k e u n.
I can also be seen ha all ca ego ies o unds had he lowes a e age mon hly e u ns om Janua y
2005 o July 2007 and he highes a e age mon hly e u ns om Augus 2007 o Decembe 2012. This
may be due o an e en known as " ly- o-sa e y". This is a inancial e m used o desc ibe a
phenomenon in which in es o s mo e money ou o ola ile asse s and in o sa e in es men s du ing
pe iods o economic unce ain y o ma ke u bulence. This ligh o sa e y is o en seen as a sign o
ma ke dis ess and can be igge ed by a ious e en s such as poli ical ins abili y, na u al disas e s, o
economic down u ns. Du ing such e en s, in es o s may sell hei equi ies, which a e ypically seen as
iskie asse s, and buy sa e in es men s such as go e nmen bonds o gold. This ligh o sa e y can
cause he p ices o hese sa e asse s o ise, while he p ices o iskie asse s may all. This is aligned
wi h he li e a u e e iew conduc ed (Filip e al., 2015).
In ola ili y e ms and excep o he mu ual bond unds and g een bond unds in he p e-c isis pe iod,
and he mu ual bond unds in he pos -c isis pe iod, all h ee ca ego ies o bond unds epo small
s anda d de ia ion han he benchma k. Bond ETFs had he lowes s anda d de ia ion in he p e-c isis
and c isis pe iods (1.280% and 2.013%, espec i ely), while in he pos -c isis pe iod i was g een bond
unds ha had a lowe s anda d de ia ion (1.545%), making hem he leas isky unds in each
subpe iod. In con as , in he p e-c isis pe iod, g een bond unds a e he iskies ela i e o o he s
(1.583%), and in he c isis and pos -c isis pe iods, bond mu ual unds a e he iskies , wi h a s anda d
de ia ion o 2.373% and 1.875% espec i ely.
18
4. RESULTS AND DISCUSSION
This chap e p esen s he esul s o he e e ed isk-adjus ed measu es. The compa a i e analysis o
he pe o mance is p esen ed in sec ion 4.1 using he Sha pe a io; in sec ion 4.2 using he T eyno
a io; and in sec ion 4.3 using he Jensen’s Alpha. The aim is o see whe e can be obse ed he bes
pe o mance in e ms o hese h ee indica o s.
Some no es o keep in mind o he nex sec ions:
• No e ha om he e on, he wo d "bond" can be omi ed, i.e., "bond mu ual unds" can simply
be called "mu ual unds", "bond ETFs" can be called "ETFs" and "g een bond unds" can be
called "g een unds";
• All calcula ions assume a 5% s a is ical signi icance le el.
A e ob aining he esul s o hese pe o mance measu es, s a is ical es s we e ca ied ou . The F-
es is used o compa e wo a iances (Da ies e al., 1968) and, acco ding o ha , he T- es s a e used
o compa e he means. I he esul o he F- es was ha he a iances we e equal, hen he wo-
sample T- es was used (S uden , 1908), o he wise i he a iances we e unequal, hen he Welch wo-
sample T- es was used (Welch, 1947). The esul s o all es s a e displayed in appendices A, B and C.
4.1. SHARPE RATIO
Table 4.1 shows he Sha pe a io esul s o bond mu ual unds, bond ETFs and g een bond unds in
each subpe iod, calcula ed using equa ion (5).
Table 4.1 - Sha pe a io esul s
P e
Du ing
Pos
To al pe iod
A e age
Mu ual
0.294
0.247
0.125
0.191
ETF
-0.100
0.231
0.116
0.140
G een
0.001
0.196
0.034
0.078
Posi i e (%)
Mu ual
53.8
89.2
77.1
76.5
ETF
27.6
91.2
81.7
81.6
G een
56.2
100.0
50.0
66.0
Nega i e (%)
Mu ual
46.2
10.8
22.9
23.5
ETF
72.4
8.8
18.3
18.4
G een
43.8
0.0
50.0
34.0
To al o unds
Mu ual
1,043
1,729
3,100
3,100
ETF
29
160
328
328
G een
16
31
56
56
In his able “P e” ep esen s he p e-c isis pe iod (Jan2005-Jul2007), “Du ing” ep esen s he c isis pe iod (Aug2007-
Dec2012), “Pos ” ep esen s he pos -c isis pe iod (Jan2013-Dec2019) and “To al pe iod” ep esen s he en i e pe iod unde
s udy (Jan2005-Dec2012). “A e age” ep esen s he mon hly a e age o he Sha pe a io, “Posi i e (%)” and “Nega i e (%)”
ep esen he p opo ion o unds wi h a posi i e o nega i e Sha pe a io espec i ely and “To al o unds” ep esen s he
o al numbe o unds conside ed.
25
5. CONCLUSIONS AND FURTHER DEVELOPMENTS
The e has been an inc ease in ESG in es men in ecen yea s, as bo h in es o s and inancial
ins i u ions ha e conside ed he impo ance o sus ainabili y in hei in es men s. Ne e heless, ESG
in es ing is s ill a wo k in p og ess. As s a ed abo e, he main d i e o manage s and in es o s o
conside sus ainable s a egies is he expec a ion o highe long- e m e u ns.
F om he ESG p inciples, he en i onmen al ac o is he one ha e lec s he impac o a company in
he en i onmen and i s e o s o educe i s ca bon oo p in . The g een bond unds a e inse ed in o
his ac o , and i is impo an o unde s and i his ca ego y o unds p esen a highe e u n han hei
con en ional coun e pa ies. The e o e, his disse a ion p o ides a compa a i e pe o mance analysis
be ween g een bond unds, bond mu ual unds and bond ETFs.
To his end, mon hly NAVs o 3,484 unds we e ex ac ed om Janua y 2005 o Decembe 2019. The
da a consis s in 3,100 bond mu ual unds, 328 bond ETFs and 56 g een bond unds, all domiciled in
Eu ope and p e iously con e ed o Eu o cu ency. Fo analysis pu poses, he da a was di ided in
h ee dis inc subpe iods: he p e-c isis pe iod (Janua y 2005 - July 2007), c isis pe iod (Augus 2007 -
Decembe 2012) and he pos -c isis pe iod (Janua y 2013 - Decembe 2019).
The s udy s a ed by compu ing he mon hly e u ns o all unds and conduc ing a p elimina y s a is ic
o assess, among o he s, he a e age e u ns as well as he le el o isk o each, i.e., he s anda d
de ia ion o each ca ego y o und ac oss he h ee subpe iods. I was du ing he c isis ha he highe
ola ili y was obse ed, bu also whe e he highe a e age e u ns we e achie ed, h oughou all und
ca ego ies.
In o de o assess he pe o mance o each und ca ego y, h ough all he men ioned subpe iods, some
isk-adjus ed measu es ha e been applied o each und, such as he Sha pe a io, he T eyno a io
and he Jensen’s alpha, along wi h he F- es s o compa e wo a iances and wo-sample T- es s o
he means.
The indings e eal ha in he p e-c isis pe iod, bond mu ual unds had he highes a e age
pe o mances o all he calcula ed isk-adjus ed measu es, compa ed o he o he wo und
ca ego ies. The ob ained T- es s showed ha bond mu ual unds a e age e u ns we e s a is ically
signi ican o bo h Sha pe and T eyno a ios a he 95% con idence le el when compa ed wi h he
o he wo und ca ego ies. This is only he case o Jensen's alpha when compa ed o bond ETFs. In he
c isis pe iod, al hough bond mu ual unds had he highe a e age Sha pe and T eyno a ios, and g een
bond unds had he highe a e age Jensen's alpha, he e is no s a is ical e idence ha hese bond
unds ou pe o med he o he und ca ego ies in his pe iod. Finally, in he pos -c isis pe iod, he e is
e idence o conclude ha g een bond unds unde pe o med he o he wo ca ego ies, o bo h
Sha pe a io and Jensen’s alpha. In addi ion, in e ms o Jensen’s alpha, bond mu ual unds
ou pe o med bond ETFs. Abou he T eyno a io, no hing could be concluded, since he none o he
T- es s e eal s a is ical di e ences be ween he means.
Analysing each und ca ego y sepa a ely o e he di e en subpe iods, he ollowing conclusions can
be d awn. All und ca ego ies ha e a consis en beha iou o e ime, e idencing i s bes a e age
pe o mance in he c isis pe iod.

26
O e all, bond mu ual unds a e conside ed o be he bes pe o me s, ollowed by g een bond unds
and bond ETFs wi h a simila pe o mance. As men ioned abo e, all und ca ego ies had pe o med
be e in he c isis pe iod, sugges ing ha in es o s seek o less isky in es men s in imes o high
ola ili y and, he e o e, op o sa e in es men s. This can be conside ed a ly- o-sa e y e en , whe e
he p ices o sa e asse s ise, and iskie asse s’ p ices ha e a downwa d endency.
Al hough g een bond unds ou pe o med in some momen s, he small numbe o unds in he sample
may ha e led o he conclusions no being s a is ically signi ican . Addi ionally, in es o s seek a highe
e u n wi h he lowes possible isk, and he e is a wide ange o in es men op ions. Al hough g een
bond unds o e lowe e u ns han o he ypes o unds, some en i onmen ally conscious in es o s
may p e e o in es in hese asse s due o he con ibu ion ha hese unds p o ide o he
en i onmen .
As men ioned in chap e 1, he main objec i e o his s udy was o unde s and whe he g een bond
unds would ou pe o m hei pee s, ei he bond mu ual unds o bond ETFs. Howe e , one o he
main limi a ions was ob aining da a. The bes ools equi e an annual o mon hly subsc ip ion and hei
in e aces a e no use - iendly o ob aining in o ma ion, as is he case wi h Bloombe g's e minal.
Addi ionally, by selec ing only he subse o he popula ion ha "su i ed" o pe sis ed o e he en i e
sample pe iod, one excludes he subse o unds ha we e discon inued due o poo pe o mance, and
his can cause a su i o ship bias. Igno ing his poo pe o mance o some unds can esul in o a
misleading unde s anding o his o ical e u ns and an inaccu a e ep esen a ion o po olio isk and
e u n measu es. To add ess his, he ull subse o he popula ion should be conside ed, i.e., including
hose ha pe sis ed o e ime, bu also hose ha ceased o exis a some poin du ing he sample
pe iod.
Fo he pu poses o his s udy, only he ne alues we e analysed, i.e., a e deduc ing commissions.
All he analysis was ca ied ou om he in es o 's poin o iew, bu i migh be in e es ing o do he
analysis om he manage 's pe spec i e, aking in o accoun he g oss alues, i.e., be o e deduc ing
commissions. Ano he in e es ing u u e analysis would be o look a a di e en geog aphical a ea
whe e he unds a e domiciled. This esea ch ocuses only on unds domiciled in Eu ope, bu looking
a unds domiciled in o he egions, o e en wi hou any es ic ion on domicile, could lead o di e en
esul s. And ano he possible app oach in his s udy would be o model ola ili y using ARCH, GARCH
o EGARCH me hods ins ead o calcula ing his o ical s anda d de ia ion.
As a inal sugges ion, i would be pe inen o conside o he ime pe iods. This esea ch has only
ocused on he 2007-2008 inancial c isis, bu a e he Co id-19 epidemic c isis and he wa in Uk aine,
which caused an excep ional inc ease in he in la ion a e, i will also be impo an o unde s and how
his ca ego y o unds beha es nowadays.
27
BIBLIOGRAPHY
Abou Eu ibo ® | The Eu opean Money Ma ke s Ins i u e (EMMI). (2023). emmi-Benchma ks.eu.
h ps://www.emmi-benchma ks.eu/benchma ks/eu ibo /
Aglia di, E., & Chechulin, V. (2020). G een Bonds s Regula Bonds: Deb Le el and Co po a e
Pe o mance. Jou nal o Co po a e Finance Resea ch / Корпоративные Финансы | ISSN: 2073-0438,
14(2), 83–99. h ps://doi.o g/10.17323/j.jc .2073-0438.14.2.2020.83-99
Bacon, C. R. (2008). P ac ical po olio pe o mance measu emen and a ibu ion. Chiches e Wiley.
Be na do, A. E., & Ledoi , O. (2000). Gain, Loss, and Asse P icing. Jou nal o Poli ical Economy, 108(1),
144–172. h ps://doi.o g/10.1086/262114
Biglo a, A., O obelli, S., Rache , S. T., & S oyano , S. (2004). Di e en App oaches o Risk Es ima ion
in Po olio Theo y. The Jou nal o Po olio Managemen , 31(1), 103–112.
h ps://doi.o g/10.3905/jpm.2004.443328
Black, F. (1972). Capi al Ma ke Equilib ium wi h Res ic ed Bo owing. The Jou nal o Business, 45(3),
444. h ps://doi.o g/10.1086/295472
Bodie, Z., Kane, A., & Ma cus, A. J. (2014). Essen ials o in es men s (10 h ed.). Mcg aw-Hill Educa ion.
Bolle sle , T. (1986). Gene alized au o eg essi e condi ional he e oskedas ici y. Jou nal o
Econome ics, 31(3), 307–327.
Chang, C. E., Nelson, W. A., & Doug Wi e, H. (2012). Do g een mu ual unds pe o m well?
Managemen Resea ch Re iew, 35(8), 693–708. h ps://doi.o g/10.1108/01409171211247695
Choudh y, M. (2010). An in oduc ion o bond ma ke s. Wiley.
Conno , G., & Ko ajczyk, R. A. (1986). Pe o mance measu emen wi h he a bi age p icing heo y.
Jou nal o Financial Economics, 15(3), 373–394. h ps://doi.o g/10.1016/0304-405x(86)90027-9
Da ies, O. L., Snedeco , G. W., & Coch an, W. G. (1968). S a is ical Me hods. 6 h Edi ion. Applied
S a is ics, 17(3). h ps://doi.o g/10.2307/2985653
De ibew, H.N. (2017). Risks and Re u ns o G een Bonds. Mas e ’s Thesis.
Dixon-Fowle , H., Sla e , D., Johnson, J., Ells and, A., & Romi, A. (2013). Beyond “‘Does i Pay o be
G een?’” A Me a-Analysis o Mode a o s o he CEP-CFP Rela ionship. Jou nal o business e hics,
112(2), 353-366.
Engle, R. F. (1982). Au o eg essi e Condi ional He e oscedas ici y wi h Es ima es o he Va iance o
Uni ed Kingdom In la ion. Econome ica, 50(4), 987–1007. h ps://doi.o g/10.2307/1912773
ESG unds ha e lec wha ma e s mos o you | Vangua d. (2022). In es o . angua d.com.
h ps://in es o . angua d.com/in es men -p oduc s/esg
28
Fama, E. F. (1972). Componen s o In es men Pe o mance. The Jou nal o Finance, 27(3), 551–567.
h ps://doi.o g/10.1111/j.1540-6261.1972. b00984.x
Filip, A.-M., Pece, A., & Lăcă uş, V. (2015). Risk-adjus ed Pe o mance o Romanian Bond Funds du ing
he Global Economic C isis. P ocedia Economics and Finance, 32, 1407–1413.
h ps://doi.o g/10.1016/s2212-5671(15)01517-8
Gemmill, G., Hwang, S., & Salmon, M. (2006). Pe o mance measu emen wi h loss a e sion. Jou nal
o Asse Managemen , 7(3-4), 190–207. h ps://doi.o g/10.1057/palg a e.jam.2240213
G eenhouse Gas Bulle in: Ano he Yea Ano he Reco d | Wo ld Me eo ological O ganiza ion. (2021,
Oc obe 25). h ps://public.wmo.in /en/media/p ess- elease/g eenhouse-gas-bulle in-ano he -yea -
ano he - eco d
Hull, J. (2018). Risk managemen and inancial ins i u ions (5 h ed.). John Wiley & Sons, Inc.
Inde s , G., & S ewa , F. (2018). Inco po a ing En i onmen al, Social and Go e nance (ESG) Fac o s
in o Fixed Income In es men . SSRN Elec onic Jou nal. h ps://doi.o g/10.2139/ss n.3175830
Is aëlsen, C. L. (2005). A e inemen o he Sha pe a io and in o ma ion a io. Jou nal o Asse
Managemen , 5(6), 423–427. h ps://doi.o g/10.1057/palg a e.jam.2240158
Jaksic, M., Leko ic, M., & Milano ic, M. (2015). Measu ing he pe o mance o mu ual unds: A case
s udy. Indus ija, 43(1), 37–51. h ps://doi.o g/10.5937/indus ija43-6677
Jensen, M. C. (1968). The Pe o mance o Mu ual Funds in he Pe iod 1945-1964. Jou nal o Finance,
23(2). h ps://doi.o g/10.2139/ss n.244153
Kea ing, C., & Shadwick, W. F. (2002). A Uni e sal Pe o mance Measu e. Jou nal o Pe o mance
Measu emen , 6, 59–84.
Ma kowi z, H. (1952). Po olio Selec ion. The Jou nal o Finance, 7(1), 77–91.
Nelson, D. B. (1991). Condi ional He e oskedas ici y in Asse Re u ns: A New App oach. Econome ica,
59(2), 347. h ps://doi.o g/10.2307/2938260
Néme h-Du kó, E., & Hegedűs, A. (2021). Clima e Change in he Capi al Ma ke s: A S udy o Ac i ely
Managed G een Bond Funds. Financial and Economic Re iew, 20(4), 38–64.
h ps://doi.o g/10.33893/ e .20.4.3864
Nizam, E., Ng, A., Dewanda u, G., Nagaye , R., & Nkoba, M. A. (2019). The impac o social and
en i onmen al sus ainabili y on inancial pe o mance: A global analysis o he banking sec o . Jou nal
o Mul ina ional Financial Managemen , 49, 35–53. h ps://doi.o g/10.1016/j.mul in.2019.01.002
Pa adino ic, I. (2017). Compa ison o he pe o mance o Islamic, S i and g een mu ual unds.
Run.unl.p . h ps:// un.unl.p /handle/10362/25469
Pézie , J., & Whi e, A. (2006). The Rela i e Me i s o In es able Hedge Fund Indices and o Funds o
Hedge Funds in Op imal Passi e Po olios. h ps://co e.ac.uk/download/pd /6560807.pd
29
Sama bakhsh, L., & Shah, M. (2021). Fixed income mu ual und pe o mance du ing and a e a c isis:
a Canadian case. Jou nal o Economics and Finance. h ps://doi.o g/10.1007/s12197-021-09541-z
Sampei, H. (2018, Sep embe 12). ESG awa eness is an endu ing legacy o he global inancial c isis.
Fideli y In e na ional. h ps://www. ideli yin e na ional.com/edi o ial/blog/pesg-awa eness-is-an-
endu ing-legacy-o - he-global- inancial-c isisp-a5a9 2-en5/
Sha pe, W. F. (1963). A Simpli ied Model o Po olio Analysis. Managemen Science, 9(2), 277–293.
h ps://doi.o g/10.1287/mnsc.9.2.277
Sha pe, W. F. (1966). Mu ual Fund Pe o mance. The Jou nal o Business, 39(1), 119–138.
h ps://doi.o g/10.1086/294846
Sha pe, W. F. (1994). The Sha pe Ra io. The Jou nal o Po olio Managemen , 21(1), 49–58.
h ps://doi.o g/10.3905/jpm.1994.409501
So ino, F., an de Mee , R. A. H., & Plan inga, A. (1999). The Du ch iangle - A amewo k o measu e
upside po en ial ela i e o downside isk. Jou nal o Po olio Managemen , 26(1), 50–58.
S i as a a, S. C., & Essayyad, M. (1994). In es iga ing a new me hodology o anking in e na ional
mu ual unds. Jou nal o Economics and Finance, 18(3), 241–260. h ps://doi.o g/10.1007/b 02920485
S anke ičienė, J., & Pe onienė, I. (2019). Bond Mu ual Funds s. Bond Exchange T aded Funds:
E alua ion o Risk Adjus ed Pe o mance. Adminis a i e Sciences, 9(2), 31.
h ps://doi.o g/10.3390/admsci9020031
S uden . (1908). The P obable E o o a Mean. Biome ika, 6(1), 1. h ps://doi.o g/10.2307/2331554
T eyno , J. L. (1965). How o a e managemen o in es men unds. Ha a d business e iew, 43(1),
63-75.
Vaidya, D. (2017). Risk Adjus ed Re u n | Top 6 Risk Ra ios You mus Know! Walls ee Mojo.
h ps://www.walls ee mojo.com/ isk-adjus ed- e u ns/
Welch, B. L. (1947). The Gene aliza ion o `S uden ’s’ P oblem when Se e al Di e en Popula ion
Va iances a e In ol ed. Biome ika, 34(1/2), 28. h ps://doi.o g/10.2307/2332510
30
APPENDICES
A. Sha pe a io s a is ical es s esul s
Table 1A. F- es o compa e wo a iances o he a e age Sha pe a ios (by subpe iod)
𝑯𝟎
SD o x
SD o y
D o x
D o y
F-S a is ic
P- alue
Me hod
Decision
σ𝑀𝑃𝑟=σ𝐸𝑃𝑟
0.982
0.194
1,042
28
25.60
3.6E-15
F es
Rejec 𝐻0
σ𝑀𝑃𝑟=σ𝐺𝑃𝑟
0.982
0.136
1,042
15
52.01
6.5E-11
F es
Rejec 𝐻0
σ𝐸𝑃𝑟=σ𝐺𝑃𝑟
0.194
0.136
28
15
2.03
0.151
F es
No ejec 𝐻0
σ𝑀𝐷=σ𝐸𝐷
0.479
0.238
1,728
159
4.05
0.0E+00
F es
Rejec 𝐻0
σ𝑀𝐷=σ𝐺𝐷
0.479
0.199
1,728
30
5.82
2.2E-07
F es
Rejec 𝐻0
σ𝐸𝐷=σ𝐺𝐷
0.238
0.199
159
30
1.44
0.242
F es
No ejec 𝐻0
σ𝑀𝑃𝑡=σ𝐸𝑃𝑡
0.203
0.210
3,099
327
0.93
0.376
F es
No ejec 𝐻0
σ𝑀𝑃𝑡=σ𝐺𝑃𝑡
0.203
0.208
3,099
55
0.95
0.750
F es
No ejec 𝐻0
σ𝐸𝑃𝑡=σ𝐺𝑃𝑡
0.210
0.208
327
55
1.02
0.957
F es
No ejec 𝐻0
This able p esen s he F- es o compa e wo a iances o he a e age Sha pe a ios, whe e 𝐻0 is he null hypo hesis,
ep esen ing ha he di e ences in he a iances a e ze o and he al e na i e hypo hesis, 𝐻1, ep esen s ha he di e ences
in he a iances a e di e en om ze o. The “decision” was made a he 5% signi icance le el. “σ” ep esen s he popula ion
a iance. “M”, “E” and “G” ep esen s he bond mu ual, bond ETF and g een bond unds, espec i ely. “P ” ep esen s he
p e-c isis pe iod (Jan2005-Jul2007), “D” ep esen s he c isis pe iod (Aug2007-Dec2012) and “P ” ep esen s he pos -c isis
pe iod (Jan2013-Dec2019). “SD o x” and “SD o y” ep esen s he s anda d de ia ion o he a iables and “D o x” and “D
o y” ep esen s he deg ees o eedom o he a iables.
Table 1B. Two-sample T- es o he a e age Sha pe a ios (by subpe iod)
𝑯𝟎
Mean
o x
Mean
o y
T-
S a is ic
P- alue
D
Me hod
Decision
𝜇𝑀𝑃𝑟=𝜇𝐸𝑃𝑟
0.294
-0.100
8.36
1.5E-12
80.9
T es (Unequal a )
Rejec 𝐻0
𝜇𝑀𝑃𝑟=𝜇𝐺𝑃𝑟
0.294
0.001
6.42
5.6E-08
48.0
T es (Unequal a )
Rejec 𝐻0
𝜇𝐸𝑃𝑟=𝜇𝐺𝑃𝑟
-0.100
0.001
-1.84
0.073
43.0
T es (Equal a )
No ejec 𝐻0
𝜇𝑀𝐷=𝜇𝐸𝐷
0.247
0.231
0.71
0.478
296.6
T es (Unequal a )
No ejec 𝐻0
𝜇𝑀𝐷=𝜇𝐺𝐷
0.247
0.196
1.35
0.184
36.6
T es (Unequal a )
No ejec 𝐻0
𝜇𝐸𝐷=𝜇𝐺𝐷
0.231
0.196
0.77
0.443
189.0
T es (Equal a )
No ejec 𝐻0
𝜇𝑀𝑃𝑡=𝜇𝐸𝑃𝑡
0.125
0.116
0.69
0.488
3,426.0
T es (Equal a )
No ejec 𝐻0
𝜇𝑀𝑃𝑡=𝜇𝐺𝑃𝑡
0.125
0.034
3.29
0.001
3,154.0
T es (Equal a )
Rejec 𝐻0
𝜇𝐸𝑃𝑡=𝜇𝐺𝑃𝑡
0.116
0.034
2.70
0.007
382.0
T es (Equal a )
Rejec 𝐻0
This able p esen s he wo-sample T- es o he a e age Sha pe a ios, whe e 𝐻0 is he null hypo hesis, ep esen ing ha
he di e ences in he means a e ze o and he al e na i e hypo hesis, 𝐻1, ep esen s ha he di e ences in he means a e
di e en om ze o. The “decision” was made a he 5% signi icance le el. “𝜇” ep esen s he popula ion mean. “M”, “E” and
“G” ep esen s he bond mu ual, bond ETF and g een bond unds, espec i ely. “P ” ep esen s he p e-c isis pe iod (Jan2005-
Jul2007), “D” ep esen s he c isis pe iod (Aug2007-Dec2012) and “P ” ep esen s he pos -c isis pe iod (Jan2013-Dec2019).
“Mean o x” and “Mean o y” ep esen s he mean o he a iables and “D ” ep esen s he deg ees o eedom.

31
Table 2A. F- es o compa e wo a iances o he a e age Sha pe a ios (by und ca ego y)
𝑯𝟎
SD o x
SD o y
D o x
D o y
F-S a is ic
P- alue
Me hod
Decision
σ𝑀𝑃𝑟=σ𝑀𝐷
0.982
0.479
1,042
1,728
4.20
0.0E+00
F es
Rejec 𝐻0
σ𝑀𝑃𝑟=σ𝑀𝑃𝑡
0.982
0.203
1,042
3,099
23.43
0.0E+00
F es
Rejec 𝐻0
σ𝑀𝐷=σ𝑀𝑃𝑡
0.479
0.203
1,728
3,099
5.58
0.0E+00
F es
Rejec 𝐻0
σ𝐸𝑃𝑟=σ𝐸𝐷
0.194
0.238
28
159
0.66
0.203
F es
No ejec 𝐻0
σ𝐸𝑃𝑟=σ𝐸𝑃𝑡
0.194
0.210
28
327
0.85
0.633
F es
No ejec 𝐻0
σ𝐸𝐷=σ𝐸𝑃𝑡
0.238
0.210
159
327
1.28
0.061
F es
No ejec 𝐻0
σ𝐺𝑃𝑟=σ𝐺𝐷
0.136
0.199
15
30
0.47
0.124
F es
No ejec 𝐻0
σ𝐺𝑃𝑟=σ𝐺𝑃𝑡
0.136
0.208
15
55
0.43
0.073
F es
No ejec 𝐻0
σ𝐺𝐷=σ𝐺𝑃𝑡
0.199
0.208
30
55
0.91
0.802
F es
No ejec 𝐻0
This able p esen s he F- es o compa e wo a iances o he a e age Sha pe a ios, whe e 𝐻0 is he null hypo hesis,
ep esen ing ha he di e ences in he a iances a e ze o and he al e na i e hypo hesis, 𝐻1, ep esen s ha he di e ences
in he a iances a e di e en om ze o. The “decision” was made a he 5% signi icance le el. “σ” ep esen s he popula ion
a iance. “M”, “E” and “G” ep esen s he bond mu ual, bond ETF and g een bond unds, espec i ely. “P ” ep esen s he
p e-c isis pe iod (Jan2005-Jul2007), “D” ep esen s he c isis pe iod (Aug2007-Dec2012) and “P ” ep esen s he pos -c isis
pe iod (Jan2013-Dec2019). “SD o x” and “SD o y” ep esen s he s anda d de ia ion o he a iables and “D o x” and “D
o y” ep esen s he deg ees o eedom o he a iables.
Table 2B. Two-sample T- es o he a e age Sha pe a ios (by und ca ego y)
𝑯𝟎
Mean
o x
Mean
o y
T-
S a is ic
P- alue
D
Me hod
Decision
𝜇𝑀𝑃𝑟=𝜇𝑀𝐷
0.294
0.247
1.44
0.149
1,346.1
T es (Unequal a )
No ejec 𝐻0
𝜇𝑀𝑃𝑟=𝜇𝑀𝑃𝑡
0.294
0.125
5.53
4.1E-08
1,072.1
T es (Unequal a )
Rejec 𝐻0
𝜇𝑀𝐷=𝜇𝑀𝑃𝑡
0.247
0.125
10.12
1.6E-23
2,079.2
T es (Unequal a )
Rejec 𝐻0
𝜇𝐸𝑃𝑟=𝜇𝐸𝐷
-0.100
0.231
-7.07
2.9E-11
187.0
T es (Equal a )
Rejec 𝐻0
𝜇𝐸𝑃𝑟=𝜇𝐸𝑃𝑡
-0.100
0.116
-5.35
1.6E-07
355.0
T es (Equal a )
Rejec 𝐻0
𝜇𝐸𝐷=𝜇𝐸𝑃𝑡
0.231
0.116
5.42
9.4E-08
486.0
T es (Equal a )
Rejec 𝐻0
𝜇𝐺𝑃𝑟=𝜇𝐺𝐷
0.001
0.196
-3.52
0.001
45.0
T es (Equal a )
Rejec 𝐻0
𝜇𝐺𝑃𝑟=𝜇𝐺𝑃𝑡
0.001
0.034
-0.61
0.542
70.0
T es (Equal a )
No ejec 𝐻0
𝜇𝐺𝐷=𝜇𝐺𝑃𝑡
0.196
0.034
3.53
0.001
85.0
T es (Equal a )
Rejec 𝐻0
This able p esen s he wo-sample T- es o he a e age Sha pe a ios, whe e 𝐻0 is he null hypo hesis, ep esen ing ha
he di e ences in he means a e ze o and he al e na i e hypo hesis, 𝐻1, ep esen s ha he di e ences in he means a e
di e en om ze o. The “decision” was made a he 5% signi icance le el. “𝜇” ep esen s he popula ion mean. “M”, “E” and
“G” ep esen s he bond mu ual, bond ETF and g een bond unds, espec i ely. “P ” ep esen s he p e-c isis pe iod (Jan2005-
Jul2007), “D” ep esen s he c isis pe iod (Aug2007-Dec2012) and “P ” ep esen s he pos -c isis pe iod (Jan2013-Dec2019).
“Mean o x” and “Mean o y” ep esen s he mean o he a iables and “D ” ep esen s he deg ees o eedom.
32
B. T eyno a io s a is ical es s esul s
Table 1A. F- es o compa e wo a iances o he a e age T eyno a ios (by subpe iod)
𝑯𝟎
SD o x
SD o y
D o x
D o y
F-S a is ic
P- alue
Me hod
Decision
σ𝑀𝑃𝑟=σ𝐸𝑃𝑟
0.175
0.010
850
27
297.36
0.0E+00
F es
Rejec 𝐻0
σ𝑀𝑃𝑟=σ𝐺𝑃𝑟
0.175
0.008
850
11
507.40
1.1E-13
F es
Rejec 𝐻0
σ𝐸𝑃𝑟=σ𝐺𝑃𝑟
0.010
0.008
27
11
1.71
0.353
F es
No ejec 𝐻0
σ𝑀𝐷=σ𝐸𝐷
0.862
0.101
1,034
135
73.02
0.0E+00
F es
Rejec H0𝐻0
σ𝑀𝐷=σ𝐺𝐷
0.862
0.083
1,034
23
107.44
0.0E+00
F es
Rejec 𝐻0
σ𝐸𝐷=σ𝐺𝐷
0.101
0.083
135
23
1.47
0.282
F es
No ejec 𝐻0
σ𝑀𝑃𝑡=σ𝐸𝑃𝑡
0.268
0.058
2,346
302
21.45
0.0E+00
F es
Rejec 𝐻0
σ𝑀𝑃𝑡=σ𝐺𝑃𝑡
0.268
0.075
2,346
48
12.76
0.0E+00
F es
Rejec 𝐻0
σ𝐸𝑃𝑡=σ𝐺𝑃𝑡
0.058
0.075
302
48
0.59
0.010
F es
Rejec 𝐻0
This able p esen s he F- es o compa e wo a iances o he a e age T eyno a ios, whe e 𝐻0 is he null hypo hesis,
ep esen ing ha he di e ences in he a iances a e ze o and he al e na i e hypo hesis, 𝐻1, ep esen s ha he di e ences
in he a iances a e di e en om ze o. The “decision” was made a he 5% signi icance le el. “σ” ep esen s he popula ion
a iance. “M”, “E” and “G” ep esen s he bond mu ual, bond ETF and g een bond unds, espec i ely. “P ” ep esen s he
p e-c isis pe iod (Jan2005-Jul2007), “D” ep esen s he c isis pe iod (Aug2007-Dec2012) and “P ” ep esen s he pos -c isis
pe iod (Jan2013-Dec2019). “SD o x” and “SD o y” ep esen s he s anda d de ia ion o he a iables and “D o x” and “D
o y” ep esen s he deg ees o eedom o he a iables.
Table 1B. Two-sample T- es o he a e age T eyno a ios (by subpe iod)
𝑯𝟎
Mean
o x
Mean
o y
T-
S a is ic
P- alue
D
Me hod
Decision
𝜇𝑀𝑃𝑟=𝜇𝐸𝑃𝑟
0.024
-0.003
4.33
1.7E-05
777.1
T es (Unequal a )
Rejec 𝐻0
𝜇𝑀𝑃𝑟=𝜇𝐺𝑃𝑟
0.024
0.001
3.59
3.7E-04
440.0
T es (Unequal a )
Rejec 𝐻0
𝜇𝐸𝑃𝑟=𝜇𝐺𝑃𝑟
-0.003
0.001
-1.31
0.197
38.0
T es (Equal a )
No ejec 𝐻0
𝜇𝑀𝐷=𝜇𝐸𝐷
0.085
0.036
1.74
0.083
1,163.9
T es (Unequal a )
No ejec 𝐻0
𝜇𝑀𝐷=𝜇𝐺𝐷
0.085
0.040
1.43
0.153
246.3
T es (Unequal a )
No ejec 𝐻0
𝜇𝐸𝐷=𝜇𝐺𝐷
0.036
0.040
-0.16
0.877
158.0
T es (Equal a )
No ejec 𝐻0
𝜇𝑀𝑃𝑡=𝜇𝐸𝑃𝑡
0.011
0.007
0.52
0.600
2,159.1
T es (Unequal a )
No ejec 𝐻0
𝜇𝑀𝑃𝑡=𝜇𝐺𝑃𝑡
0.011
0.011
0.00
0.998
76.9
T es (Unequal a )
No ejec 𝐻0
𝜇𝐸𝑃𝑡=𝜇𝐺𝑃𝑡
0.007
0.011
-0.30
0.762
57.6
T es (Unequal a )
No ejec 𝐻0
This able p esen s he wo-sample T- es o he a e age T eyno a ios, whe e 𝐻0 is he null hypo hesis, ep esen ing ha
he di e ences in he means a e ze o and he al e na i e hypo hesis, 𝐻1, ep esen s ha he di e ences in he means a e
di e en om ze o. The “decision” was made a he 5% signi icance le el. “𝜇” ep esen s he popula ion mean. “M”, “E” and
“G” ep esen s he bond mu ual, bond ETF and g een bond unds, espec i ely. “P ” ep esen s he p e-c isis pe iod (Jan2005-
Jul2007), “D” ep esen s he c isis pe iod (Aug2007-Dec2012) and “P ” ep esen s he pos -c isis pe iod (Jan2013-Dec2019).
“Mean o x” and “Mean o y” ep esen s he mean o he a iables and “D ” ep esen s he deg ees o eedom.
33
Table 2A. F- es o compa e wo a iances o he a e age T eyno a ios (by und ca ego y)
𝑯𝟎
SD o x
SD o y
D o x
D o y
F-S a is ic
P- alue
Me hod
Decision
σ𝑀𝑃𝑟=σ𝑀𝐷
0.175
0.862
850
1,034
0.04
0.0E+00
F es
Rejec 𝐻0
σ𝑀𝑃𝑟=σ𝑀𝑃𝑡
0.175
0.268
850
2,346
0.43
4.7E-44
F es
Rejec 𝐻0
σ𝑀𝐷=σ𝑀𝑃𝑡
0.862
0.268
1,034
2,346
10.37
0.0E+00
F es
Rejec 𝐻0
σ𝐸𝑃𝑟=σ𝐸𝐷
0.010
0.101
27
135
0.01
5.2E-22
F es
Rejec 𝐻0
σ𝐸𝑃𝑟=σ𝐸𝑃𝑡
0.010
0.058
27
302
0.03
7.2E-16
F es
Rejec 𝐻0
σ𝐸𝐷=σ𝐸𝑃𝑡
0.101
0.058
135
302
3.05
1.3E-15
F es
Rejec 𝐻0
σ𝐺𝑃𝑟=σ𝐺𝐷
0.008
0.083
11
23
0.01
9.4E-10
F es
Rejec 𝐻0
σ𝐺𝑃𝑟=σ𝐺𝑃𝑡
0.008
0.075
11
48
0.01
1.9E-09
F es
Rejec 𝐻0
σ𝐺𝐷=σ𝐺𝑃𝑡
0.083
0.075
23
48
1.23
0.532
F es
No ejec 𝐻0
This able p esen s he F- es o compa e wo a iances o he a e age T eyno a ios, whe e 𝐻0 is he null hypo hesis,
ep esen ing ha he di e ences in he a iances a e ze o and he al e na i e hypo hesis, 𝐻1, ep esen s ha he di e ences
in he a iances a e di e en om ze o. The “decision” was made a he 5% signi icance le el. “σ” ep esen s he popula ion
a iance. “M”, “E” and “G” ep esen s he bond mu ual, bond ETF and g een bond unds, espec i ely. “P ” ep esen s he
p e-c isis pe iod (Jan2005-Jul2007), “D” ep esen s he c isis pe iod (Aug2007-Dec2012) and “P ” ep esen s he pos -c isis
pe iod (Jan2013-Dec2019). “SD o x” and “SD o y” ep esen s he s anda d de ia ion o he a iables and “D o x” and “D
o y” ep esen s he deg ees o eedom o he a iables.
Table 2B. Two-sample T- es o he a e age T eyno a ios (by und ca ego y)
𝑯𝟎
Mean
o x
Mean
o y
T-
S a is ic
P- alue
D
Me hod
Decision
𝜇𝑀𝑃𝑟=𝜇𝑀𝐷
0.024
0.085
-2.21
0.027
1,136.8
T es (Unequal a )
Rejec 𝐻0
𝜇𝑀𝑃𝑟=𝜇𝑀𝑃𝑡
0.024
0.011
1.71
0.087
2,302.2
T es (Unequal a )
No ejec 𝐻0
𝜇𝑀𝐷=𝜇𝑀𝑃𝑡
0.085
0.011
2.73
0.006
1,122.9
T es (Unequal a )
Rejec 𝐻0
𝜇𝐸𝑃𝑟=𝜇𝐸𝐷
-0.003
0.036
-4.42
1.9E-05
146.8
T es (Unequal a )
Rejec 𝐻0
𝜇𝐸𝑃𝑟=𝜇𝐸𝑃𝑡
-0.003
0.007
-2.60
0.010
239.1
T es (Unequal a )
Rejec 𝐻0
𝜇𝐸𝐷=𝜇𝐸𝑃𝑡
0.036
0.007
3.15
0.002
176.0
T es (Unequal a )
Rejec 𝐻0
𝜇𝐺𝑃𝑟=𝜇𝐺𝐷
0.001
0.040
-2.23
0.035
23.8
T es (Unequal a )
Rejec 𝐻0
𝜇𝐺𝑃𝑟=𝜇𝐺𝑃𝑡
0.001
0.011
-0.83
0.412
51.9
T es (Unequal a )
No ejec 𝐻0
𝜇𝐺𝐷=𝜇𝐺𝑃𝑡
0.040
0.011
1.51
0.136
71.0
T es (Equal a )
No ejec 𝐻0
This able p esen s he wo-sample T- es o he a e age T eyno a ios, whe e 𝐻0 is he null hypo hesis, ep esen ing ha
he di e ences in he means a e ze o and he al e na i e hypo hesis, 𝐻1, ep esen s ha he di e ences in he means a e
di e en om ze o. The “decision” was made a he 5% signi icance le el. “𝜇” ep esen s he popula ion mean. “M”, “E” and
“G” ep esen s he bond mu ual, bond ETF and g een bond unds, espec i ely. “P ” ep esen s he p e-c isis pe iod (Jan2005-
Jul2007), “D” ep esen s he c isis pe iod (Aug2007-Dec2012) and “P ” ep esen s he pos -c isis pe iod (Jan2013-Dec2019).
“Mean o x” and “Mean o y” ep esen s he mean o he a iables and “D ” ep esen s he deg ees o eedom.
34
C. Jensen’s alpha s a is ical es s esul s
Table 1A. F- es o compa e wo a iances o he a e age Jensen's alphas (by subpe iod)
𝑯𝟎
SD o x
SD o y
D o x
D o y
F-S a is ic
P- alue
Me hod
Decision
σ𝑀𝑃𝑟=σ𝐸𝑃𝑟
0.0038
0.0025
1,042
28
2.29
0.009
F es
Rejec 𝐻0
σ𝑀𝑃𝑟=σ𝐺𝑃𝑟
0.0038
0.0016
1,042
15
5.86
3.1E-04
F es
Rejec 𝐻0
σ𝐸𝑃𝑟=σ𝐺𝑃𝑟
0.0025
0.0016
28
15
2.56
0.059
F es
No ejec 𝐻0
σ𝑀𝐷=σ𝐸𝐷
0.0050
0.0045
1,728
159
1.20
0.139
F es
No ejec 𝐻0
σ𝑀𝐷=σ𝐺𝐷
0.0050
0.0028
1,728
30
3.05
4.0E-04
F es
Rejec 𝐻0
σ𝐸𝐷=σ𝐺𝐷
0.0045
0.0028
159
30
2.54
0.004
F es
Rejec 𝐻0
σ𝑀𝑃𝑡=σ𝐸𝑃𝑡
0.0044
0.0024
3,099
327
3.37
0.0E+00
F es
Rejec 𝐻0
σ𝑀𝑃𝑡=σ𝐺𝑃𝑡
0.0044
0.0030
3,099
55
2.16
4.6E-04
F es
Rejec 𝐻0
σ𝐸𝑃𝑡=σ𝐺𝑃𝑡
0.0024
0.0030
327
55
0.64
0.021
F es
Rejec 𝐻0
This able p esen s he F- es o compa e wo a iances o he a e age Jensen’s alphas, whe e 𝐻0 is he null hypo hesis,
ep esen ing ha he di e ences in he a iances a e ze o and he al e na i e hypo hesis, 𝐻1, ep esen s ha he di e ences
in he a iances a e di e en om ze o. The “decision” was made a he 5% signi icance le el. “σ” ep esen s he popula ion
a iance. “M”, “E” and “G” ep esen s he bond mu ual, bond ETF and g een bond unds, espec i ely. “P ” ep esen s he
p e-c isis pe iod (Jan2005-Jul2007), “D” ep esen s he c isis pe iod (Aug2007-Dec2012) and “P ” ep esen s he pos -c isis
pe iod (Jan2013-Dec2019). “SD o x” and “SD o y” ep esen s he s anda d de ia ion o he a iables and “D o x” and “D
o y” ep esen s he deg ees o eedom o he a iables.
Table 1B. Two-sample T- es o he a e age Jensen’s alphas (by subpe iod)
𝑯𝟎
Mean
o x
Mean
o y
T-
S a is ic
P- alue
D
Me hod
Decision
𝜇𝑀𝑃𝑟=𝜇𝐸𝑃𝑟
0.0005
-0.0007
2.41
0.022
31.7
T es (Unequal a )
Rejec 𝐻0
𝜇𝑀𝑃𝑟=𝜇𝐺𝑃𝑟
0.0005
-0.0003
1.79
0.091
17.8
T es (Unequal a )
No ejec 𝐻0
𝜇𝐸𝑃𝑟=𝜇𝐺𝑃𝑟
-0.0007
-0.0003
-0.62
0.539
43.0
T es (Equal a )
No ejec 𝐻0
𝜇𝑀𝐷=𝜇𝐸𝐷
0.0024
0.0017
1.77
0.077
1,887.0
T es (Equal a )
No ejec 𝐻0
𝜇𝑀𝐷=𝜇𝐺𝐷
0.0024
0.0024
-0.02
0.981
33.4
T es (Unequal a )
No ejec 𝐻0
𝜇𝐸𝐷=𝜇𝐺𝐷
0.0017
0.0024
-1.18
0.244
63.9
T es (Unequal a )
No ejec 𝐻0
𝜇𝑀𝑃𝑡=𝜇𝐸𝑃𝑡
0.0005
-0.0002
4.46
9.6E-06
593.7
T es (Unequal a )
Rejec 𝐻0
𝜇𝑀𝑃𝑡=𝜇𝐺𝑃𝑡
0.0005
-0.0012
4.02
1.7E-04
59.4
T es (Unequal a )
Rejec 𝐻0
𝜇𝐸𝑃𝑡=𝜇𝐺𝑃𝑡
-0.0002
-0.0012
2.26
0.027
67.6
T es (Unequal a )
Rejec 𝐻0
This able p esen s he wo-sample T- es o he a e age Jensen’s alphas, whe e 𝐻0 is he null hypo hesis, ep esen ing ha
he di e ences in he means a e ze o and he al e na i e hypo hesis, 𝐻1, ep esen s ha he di e ences in he means a e
di e en om ze o. The “decision” was made a he 5% signi icance le el. “𝜇” ep esen s he popula ion mean. “M”, “E” and
“G” ep esen s he bond mu ual, bond ETF and g een bond unds, espec i ely. “P ” ep esen s he p e-c isis pe iod (Jan2005-
Jul2007), “D” ep esen s he c isis pe iod (Aug2007-Dec2012) and “P ” ep esen s he pos -c isis pe iod (Jan2013-Dec2019).
“Mean o x” and “Mean o y” ep esen s he mean o he a iables and “D ” ep esen s he deg ees o eedom.