Youth and jobs in rural Africa: Beyond stylized facts

Mueller, Valerie (Ed.); Thurlow, James (Ed.) Book — Published Version Youth and jobs in rural Africa: Beyond stylized facts Provided in Cooperation with: Oxford University Press (OUP) Suggested Citation: Mueller, Valerie (Ed.); Thurlow, James (Ed.) (2019) : Youth and jobs in rural Africa: Beyond stylized facts, ISBN 978-0-19-884805-9, Oxford University Press, Oxford, https://doi.org/10.1093/oso/9780198848059.003.0001 This Version is available at: https://hdl.handle.net/10419/213904 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by-nc/4.0/ OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Youth and Jobs in Rural Africa OUP CORRECTED PROOF – FINAL, 29/10/19, SPi OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Youth and Jobs in RuralAfrica Beyond Stylized Facts Edited by VALERIE MUELLER and JAMES THURLOW 1 OUP CORRECTED PROOF – FINAL, 29/10/19, SPi 1 Great Clarendon Street, Oxford, OX2 6DP, United Kingdom Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide. Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries © International Food Policy Research Institute 2019 The moral rights of the authors have been asserted First Edition published in 2019 Impression: 1 Some rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, for commercial purposes, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, by licence or under terms agreed with the appropriate reprographics rights organization. This is an open access publication, available online and distributed under the terms of a Creative Commons Attribution – Non Commercial 4.0 International licence (CC BY-NC 4.0), a copy of which is available at http://creativecommons.org/licenses/by-nc/4.0/. Enquiries concerning reproduction outside the scope of this licence should be sent to the Rights Department, Oxford University Press, at the address above Published in the United States of America by Oxford University Press 198 Madison Avenue, New York, NY 10016, United States of America British Library Cataloguing in Publication Data Data available Library of Congress Control Number: 2019952033 ISBN 978–0–19–884805–9 DOI: 10.1093/oso/9780198848059.003.0001 Printed and bound in Great Britain by Clays Ltd, Elcograf S.p.A. Links to third party websites are provided by Oxford in good faith and for information only. Oxford disclaims any responsibility for the materials contained in any third party website referenced in this work. Any opinions stated in the book are those of the author(s) and are not necessarily representative of or endorsed by IFPRI. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Preface The prospect of widespread youth unemployment in Sub-Saharan Africa (henceforth ‘Africa’) is a serious concern for governments today, both on the sub con tinent and in developed countries. Underlying this is a sense of alarm or urgency, borne out of the view that Africa’s ‘youth bulge’ is an unprecedented global challenge, and that African economies will struggle to absorb enough young job seekers in the coming decades. Concerns are particularly pronounced in rural Africa, where most of the world’s poor population reside and where farming is still the main livelihood for most households. The conventional view is that African youth do not aspire to work in agriculture, because the sector is characterized by low productivity and is far from the dynamic lifestyles offered by cities. Yet job prospects in Africa’s cities and towns are also limited, and so most young Africans will inevitably need to find work somewhere in the rural economy. While Africa’s youth bulge presents a challenge, it can also be viewed as an opportunity for rural development. A young and better-educated workforce might encourage greater use of more sophisticated farm technologies, commercial agricultural practices, and an expansion of rural nonfarm enterprises. These are crucial steps for accelerating agricultural transformation in Africa, and young men and women could be the ‘agents of change’ that the region so badly needs. The debate around youth employment in Africa is therefore one of contrasts— between urgent concern on the one hand and cautious optimism on the other. Although African youth receive greater attention today from researchers and policymakers, there are still major gaps in our knowledge. Most reports from international organizations, for example, adopt a regional perspective and identify general trends and constraints. This overlooks differences between African countries. While some studies do consider youth employment within countries, these rarely focus on the specific challenges facing youth in rural areas. As a result, many policies aimed at rural youth in Africa are based on stylized facts drawn from cross-country data and general frameworks. This book questions some of the stylized facts: Is Africa’s youth bulge unprecedented? Are youth more likely than adults to adopt modern farm technologies and practices? Are youth more likely to engage in rural nonfarm activities or migrate to urban centres? Are policymakers adequately responding to the youth employment challenge, and are rural youth themselves mobilizing and demanding policy reforms from their governments? To answer these questions, this book presents a series of thematic and country case studies that analyse household and firm surveys across a range of country OUP CORRECTED PROOF – FINAL, 29/10/19, SPi vi Preface contexts. The book’s country focus and use of survey data better reflects the wide variations in trends and constraints observed across and within African countries. The book’s focus on rural Africa and the participation of youth in agricultural transformation fills an important gap in our understanding. This book finds that a balance between alarm and optimism is warranted. Addressing youth employment in Africa is a global challenge, but it is one that was overcome by other developing regions when they underwent similar demographic transitions three decades ago. The pressure to create jobs in rural areas is acute, given that Africa’s rural population is growing, and its rural economy is underdeveloped. Yet evidence also suggests that agriculture is transforming in many countries, albeit slowly, and that youth are often participating in this process. Unfortunately, the idea that youth are better positioned than adults to adopt new farm technologies or run successful nonfarm businesses is not borne out in most of the book’s case study countries. Even where there is evidence that youth are leading agricultural transformation, the differences between adults and youth are small or the transformation process itself is modest. More needs to be done by governments to help youth in rural Africa. However, the book finds that, while youth employment is a major policy goal today, policies themselves often fall short of addressing the constraints facing young job seekers. This partly reflects a lack of understanding about country-specific constraints and opportunities—a gap that this book only begins to address. Fortunately, while the policy reforms and actions needed to address Africa’s youth bulge are daunting, the book finds that there is increasing alignment between African governments, who have made youth employment a policy priority, and African youth, who are demanding policies to improve their job prospects. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Acknowledgements We are grateful to four anonymous reviewers who provided comments at different stages of the book. We also appreciate World Development and Journal of Development Studies for granting us permission to reprint the tables in Chapters 7 and 8, respectively. The research in the book was conducted as part of and funded by the CGIAR Research Programme on Policies, Institutions, and Markets (PIM), which is led by the International Food Policy Research Institute (IFPRI) and carried out with support from the CGIAR Trust Fund and through bilateral funding agreements. The United States Agency for International Development (USAID) provided funding for the Ghana chapter via its support for IFPRI’s Ghana Strategy Support Programme (grant number: EEM-G-00-04-00013). The United Kingdom’s Department for International Development (DFID) provided funding for the Malawi chapter via its support for IFPRI’s Malawi Strategy Support Programme (grant number: 203824-106). The International Growth Centre provided funding for the Tanzania chapter (grant number: 1-VCC-VTZA-VXXXX-40414). OUP CORRECTED PROOF – FINAL, 29/10/19, SPi xiv List of Tables 5.5. Labour type, by location and age cohort, 2013 118 5.6. Average time worked per year by type of work, by age cohort, weeks 119 5.7. Agricultural household-level characteristics in rural and small town areas, byagecohort of household head, means 121 5.8. Multinomial models of determinants of type of labour engagement for ruralworkers in Ethiopia, by age cohort 127 5.A1. Water and wood collectors, numbers (in thousands) and percentage of the economically active population, by age cohort, sex, and rural/urban, 2013 131 5.A2. Youth labour type, by location and age cohort, 2013, present 131 5.A3. Average time worked per year by type of work, by youth age cohort, weeks 132 5.A4. Profile of rural and small town workers (ages 15–64) by employment type, meancharacteristics 132 6.1. Sample size and period of administration of Malawi Integrated Household Surveys used 142 6.2. Malawi, change in size of employment categories by age cohort, disaggregated by rural and urban and by male and female 147 6.3. Changes in educational attainment among working age individuals in Malawi, 2004 to 2013 152 6.4. Change in category of employment between 2010 and 2013 for working ageindividuals in Malawi, row totals in per cent 154 6.5. Dependent and explanatory variables for multinomial logit analysis of determinants of an individual being a member of a particular employment category, working age sample 159 6.6. Determinants of employment category for working individuals in Malawi, multinomial logit results presented as relative risk ratios 163 7.1. Distribution of rural households by members’ primary employment in Ghana (Columns (1)–(4) sum to 100 in each survey year) 178 7.2. Distribution of rural households by agricultural, non-agricultural, and mixed occupations across district groups—rural total versus youth-headed households 181 7.3. Marginal effects of probit model regressions on factors affecting being a non-agricultural household, pooled data of GLSS5 and GLSS6 184 7.4. Types of different non-agricultural households according to family members’employment 188 7.5. Marginal effect of probit model regressions on factors affecting agricultural input use, pooled data of GLSS5 and GLSS6 190 7.6. Marginal effects of probit model regressions on factors affecting being a nonpoor or a middle-class household in rural Ghana, pooled data of GLSS5and GLSS6 196 7.A1. Marginal effects in the probit estimations on the determinants of being a non-agricultural household, pooled data of Census 2000 and 2010 200 OUP CORRECTED PROOF – FINAL, 29/10/19, SPi List of Tables xv 8.1. Rural population and agricultural employment shares and annual growthrates 209 8.2. Contribution to new employment by sector, formal and informal, 2002–2012 210 8.3(a) Rural and urban MSME summary statistics 213 8.3(b) Rural and urban MSME summary statistics: youth 215 8.3(c) Rural and urban MSME summary statistics: other adults 216 8.4. Distribution of three types of rural households in Tanzania in 2012 221 8.5. The marginal effect of probit regression results using 2012 HBS data 223 8.6. Distribution of population and MSMEs (weighted, percentage) 229 8.7. Sectoral distribution of rural and urban MSME firms in the survey (weighted, percentage) 229 8.8. Occupation prior to starting business of MSMEs (weighted, percentage) 231 8.9. Reasons for business choice by broad sector in MSME survey (weighted, percentage) 233 8.10. Job satisfaction in MSME survey (weighted, percentage) 235 8.A1. Summary statistics of main variables of 2012 HBS used in the regression:rural 244 8.A2. Summary statistics of main variables of 2012 HBS used in the regression:urban 244 8.A3. Summary statistics of main variables of 2012 HBS used in the regression, bytypes of rural households 245 8.A4(a). Probit results for probability of being in-between rural and urban enterprises, marginal effect: owner’s personal characteristics 246 8.A4(b). Probit results for probability of being in-between rural and urban enterprises, marginal effect: business characteristics 247 8.A4(c). Probit results for probability of being in-between rural and urban enterprises, marginal effect—infrastructure, technology, and financial services 248 9.1. Employment in agriculture and the rural nonfarm economy 261 9.2. Migration patterns of youth leaving agriculture 265 9.A1. Individual characteristics of rural youth employed according to sector ofemployment 271 9.A2. Households with youth characteristics according to sector of employment 272 9.A3. Characteristics of migrated youth and their households according tolabourtransition 273 9.A4. Sector of employment of rural households by remittances status 273 OUP CORRECTED PROOF – FINAL, 29/10/19, SPi OUP CORRECTED PROOF – FINAL, 29/10/19, SPi List of Abbreviations and Acronyms ACLED Armed Conflict Location and Event Data ADLI Agricultural development led industrialization AfDB African Development Bank ANPEJ National Agency for the Promotion of Youth Employment ANSD Senegalese National Agency for Statistics and Demography ARD Agricultural research and development AU African Union BRELA Business Registration and Licensing Agency CAADP Comprehensive Africa Agriculture Development Programme CSA Central Statistical Agency (of Ethiopia) CTA Technical Centre for Agricultural and Rural Cooperation EA Enumeration areas EAP East Asia and the Pacific ERSS Ethiopia Rural Socioeconomic Survey ESAM Senegal Household Survey ESPS Senegal Poverty Monitoring Survey ESS Ethiopia Socioeconomic Survey FAO Food and Agriculture Organization FEES Formal Employment and Earning Survey FEP Food for Education Programme FEWSNET Famine Early Warning Systems Network FISP Farm Input Subsidy Programme GDP Gross domestic product GGDC Groningen Growth and Development Centre GLSS Ghana Living Standards Survey GSS Ghana Statistical Service Ha Hectare HBS Household Budget Survey HH Household HIES Household Income and Expenditure Survey HIPC Heavily indebted poor countries HIV Human immunodeficiency virus HQ Headquarters HR High-return IIA Independent of irrelevant alternatives ICA Integrated Country Approach ID Identification IEG (World Bank) Independent Evaluation Group IFAD International Fund for Agricultural Development OUP CORRECTED PROOF – FINAL, 29/10/19, SPi xviii List of Abbreviations and Acronyms IFI International financial institution IFPRI International Food Policy Research Institute IHPS Integrated Household Panel Survey IHS (Malawi) Integrated Household Survey ILC International Labour Conference ILFS Integrated Labour Force Survey ILO International Labour Organization ILS International labour standards IMF International Monetary Fund IOM International Organization for Migration IPUMS Integrated Public Use Microdata Series IRR Internal rate of return ISI Import substitution industrialization ISIC International Standard Industrial Classification JFFLS Junior farmer field and life schools KILM Key indicators of the labour market Kg Kilogram Km Kilometre LAC Latin America and the Caribbean LFS Labour Force Survey LL Lower limit LOASP Agro-Sylvo-Pastoral Orientation Act LPI Lived Poverty Index LPM Linear probability model LR Low-return LSMS-ISA Living Standards Measurement Study—Integrated Surveys on Agriculture MDG Millennium Development Goals MIJARC International Movement of Catholic Agricultural and Rural Youth MNL Multinomial logit MOE Ministry of Education MRHS Migration and Remittances Household Survey MSME Micro, small, and medium enterprises NBS National Bureau of Statistics NEPAD New Partnership for Africa’s Development NFE Nonfarm enterprise/economy NGO Nongovernmental organization NLFS National Labour Force Survey NSGRP National Strategy for Growth and Reduction of Poverty NSO National Statistics Office ODA Official development assistance OLS Ordinary least squares OSH Occupational safety and health PNAD Brazilian National Household Sample Survey PPEJMR Politique de Promotion de l’Emploi des Jeunes en Milieu Rural PPP Purchasing power parity OUP CORRECTED PROOF – FINAL, 29/10/19, SPi List of Abbreviations and Acronyms xix PRSP Poverty reduction strategy papers PSE Emerging Senegal Plan PSU Primary sampling unit RIGA Rural Income Generative Activities RNFE Rural nonfarm economy RRR Relative risk ratio SA South Asia SAP Structural adjustment programmes S.D. Standard deviation SE Standard error SNNP Southern nations, nationalities, and peoples SSA Sub-Saharan Africa TZS Tanzanian shillings UL Upper limit UN United Nations UN DESA United Nations Department of Economic and Social Affairs UNGA United Nations General Assembly US$ United States Dollar VCD Value chain development WB World Bank WDI World Development Indicators WGI Worldwide Governance Indicators YEI Youth employment inventory OUP CORRECTED PROOF – FINAL, 29/10/19, SPi OUP CORRECTED PROOF – FINAL, 29/10/19, SPi List of Contributors Bob Baulch is a Senior Research Fellow in the Development Strategy and Governance division at the IFPRI, and the Country Programme Coordinator for IFPRI’s Malawi Strategy Support Programme in Lilongwe, Malawi. Firew Bekele Woldeyes is a Research Fellow in the Macroeconomics and Trade Policy Department at the Policy Studies Institute in Addis Ababa, Ethiopia. Todd Benson is a Senior Research Fellow in the Development Strategy and Governance division at IFPRI in Washington DC, U.S.A. Xinshen Diao is a Senior Research Fellow and Deputy Division Director of the Development Strategy and Governance division at IFPRI in Washington DC, U.S.A. Alvina Erman is an Economist in Disaster Risk Management at the World Bank in Washington DC, U.S.A., and a former Senior Research Assistant Analyst at IFPRI. Elisenda Estruch is an Economist in the Department of Sectoral Policies at the International Labour Organization in Genève, Switzerland. Peixun Fang is a Research Analyst in the Development Strategy and Governance div ision at IFPRI in Washington DC, U.S.A. Ileana Grandelis is a Rural Employment Officer in the Decent Rural Employment Team at the United Nations Food and Agricultural Organization in Rome, Italy. Hak Lim Lee is a Senior Research Analyst at the Legal Services Corporation in Washington DC, U.S.A., and a former Research Analyst at IFPRI. Eduardo Magalhaes (deceased) was an independent consultant and former Research Analyst at IFPRI. Ian Masias is a Programme Manager in the Development Strategy and Governance division at IFPRI in Washington DC, U.S.A. Margaret McMillan is a Professor of Economics at Tufts University in Medford MA, U.S.A., and a Senior Research Fellow at IFPRI. Valerie Mueller is an Assistant Professor in the School of Politics and Global Studies at Arizona State University in Tempe AZ, U.S.A., and a Nonresident Fellow at IFPRI. Stefan Pahl is a PhD student in the Faculty of Economics and Business at the University ofGroningen in the Netherlands. Josee Randriamamonjy is a Senior Research Analyst in the Development Strategy and Governance division at IFPRI in Washington DC, U.S.A. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi xxii List of Contributors Danielle Resnick is a Senior Research Fellow and Governance Theme Leader in the Development Strategy and Governance division at IFPRI in Washington DC, U.S.A. Gracie Rosenbach is a Research Analyst in the Development Strategy and Governance division at IFPRI in Washington DC, U.S.A. Emily Schmidt is a Research Fellow in the Development Strategy and Governance div ision at IFPRI in Washington DC, U.S.A. David Schwebel is a Community Advisor at Cargill in Amsterdam, Netherlands, and a former Consultant at the United Nations Food and Agricultural Organization in Rome, Italy. Jed Silver is PhD student in the Department of Agricultural and Resource Economics at the University of California, Berkeley CA, U.S.A., and a former Senior Research Assistant at IFPRI. James Thurlow is a Senior Research Fellow in the Development Strategy and Governance division at IFPRI in Washington DC, U.S.A. Lisa Van Dijck is a Consultant in the Decent Rural Employment Team at the United Nations Food and Agricultural Organization in Rome, Italy. Peter Wobst is a Senior Economist in the Social Policies and Rural Institutions division atthe United Nations Food and Agricultural Organization in Rome, Italy. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Valerie Mueller, James Thurlow, Gracie Rosenbach, and Ian Masias, Africa’s Rural Youth in the Global Context In:Youth and Jobs in Rural Africa: Beyond Stylized Facts. Edited by: Valerie Mueller and James Thurlow, Oxford University Press (2019). © International Food Policy Research Institute. DOI: 10.1093/oso/9780198848059.003.0001 1 Africa’s Rural Youth in the Global Context Valerie Mueller, James Thurlow, Gracie Rosenbach, and Ian Masias 1.1 Introduction Governments in Sub-Saharan Africa are under enormous pressure to create more and better jobs for the region’s young and rapidly growing population.1 Africa is undergoing a ‘youth bulge’ in which the share of young people in the working age population is peaking due to past declines in mortality coupled with persistently high fertility (Canning, Raja, and Yazbeck2015). This demographic transition has created a sense of urgency, and even anxiety, within national governments and the international development community (Resnick and Thurlow2015). With the advent of the Sustainable Development Goals (UNDESA2016), most policies and strategies in Africa today focus on promoting ‘inclusive growth’, which means that the population, especially the poor, should not only benefit from, but also participate in, the development process. This has made job creation a major policy objective, alongside the more traditional goals of accelerating economic growth and reducing poverty and hunger. The successes of other developing countries, especially in Asia, provides African governments with what is sometimes considered a ‘blueprint’ for inclusive growth. Rapid economic growth in East Asia, for example, was accompanied by a process of ‘structural change’ in which the share of workers employed in agriculture declined as jobs were created in more productive and remunerative industrial sectors (McMillan, Rodrik, and Verduzco-Gallo 2014). This led to substantial poverty reduction, in large part because poor workers, especially farmers and their families, were able to take advantage of better job opportunities, often by migrating to cities and towns (Ravallion et al. 2007). Urbanization and structural transform ation were supported by rising agricultural productivity (Ravallion2009). This allowed workers to leave farming without raising food prices and urban wages, which might have jeopardized industrialization (Zhang, Yang, and Wang 1 Unless stated otherwise, the terms ‘Sub-Saharan Africa’ and ‘Africa’ will be used interchangeably in the book. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi 8 Valerie Mueller et al. transitions. Given their similar stages of development, it is not surprising then than alarge share of Africa and South Asia’s workers continued to be self-employed in rural agriculture 15 years after their youth bulges peaked. It is also in these two regions where most of the world’s poor population are concentrated today (Thurlow, Dorosh, and Davies 2019). This underscores the importance of creating jobs and income opportunities in rural Africa. The table also compares education levels across regions. Again, we find that Africa has more in common with South Asia. Both regions had low primary and secondary school enrolment at the peak of their youth bulges, and even though enrolment increased over the next 15 years, much of these gains were achieved by closing primary school enrolment gaps. In contrast, East Asia and Latin America started with much higher school enrolment and were far more successful in closing secondary school enrolment gaps. The quality of education in Africa vis-à-vis other regions three decades ago is difficult to assess. Nevertheless, Africa has moved closer towards achieving universal primary schooling, and this highlights the better educational attainment of young Africans compared to adults. Africa also has much higher labour force participation. This is because more women are part of Africa’s workforce today than they were in other regions, except for East Asia. Of course, high participation rates mean that more jobs will be needed as Africa’s population grows. However, it also means that a larger share of the African population is participating in, and hopefully benefiting from, the region’s growth process. Finally, we compare the pace of economic growth and structural change during countries’ demographic transitions. As mentioned earlier, successful economic development is usually accompanied by a falling share of workers in agriculture, and a shift in employment towards more productive sectors, leading to faster economic growth. Figure1.3 uses employment data collected from national popu lation censuses and household and labour force surveys around the period when countries’ youth bulges peaked. Unfortunately, not all countries have such data, especially those whose youth bulges occurred during the 1960s and 1970s when surveys were conducted less frequently or not at all. It is also not possible to estimate comparable changes in employment patterns for countries that have only recently (or not yet) undergone their peak youth bulge. As a result, the figure only includes information for about half of all developing countries, and so regional averages are not reported. Despite limited country coverage, it is still possible to discern regional patterns from the figure. East Asian countries, for example, generally experienced strong economic growth (horizontal axis) as well as a rapid decline in agricultural employment shares (vertical axis) (see China and Indonesia). In contrast, Latin American countries experienced more modest, or even negative, economic growth, and a more gradual exit from agricultural employment (see Mexico). Again, there OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Africa’s Rural Youth in the Global Context 9 is especially wide variation across African countries. A few fast-transforming economies experienced rapid growth and structural change similar to East Asia (see Botswana and Rwanda). However, economic growth in most African countries is slow, and workers are only gradually leaving agriculture. This is consistent with projections suggesting that most of the jobs created in rural Africa until 2030 will be in agriculture (Filmer and Fox2014; Thurlow2015). In summary, the lateness and absolute size of Africa’s demographic transition isunique. The region will soon become the main driver of growth in the global workforce, and African economies will need to create large numbers of jobs just to keep pace with rapid population growth. Fortunately, African economies are growing, but, except for a few countries, they are not matching East Asia’s high rates of economic growth and structural change. Moreover, Africa’s rural popu lation continues to expand, despite rapid urbanization. Together, these trends indicate that creating rural employment, including in agriculture, will be crucial in ensuring that African economies can absorb enough job seekers into the workforce and avoid rising unemployment. At the same time, Africa will need to Angola Bangladesh Bolivia Botswana China Cuba Ethiopia India Indonesia Iraq Laos Malaysia Mexico Nigeria Pakistan Rwanda Sri Lanka –1.0 –0.5 0.0 0.5 1.0 1.5 2.0 –4 –2 02468 Annual decline in agriculture’s share of total employment (%-point) Average annual GDP per capita growth rate (%) Sub-Saharan Africa (SSA) South Asia (SA) Latin America/Caribbean (LAC) East Asia/Pacific (EAP) Middle East/North Africa (MENA) Figure 1.3. Rates of economic growth and structural change during the 15 years following countries’ peak youth bulge years Notes: Sample includes 62 low-and middle-income countries (32 in SSA, 6 in SA, 12 in LAC, 8 in EAP and 4 in MENA). Reported changes are for the 15 years immediately after the year in which a country’s youth bulge peaked. GDP is measured in constant 2010 US dollars and not adjusted for purchasing power parity. Source: Authors’ calculations using data from ILO (2018), Timmer et al. (2015), and World Bank (2018). OUP CORRECTED PROOF – FINAL, 29/10/19, SPi 10 Valerie Mueller et al. provide better jobs for its adult workforce, who also aspire to improved living standards and working conditions. In that regard, Africa does not necessarily face a youth challenge, but rather the broader challenge of promoting inclusive growth and decent employment in today’s competitive global economy. 1.3 Framing Agricultural and Rural Transformation This book focuses on the participation of rural youth in national development. Agricultural and rural transformation are therefore important concepts that help structure the research questions and analysis. As discussed earlier, economic development is strongly associated with structural change, which occurs when workers leave agriculture for more productive jobs in other sectors (Johnston and Kilby1975; Chenery and Syrquin1975). However, structural change is not the only driver of economic growth. Economy-wide labour productivity also rises when workers within a sector become more productive without needing to move to other sectors of employment. Agricultural transformation refers to a process in which farm productivity rises, leading to growth in the broader rural economy. Timmer (1988) provides a framework with four stages that are summarized inFigure1.4. During the first stage (subsistence agriculture), most rural inhabitants are farmers engaged in food production for their own consumption and use rudimentary technologies and farming practices. The focus for policy at this stage is raising farm productivity, such as through greater use of improved seeds, chemical fer til izers, and soil and water management. Land and labour resources at this stage are likely to be underemployed and it may not matter if technological improvements are labouror land-saving. Since youth in Africa are better educated than adults, many expect that they are more likely to adopt improved farm technologies (see Sheahan and Barret2017). The country chapters in this book assess the contribution of youth to ongoing changes in the farming sector. During the second stage, there is an expansion of farm-nonfarm linkages, as farm productivity rises and farmers begin to produce marketable surpluses. This leads to growth in goods that are produced in rural areas and primarily sold to other rural households. The rise of rural markets creates nonfarm jobs linked to agriculture, such as traders and transporters. Nonfarm workers may live in rural market centres, where agriculture indirectly supports an even wider range of Subsistence agriculture Farm-nonfarm linkages Rural-urban linkages Modernized agriculture 1 2 3 4 Figure 1.4. Timmer’s four stages of agricultural transformation Source: Authors’ interpretation of Timmer (1988). OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Africa’s Rural Youth in the Global Context 11 occupations. New job opportunities encourage farmers to diversify incomes or exit agriculture entirely. However, at this stage of agricultural transformation, farming remains the primary driver of national growth and job creation. Youth are again expected to play a key role during this stage. Better-educated people are more likely to run rural nonfarm businesses (Naglar and Naude 2017), and emerging land constraints may mean that it is youth who are more likely to seek off-farm work (Bezu and Holden2014). The country chapters examine the links between youth, education, and rural nonfarm employment. The third stage of agricultural transformation involves a strengthening of rural-urban linkages. Nonagricultural sectors, particularly in cities and towns, become drivers of national development. Agriculture increasingly supplies urban consumers and rural inhabitants migrate in search of urban job opportunities. Migrant workers may remit incomes back to their rural families or occasionally return to rural areas to alleviate seasonal labour shortages. Outmigration may require labour-saving technological improvements in agriculture to prevent food prices and urban wages from rising and stalling structural change. At this stage, urban nonagricultural growth drives national development and pulls agriculture behind it. Chapter2 in this book specifically addresses the role of youth in migration decisions, and the various country chapters examine the links between youth and urbanization. The final stage is the transition to modernized agriculture. This is most relevant for today’s developed countries, where high rural-urban inequality and concerns about national food security may prompt governments to subsidize agriculture and protect ‘rural lifestyles’. Few, if any, African countries, or even areas within these countries, have reached this late stage of agricultural transformation. Although Timmer’s framework was developed three decades ago and is grounded in the Asian experience, it still provides a useful device for analysing the pace and participation of youth in Africa’s agricultural and rural economies. It underpins the view that, despite global developments since the 1980s, agricultural transformation is still essential for economic development in Africa (Diao, Hazell, and Thurlow 2010; Timmer and Akkus2008). Not only will Africa’s youth need to find jobs in agriculture and rural areas, but they could help drive the trans form ation process. Chapters3 and4 in this book examine whether the needs and potential of youth are reflected in national policies, and whether youth are more politically active and demanding of their governments. 1.4 Evidence of Agricultural Transformation in Africa Most African countries today, or at least most rural populations within African countries, are in the second or third stage of Timmer’s transformation framework. Farmers still grow some of the food they consume, but most now sell at OUP CORRECTED PROOF – FINAL, 29/10/19, SPi 12 Valerie Mueller et al. least some of their output in local markets (Carletto, Corral, and Guelfi 2017). This marketable surplus is the result of rising farm production levels over the last 15 years. Figure1.5 reports the growth in agricultural land and labour productivity that occurred after Africa’s youth bulge peaked in 2003. The figure shows that both land and labour productivity increased for most African countries, including the region as a whole. However, land productivity growth exceeded labour productivity growth in almost all countries, implying that agricultural labour grew faster than agricultural land. This reflects growing concerns about rising rural population densities and the ability of available lands to support the livelihoods of a rapidly growing rural workforce. The case study chapters in this book examine the contribution of youth to agricultural transformation in five countries: Ethiopia, Ghana, Malawi, Senegal, and Tanzania. These countries were selected to capture the variation in trends observed across Africa, although data availability was also a consideration. As indicated in the figure, Ethiopian agriculture is transforming rapidly, whereas Malawian and Senegalese agriculture are not. Ghana and Tanzania are close to the African average. The case studies allow us to examine the role of youth in Individual countries Case study countries Sub-Saharan Africa (Weighted) Sub-Saharan Africa (Unweighted) Ethiopia Ghana Malawi Senegal Tanzania –4 –2 0 2 4 6 8 10 –4 –2 0246 81 0 Average annual growth agricultural GDP per hectare of agricultural land (%) Average annual growth in agricultural GDP per worker (%) Figure 1.5. Agricultural productivity growth in sub-Saharan Africa, 2003–2016 Notes: Sample includes 42 Sub-Saharan African countries (excluded are Eritrea, Sao Tome and Principe, Seychelles, South Africa, South Sudan, and Sudan). GDP is measured in constant 2010 US dollars unadjusted for purchasing power parity. Agricultural land includes lands used crop cultivation and animal husbandry. Source: Authors’ calculations using data from ILO (2018), FAO (2018), and UNSD (2018). OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Africa’s Rural Youth in the Global Context 13 raising farm productivity–the first stage in Timmer’s framework–taking account of how this may vary across African countries. Although Africa’s rural economy is dominated by agriculture, a large share of rural incomes is earned in the rural nonfarm economy (Carletto, Corral, and Guelfi 2017). This is important for the second stage of Timmer’s framework, when farm-nonfarm linkages expand. Household surveys suggest that more than a third of rural incomes in Africa are generated through nonfarm employment (Haggblade, Hazell, and Reardon 2007), and that most rural households engage in some form of nonfarm activity (Davis, Di Giuseppe, and Zezza 2017). As mentioned earlier, most of the structural change in Africa in recent years was driven by workers leaving agriculture to work in informal services. Many of these services are agriculture-related, such as the trading and transport of food and agricultural products. While these are not the kinds of high productivity industrial jobs that dominated the East Asian experience, their growth has helped reduce poverty in many parts of Africa (Dorosh and Thurlow2016). Figure 1.6 reports changes in urban population shares and nonagricultural employment shares since 2003. For Sub-Saharan Africa as a whole, the decline in the rural population was matched by a decline in agricultural employment (i.e. the Ethiopia Ghana Malawi Senegal Tanzania –1.0 –0.5 0.0 0.5 1.0 1.5 –1.0 –0.5 0.0 0.5 1.0 1.5 Annual change in non-agriculture’s share of total employment (%-point) Annual change in urban population share (%-point) Individual countries Case study countries Sub-Saharan Africa (Weighted) Sub-Saharan Africa (Unweighted) Figure 1.6. Urban population and nonfarm employment shares in sub-Saharan Africa, 2003–2016 Notes: Sample includes 44 Sub-Saharan African countries (excluded are Sao Tome and Principe, Seychelles, South Africa, South Sudan, and Sudan). Official definitions of urban areas are used. Source: Authors’ calculations using data from ILO (2018) and UNDESA (2018). OUP CORRECTED PROOF – FINAL, 29/10/19, SPi 14 Valerie Mueller et al. regional average lies close to the diagonal line in the figure). This suggests that there was no significant change in the share of rural nonfarm employment in the region. Again, we find that regional averages hide wide variation across countries– differences that are captured in our choice of country case studies. The exit from agriculture in Ethiopia, for example, greatly exceeds the pace of urbanization, suggesting that many of the workers that left farming found employment in the rural nonfarm economy. The opposite is true for Senegal, where agricultural employment has risen, despite urbanization. This suggests that some of Senegal’s rural nonfarm workers are returning to agriculture. Tanzania and Ghana are again closer to the African average, and there was little agricultural trans form ation taking place in Malawi. Detailed household surveys allow the country chapters to investigate whether it is youth or adults, or young men or women, who are more actively engaged in the rural nonfarm economy. The third stage of Timmer’s framework is characterized by a strengthening of rural-urban linkages. There is some evidence that Africa’s urban consumers are increasingly driving demand for agricultural products (Tshirley et al. 2015). As mentioned earlier, rapid urbanization is a defining feature of African development. Moreover, expanding urban populations and migration within rural areas has meant that many of Africa’s rural inhabitants today reside in ‘peri-urban areas’ adjacent to major urban agglomerations (FAO2017). It is in peri-urban areas where rural-urban linkages are expected to be strongest and where agricultural trans form ation should be most advanced (Dorosh and Thurlow2014). Cross-country data suggests that young African men may be more likely to migrate than either adults or young women. Figure1.7 estimates the relative speed of urbanization for youth and adults (horizontal axis) and for young men and women (vertical axis). This is measured by estimating the gap between average annual urban and rural population growth rates for each population subgroup. Alarge positive number means that the subgroup’s urban population is growing much faster than its rural population. The figure reports differences in the speed of urbanization between two population subgroups. For example, the horizontal axis focuses on the differences between youth and adults. The relative speed of urbanization is generally positive, implying that African youth are concentrating in urban areas faster than African adults. This is consistent with findings in other studies (see De Brauw, Mueller, and Lee 2014; Holden and Otsuka2014), but it hides how most urban migration is to smaller towns, rather than bigger cities (Mueller et al. 2019). The tendency for youth to urbanize faster than adults is most pronounced in Ghana and Malawi, but it is negligible in the other three case study countries. Similarly, there is some evidence that young men are urbanizing faster than young women. Our case studies capture variation across African countries. This variation may reflect differences in education or other factors that influence the decision to migrate. A thematic chapter in this book analyses youth migration decisions using detailed household surveys rather than country-level data. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Africa’s Rural Youth in the Global Context 15 In summary, African agriculture is transforming, albeit slowly and with some cause for concern. Agricultural land and labour productivity are growing, but so too are rural population densities. This suggests that agriculture’s contribution to future job creation may be constrained as lands become scarce. Africa’s rural nonfarm economy is also expanding, although in many countries it is not keeping pace with urbanization. On average, workers are leaving agriculture and moving to urban areas faster than they are finding work in the rural nonfarm economy. As urban centres become congested, more of the population are likely to reside in peri-urban (or peri-rural) areas where ruralurban linkages are often strongest (Thurlow, Dorosh, and Davies 2018). Recent estimates suggest that one third of rural Africans already live within one-hour travel time of cities with populations of 50,000 people or more (SOFA 2017). The nonfarm economy surrounding cities and towns will therefore play an important role in creating work for rural job seekers, including youth. This means that, while agricultural transformation is proceeding in Africa, it is not only uneven across countries, but also across areas within countries. This underscores the need for detailed country case studies and cautions against an overreliance on country-level data. Ethiopia Ghana Malawi Senegal Tanzania –0.5 0.0 0.5 1.0 –1.0 –0.5 0.0 0.5 1.0 1.5 2.0 Speed of urbaniation: young men minus young women (%-point) Speed of urbanization: youth minus adult (%-point) Individual countries Case study countries Sub-Saharan Africa (Weighted) Sub-Saharan Africa (Unweighted) Figure 1.7. Speed of urbanization for youth and adults in sub-Saharan Africa, 2003–2015 Note: Sample includes 44 Sub-Saharan African countries (excluded are Sao Tome and Principe, Seychelles, South Africa, South Sudan, and Sudan). Speed of urbanization is the difference between average annual urban and rural population growth rates, i.e., a number greater than one implies that the urban population share is rising over time). Figure compares the speed of urbanization for different population groups, i.e., a number greater than one means that the first group is urbanizing faster (or deurbanizing slower) than the second group. Source: Authors’ calculations using historical population data from ILO (2018). OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Table 1.2. Country case studies Sub-Saharan Africa or Case Study SSA Ethiopia Ghana Malawi Senegal Tanzania (year when youth bulge peaked) (2003) (2014) (1987) (2010) (2000) (2000) Population Population, 2016 (millions) 1,033 102 28 18 15 56 Rural share (%) 61.0 80.1 45.3 83.5 53.7 67.7 Population growth, 2006–16 (%) 2.8 2.6 2.5 3.0 2.9 3.2 Rural areas 1.9 2.2 1.0 2.9 2.2 2.2 Urban areas 4.2 5.0 3.8 3.9 3.9 5.7 GDP GDP per capita, 2016 ($) 1,173 426 1,617 448 962 792 Agriculture share (%) 22.3 35.8 23.9 29.9 15.5 25.2 GDP per capita growth, 2006–16 (%) 3.1 7.5 4.4 2.5 1.5 3.3 Agriculture 2.8 3.8 1.2 1.1 1.6 0.3 Employment Labour force participation, 2016 (%) 68.2 82.3 76.7 76.8 57.0 83.3 Youth (15–24) 48.6 75.0 53.6 63.0 41.3 72.0 Unemployment rate, 2016 (%) 7.3 5.1 2.3 5.9 4.8 2.2 Youth (15–24) 13.8 7.3 4.7 7.8 5.4 3.8 Agricultural employment, 2016 (%) 57.4 69.0 41.2 84.7 53.6 67.2 Annual change, 2006–16 (%-point) –0.53 –1.10 –0.38 0.04 1.38 –0.74 Note: GDP is measured in constant 2010 US dollars unadjusted for purchasing power parity. Source: Authors’ calculations using GDP data from UNSD (2018) and other data from the World Bank (2018). OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Africa’s Rural Youth in the Global Context 17 Our country case studies reflect some of the important variations observed across Sub-Saharan Africa. Table1.2 provides current statistics for the five countries and the region. By design, our cases are either lowor lower-middle-income countries, often with a greater dependence on agriculture and with a larger share of the population in rural areas. Youth unemployment is lower amongst our case study countries than in Africa as a whole, which partly reflects our focus on agrarian economies, which have lower unemployment rates than more miningbased economies like Nigeria or South Africa. Ghana and Ethiopia are two of Africa’s fastest transforming countries, but Ghana is at a later stage of development (i.e. GDP per capita is higher and the population is more urbanized). Ghana is also one of the earliest African countries to experience a demographic transition (i.e. its youth bulge peaked in 1987), whereas Ethiopia is one of the last countries. Malawi is at a similar stage of development as Ethiopia, although Malawi, like Senegal, is not experiencing rapid economic growth and the share of employment in agriculture is not falling. Labour force participation in Senegal is one of the lowest in Sub-Saharan Africa, largely because women are less likely to work and because international migration is particularly important for Senegal. Finally, Tanzania provides an intermediate case. The country is transforming, and workers are leaving agriculture, often for urban areas, but the economy, particularly agriculture, is growing much slower than in Ethiopia or Ghana. Our five case studies therefore reflect the diversity of African countries and allow us to gain a more nuanced understanding of youth in rural areas. 1.5 Organization of the Book There is a large body of research on agricultural transformation and structural change in Africa (see, for example, Diao et al.2007; McMillan, Rodrik, and Sepúlveda 2016). Few studies, however, examine employment through a youth lens and with a focus on rural Africa. This book provides new empirical evidence on the participation of rural youth in national development processes. Cross-country evidence is informative, but cannot substitute for detailed case studies that use micro-level data to reveal countries’ unique characteristics and challenges. It is only through the collection of robust country-specific evidence that we can move beyond stylized facts and determine to what extent African youth should be a source of optimism or a cause for concern. The book is separated into two parts. Part I includes three thematic chapters that cover important under-researched areas for youth employment. Rising popu la tion densities in rural areas has raised concerns about the future role agriculture in job creation and the prospect of accelerated urbanization. Chapter2 uses new household survey data to investigate youth migration patterns in four African countries, paying particular attention to the effect of land scarcity on young OUP CORRECTED PROOF – FINAL, 29/10/19, SPi OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Valerie Mueller and Hak Lim Lee, Can Migration be a Conduit for Transformative Youth Employment? In: Youth and Jobs in Rural Africa: Beyond Stylized Facts. Edited by: Valerie Mueller and James Thurlow, Oxford University Press (2019). © International Food Policy Research Institute. DOI: 10.1093/oso/9780198848059.003.0002 2 Can Migration be a Conduit for Transformative Youth Employment? Valerie Mueller and Hak Lim Lee 2.1 Introduction Migration has traditionally been considered a necessary component of the transformation process (de Brauw, Mueller, and Lee 2014). Rural workers are attracted to higher earning potential in the manufacturing or rural non-farm sectors (Harris and Todaro1970). The latter process, which occurred in India, for ex ample, was primarily driven by innovation and shifts in rural worker prod uct iv ity. Both of these factors allowed for the creation of a rural labour surplus to transfer into the modern sector, as well as generated demand for additional goods and services in rural areas by augmenting the income of farmers (Hazell and Haggblade1990). Given demographic trends, African youth will be responsible for spearheading economic growth. Yet, they face more substantive barriers than their predecessors: declines in arable land (Jayne, Mather, and Mghenyi2010, Muyanga and Jayne2014), a lack of Green Revolution (Headey, Bezemer, and Hazell2010, Nin-Pratt and McBride2014) or government-sponsored industrialization (Jedwab and Vollrath2015), and competition from the global economy (Headey, Bezemer, and Hazell2010). In this chapter, we examine whether migration offers youth (ages 15–24, 25–34) access to more transformative forms of employment in four African countries, following the traditional pathways to structural change. While a few seminal youth migration studies have raised awareness of orphanhood in Africa (Beegle, De Weerdt, and Dercon 2006, Beegle et al.2010), applications which demonstrate whether migration is a conduit for diversification and productive employment among youth are rare. In what follows, we first establish the knowledge gaps in the literature with respect to the relationship between migration and sector-specific youth employment in Africa. We then focus on addressing a few of the highlighted knowledge gaps using descriptive evidence in four countries. First, we present statistics on the level of engagement in exclusive non-agricultural employment and joint non-agricultural and agricultural employment by youth migration status. Second, we illustrate whether migration allows youth to generate greater returns to OUP CORRECTED PROOF – FINAL, 29/10/19, SPi 26 Valerie Mueller and Hak Lim Lee production. In particular, we compare the agricultural income per capita of youth migrants and non-migrants over time. We further disentangle whether migrants are more likely to move into high-return versus low-return non-agricultural occupations to supplement the income analysis. 2.2 Literature Review 2.2.1 Youth Engagement in the Agricultural Sector Recent empirical evidence in Africa suggests declining trends in the size of landholdings of rural households (Jayne, Mather, and Mghenyi 2010, Muyanga and Jayne2014). In a few concentrated countries, these associations are driven by the underutilization of land due to conflict, forested area, or remoteness and isolation (Chamberlain, Jayne, and Headey 2014). Other African countries, specifically those covered in the LSMS–ISA, suffer from limited surplus of land and high population pressure. In light of the emerging scarcity of arable land, there is a growing research interest to uncover whether diminishing landholdings hasbeen accompanied by increased agricultural intensification to maintain or enhance yields. Sheahan and Barrett (2014) examine various input practices (use of fertilizer, improved seeds, agro-chemicals, animal traction, and mechanized equipment) among households in the LSMS-ISA countries. Although modern input use is relatively low in aggregate, the application of inorganic fertilizer and agro-chemicals has become more common in Ethiopia, Malawi, and Nigeria than documented in previous work by Minot and Benson (2009). Using alternative data sources, Headey and Jayne (2014) and Muyanga and Jayne (2014) show the application of the aforementioned inputs is positively related to changes in population density. The intersection between input intensification, land size, and labour use is of notable importance for understanding employment trends more broadly and youth employment patterns specifically. Thus far, multiple studies find negative relationships between farm size and input use (Barrett, Bellemare, and Hou 2010, Bellemare2013, Carletto, Savastano, and Zezza 2013, Headey, Dereje, and Taffesse 2014, Larson et al.2014, Sheahan and Barrett2014), which suggests intensifying farming practices may be used to overcome land constraints to productivity. For the case of Ethiopia, Headey, Dereje, and Taffesse (2014) find a small, positive correlation between farm size and hired labour, but a much stronger negative relationship with family labour and farm size. Their interpretation of the results is that small farms use land more intensively while large farms are labour constrained. Complementary relationships between labour and input use, particularly for small farms, could suggest an increase in the demand for family labour and perhaps youth employment. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Can Migration be a Conduit 27 Projections of youth employment in agriculture will depend not only on farm size but the substitutive and complementary nature of modern inputs and labour by stage of one’s life cycle status. The existence of agricultural wage labour markets and land rental markets may provide additional forums for youth to continue engaging in agriculture, under sparse opportunities for landownership (for example, as shown in Ethiopia by Bezu and Holden (2014)). With respect to the latter, Deininger, Xia, and Savastano (2015) show land-poor households and households with younger heads are more likely to take advantage of these opportunities to access land in Malawi, Nigeria, Niger, Tanzania, and Uganda. Dillon and Barrett (2014) foreshadow limitations to off-farm employment opportunities in the agricultural sector given existing market failures. First, in most cases, the percentage of households hiring workers for non-harvest types of employment exceeds the percentage of households hiring workers for harvest employment. Second, as the number of acres per household member increases, the hiring of outside workers does not increase proportionally. Economies of scale of labour, or credit market failures possibly explain these patterns. The above studies imply youth participation in agricultural employment will depend on at least two factors. First, if land-constrained households are driven to intensify their land to overcome productivity constraints, then youth employment on family farms will depend on the complementary nature between those inputs and youth labour. Furthermore, whether youth self-select into on-farm or off-farm agricultural jobs will depend on the factor-bias of the input technology adopted (Bustos, Caprettini, and Ponticelli 2016). Second, increased access to land and opportunities for employment off of the farm may allow youth to continue working in agriculture. Understanding the nature of local land rental and sales markets will be important in measuring the determinants of youth employment in the agricultural sector, as well as the composition of medium and largescale farms to gauge demand for agricultural wage labour. 2.2.2 Youth Diversification Out of Agriculture A few stylized facts regarding diversification trends out of agriculture have emerged from the Sub-Saharan Africa employment literature. While diversification out of agriculture is on the rise (Jones and Tarp2012), agriculture continues to absorb a significant share of the workforce (Jones and Tarp 2012, Page 2012, Davis, Di Giuseppe, and Zezza2014, McCullough2017). The nonfarm wage sector (private and public) has grown but participation remains less common (Jones and Tarp 2012, Fox and Sohnesen 2012); instead, the informal sector is the principal locus of new job creation (Jones and Tarp2012, Fox and Sohnesen2012, De Vreyer and Roubaud2013, Nagler and Naudé2014). Household entrepreneurship OUP CORRECTED PROOF – FINAL, 29/10/19, SPi 28 Valerie Mueller and Hak Lim Lee has the potential to increase the marginal productivity of labour and enhance welfare (Reardon1997, Grimm, Knorringa, and Lay 2012, Nagler and Naudé2014, McCollough2017). Earlier household analyses emphasize the importance of household demographic composition and household head’s age on diversification out of agriculture. Jones and Tarp (2012) and Nagler and Naudé (2014) find the percentages of adults and young male workers (Jones and Tarp only) are negatively associated with spe cializa tion in agriculture in Mozambique. Bezu and Barrett (2012) monitor employment transitions into and between lowand high-return rural nonfarm employment using panel data from Ethiopia (1999, 2004). They find transitions from lowreturn to high-return rural nonfarm employment are positively correlated with the number of children aged 5 to 14 in 1999. The authors posit that children may not directly engage in rural nonfarm employment, but serve as substitutes for adult household labour. Older heads are also found to be more likely to diversify out of agriculture (Nagler and Naudé2014), but other studies have shown older heads can also revert back to agriculture after operating an enterprise (Bezu and Barrett2012). A logical next question is how will these trends affect youth. Researchers have first focused on educational trends to understand whether youth have different earning potential than previous generations. While education levels have increased, they remain low (Filmer and Fox2014, Garcia and Fares2008, Elder and Kone2014). Rural youth are much less likely to be in school than their urban counterparts (Filmer and Fox2014, Garcia and Fares2008). Although labour participation remains high (McCullough2017, Gracia and Fares 2008, Jones and Tarp2012), underemployment is rife (Shehu and Nilsson2014, Jones and Tarp 2012). Opportunities to diversify out of agriculture, particularly into high-return activities, may be low given extant skill deficits (Filmer and Fox2014). Elder and Kone (2014) report findings from the ILO’s School-to-Work Transition surveys (2012–13) covering 15–29 year old individuals at the national level for eight countries in SSA (Benin, Liberia, Madagascar, Malawi, Tanzania, Togo, Uganda, and Zambia) with an average sample size of 3,300 persons. Forty-six per cent of the unemployed youth indicate employment searches lasting longer than a year, mainly in pursuit of establishing their own business or farm, or finding a job in the public or private sector. The biggest obstacle to finding employment was articulated to be a paucity of jobs, as well as insufficient qualifications for existing jobs. They find working youth tended to engage in skilled agricultural and fishery occupations (35.7 per cent), followed by service (25.7 per cent), shop and market sales work (18.3 per cent), elementary occupations (18.3 per cent), and craft and related trade work (10 per cent). The aforementioned studies contend relatively high youth participation rates with a concentration in the agricultural sector in rural areas. One open question is to what extent are these rates influenced by the mobility of youth. Youth OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Can Migration be a Conduit 29 migration patterns motivated by education (de Brauw, Mueller, and Woldehanna 2013) and orphanhood (Beegle, De Weerdt, and Dercon 2006, Beegle et al.2010) have been documented in various African contexts. The omission of youth migrants from surveys could influence how labour participation rates and shifts in employment are perceived in the broader literature. Another discrepancy in the literature arises from the lack of detailed information on agricultural occupations and youth productivity. McCullough (2017) shows that the measure of productivity can affect marginal productivity of labour estimates. For example, when measuring output per person per year, the prod uctiv ity of workers receiving wages in industry, agriculture, and enterprises are higher than on farms in Malawi, Tanzania, and Uganda. However, when the prod uct iv ity ratio is based on output per hour, then she finds farm productivity is higher than for all other sectors in Ethiopia and Malawi but not Tanzania and Uganda. The differences are largely due to a higher number of hours supplied to nonfarm work. It is possible for youth to remain in agriculture, but they are positioned todrive a structural transformation in agriculture with respect to being more productive, more likely to work in modern agricultural jobs or jobs at higher stages of the value chain. In this context, migration may still be utilized by youth in order to access land to facilitate entry into more commercialized agricultural self-employment. 2.3 Data We use the Living Standards Measurement Study-Integrated Surveys on Agriculture (LSMS–ISA) in Ethiopia (2011–12, 2013–14), Malawi (2010–11, 2012–13), Nigeria (2010–11, 2012–13), and Tanzania (2008–19, 2010–11) (World Bank 2016a,b,c,d) to construct an individual dataset of youth ages 15–34 for descriptive statistics on youth migration between baseline and endline per country. Approximately, 5,364 observations in Ethiopia, 4,060 in Malawi, 7,383 in Nigeria, and 4,618 in Tanzania are used to create the migration statistics. Otherwise, when we focus on rural non-migrants, rural–rural migrants, and rural–urban migrants, we have 4,732 observations in Ethiopia, 2,960 in Malawi, 5,179 in Nigeria, and 3,101 in Tanzania, respectively. The analyses using the employment and income outcomes draw from smaller youth samples since we are missing individual responses for those outcomes over time. We define a person as a migrant if he was a member of the household at baseline and departed the household in the follow-up survey. Different instruments were used to detect migration in the surveys. For the Malawi and Tanzania surveys, themigration definition is based on the diligent tracking of split-off households. In Ethiopia and Nigeria, we rely on information reported by the proxy respondent in the follow-up survey on the whereabouts of each household member from OUP CORRECTED PROOF – FINAL, 29/10/19, SPi 30 Valerie Mueller and Hak Lim Lee the baseline roster. Since the baseline and follow-up rounds are two years apart, our measure of migration is over a two-year period and therefore considered a permanent move. However, a key limitation in ourinterpretation of migration across regions will be our inability to disentangle the variation in mobility that stems from differences in contexts across countries from the vari ation in mobility due to measurement differences associated with the use of different survey instruments and interview times across countries. We additionally define the migrant by origin and destination using rural and urban classifications established in the surveys: rural–rural, rural–urban, urban– urban, urban–rural. We are unable to compute representative urban–urban and urban–rural migration rates for Ethiopia, because the baseline survey did not sample large towns until the second round and therefore are omitted from the sample. The definition of urban in Ethiopia typically consists of small (population less than 10,000) and large (population greater than 10,000) towns. Thus, the urban–urban and urban–rural migration rates constructed in this chapter reflect migration within and to smaller towns rather than within and to metropolitan areas. Detailed information on individual employment was extrapolated from the wage, agricultural on–farm labour, and non-agricultural enterprise modules of the surveys. These modules document any engagement in wage or self-employment activities over a 12-month period. We focus on labour participation rather than hours supplied. In our descriptive statistics, a youth is considered to have engaged in a specific activity irrespective of the number of hours reported. This allows us to avoid measurement issues associated with missing hours in the supplied values, but of course fails to account for differences in partial versus full employment. For the purpose of the analysis, we define employment portfolios into four categories: exclusively agriculture, exclusively non-agriculture, mixed agriculture and non-agriculture, and student. For brevity, an individual who was actively a student is automatically placed in the last category, despite evidence of engagement in farm or off-farm activities. One of the aims of this piece is to evaluate whether relocation offers youth opportunities to diversify employment or improve their agricultural production prospects. This requires knowledge of employment of migrants and non-migrants over time. Since tracking at the individual level was only performed in Malawi and Tanzania, detailed descriptive statistics on employment and income patterns by migration status are only available for these countries. We compute income by source for each household to illustrate qualitatively whether migration improved the prospects of youths (ages 15–34). The recall period for income dates 12 months prior to the interview for the Tanzania surveys and 12–18 months prior to the interview for the Malawi survey. Incomes are winsorized at the 5 per cent level to remove influences from outliers and measurement error on descriptive statistics. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Can Migration be a Conduit 31 2.4 Results 2.4.1 Youth Migration Figure2.1 provides the youth migration rates by country and gender for two cohorts: 15–24 and 25–34. We focus on those splitting from their household, originating from rural areas given the premise of the book (see Figure 2.2).1 The percentage of young men in rural areas migrating to other rural areas is as high as 13.6 per cent in Ethiopia and 17.6 per cent in Malawi. Rural–rural migration rates for women in these two countries are slightly greater than the rates of young men at 17.5 per cent and 23.3 per cent. Rural–urban migration rates are considerably lower. The highest migration to cities occurs among Ethiopian young men (9.7 per cent) and women (8.4 per cent). The remaining countries have rural–urban migration rates within the range of 2 to 5 per cent. Young adults (25–34) are less mobile than the younger cohort, particularly in Ethiopia and Malawi. Figure2.3 displays the primary motivation for youth migration. Interestingly, the rates of rural youth claiming to move to other rural areas for employment reasons are quite similar across cohorts with the exception of Malawi. For ex ample, 13.7 per cent of 15–24 year old migrants report moving for employment in Nigeria compared to 12.6 per cent of 25–34 year old migrants. The distinctions in reasons for moving are more pronounced among classes of rural–urban migration. The most drastic example takes place in Ethiopia. Approximately, 31.9 per cent of 15–24 year olds state having moved for work, while 55.4 per cent indicate having moved for education. This can be compared to 49.6 per cent of rural–urban young adult migrants reporting having moved for employment and a mere 12.5 per cent for education. We further compare the distances travelled by migration pattern (not shown here). The median distance that young (15–24) rural–rural male (female) migrants travel is 1.4 (1.4) kilometres in Malawi, and 0.2 (1.6) kilometres in Tanzania. These figures can be compared to those obtained for rural–urban male (female) migrants who undergo median travel distances of 68.1 (59.8) kilometres in Malawi, and 90.3 (54.5) kilometres in Tanzania. These figures are qualitatively comparable for the older youth cohort. For example, the median distance that mature youth (25–34) rural–rural male (female) migrants travel is 1.2 (0.3) kilometres in Malawi, and 0.3 (3.4) kilometres in Tanzania. These figures can be compared to those obtained for rural–urban male (female) migrants who undergo median travel distances of 74.8 (25.2) kilometres in Malawi and 61.9 (37.7) kilometres in Tanzania. Workers 1 Figure 2.2 illustrates youth migration rates for those departing with the entire household. Rates of migration are much lower, although still noteworthy in size for rural populations in Malawi and Tanzania. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi 32 Valerie Mueller and Hak Lim Lee Migration Rates: Youth (Ages 15–24) Splits Ethiopia Urban-Urban 0.2 Malawi Nigeria Tanzania Urban-Rural Rural-Urban Rural-Rural 3.7 5.1 0.1 1.4 8.4 3.2 17.5 23.3 Migration Rates: Youth (Ages 25–34) Splits Ethiopia Urban-Urban 0 Malawi Nigeria Tanzania Urban-Urban Urban-Rural Rural-Urban Rural-Rural Urban-Rural Rural-Urban Rural-Rural 4.4 0.2 Urban-Urban 5.6 Urban-Rural Rural-Urban Rural-Rural 05 10 15 20 0 Percentage 5 10 15 20 0.0 2.5 5.0 7.5 10.0 12.5 Percentage 0.0 2.5 5.0 7.5 10.0 12.5 4.8 3.4 2.5 0.1 9.7 0.8 2.8 17.613.6 0.9 0.7 2.2 2.1 8.7 1.3 4.6 3.7 1.7 3.5 1.5 4.2 0 1.7 4.3 1 1.6 11.4 0.1 0 2.2 7.6 1.3 1 12.3 3.9 1.1 1.7 2.2 1.4 0.7 3.7 9.1 5.9 2.5 0.8 2.7 2.6 0.5 1.1 8.1 Female Male 3.5 Figure 2.1. Youth migration rates Note: Sampling weights used to calculate statistics OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Can Migration be a Conduit 33 Migration Rates: Youth (Ages 15–24) Household Movers Ethiopia Malawi Migration Rates: Youth (Ages 25–34) Household Movers Ethiopia Malawi Nigeria Tanzania Nigeria Tanzania Urban-Urban Urban-Rural Rural-Urban Rural-Rural Urban-Urban Urban-Rural Rural-Urban Rural-Rural 02 4 6 8 0 Percentage 2 4 6 8 0.0 2.6 5.0 7.5 0.0 Percentage 2.5 5.0 7.5 0 0 0 0.5 3 1 1.2 7.6 0 0 0 0.1 1.1 0 0.8 6.6 3.2 0.2 0.4 0.8 0.6 0.4 4.2 4.2 1.4 0.4 0.7 0.1 0.1 1.8 3.2 Urban-Urban Urban-Rural Rural-Urban Rural-Rural Urban-Urban Urban-Rural Rural-Urban Rural-Rural 0 0 0 0.9 2.9 0.7 1.2 7.5 0.1 0 0.5 0.8 4.3 0.9 2.1 9.1 3.4 0.3 0.5 1 5.6 0.6 1.1 5.1 4.3 0.4 0.7 0.5 1 1 7 6.3 Female Male 0.3 Figure 2.2. Migration rates of youth moving with their entire household Note: Sampling weights used to calculate statistics OUP CORRECTED PROOF – FINAL, 29/10/19, SPi 40 Valerie Mueller and Hak Lim Lee we demonstrate whether migrants are more likely to move from presumably low-return agriculture to high-return nonfarm employment. This requires developing a typology similar to Bezu and Barrett (2012), categorizing the occupations of youth into low-return and high-return non-agricultural activities. We then show how the employment transitions between waves 1 and 2 vary by whether youth stayed in their baseline location, moved to a rural destination, or moved to an urban destination in the Malawi and Tanzania surveys. Table2.1 displays the employment transitions between waves by migration status in Malawi and Tanzania. In rural areas, 61.7 per cent and 53.6 per cent of the youth non-migrant population remained engaged in agricultural employment throughout the two waves in Malawi and Tanzania, respectively. These figures are comparable to the employment rates of the adult non-migrant population, which we also include in Table2.1 as a reference but leave out of the discussion hereafter. The figures are only slightly reduced for the rural–rural migrant population (49.5 per cent in Malawi and 40.0 per cent in Tanzania). In short, the majority of the youth population who remains or moves within rural areas stays in the agricultural sector. In both countries, rural–urban and rural–rural migration offer more pos si bilities for youth to engage in high-return employment which confers the observed positive income change associated with mobility. A greater percentage of rural–urban migrants (17.3 per cent in Malawi and 15.5 per cent in Tanzania) and rural–rural migrants (13.7 per cent in Malawi and 10.1 in Tanzania) specialize in high-return wage/enterprise activities compared to rural non-migrants (7.6per cent in Malawi and 6.8 per cent in Tanzania). Further, 22.2 (19.6) per cent of rural–urban migrants and 15.3 (8.4) per cent of rural–rural migrants in Malawi (Tanzania) transition out of agriculture compared to 10.2 (4.8) per cent of rural non-migrants. Relocation also offers the unemployed additional job opportunities in lowand high-return non-agricultural activities. Less than 1 per cent of rural non-migrants were unemployed and obtained a low-return or high-return wage or enterprise job in the later round in Malawi (Tanzania) relative to 7.5 (7.7) per cent of rural–urban migrants and 2.8 (2.2) per cent of rural–rural migrants. While the prob abil ity of obtaining a job in a high-return activity for the unemployed is marked higher for youth moving to urban areas, such movement comes with an add ition al risk of unemployment. Approximately, 4.6 and 12.7 per cent of rural–urban migrants were unemployed in Malawi and Tanzania, respectively, in both rounds, compared to and 1.7 and 4.0 per cent of rural–rural migrants and 0.8 and 3.3 per cent of rural non-migrants. However, the difference in the proportions of migrant youth that remain unemployed in both rounds (relative to non-migrant rural youth) is only statistically significant when comparing rural–rural migrant employment rates with those of rural non-migrants. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Table 2.1. Employment rates by migration status Malawi Tanzania Nonmigrant (35–60) Nonmigrant (15–34) Rural-urban Rural-rural Nonmigrant (35–60) Nonmigrant (15–34) Rural-urban Rural-rural Employment Transitions migrant migrant migrant migrant Agriculture to student 0.000 0.016 0.032*0.007 0.003 0.015 0.013 0.010 Agriculture to unemployed 0.021 0.031 0.121** 0.055** 0.013 0.037 0.208*** 0.120*** Agriculture to agriculture 0.617 0.617 0.172*** 0.495*** 0.627 0.536 0.097*** 0.400*** Agriculture to LR wage or enterprise 0.046 0.056 0.178 0.062*** 0.016 0.016 0.129 0.032** Agriculture to HR wage or enterprise 0.080 0.046 0.044*** 0.090 0.031 0.032 0.067 0.052 LR wage or enterprise to student 0.000 0.000 0.000 0.000 0.002 0.001 0.000 0.002* LR wage or enterprise to unemployed 0.000 0.001 0.000 0.000 0.006 0.005 0.019 0.011 LR wage or enterprise to agriculture 0.030 0.030 0.011*0.018 0.115 0.075 0.000 0.062*** LR wage or enterprise to LR wage or enterprise 0.026 0.022 0.035** 0.009 0.008 0.010 0.022 0.008 LR wage or enterprise to HR wage or enterprise 0.016 0.004 0.000 0.009*** 0.016 0.009 0.031 0.010 HR wage or enterprise to student 0.000 0.000 0.000 0.000 0.001 0.001 0.000 0.000 HR wage or enterprise to unemployed 0.004 0.002 0.029 0.003 0.002 0.002 0.000 0.007** HR wage or enterprise to agriculture 0.061 0.031 0.000*0.016*** 0.062 0.041 0.000*0.021*** HR wage or enterprise to LR wage or enterprise 0.008 0.006 0.009 0.008 0.005 0.006 0.000 0.003*** Continued OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Malawi Tanzania Nonmigrant (35–60) Nonmigrant (15–34) Rural-urban Rural-rural Nonmigrant (35–60) Nonmigrant (15–34) Rural-urban Rural-rural HR wage or enterprise to HR wage or enterprise 0.070 0.030 0.061 0.024 0.038 0.016 0.011 0.014 Unemployed to student 0.000 0.004 0.000 0.001** 0.000 0.006 0.001*** 0.000** Unemployed to unemployed 0.007 0.008 0.046 0.017*0.012 0.033 0.127 0.040** Unemployed to agriculture 0.012 0.029 0.020*** 0.079 0.029 0.076 0.066*** 0.132 Unemployed to LR wage or enterprise 0.000 0.003 0.023*0.010 0.001 0.003 0.017 0.005 Unemployed to HR wage or enterprise 0.001 0.004 0.052*0.018*0.004 0.008 0.060 0.017* Student to student 0.000 0.012 0.034** 0.005 0.003 0.025 0.016*** 0.001 Student to unemployed 0.000 0.007 0.042 0.005 0.000 0.009 0.070 0.013** Student to agriculture 0.000 0.043 0.038 0.052 0.004 0.034 0.000 0.028*** Student to LR wage or enterprise 0.000 0.000 0.038** 0.012*0.000 0.001 0.000 0.003 Student to HR wage or enterprise 0.000 0.002 0.016 0.005 0.001 0.003 0.046 0.008* Individuals 1353 2276 84 600 1886 2612 96 388 Notes: * p < 0.1, ** p < 0.05, *** p < 0.1. T statistics test the difference between the proportion of migrants and non-migrants in each employment transition category for the consolidated 15–34 youth group. Rural–rural and rural–urban migrants are 15–34 years old. Table 2.1. Continued OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Can Migration be a Conduit 43 2.5 Conclusion The LSMS–ISA surveys are one of the few data collection efforts in Africa that enable researchers to view a more detailed snapshot of youth migration and employment transitions. We find young individuals, ages 15 to 24, are highly mobile in all four countries. The individual tracking protocols performed in the Malawi and Tanzania surveys allowed us to monitor transitions in income and employment between migrant and non-migrant youth samples. This analysis suggests migration is potentially welfare–enhancing both in terms of income improvements and employment prospects particularly for those migrants who were previously unemployed. One thing to note is that, like any decision, youth face different tradeoffs when contemplating where to relocate. Rural–urban migration facilitates movement out of agriculture with a greater tendency towards high-return activities in Malawi and Tanzania, yet in absolute numbers this affects a small portion of youth. Rural–rural youth migration, in contrast, attracts a greater percentage of youth. It may be the most formative mobility pattern in the transformation process by encouraging youth to diversify from exclusive employment in the agricultural sector. This chapter focuses exclusively on the migration and employment patterns ofyouth individuals. However, these patterns likely arise from decision-making behaviour at the household level. Of future interest will be to decipher the extent households spatially allocate young members to access income for agricultural intensification or farm expansion purposes, and further how these household decisions might be beneficial or harmful to youth in the long term. Understanding such household dynamics requires knowledge of whether youth migration patterns are driven by a household wealth effect, for example, from increases in investments (Deininger et al.2008), or a substitution effect between youth labour and investments. These complementarities are important as the former (not the latter) has the potential to be transformative for youth. References Barrett, C., M.F.Bellemare, and J.Y.Hou. 2010. Reconsidering conventional ex planations of the inverse productivity-size relationship. World Development 38 (1): 88–97. Beegle, K., J.De Weerdt, and S.Dercon. 2006. Orphanhood and the long-run impact on children. American Journal of Agricultural Economics 88 (5): 1266–72. Beegle, K., D.Filmer, A.Stokes, and L.Tiererova. 2010. Orphanhood and the living arrangements of children in Sub-Saharan Africa. World Development 38 (12): 1727–46. Bellemare, M.F. 2013. The productivity impacts of formal and informal land rights: Evidence from Madagascar. Land Economics 89 (2): 272–90. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi 44 Valerie Mueller and Hak Lim Lee Bezu, S., and C.Barrett. 2012. Employment dynamics in the rural nonfarm sector in Ethiopia: Do the poor have time on their side? Journal of Development Studies 48 (9): 1223–40. Bezu, S., and S.Holden. 2014. Are rural youth in Ethiopia abandoning agriculture? World Development 64: 259–72. Bustos, P., B, Caprettini, and J.Ponticelli. 2016. Agricultural productivity and structural transformation. Evidence from Brazil. American Economic Review 106 (6): 1320–65. Carletto, C., S.Savastano, and A.Zezza. 2013. Fact or artifact: The impact of measurement errors on the farm size–productivity relationship. Journal of Development Economics 103: 254–61. Davis, B., S.Di Giuseppe, and A.Zezza. 2014. Income diversification patterns in rural Sub-Saharan Africa: Reassessing the evidence. Policy Research Working Paper 7108: Washington, DC, U.S.A.: World Bank. de Brauw, A., V.Mueller, and H.K.Lee. 2014. The role of rural-urban migration in the structural transformation of Sub-Saharan Africa. World Development 10 (63), 33–42. de Brauw, A., V.Mueller, and T.Woldehanna. 2013. Motives to remit: Evidence from tracked internal migrants in Ethiopia. World Development 50: 13–23. De Vreyer, P., and F.Roubaud. 2013. Urban labor markets in Sub-Saharan Africa. Washington, DC, U.S.A.: World Bank. Deininger, K., D.Ali, S.Holden, and J.Zevenbergen. 2008. Rural land certification in Ethiopia: Process, initial impact, and implications for other African countries. World Development 36 (10): 1786–812. Deininger, K., F.Xia, and S.Savastano. 2015. Smallholders’ land ownership and access in Sub-Saharan Africa: A new landscape? Policy Research Working Paper 7285. Washington, DC, U.S.A.: World Bank. Dillon, B., and C.B.Barrett. 2014. Agricultural factor markets in Sub-Saharan Africa An updated view with formal tests for market failure. Food Policy 67: 64–77. Elder, S., and K.S.Kone. 2014. Labour market transitions of young women and men in Sub-Saharan Africa. Work4Youth, No. 9. Geneva, Switzerland: International Labour Office. Filmer, D., and L.Fox. 2014. Youth employment in Sub-Saharan Africa. Washington, DC, U.S.A.: World Bank. Fox, L., and T.P.Sohnesen. 2012. Household enterprises in Sub-Saharan Africa: Why they matter for growth, jobs, and livelihoods. Policy Research Working Paper 6184. Washington, DC, U.S.A.: World Bank. Garcia, M., and J.Fares. 2008. Youth in Africa’s labor market. Washington, DC, U.S.A.: World Bank. Grimm, M., P.Knorringa, and J.Lay. 2012. Constrained gazelles: High potentials in West Africa’s informal economy. World Development 40 (7): 1352–68. Harris, J.R., and M.Todaro. 1970. Migration, unemployment and development: Atwo-sector analysis. American Economic Review 60 (1), 126–42. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Can Migration be a Conduit 45 Hazell, P.B. and S.Haggblade. 1990. Rural–urban growth linkages in India. Working Paper Series 430. Washington, DC: World Bank. Headey, D., D.Bezemer, and P.Hazell. 2010. Agricultural employment trends in Asia and Africa: Too fast or too slow? World Bank Research Observer 25, 57–89. Headey, D., and T.S.Jayne. 2014. Adaptation to land constraints: Is Africa different? Food Policy 48: 18–33. Headey, D., M.Dereje, and A.S. Taffesse. 2014. Land constraints and agricultural intensification in Ethiopia: A village-level analysis of high-potential areas. Food Policy 48: 129–41. Honwana, A.M. 2012. The time of youth: Work, social change, and politics in Africa. London, U.K.: Kumarian Press. Jayne, T.S., D.Mather, and E.Mghenyi. 2010. Principal challenges confronting smallholder agriculture in Sub-Saharan Africa. World Development 38 (10): 1384–1398. Jedwab, R., and D. Vollrath. 2015. Urbanization without growth in historical perspective. Explorations in Economic History 58, 1–21. Jones, S., and F.Tarp. 2012. Jobs and welfare in Mozambique. WIDER Working Paper 2013/045. Copenhagen, Denmark: UNU-WIDER. Larson, D.F., K.Otsuka, T.Matsumoto, and T.Kilic. 2014. Should African rural development strategies depend on smallholder farms? An exploration of the inverse-productivity hypothesis. Agricultural Economics 45 (3): 355–67. McCullough, E.B. 2017. Labor productivity and employment gaps in Sub-Saharan Africa. Food Policy 67: 133–52. Minot, N., and T. Benson. 2009. Fertilizer subsidies in Africa: Are vouchers the answer? Issue Brief 60. Washington, DC, U.S.A.: International Food Policy Research Institute. Muyanga, M., and T.S.Jayne. 2014. Effects of rising rural population density on smallholder agriculture in Kenya. Food Policy 48: 98–13. Nagler, P., and W.Naudé. 2014. Non-farm enterprises in rural Africa: New empirical evidence. World Bank Policy Research Working Paper 7066: Washington, DC, U.S.A.: World Bank. Nin-Pratt, A., and L.McBride. 2014. Agricultural intensification in Ghana: Evaluating the optimist’s case for a Green Revolution. Food Policy 48, 153–167. Page, J. 2012. Youth, jobs, and structural change: Confronting Africa’s ‘employment problem’. African Development Bank Working Paper 155. Tunis, Tunisia: African Development Bank. Reardon, T. 1997. Using evidence of household income diversification to inform study of the rural nonfarm labor market in Africa. World Development 25 (5): 735–47. Sheahan, M., and C.B.Barrett. 2014. Understanding the input landscape in Sub-Saharan Africa: Recent plot, household, and community-level agricultural evidence. Policy Research Working Paper 7014. Washington, DC, U.S.A.: World Bank. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi 46 Valerie Mueller and Hak Lim Lee Shehu, E., and B.Nilsson. 2014. Informal employment among youth: Evidence from 20 school-to-work transition surveys. Work4Youth Publication Series 9. Geneva, Switzerland: International Labour Organisation. Smirnov, N.V. 1933. Estimate of deviation between empirical distribution functions in two independent samples. Bulletin Moscow University 2: 3–16. World Bank. 2016a. Living Standards Measurement Study—Integrated Surveys on Agriculture, Ethiopia. Washington, DC, U.S.A.http://surveys.worldbank.org/lsms/ programs/integrated-surveys-agriculture-ISA/ethiopia#bootstrap-panel--4. Accessed October 19, 2017. World Bank. 2016b. Living Standards Measurement Study—Integrated Surveys on Agriculture, Malawi. Washington, DC, U.S.A.http://surveys.worldbank.org/lsms/ programs/integrated-surveys-agriculture-ISA/malawi#bootstrap-panel--4. Accessed October 19, 2017. World Bank. 2016c. Living Standards Measurement Study—Integrated Surveys on Agriculture, Nigeria. Washington, DC, U.S.A.http://surveys.worldbank.org/lsms/ programs/integrated-surveys-agriculture-ISA/nigeria#bootstrap-panel--4. Accessed October 19, 2017. World Bank. 2016d. Living Standards Measurement Study—Integrated Surveys on Agriculture, Tanzania. Washington, DC, U.S.A.http://surveys.worldbank.org/lsms/ programs/integrated-surveys-agriculture-ISA/tanzania#bootstrap-panel--4. Accessed October 19, 2017. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi David Schwebel, Elisenda Estruch, Peter Wobst, and Ileana Grandelis, Policies for Youth Employment in Sub-Saharan Africa In: Youth and Jobs in Rural Africa: Beyond Stylized Facts. Edited by: Valerie Mueller and James Thurlow, Oxford University Press (2019). © International Food Policy Research Institute. DOI: 10.1093/oso/9780198848059.003.0003 3 Policies for Youth Employment in Sub-Saharan Africa David Schwebel, Elisenda Estruch, Peter Wobst, and Ileana Grandelis1 3.1 Introduction The global trend of increased youth unemployment has led many governments and international organizations to develop youth-targeted policies and strategies. The 2030 Agenda for Sustainable Development in its Goal 8 commits to ‘promote sustained, inclusive, and sustainable economic growth, full and productive employment, and decent work for all’. Specific targets were incorporated into this goal, including on achieving full employment for young people (8.5); on substantially reducing the proportion of youth not in employment, education, or training (8.6); as well as on developing and operationalizing a global strategy for youth employment by 2020 (8.b) (UNGA2015). At the regional level, the African Union (AU) has also a number of initiatives to promote youth employment. In its Agenda 2063, the AU commits to speed up actions to support young people through strategies that combat youth unemployment and underemployment (AU Commission2015). In its Ouagadougou Declaration, the AU sets an overall regional framework for employment promotion by all AU member states, emphasizing youth and women. The Action Plan of this Declaration underlined the importance of promoting agricultural and rural development. This was followed by the Ouagadougou+10 Declaration on Employment, Poverty Eradicationand Inclusive Development in Africa in January 2015, thus reiterating the importance of placing employment at the centre of development strategies (AU Commission2015). The Malabo Declaration on Accelerated Agricultural Growth and Transformation for Shared Prosperity and Improved Livelihoods includes a specific target to create job opportunities for at least 30 per cent of the youth in agricultural value chains (AU Summit 2014). Similarly, the Comprehensive Africa Agriculture Development Programme 1 The views expressed in this chapter are the authors’ own and do not necessarily reflect the views of the International Labour Organization (ILO) or the Food and Agriculture Organization (FAO). OUP CORRECTED PROOF – FINAL, 29/10/19, SPi 48 David Schwebel ET al. (CAADP) Results Framework (2015–25) proposes expanding local agro-industry and value chain development (VCD) inclusive of women and youth (NEPAD 2015). At the national level, many countries of Sub-Saharan Africa (SSA) have included objectives on youth employment promotion in their policies. Yet, much effort is still needed at the policy level to push support on rural youth employment to a scale commensurate with the magnitude of the challenge. In particular, additional efforts are needed to foster policy coherence towards more youth-friendly approaches for agriculture and food systems’ development. Policy coherence should especially be encouraged between employment and youth policies, as well as agricultural and rural development policies. This chapter presents a comparative qualitative policy analysis of national policies in Sub-Saharan Africa (SSA), based on a framework that incorporates the main constraints affecting the quantity and quality of rural youth employment. Whilst youth employment promotion in agriculture and rural areas is high in the regional and national agendas, few policy analytical frameworks and inventories include rural youth as a target group and this prevents analysing in a systematic and structured manner how the issue is being addressed in existing policies. Hence, the chapter builds on existing frameworks, which acknowledge the need for integrated policy approaches to youth employment, and further expands them by adding specific attention to rural youth and to the linkages between employment and rural development. The chapter follows by applying the analytical framework to 47 policies from 13 SSA countries from 1996 to 2016. The analysis follows the policy discourse analysis literature and focuses on the formulation stage of the policymaking process, therefore reviewing if the policy documents address main constraints to rural youth employment. The policies examined include development, agricultural, rural development, youth, and employment policies. With the analytical framework, this chapter contributes to a more systematic and structured approach to raise awareness among policymakers and the development community about existing gaps in addressing the constraints to rural youth employment at policy level. This framework allows for the first time to systematically assess policies of SSA with a youth employment lens associated to the different pillars of the Decent Work Agenda. The importance of developing such a framework lies in the fact that previous policy reviews showed the prevalence of actions focused on labour supply, and the need to have a stronger focus on interventions addressing the labour demand. The policy analysis conducted reveals several areas for improvement to create better employment opportunities for rural youth in SSA. In particular, the main findings show that policies focus more on promoting labour supply strategies— such as training programmes on entrepreneurship skills, rather than demandside ones—such as reducing the constraints to business development and job creation at the sectoral level. In particular, the unfavourable agribusiness environment for youth in rural areas was the constraint to rural youth employment least addressed by the policies OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Policies for Youth Employment 49 analysed. Also, some constraints related to the quality of employment (labour regulations, social protection, and social dialogue) were insufficiently addressed. Yet, a sound institutional and regulatory framework is crucial to protect workers’ rights and vulnerable youth groups. Other relevant aspects often not addressed were access to social protection and youth representation in social and policy dialogue. 3.2 Rural Labour Market Dynamics and Effects on Youth Employment Employment dynamics in rural labour markets are different from urban areas. They are generally characterized for a sub-optimal allocation of labour and lower income of workers, which leads to limited rural development (Tocco, Davidova, and Bailey2012). The main employment challenge in rural labour markets is not unemployment, but a higher incidence of underemployment, especially through self-employment and casual wage employment in the informal sector. This is the result of structural constraints in rural labour markets that particularly affect rural youth participation in the labour force. In rural areas of developing countries, the lack of infrastructure, investments, farm inputs, and policy support has led to low levels of human capital, an agricultural sector with low productivity, and limited non-agricultural employment opportunities. It is therefore important to better understand how the conditions affecting labour supply interact with those affecting labour demand across rural labour markets (ILO2008). The ILO proposes a comprehensive rural labour market framework (Table3.1) in terms of supply, demand, and institutions which is useful to analyse its impacts on rural youth in SSA. The supply side is mainly determined by demographics, access to productive assets, education levels, and social norms. The young population in SSA is expected to continue growing in the next decades, leading to approximately 370 million young people joining labour markets in the next 15years (AfDB et al. 2015). This can produce an oversupply of unskilled labour in rural areas with limited employment opportunities in farm and nonfarm activities. One of the main limiting factors is the low access to productive assets through financial services for rural youth—including credit, savings, and insurance—to start their own business. Another factor is low levels of education and limited skills that curb the productivity of rural youth and hinder their entrepreneurial abilities. Social norms, which define the role that rural youth should play in a community or household, can also impair their ability to find a job or start a business, especially for young women. The demand side is affected by economic growth, investment levels, and market access. Low public and private investments in rural areas and agriculture causes limited rural enterprise growth and job creation, which contributes to widespread underemployment and offers young school leavers few viable employment OUP CORRECTED PROOF – FINAL, 29/10/19, SPi 56 David Schwebel ET al. of the policy areas is based on existing frameworks adapted to the reality of rural labour markets. In particular, they reflect the pillars of the Decent Work Agenda emphasizing the key role that the development of the agricultural sector plays in rural areas as it still occupies the vast majority of the labour force, reaching 75 per cent or above in certain Sahel and East African countries (Losch2017). Within each of the policy areas, the selected constraints reflect the main bottlenecks that prevent young people from accessing decent jobs in rural areas. The policy framework was used to analyse the content of 47 policies from 13countries of SSA (see Table 3.A1 for complete list of policies per country). Its main purpose was to determine whether or not key constraints to rural youth employment were addressed in the policy documents. Depending on availability, four types of policy were taken by country: 1) development policy, vision, or strategy to reduce poverty; 2) employment policy; 3) youth policy; and 4) agricultural or rural development policy. The rationale for selecting these four types was to assess the consistency of interventions across policies and policy coherence towards rural youth employment promotion. 3.3.3 Policy Discourse Analysis It is important to begin by explaining what is meant by policy analysis as well as the selected theoretical background and methodology. Policy analysis emerged as a technique to better understand the policymaking process and provide decision makers with reliable knowledge and information on pressing economic and social problems (Fischer, Miller, and Sidney2007). There are multiple quantitative and qualitative methods that can be used to analyse policies depending on the area of interest and purpose of the analysis. The methodology could also vary depending on the stage in the policy cycle being observed: either the formulation, implementation, or evaluation of a policy.2 As the purpose of the current policy analysis is to examine if the constraints to rural youth employment are addressed in selected policies, we focus on the formulation stage of the policymaking process. During the policy formulation, the objectives are defined based on the priorities of a government and the development needs of a country. At this stage, it is essential to conduct a thorough analysis of the socioeconomic challenges faced in a country and engage all key stakeholders in order to define the actual priorities and needs (ILO 2012b). There are however vulnerable groups that are sometimes not included in the consultations, as often happens with rural youth. It is therefore crucial to assess if policies are considering rural youth and their employment needs. To carry out an assessment of this type a sound methodology is needed to clearly analyse the content of a policy. 2 Although policies are developed in several standard steps, there is no universal model of the policy cycle and variations might emerge depending on particular contexts and institutional arrangements. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Policies for Youth Employment 57 The applied methodology to conduct this comparative policy analysis was based on the discourse theory for policy analysis. A discourse analysis focuses on the use of language in a speech or text (here it will be policies) within a specific context. An important component of policy analysis is the examination of the discourse, in this case policy statements, from a qualitative or quantitative perspective. We therefore consider the discourse as an integral part of the policymaking process. In particular, we focus on the poststructuralist3 interpretation of the discourse theory that emphasizes the ways in which language materializes in practices (Paul2009). In other words, we analyse how governmental institutions state in their policy objectives how they will pursue specific actions that will translate into institutional practices. Due to limited availability of information, this analysis focuses on the language throughout the policy design without reaching the point of corroborating if the discourse is actually translated into practice. The rationale behind selecting discourse theory over other methods for policy analysis is that the main goal of this assessment is to clearly identify if the language being used in the policy statements properly addresses the main constraints to rural youth employment. The main unit of analysis is thus the full text of the policies in which specific words were searched. Then, the linguistic meanings of the policy statements were assessed to see if they appropriately reflected the selected constraints. It is assumed that if the policy statements address the constraints, they will eventually be transformed into a political discourse and consecutively into actions, in this case, to promote rural youth employment. A caveat in the present analysis is that it is confined to the assessment of the explicit wording of the respective policy documents, and thus does not interpret the wording in terms of policy change induced by the policy or overall country performance on rural youth employment. Being a qualitative methodology, policy analysis inevitably entails a certain risk of subjectivity, mainly with regards to the interpretation of words. Hence, in order to mitigate related errors and biases, a methodological approach was applied to the systematic and structured review of the policy documents in relation to how main constraints to rural youth employment are or are not addressed. 3.3.4 Scoring Methodology A scoring methodology was adopted to conduct desk review of policy documents for all countries and policy areas under consideration. A review of different scoring systems was carried out in order to choose the appropriate method 3 Poststructuralist Discourse Analysis has been instrumental in developing a more dynamic and historically-sensitive mode of critical inquiry claiming that texts are multiply implicated in their social contexts and, thereby, come to shape various forms of knowledge and identity (Chouliaraki2008). OUP CORRECTED PROOF – FINAL, 29/10/19, SPi 58 David Schwebel ET al. to assess how rural youth employment is addressed in the policy statements. One system is constructing a discourse quality index based on particular indicators that measure different dimensions of a political discourse (Steenbergen et al.2003). Another method is using different scales to measure the quality of the discourse, for example a five-point scale ranging from ‘very favourable’ to ‘very unfavourable’ for a specific policy issue (Stromer-Galley2007). A third system is to develop a binary indicator to capture the positive or negative quality of discourse within a policy. In this case it was decided to use the binary measurement system (1 or 0) in order to appreciate in a simple and clear way if the policies did or did not address the main constraints to rural youth employment. The binary criterion to qualify policy statements stems from the fact that general interventions to improve labour market outcomes for rural workers could also contribute to address particular constraints that rural youth face. As they are the predominant cohort facing underemployment, especially in rural areas in SSA, it would be expected that policies addressing employment issues will inadvertently be also covering or targeting the rural youth. However, given the aim of our analysis, our approach gives more weight when rural youth are explicitly mentioned. The following describes the steps taken to perform the policy discourse analysis. A desk review was first conducted to identify key policies by country. The consulted sources to collect the policies included websites of government ministries, as well as policy inventories and databases. The most recent policies were selected, including the ones that are still under approval in national parliaments. The structure, length, and content of each one varied considerably depending on the country and type of policy. To facilitate comparison, policy statements were selected that expressed the aim or objective of addressing any of the identified constraints. Within the policies, keywords were searched and subsequently assessed consistently. The criterion was assigning 1 on each policy statement if it explicitly mentioned a constraint or 0 if it did not. After scoring the nine constraints per policy, weighted averages were calculated for each policy area per country. It was decided to calculate weighted means to avoid skewness, as some policy areas have one constraint while others have multiple. Since the highest score a country can get is 100, each of the five policy areas has a weight of 20divided by the number of constraints. An illustrative example is the analysis of Tanzania’s policies. In the case of Tanzania’s National Agricultural Policy (2013), looking at the sectoral development policy area, specific words were searched linked to the constraint of unfavourable agribusiness environment for youth. The related policy statement found was: ‘The Government in collaboration with private sector shall create conducive environment for youth to settle in rural areas through improvement of social services, infrastructure, and promote rural development.’ As can be seen, it clearly makes OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Policies for Youth Employment 59 reference on how a favourable environment is necessary to attract rural youth to agriculture. Therefore a value of 1 was assigned to this particular constraint. The same process was followed for Tanzania’s four policies, assigning 1 or 0 for each constraint depending on whether they were addressed or not based on the keyword search (see Table3.3). Once the four policies had binary scores, the weighted averages for each constraint were calculated within each policy area. For example, for the policy area Sectoral Development the three constraints were scored in each of the four policies. The constraint unfavourable agribusiness environment for youth was addressed in three out of the four policies, its weighted average is therefore 5.00. Once the weighted averages were calculated, they were summed to obtain the overall policy score for Tanzania, which is 73. The scoring of the policies was not carried out arbitrarily as all policy documents were systematically reviewed based on the proposed framework in order to assess to what extent the main constraints on rural youth employment were being addressed. There are however some limitations in this analysis. First of all, it focuses on the policy design and discourse, and thus not on the implementation of the policies. On the latter, it was only verified that the policy was available online, that it was complemented with an action plan, and that it mentioned specific programmes and projects. However, we did not get to the step of corroborating whether the policies were actually translated into particular actions, nor on their ultimate impacts. The policy analysis was mainly based on a desk review and expert assessment from FAO’s Decent Rural Employment Team, as well as building on expertise generated through FAO’s field programme on the subject matter. Moreover, we did not look at the interactions and policy coherence among different government ministries and sectors due to limited availability of information. Finally, the policies under consideration covered the period between 1996 and 2016. 3.3.5 Context Indicators for Selected SSA Countries Before turning into the results of the policy analysis, it is relevant to have a general idea of the socioeconomic and political context in each of the countries. Thirteen countries from SSA were selected based on the availability of data and policies at the national level. The selection captures diversity in terms of social and economic conditions as well as income, geographic area, size, and population. It should be noted that cross-country comparisons can sometimes be misleading as many of the indicators are measured based on national definitions that may vary from country to country. Also, South Africa does not generally follow the trends of other SSA countries due to particular socioeconomic conditions. Table3.4 presents eleven key indicators that have an impact on rural labour markets and rural youth employment at the macro level. Due to the lack of available OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Table 3.3. Policy scoring example of Tanzania Policies Sectoral development Self-employment capacity, employability, and skills development Labour market institutions and regulations Social protection Social and policy dialogue Unfavourable agribusiness environment for youth Low investments in ARD Lack of labour demand Job–relevant skills, constraints, and lack of education Job search, info. and business start-up constraints Weak regulations, standards, and rights at work Social constraints Limited social protection Limited social dialogue and youth rep. National Strategy for Growth and Reduction of Poverty II 0 1 1 1 1 0 0 0 1 National Employment Policy 1 1 1 1 1 1 0 0 0 National Youth Development Policy 1 1 1 0 1 1 1 1 1 National Agriculture Policy 1 1 1 1 1 1 1 1 1 Weighted mean 5.00 6.67 6.67 7.50 10.00 7.50 5.00 10.00 15.00 73.33 OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Table 3.4. Key indicators of selected countries Country GDP per capita, PPP (US$)1 Pop. aged 15–34 (%)2 Rural pop. (%)3 Rural poverty (%)4 Agri. emp. (%)5 Wage emp. (%)5 Self-emp. (%)5 Vulner. emp.5 Youth unemp. rate (%)6 Political stability score7 Policy score8 Benin 2,272 33.9 53.2 39.7 43.2 8.1 88.9 87.7 5.2 0.05 46 Ethiopia 1,899 36.3 79.7 30.4 68.2 10.0 89.5 88.8 7.4 −1.24 63 Ghana 4,492 34.8 44.6 37.9 40.7 18.2 81.7 76.8 4.9 −0.13 53 Kenya 3,285 36.1 73.4 49.1 38.0 33.4 63.4 77.7 26.2 −1.27 64 Liberia 1,283 33.8 49.3 67.7 43.0 18.1 81.7 78.7 3.3 −0.63 69 Malawi 1,202 35.8 83.3 56.6 84.7 16.1 83.9 83.9 7.8 0.12 52 Nigeria 5,875 32.9 50.5 52.8 36.6 N/A N/A N/A 13.4 −2.11 44 Senegal 3,450 34.3 53.3 57.1 53.4 22.3 58.3 58.0 5.5 −0.13 40 South Africa 13,498 35.9 34.2 77.0 5.6 85.9 13.6 9.3 53.5 −0.08 51 Tanzania 2,946 33.2 66.9 33.3 66.7 16.2 75.9 74.0 3.9 −0.54 73 Togo 1,660 34.1 58.8 73.4 37.8 10.9 89.1 89.1 2.8 −0.16 59 Uganda 1,864 34.0 76.8 22.4 69.0 19.6 80.2 78.9 2.9 −0.93 72 Zambia 4,024 35.0 57.0 77.9 53.3 20.4 79.3 79.0 15.94 0.21 55 Sources: 1 WB WDI. 2017. Gross domestic product per capita based on purchasing power parity (PPP). 2 UNDESA, World Population Prospects. 2017. The African Union’s definition of youth covers the age range 15–34. 3 UN World Urbanization Prospects. 2018. Rural population refers to people living in rural areas as defined by national statistical offices. 4 WB WDI. Latest available year. Rural poverty headcount ratio is the percentage of the rural population living below the national poverty lines. 5 ILOSTAT 2017. Status in employment distinguishes between two categories of the employed—(a) wage and salaried workers and (b) self-employed workers. The vulnerable employment rate is calculated as the sum of contributing family workers and own-account workers as a percentage of total employment. 6 ILOSTAT 2017. The unemployment rate indicates the proportion of the labour force that does not have a job and is actively looking and available for work covering persons aged 15 to 24. 7 WB WGI. 2014. Political Stability and Absence of Violence/Terrorism measures perceptions of the likelihood of political instability and/or politically-motivated violence, including terrorism. Units range from –2.5 to 2.5, with higher values corresponding to better governance. 8 Own calculations based on rural youth employment policy framework; units range from 0 to 100. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi 62 David Schwebel ET al. data on rural employment, general employment data was included. The policy score obtained from the discourse analysis was also added to appreciate its relationship with the other indicators. As can be observed in the socioeconomic indicators, the GDP per capita varies between countries: eight are low-income economies, four are lower-middle-income economies, and only South Africa is an upper-middle-income economy. In all countries, young people (aged 15–34) comprise between 30 and 40 per cent of the total population. In 10 out of 13 of the countries more than half of the population lives in rural areas; on the one side Ethiopia has 80 per cent of its population living in rural areas, on the other side South Africa has only 34 per cent (UN DESA 2018). Likewise, rural poverty prevails in many of the countries reflecting low agricultural incomes, with Zambia having around 78 per cent of the rural population living below the poverty line and Uganda with around 22 per cent. With respect to labour market indicators, agriculture remains the main sector of employment in most countries, employing more than half of the population in 7 out of 13 of the countries. For example, in Ethiopia around 73 per cent of the population works in agriculture against only 4.6 per cent in South Africa. Most of the economically active population is self-employed, including employers, ownaccount workers, members of producers’ cooperatives, and contributing family workers. Also, most of the employment is considered to be vulnerable as the majority of the employed population includes own-account workers and contributing family workers, who are less likely to have formal work arrangements, and are therefore more likely to lack decent working conditions (ILO2013). Youth unemployment rates also vary considerably between countries; on the one hand Benin has a youth unemployment rate of only 2 per cent, on the other hand 50.7 per cent of South Africa’s youth is unemployed. With regard to the political context, some countries have a low score in political stability and absence of violence/terrorism, which can considerably affect the policy and institutional environment in a country. For example, Nigeria has the lowest score given the current political instability and terrorism in the country, which could hamper the implementation of policies. Finally, the policy score is the result of the analysis carried out with the proposed policy framework reflecting the average of the scored constraints per country. The country that received the highest score is Tanzania with 73, while the one with the lowest is Senegal with 33. As can be seen, the policy scores do not necessarily correspond with the other indicators presented; for example, a low policy score does not translate into a high rural poverty headcount. The reason for this is that the policy assessment only focuses on the policy design without checking whether the policy statements were indeed transformed into concrete actions that have an impact on the economic performance and on rural youth employment in a given country. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Policies for Youth Employment 63 3.4 Main Findings of the Comparative Policy Analysis This section presents the main results of the comparative policy analysis conducted with the rural youth employment policy framework in the selected SSA countries. The implementation of the methodology through the proposed policy framework provided revealing findings which require an in-depth discussion. The results of scoring policy statements within the four types of policies are presented at the policy area, constraint, and country level. The discussion of the findings highlights the issues with lower scores, interprets the results, and proposes strategies to overcome these challenges. 3.4.1 Policy Area Analysis The scores for the five policy areas let us appreciate the differences between countries in addressing the constraints to rural youth employment within the policies. In Figure 3.1, we can see which policy areas received the highest and the lowest scores on average, and the performance of each country across the five policy areas. In each policy area the scores for the thirteen countries range from 0 to 100, with the average marked with a line. As previously explained, the score for each country within a policy area is the average that resulted in the binary assessment of the nine constraints that were grouped into the five policy areas for each of the analysed policies. The policy area that received the highest score is self-employment, employability, and skills development (74), supporting the argument that most policies focus on labour supply. Liberia and Benin had the lowest score (33) while South Africa obtained the highest (100). South Africa received the highest possible score because the three analysed policies explicitly addressed the two main constraints on labour supply for rural youth, namely: lack of skills and education, as well as inadequate job matching services, information, and business start-up resources. For example, South Africa’s National Youth Policy (2015) states that training young people in skills relevant to agriculture and the agricultural value chain will also help to attract young people to the sector and promote agriculture and agroprocessing. The second policy area is sectoral development (56), linked with labour demand, which includes ARD investments, rural labour demand, and agribusiness environment for youth. Although many policies include this type of intervention, they do it at a lesser extent than labour supply interventions. Tanzania received the highest score in this policy area (92), while Nigeria and Malawi had the lowest score (33). Kenya has a score in sectoral development on the average (56) given that the three analysed policies addressed only some of the constraints to the labour demand of rural youth. For instance, Kenya’s Agricultural OUP CORRECTED PROOF – FINAL, 29/10/19, SPi 64 David Schwebel ET al. Sector Development Policy (2010) commits to empower the rural youth by sensitizing them on lucrative ventures in the agricultural sector and establishing processing plants for value addition in rural areas to provide employment opportunities for youth. The third policy area is labour market institutions and regulations (53), indicating that labour standards and social constraints is an issue that needs to be better addressed. Uganda received the highest score (88), while Senegal had the lowest (17). Social constraints are a major challenge for rural youth. Local traditions and social norms prevent young people, and especially young women, from accessing the necessary productive resources. Particularly young women have lower incomes since they are less likely than their male counterparts to own land. Malawi’s Youth National Policy (2013) addresses this constraint with the goal of providing access to productive agricultural land in adequate proportion and other factors of production for the youth who fail to access these resources due to culture, gender, and/or other socioeconomic factors. Malawi Ghana Malawi Ethiopia Ghana Malawi Zambia Benin Tanzania Zambia Zambia Zambia Togo Liberia Ethiopia Zambia Tanzania Benin Uganda Uganda South Africa Kenya Uganda Tanzania Uganda Tanzania Uganda Uganda Ethiopia Tanzania Ethiopia Zambia Nigeria Nigeria Malawi Nigeria Nigeria Ghana Senegal Senegal Senegal Ghana Tanzania Nigeria South AfricaSouth Africa South AfricaSouth Africa South Africa Benin Ghana Ethiopia Benin Malawi Togo Togo Liberia Liberia Benin LiberiaLiberia Liberia Togo Kenya Kenya Senegal Kenya Kenya Togo South Africa Senegal 0 10 20 30 40 50 60 70 80 90 100 Sectoral Development Self-employment, Employability, and Skills Development Labour Market Institutions and Regulations Social Protection Social and Policy Dialogue Figure 3.1. Total compliance by policy area OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Policies for Youth Employment 65 In third place is also the social protection policy area (53), showing that policies partially cover social security schemes such as cash transfers, minimum wages, tax exemptions, health services, and so on. All of Liberia’s policies adequately addressed the access to social protection for rural youth, it thus received the highest score (100). In contrast, only one out of four of Nigeria’s policies addressed social protection issues for vulnerable groups, so it received the lowest score (25). One of the policies that explicitly mentioned young people is the National Youth Policy of Ghana (2010) which commits to provide social protection for the vulnerable and excluded youth. Finally, the least addressed policy area is social and policy dialogue (51) as promoting the organization of rural young workers to increase their bargaining power was only included in around half of the policies. Some policies did commit to promote social dialogue and tripartism, however few explicitly mentioned the importance of engaging rural youth in decision-making processes. For example, the Second National Youth Policy Document of the Federal Republic of Nigeria (2009) declares that governments should always lend support to and be willing to engage in dialogue with youth-led organizations and work with a broad range of the youth population. In contrast, none of Senegal’s policies mentioned the importance of involving excluded young people in social dialogue. As can be seen, the systematic review of policies shows a clear trend to address one policy area, that is, promoting self-employment, employability, and skills development, over the other four policy areas. This finding indicates in turn that policies would focus more on the supply side of the labour market, paying less attention to other constraints that also affect the access of rural youth to decent employment. It is also noteworthy that the other four policy areas had a similar score of just over 50, showing that around half of the policies would be addressing in rather similar ways these policy areas. 3.4.2 Constraint Analysis The analysis at the constraint level adds further elements to our comparative policy analysis to examine how the policies across 13 SSA countries addressed the main challenges to rural youth employment. Table3.5 presents the constraints with the lowest scores per country. To capture the binding character of the constraints, countries were bound by the lowest score. As can be seen, the constraint with the lowest score in most countries is unfavourable agribusiness environment for youth within the sectoral development policy area. This shows that most policies do not explicitly focus on creating an enabling environment for youth agribusiness development. Many policies did mention the intention to enhance a conducive business environment to promote private sector growth, but they neither focus on the agricultural sector nor on rural youth. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Table 3.A2. Scores for constraints per country Country Sectoral development Self-employment capacity, employability, and skills development Labour market institutions and regulations Social protection Social and policy dialogue Unfavorable agribusiness environment for youth Low investments in ARD Lack of labour demand Job-relevant skills constraints and lack of education Job search, information, and business start-up constraints Weak regulations, standards, and rights at work Social constraints Limited social protection Limited social dialogue and youth representation Benin 25 75 25 75 25 075 50 50 Ethiopia 0 75 50 100 50 25 75 75 75 Ghana 075 50 25 100 50 75 75 25 Kenya 0 67 100 67 100 67 100 33 67 Liberia 33 100 100 100 0 0 100 100 67 Malawi 075 075 50 25 50 75 50 Nigeria 0 75 25 100 50 50 75 25 25 Senegal 33 67 100 100 67 0 33 33 0 South Africa 67 33 67 100 100 0 67 33 33 Tanzania 75 100 100 75 100 75 50 50 75 Togo 33 100 100 100 33 67 33 33 67 Uganda 0100 75 100 75 100 75 50 75 Zambia 25 50 75 100 50 50 50 50 50 OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Policies for Youth Employment 73 References African Development Bank, OECD, UNDP, and UNECA. 2015. African economic outlook 2015: Regional development and spatial inclusion. Paris, France: Organisation for Economic Co-operation and Development. African Union Commission. 2006. African youth charter. Banjul, the Gambia: African Union. African Union Commission. 2015. Agenda 2063: The Africa we want. Addis Ababa, Ethiopia: African Union. African Union Summit. 2014. The Malabo declaration on accelerated agricultural growth and transformation for shared prosperity and improved livelihoods. Malabo, Equatorial Guinea: African Union. Chouliaraki, L. 2008. Discourse analysis. In The SAGE handbook of cultural analysis, ed. T.Bennett, and J. Frow. London, U.K.: SAGE Publications. FAO. 2016. Incorporating decent rural employment in the strategic planning for agricultural development. Guidance Material #3. Rome, Italy: Food and Agriculture Organization of the United Nations. FAO, IFAD, and CTA. 2014. Youth and agriculture: Key challenges and concrete solutions. Rome, Italy: Food and Agriculture Organization of the United Nations. Fischer, F., G.J.Miller, and M.S.Sidney. 2007. Handbook of public policy analysis: Theory, politics, and methods. Boca Raton, FL, U.S.A.: CRC Press. ILO. 2008. Report IV: Promotion of rural employment for poverty reduction. International Labor Conference, 97th Session. Geneva, Switzerland. ILO. 2012a. The youth employment crisis: A call for action. Resolution and conclusions of the 101st Session of the International Labor Conference. Geneva, Switzerland: International Labor Organization. ILO. 2012b. Guide for the formulation of national employment policies. Geneva, Switzerland: International Labor Organization. ILO. 2013. The informal economy and decent work: A policy resource guide, supporting transitions to formality. Geneva, Switzerland: International Labor Organization. Independent Evaluation Group (IEG). 2012. Youth employment programs: An evaluation of World Bank and International Finance Corporation support. Washington, DC, U.S.A.: World Bank. Leavy, J., Smith, S. 2010. Future farmers? Exploring youth aspirations for African agriculture. Futures Agricultures Consortium Policy Brief 037. Losch, B. 2017. +789 Million and Counting: the Sub-Saharan African Equation. Employment Research Brief. Geneva, Switzerland: International Labor Organization. MIJARC, IFAD, and FAO, 2012. Facilitating access of rural youth to agricultural activities. Summary of the findings of the project implemented by MIJARC in collaboration with IFAD and FAO. Rome, Italy: International Fund for Agricultural Development. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi 74 David Schwebel ET al. New Partnership for Africa’s Development (NEPAD). 2015. The CAADP results framework (2015–2025). Midrand, South Africa: New Partnership for Africa’s Development. Paul, K.T. 2009. Discourse analysis: An exploration of methodological issues and a call for methodological courage in the field of policy analysis. Critical Policy Studies 3 (2): 240–53. Proctor, F., Lucchesi, V. 2012. Small-scale farming and youth in an era of rapid rural change. London/The Hague: IIED/HIVOS. Seck, A. 2016. Fertilizer subsidy and agricultural productivity in Senegal. AGRODEP Working Paper 0024. Washington, DC, U.S.A.: International Food Policy Research Institute. Steenbergen, M.R., A.Bächtigerb, M.Spörndlib, and J.Steinera. 2003. Measuring political deliberation: A Discourse Quality Index. Comparative European Politics 1 (1): 21–48. Stromer-Galley, J. 2007. Measuring deliberation’s content: A coding scheme. Journal of Public Deliberation. 2 (1): 1–35. Tocco, B., S.Davidova, and A.Bailey. 2012. Key issues in agricultural labor markets: A review of major studies and project reports on agriculture and rural labor markets. Factor Markets Working Paper No. 20. Brussels, Belgium: Centre for European Policy Studies. UN Department of Economic and Social Affairs (UN DESA). 2018. World urbanization prospects: The 2018 revision. New York, NY, U.S.A.: United Nations. UN General Assembly (UNGA). 2015. Transforming our world: The 2030 Agenda for Sustainable Development. Resolution 70/1, 25 September 2015. New York, NY, U.S.A.: United Nations. World Bank. 2007. Miles to go: A quest for an operational labor market paradigm for developing countries. Washington DC, U.S.A.: World Bank. Youthpolicy.org. 2014. The state of youth policy in 2014. Berlin, Germany: Youth Policy Press. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Danielle Resnick, Troublemakers, Bystanders, and Pathbreakers: The Political Participation of African Youth In:Youth and Jobs in Rural Africa: Beyond Stylized Facts. Edited by: Valerie Mueller and James Thurlow, OxfordUniversity Press (2019). © International Food Policy Research Institute. DOI: 10.1093/oso/9780198848059.003.0004 4 Troublemakers, Bystanders, and Pathbreakers The Political Participation of African Youth Danielle Resnick 4.1 Introduction Creating decent jobs for African youth not only is critical for improving their economic welfare but also has political salience given the historic ability of this constituency to disrupt established governance structures. On the one hand, African youth have been viewed as progressive and pro-democratic. African youth, consisting of secondary school and university students, played a significant role in the anti-colonial movements of the 1950s and 1960s (Allman1990, Burgess 2005). Initially motivated by teaching shortages, high food prices, and poor study facilities, they were similarly at the vanguard of protests in the late 1980s and 1990s in more than a dozen African countries, which heralded a wave of transitions from one-party to democratic rule (Bratton and van de Walle1992). In more recent years, youth groups in countries as diverse as Angola, Burkina Faso, Senegal, South Africa, and Sudan have proved important actors in protesting against violations of the constitution and the rule of law by leaders in those countries (Alexander2010, Hamilton2010, Wonacott2012). On the other hand, there has been concern that unemployed youth are especially prone to radicalization and anti-government behaviour, particularly if they are unemployed. Kaplan (1996) famously suggested that African youth are ‘out of school, unemployed, loose molecules in an unstable social fluid that threatened to ignite’. More generally, some research suggests that countries with youth bulges have a higher likelihood of experiencing political violence since high unemployment creates low opportunity costs for this group (Collier2007, Leahy et al.2007, Urdal2006). The role of youth militias at the forefront of some of Africa’s civil wars in the 1990s, ranging from Liberia, Rwanda, and Sierra Leone, bolstered this alarmist view. More recently, the enduring presence of youth vigilante groups, such as Nigeria’s Bakassi Boys or Côte d’Ivoire’s ‘microbe’ criminal gangs in Abidjan, further creates a sense of urgency about the social implications of youth unemployment. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi 76 Danielle Resnick Whether peaceful or violent, inspirational or exploitative, these contrasting narratives have played an important role in placing youth and youth employment back on the agenda of African governments, epitomized by the African Union’s 2006 Youth Charter (AU2006) and the more than two dozen African countries that drafted youth policies during the 2000s onwards.1 Since particularly urban youth are seen as potentially more disruptive, and rural youth are perceived as more deprived, practical policy responses have included efforts to improve the attractiveness of agriculture to encourage young people to either stay in rural areas or return to them (Sumberg et al.2015). Examples include the ‘Return to Agriculture’ Plan launched by former President Wade of Senegal, and the block farm programmes in countries such as Ghana and Zambia (Benin et al.2013, Sall 2012).2 Yet, do African youth actually mobilize for change through extra-institutional channels, such as protest, more than their older counterparts? Is unemployment their main policy preoccupation, or do other concerns take precedence? Are these behaviours and preferences consistent over time or dependent on the wider economic and political context in which young people are embedded? This chapter addresses these questions in detail by first discussing the literature on youth and protest as well as developments in Africa more recently. In doing so, I look at whether African youth are more likely to protest today than in the past by using the Armed Conflict Location and Event Data (ACLED), which spans the 1997–2015 period and includes all Sub-Saharan African countries. While this analysis provides a macro perspective on trends over time, the chapter subsequently provides a more microanalysis by employing Afrobarometer public opinion data for 16 countries. Focusing on six age cohorts between 2003 and 2014, the analysis probes rural and urban youth’s socioeconomic status, policy preferences, political awareness and trust in institutions, and their political participation, including in protest activities. This is followed by a multivariate analysis of the micro-level drivers of protest behaviour. The findings reveal that while slightly higher among the youth, protesting is a form of political participation engaged in by older Africans as well. However, results from surveys conducted in 2014 show that the drivers of youth protest vary from a decade earlier. Specifically, the more recent findings indicate that among the youth, protesters can be characterized as ‘frustrated activists’ who have higher levels of education and who are more engaged in their communities than nonprotesters but who are also unemployed, experience higher levels of de priv ation, and have less trust in political institutions. This suggests that for governments intent on minimizing protests, one pathway is to ensure that employment creation projects match young people’s skills and aspirations. Moreover, governments 1 See: http://www.youthpolicy.org/nationalyouthpolicies/. 2 On Z ambia, see: http://agrf.org/agrf2015/zambia-targets-one-million-hectares-of-farm-block-irrigationfor-youth-led-agriculture-development/. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Troublemakers, Bystanders, and Pathbreakers 77 need to generate greater trust that youth policies and initiatives are aimed at enhancing this constituency’s long-term potential rather than simply mobilizing their short-term political support. 4.2 Macro Trends in Protest Activity in Africa Political participation consists of an array of ‘actions by citizens which are aimed at influencing decisions which are, in most cases, ultimately taken by public representatives and officials’ (Parry, Moyser, and Day1992). Globally, young people are associated with certain trends in political participation. On average, younger people are less likely to vote than older ones (Franklin2004, Wattenberg2008), but considered more likely to protest. Historically, one reason that they are more likely to protest is because of a lack of career or familial responsibilities (Parry, Moyser, and Day1992). More recently, others have noted that with the spread of new technologies and shifts in political ideologies, young people are more likely to view themselves as ‘self-actualizing’ or ‘engaged’ citizens rather than ‘dutiful’ citizens (Bennett2008, Dalton2008a). While the latter favours conventional forms of participation and the ‘politics of loyalties’, including voting, the former emphasizes more direct actions and the ‘politics of choice’ (Norris2002), which manifests via protests, demonstrations, and boycotts. Since the late 2000s, and especially from 2011 onwards, Africanists have observed a wave of protests across the continent, particularly in urban areas (Branch and Mampilly 2015). However, the trend is not specific to Africa. Economic crisis and de-alignment with traditional political parties have renewed the salience of protests globally as a major form of political participation (Rüdig and Karyotis 2013). In developing countries, Valenzuela, Arriagada, and Scherman (2012) observe that three key elements have characterized these protests: organization by the masses rather than by political parties, the central role of social media, and the dominant role played by youth. Consequently, two patterns have been suggested. One is that protests are generally on the rise again and secondly, that protests disproportionately are a modality of participation pursued by the youth. The first idea, but not necessarily the second, appears to be borne out by aggregate data for the African region. Specifically, Figure4.1 utilizes data on protest events available from ACLED to analyse trends over time.3 This coding involved first isolating media reports on protest events that mentioned involvement of ‘youth’, ‘young people’, ‘students’, and/or ‘teenagers’. Subsequently, I examined the full description of the event to 3 Protest events are not equivalent to outright violence and Raleigh (2015) has found that the locus of violent conflict in Africa has increased considerably over the last 20 years from rural to urban areas. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi 78 Danielle Resnick ensure that it was coding the intended outcome correctly, and I excluded events that mentioned youth being a victim of a crime rather than youth participating in a protest movement. Duplicate events being reported by the same source were removed. However, duplicates were retained if they were referring to events that occurred on the same day but in different locations or if they were the same event dispersed over multiple days (for example, student strike). Importantly, Figure4.1 demonstrates that the total number of youth protests has increased dramatically between 1997 and 2015, largely supporting Branch and Mampilly’s (2015) observation of a new wave of protest activities in the region. Aseries of major mobilizing events partially accounts for this trend. Forinstance, rising costs of electricity, fuel, and schooling motivated Nigeria’s Occupy Movement, South Africa’s ‘Fees Must Fall’ campaign, Ghana’s Red Friday protests, and Uganda’s ‘Walk to Work’ protests. Electoral malfeasance and attempts to change constitutions were major precipitators in Senegal’s Fed Up movement, and in anti-government protests in Burkina Faso and Democratic Republic of Congo. At the same time, however, as a share of total protests, youth protests tend to wax and wane and have never constituted more than 32 per cent of total protests. Key spikes in activity have occurred in 1999, 2003, 2010, and 2015. This observation suggests that protest is not just a modality of political participation that is intrinsically tied to age; rather, youth protests may simply coincide with broader periods of social and economic discontent. In other words, protest activity may not simply be a life cycle event whereby younger people are more inclined to go into the streets, but a form of political participation that is more (or less) pronounced depending on the broader contextual environment. 0 5 10 15 20 25 30 35 199 7 1998 1999 2000 200 1 2002 2003 2004 2005 2006 200 7 2008 200 9 2010 2011 2012 2013 2014 2015 0 100 200 300 400 500 600 700 800 900 Share of youth protests in total (%) Number of youth protests Figure 4.1. Youth protests over time in sub-Saharan Africa OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Troublemakers, Bystanders, and Pathbreakers 79 4.3 Identifying Age Cohorts Over Time The latter observation suggests that three potential effects are relevant for research on youth in Africa and elsewhere: life-cycle, generational, and period effects. Lifecycle, or age, effects imply that individuals’ behaviours and characteristics change and mirror those of their older peers as they age (Nie, Verba, and Kim1974). With the accumulation of more experience, individuals always alter their behaviours over time. Examples of period effects would include if a survey is taken during an election, a food crisis, or a drought (Neundorf and Niemi2014, Yang and Land 2013). The events affect all age groups at the same time but, the level of impact may differ depending on where one is located in the life-cycle (Neundorf and Niemi 2014). Generational effects imply that period effects disproportionately affect those at a certain stage of life, particularly during late adolescence and early adulthood (Dalton1988, Markus1983, Ryder1965). In other words, while all age groups may be exposed to a civil war, it may leave a deeper impression on younger people that continues to affect their behaviours and outlook as they age. Given the range of political transformations and shifts in economic ideology in Africa since independence, taking all these effects into account is essential for better understanding whether African youth are distinctly different depending on their birth cohort. The Afrobarometer public opinion data help analyse distinctions between life-cycle, period, and generational effects and their attendant implications for pol it ical participation. The Afrobarometer project includes six rounds of data collection.4 The analysis here employs survey Rounds 2 and 6, which were taken in 2002–4 and 2014–15, respectively. Consequently, the data spans at least a decade. This is necessary since if repeated cross-sectional data is less than 10 years apart, cohort and age effects become increasingly correlated (Smets and Neundorf 2014). The Afrobarometer dataset used here includes sixteen countries with a range of political regimes and economic development: Botswana, Cape Verde, Ghana, Kenya, Lesotho, Malawi, Mali, Mozambique, Namibia, Nigeria, Senegal, South Africa, Tanzania, Uganda, Zambia, and Zimbabwe. Collectively, the data captures 24,301 observations in Round 2 and 29,972 observations in Round 6. Cohorts refer to a set of individuals who have shared experiences of socialization (Glenn2005), and they often are operationalized by individuals’ birth years (Neundorf and Niemi2014). In creating cohorts, six age groups were considered. These include three ‘youth’ groups that span ages 18–24, 25–29, and 30–34. Doing so allows for including both the United Nations’ upper youth threshold of 24 and the African Union’s more expansive upper threshold of 34. At the same time, 25–29 year olds sandwiched between these two decadal cut-offs may have finished 4 See http://www.afrobarometer.org/. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi 80 Danielle Resnick school but not yet established families, potentially resulting in distinct priorities and modalities of participation. Table 4.1 presents the classification of the age cohorts across the two survey rounds, their respective birth years, and the broader political and economic context facing the continent at the time each cohort reached early adulthood, which in this case is symbolized by being 18 years or older. Late adolescence or early adulthood is considered to be a highly impressionable period when behaviours and preferences begin to crystallize and may persist throughout one’s life (Jennings 1996, Markus1983). In this case, 18 years old also is the minimum voting age in most African countries and therefore the time when individuals become most aware of their political environments. Notwithstanding the diversity of countries in the region, some common trends were occurring as each of these cohorts reached early adulthood. For those born in 1948 or prior to that year, they came of age at a time of transition from colonial Table 4.1. Description of age cohorts from Afrobarometer 2003 Surveys (Round 2) Age at survey year Year of birth Year turned 18 Political era Economic era 18–24 1979–1985 1997–2003 Democratic consolidation HIPC and PRSPs 25–29 1978–1974 1992–1996 Democratic transitions SAPs 30–34 1969–1973 1987–1991 Democratic liberalization SAPs 35–44 1959–1968 1977–1986 One party regimes Stabilization 45–54 1949–1958 1967–1976 One party regimes ISI 55+ 1948 and earlier 1966 Colonial transition Extractive economies 2014 Surveys (Round 6) 18–24 1990–1996 2008–2014 Born frees Resurgence 25–29 1985–1989 2003–2007 Democratic consolidation MDGs 30–34 1980–1984 1998–2002 Democratic consolidation HIPC and PRSPs 35–44 1970–1979 1988–1997 Democratic liberalization and transition SAPs 45–54 1960–1969 1978–1987 One party states Stabilization 55+ 1959 and earlier 1977 One party states ISI Notes: HIPC = Highly Indebted Poor Countries; ISI = Import Substitution Industrialization; MDGs = Millennium Development Goals; PRSPs = Poverty Reduction Strategy Papers; SAPs = Structural Adjustment Programmes. For ease of explication, the Round 2 and Round 6 surveys refer to a base year of 2003 and 2014, respectively since those are the survey years of a majority of the countries inthe sample. Source: Author’s compilation OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Troublemakers, Bystanders, and Pathbreakers 81 administrations to independent states and from extractive economic policies aimed at benefiting European countries to more inwardly-focused import substitution industrialization (ISI) policies.5 They were followed by a cohort that, with the exception of a few Southern African countries, largely matured under independent and increasingly autocratic, one-party regimes that sought to solidify their ISI strategies. Those born in the 1960s and early 1970s reached 18 when oneparty states were overwhelmingly the norm, but the flaws of ISI were leading to massive debt and macroeconomic contraction. Starting with Ghana in 1982, this period heralded the beginning of stabilization policies under international financial institutions (IFIs). A tumultuous period followed for those born mostly in the 1970s and who reached maturity as IFIs sought not only to stabilize but to actually ‘structurally adjust’ economies. The resultant austerity measures and privatizations prompted a wave of pro-democracy protests and political lib er aliza tion, starting with Benin in 1989 and Zambia in 1991 (Bratton and van de Walle1992). The successive cohort, born in the late 1970s to mid-1980s, were often voting in their country’s first democratic elections.6 Around the same time, IFIs and NGOs launched the heavily indebted poor countries (HIPC) initiative and generated a renewed focus on tackling poverty via the poverty reduction strategy papers (PRSPs). Those born in the late 1980s faced similar political circumstances, but matured at the time of major global and regional initiatives, such as the Millennium Development Goals (MDGs) and the Comprehensive African Agriculture Development Program (CAADP), which prompted large investments in education, health, and agriculture. Youth of this generation therefore began seeing a big improvement in educational access compared to youth of prior generations (Resnick and Thurlow2015). Finally, those born between 1990–6 are often referred to as the ‘born free’ generation (Mattes2012) in that they largely escaped living under purely authoritarian regimes. This cohort matured under a period of relative economic resurgence, bolstered by high commodity prices. This often resulted in dual perspectives, including optimism about Africa’s growing middle class (McKinsey2010, Ncube and Lufumpa2015) and pessimism that this resurgence did not ameliorate poverty or inequality (World Bank2016). Beyond political and economic variations, these cohorts obviously have also lived through very different communications, technology, and media environments. 5 See Nugent (2004) for more details on these periodicizations, especially prior to 2000. 6 Within the Afrobarometer sample, there are some important caveats to this characterization. Uganda finally allowed multi-party competition in 2005 and Zimbabwe allowed multi-party competition starting with the 2000 parliamentary and 2002 presidential elections. In both cases, elections have not been deemed ‘free and fair’, with an uneven playing field for the incumbent presidents. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi 88 Danielle Resnick for the future and the reality of far fewer job opportunities as economies contracted under SAP conditions. 4.5 Political Perspectives and Modes of Participation As noted earlier, today’s African youth have not only experienced different eras of economic policy, contraction, and growth but also distinctly different political environments that might condition their modalities of political participation. Two key concepts are probed here. The first reflects respect for the political status quo, which is measured here by both closeness to the incumbent party in government and trust in political institutions. Closeness to the incumbent is captured by first identifying the share of respondents who note that they are close to a pol it ical party. Among those who are, Afrobarometer subsequently asks for which specific political party the respondent feels the greatest affinity. If the party chosen was the incumbent party at the time the survey was conducted, then the respondent is coded as being close to the incumbent; otherwise, if any other party is selected instead, the respondent is close to the opposition. Political trust refers to an orientation towards government based on ‘how well the government is operating according to people’s normative expectations’ (Hetherington 1998). Some argue that the erosion in trust of political institutions can have long term negative consequences for social and political stability (Newton and Norris2000, Scholz and Lubell1998). Political trust is operationalized here through an index that encompasses nine different formal political institutions: the president, parliament, electoral commission, local government, ruling party, opposition parties, police, army, and the courts.11 The index runs from 0, denoting ‘not at all’, to 3, conveying ‘a lot’ of trust. Table4.5 suggests that life-cycle effects are fairly pronounced with a higher share of older age groups across both survey rounds likely to express that they trust institutions a lot and that they feel close to the incumbent party. Notably, though, there is a fairly dramatic increase in trust for institutions across all age groups in rural areas between 2003 and 2014, but this is specifically true amongst the youngest age group. This trend suggests important period effects, including that over the decade in question, there was a noticeable reversal of urban bias as African governments try to allocate more distributive goods in rural areas so as to mobilize rural voters for elections (Bates and Block2013, Boone and Wahman2015). The exact source of increased rural trust is difficult to pinpoint but some trends are suggestive from Table4.6, which disaggregates trust levels by country and 11 Trust in informal institutions, such as religious or traditional authorities, was excluded. In addition, trust in the tax revenue authorities was excluded from the index because the question was omitted from the Round 2 survey. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Table 4.5. Respect for political status quo (per cent of age group) Variable 2003 2014 18–24 25–29 30–34 35–44 45–54 55+ 18–24 25–29 30–34 35–44 45–54 55+ Trust institutions a lot Total 31.6 33.2 35.7 39.3 41.5 48.1 36.7 34.9 38.1 39.8 43.9 51.1 Rural 34.7 35.1 37.6 43.5 44.6 50.4 42.7 40.1 42.7 45.4 47.8 54.4 Urban 27.8 30.2 32.3 31.3 35.0 42.2 29.5 28.6 31.9 31.7 37.5 44.5 Close to incumbent party Total 31.0 34.8 34.9 37.3 39.3 40.1 28.6 30.0 31.4 35.5 37.4 38.4 Rural 35.7 39.5 40.1 41.5 44.0 41.9 32.0 33.0 36.1 38.0 40.1 39.4 Urban 25.1 27.6 25.2 29.0 29.3 35.6 24.5 26.3 25.0 31.9 32.9 36.3 Total observations 5,701 3,926 3,215 4,857 2,912 3,087 6,240 5,059 4,308 6,240 3,773 4,137 Source: Calculated from Afrobarometer, Rounds 2 and 6. All descriptives are weighted by cross-country survey sample weights. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi 90 Danielle Resnick averages across all youth groups (that is, 18–34). For instance, some of the biggest increases in trust over the 2003–14 period were in countries such as Uganda, where many new districts have been purposely created in rural areas (Grossman and Lewis2014) and in countries that have had sustained targeted input subsidy programmes that began in 2003 or soon thereafter, including in Kenya, Malawi, Tanzania, and Zambia (Jayne and Rashid2013). By contrast, the large declines in rural trust observed in Mali and Senegal coincide, respectively, with government collapse and the onset of civil war in 2012 (Bleck, Dembele, and Guindo2016) and to perceptions of growing government corruption (Sall2015).12 Beyond their perspectives on political institutions and parties, the second concept focuses distinctly on political participation, which manifests in a variety of ways. If aligned along a spectrum, the most basic measure is an intrinsic interest in current events and efforts to stay informed of such events. Those who are more informed may be more likely to pursue protest activities as they become aware of perceived injustices or of events that could serve as rallying points for mobilization (Tufekci and Wilson2012, Valenzuela, Arriagada, and Scherman2014). To examine these dynamics, I examine young people’s degree of interest in public affairs and 12 The Lesotho Round 6 survey occurred in May 2014, prior to the unexpected August 2014 coup, which may partially explain the relatively high trust in government at that time. Table 4.6. Trust in government institutions by country and rural youth (per cent trusting a lot) Country 2003 2014 Difference 18–34 18–34 Botswana 29.7 38.9 9.2 Cape Verde 28.2 33.8 5.6 Ghana 33.4 28.1 −5.4 Kenya 25.8 37.1 11.3 Lesotho 28.6 52.0 23.4 Malawi 29.4 41.9 12.5 Mali 61.4 35.9 −25.5 Mozambique 49.4 48.4 −1.0 Namibia 60.4 58.4 −2.0 Nigeria 14.3 12.0 −2.2 Senegal 77.8 60.6 −17.2 South Africa 30.0 32.2 2.2 Tanzania 37.3 51.4 14.1 Uganda 19.6 49.6 30.0 Zambia 31.9 43.1 11.2 Zimbabwe 42.8 44.4 1.6 Sources: Calculated from Afrobarometer, Rounds 2 and 6. All descriptives are weighted by country-specific survey sample weights. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Troublemakers, Bystanders, and Pathbreakers 91 how often they access the news. For the former, I focus on those survey respondents who note that they are ‘somewhat’ or ‘very’ interested in public affairs. For the latter, an index is created that combines degree of access to the news via various sources, including radio, television, newspaper, internet, and social media. The index runs from 0 to 4 with the former referring to never accessing the news through any media outlet and the latter indicating daily access to the news. For simplicity, Table4.7 indicates the share of respondents who access the news ‘many times a week’ or on a daily basis, which is equivalent to 3 and 4 on the index. In addition to just being informed, there are more proactive modes of engagement. When disgruntled about public policy at either the national or local level, there are various courses of action available. One includes actively contacting the relevant authorities. To capture this, I create an index reflecting whether a respondent has contacted any of the following four authorities either a ‘few times’ or ‘often’ over the previous year: local government, member of parliament, an official of a government agency, and/or a political party.13 Another course of action is to get together with others to raise an issue, and Table4.7 focuses on the share of those who answered that they have done this at least once during the year preceding the survey. Central to this paper’s focus, the most extreme form of political participation is to be involved in a protest march or demonstration during the previous year.14 Table4.7 reveals that interest in public affairs appears to have increased over time among all age groups, with more than 50 per cent of young people in all youth groups expressing their interest. Despite technological advances over the last decade, there is not a massive change in the share of individuals accessing the news on a frequent basis, and access remains relatively low on average. This may reflect that internet and social media in particular continue to have very little penetration across the region due to electricity outages and other infrastructure constraints. There do, however, appear to be some life-cycle dynamics such that regardless of time period, the two oldest age groups are less likely to access the news. In terms of active participation, contacting an authority, or joining others to raise an issue are more common modalities of youth participation than protests. Moreover, while the former two modalities of participation appear to become more pronounced as individuals age, the latter is not especially the reserve of a particular age group, especially in the Round 6 surveys. Indeed, while for example, 21 per cent of all protesters fall into the 18–24 year-old group, only about 9 per cent of that age group claims to have actually participated in a protest. 13 For Malawi, the index does not include contacting local authorities since this option was not included in the country questionnaire. 14 Whether one voted in the last elections in his/her country is an obvious measure of political participation. Unfortunately, however, this question was not asked in Round 2 of Afrobarometer so temporal comparisons could not be made. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Table 4.7. Level of political engagement and participation by age group (percentages) Variable 2003 2014 18–24 25–29 30–34 35–44 45–54 55+ 18–24 25–29 30–34 35–44 45–54 55+ Interest in public affairs 42.0 39.4 40.1 42.0 43.7 43.5 55.9 59.4 59.7 62.2 62.0 59.3 Access news many times a week or everyday 29.3 18.4 13.9 20.0 10.8 7.8 30.5 24.4 16.6 18.3 6.8 3.5 Contact authority (once or few times) 31.5 38.3 40.8 42.6 43.4 41.3 22.0 31.1 35.6 38.9 39.9 39.4 Join others to raise an issue 46.5 51.2 53.0 55.1 54.8 54.3 33.4 39.4 42.9 46.8 49.8 49.3 Protest 16.6 15.7 14.6 14.2 12.4 9.4 8.7 9.8 9.5 9.4 8.6 6.3 Total observations 5,701 3,926 3,215 4,857 2,912 3,087 6,240 5,059 4,308 6,240 3,773 4,137 Source: Calculated from Afrobarometer, Rounds 2 and 6. All descriptives are weighted by cross-country survey sample weights. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Troublemakers, Bystanders, and Pathbreakers 93 Figure4.3 further disaggregates these protest trends with a specific focus on rural and urban respondents and their corresponding birth cohorts. Protest is an activity that a higher share of urban than rural respondents claim to engage in, which is true across all age groups. In addition, the figure suggests that while older birth cohorts no longer protest as much as they did in the past, their levels of protest are not necessarily significantly below those of today’s younger age groups. This implies that some lingering generational effects persist, especially for urban members of the 1974–8 generation, which came of age during the onset of ‘third wave’ democratic transitions in the early to mid-1990s, often due to protests. Although protest remains relatively low on average among the various age groups in the Round 6 surveys, Table4.8 examines country-specific dynamics with a focus on urban areas where protest is generally higher. Malian youth between 18–24 were particularly mobilized, followed by their counterparts in Senegal. While South Africa also has relatively high levels of youth protest, it was concentrated more among 25–34 year olds. In Nigeria, protest activity was actually highest among older age groups, perhaps reflecting the concerns with civil service salaries and fuel subsidies that were more specific to those older age groups. 4.6 Drivers of Youth Protest In order to determine in a more rigorous manner what individual-level characteristics are associated with protest activity and whether these have changed over time, I employ separate logit, country-level fixed effects models to both the Round2 and 0 2 4 6 8 10 12 14 16 18 20 1990–1996 1985–1989 1979–1985 1978–1974 1969–1973 1959–1968 1949–1958 1948 Share of respondents (%) 2003 (Total) 2014 (Total) 2003 (Rural) 2014 (Rural) 2003 (Urban) 2014 (Urban) Figure 4.3. Participation in protest or demonstration by approximate birth cohort OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Table 4.8. Protest by age groups in urban areas from Round 6 Afrobarometer (percentages) Country Age groups Total Total urban observations 18–24 25–29 30–34 35–44 45–54 55+ Botswana 10.34 7.77 8.03 13.11 11.31 6.46 9.54 760 Cape Verde 10.23 18.7 11.01 14.24 19.63 11.92 13.86 776 Ghana 9.28 7.83 8.14 8.84 4.84 4.82 7.54 1,304 Kenya 9.59 10.18 9.47 12.37 1.02 1.84 8.96 872 Lesotho 7.07 7.2 6.09 4.55 0.0 4.87 5.13 360 Malawi 8.56 16.34 9.35 10.57 0.0 5.11 9.46 448 Mali 34.26 20.36 27.54 23.2 18.84 7.88 23.48 304 Mozambique 9.68 9.51 7.79 12.26 4.24 9.4 9.33 840 Namibia 8.79 12.77 14.72 14.36 13.49 14.98 12.56 584 Nigeria 11.75 11.71 16.09 16.78 22.79 19.02 14.68 1,048 Senegal 19.36 16.98 20.85 19.23 11.79 2.72 15.63 592 South Africa 13.27 20.63 19.78 19.14 14.59 11.15 16.61 1,627 Tanzania 1.23 5.65 2.71 7.1 6.73 4.53 4.83 836 Uganda 6.46 13.78 9.6 4.21 7.11 3.06 7.72 448 Zambia 5.87 11.36 1.51 2.72 6.71 3.4 5.31 520 Zimbabwe 2.65 1.83 2.54 4.07 3.53 1.22 2.76 888 OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Troublemakers, Bystanders, and Pathbreakers 95 Round 6 data to estimate whether one has engaged in a protest or demonstration. Three basic demographic control variables are included, which are age, gender, and whether one lives in a rural area. In addition, consideration is given to testing four alternative hypotheses that may underlie one’s propensity to protest. First, relative deprivation is considered a strong incentive to protest. Grievance theories of protest long ago stressed that poverty, unemployment, and inequality were likely to provide the substantive incentive to engage in non-formal modes of political participation (Gurr1970). The primary driver is the psychological and emotional stress created by economic deprivation, which prompts individuals to challenge the prevailing political order (Buechler2004, Opp1988). More recent studies of the relationship between economic crises and protests in Greece (Rüdig and Karyotis2013) and Iceland (Bernburg2015) have again uncovered that perceived economic deprivation is an important predicator of who goes to the streets. Key variables to test this hypothesis include the LPI, relative perceptions of socioeconomic status vis-à-vis other compatriots, and those who are not employed but looking. Secondly, and in contrast to the deprivation hypothesis, the resource mobilization school has placed greater emphasis on the need for resources in order to organize protests, including skilled and educated protest leaders (McCarthy and Zald1977, McVeigh and Smith1999). Furthermore, individuals with a broad social network, who are well-informed with the capacity to process complex political information and recognize its consequences, and who have a greater sense of civic responsibility may be more likely to protest (Dalton2008b, Rosenstone and Hansen1993).15 Relevant variables to test this hypothesis include respondents’ education levels, access to the news, interest in public affairs, contacting an authority, and joining with others to raise an issue. Thirdly, partisanship and trust in government may mitigate how one views his/ her economic circumstances. Those who are close to a party, and particularly those who are close to the ruling party, may be less critical of the party and avoid pursuing activities that question the government’s legitimacy. Relatedly, those with higher levels of trust in the government are less likely to resort to extra-institutional modalities, such as protest, in order to convey their preferences. By contrast, those with lower levels of trust are more likely to view the status quo as unresponsive and unrepresentative and to challenge political elites (Gamson1968, Inglehart and Catterberg2002). Labelled ‘disaffected radicalism’ by Norris, Walgrave, and 15 In addition, some argue that those who are more involved in civic associations or who are embedded in social networks, such as those found through trade unions or religious organizations, have a greater propensity to be mobilized for protest activities (Putnam1993, Verba, Schlozman, and Brady1995). While Afrobarometer does ask about membership in such associations, the associations included varied substantially across the two rounds. In addition, the question was not asked in Zimbabwe. As such, the variable is not included in the analyses here. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi 96 Danielle Resnick Aelst (2005), this view has more recently been challenged by those who argue that protest behaviour may actually be driven more by trust in government if protesters believe their voices will be heard by politicians and their concerns will be addressed accordingly (Dubrow, Slomczynski, and Tomescu-Dubrow2008). Including the measure of incumbent closeness and trust in political institutions allows for testing these arguments. Finally, positions on policy could very likely be the proximate driver of protest activities. Indeed, individuals are not likely to go into the streets to protest if there is not a policy lever that they intend to change. Bermeo and Bartels (2014) stress this point in their work on developed countries recently facing economic crisis whereby protesters were more likely to oppose austerity policies rather than to feel economically deprived per se. Consequently, the analysis here draws on Table4.4 and includes a dummy variable capturing whether a respondent feels that unemployment is the most important policy priority for their country and whether s/he believes that that the government is handling job creation well. The findings in Table4.9 show the results for the full sample for both the Round 2 and Round 6 surveys (models 1 and 4) as well as for sub-groups of the sample. For ease of explication and presentation, the three separate youth and nonyouth groups were collapsed into one youth group (models 2 and 5) and one non-youth group (models 3 and 6). The results suggest that there are common profiles of protesters across age groups and survey rounds as well as important period and cohort effects. In terms of demographics, age demonstrates a strong effect in general, indicating that the likelihood of protest is lower as one becomes older, and this is even true among those in the non-youth group. In other words, there are indeed some strong life-cycle effects to protest. By contrast, there is no clear pattern to other demographic variables. Despite the patterns from the descriptive statistics, rural youth are not more likely to protest than their urban counterparts over time at a statistically significant level. However, today’s older rural residents are significantly less likely to protest. Similarly, young women may have been more likely to stay away from protest ten years ago but this pattern is insignificant at the time of the Round 6 surveys. These findings counter the perception that protests in Africa are purely the reserve of young men. More broadly, the findings suggest that the deprivation and resource mobilization hypotheses are equally relevant in the African context. Particularly, protest likelihood is highest among those who are worse off on the LPI, and this trend persists across the two rounds of surveys. At the same time, those with higher levels of education, who are intrinsically interested in public affairs, who frequently access the news, and who engage in other forms of participation, such as contacting a government authority or joining others to raise an issue, demonstrate a correlation OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Table 4.9. Logit analysis of protest likelihood across survey rounds and youth/non-youth age groups Independent variables Round 2 survey (2003) Round 6 survey (2014) Full sample 18–34 35 and older Full sample 18–34 35 and older (1) (2) (3) (4) (5) (6) Age 0.0177*** (0.00199) –0.0283*** (0.00660) −0.0139*** (0.00370) 0.00917*** (0.00193) –0.0177* (0.00711) –0.0105** (0.00346) Rural 0.123* (0.0552) –0.0915 (0.0710) –0.151 (0.0888) 0.162** (0.0535) –0.0809 (0.0720) –0.231** (0.0805) Female 0.121* (0.0483) –0.133* (0.0613) −0.104 (0.0794) 0.00420 (0.0486) −0.0611 (0.0648) 0.0544 (0.0739) Lived poverty index 0.167*** (0.0300) 0.158*** (0.0394) 0.190*** (0.0466) 0.210*** (0.0302) 0.231*** (0.0408) 0.199*** (0.0452) Relative living conditions 0.0411 (0.0296) –0.0356 (0.0378) -0.0462 (0.0481) 0.0722* (0.0306) –0.0733 (0.0412) –0.0679 (0.0460) Not employed and looking –0.0606 (0.0547) –0.0750 (0.0660) –0.0236 (0.0987) 0.223*** (0.0632) 0.338*** (0.0834) 0.0835 (0.0987) Education level 0.0738*** (0.0184) 0.118*** (0.0250) 0.0334 (0.0280) 0.0891*** (0.0190) 0.0825** (0.0271) 0.0962*** (0.0273) News access index 0.138*** (0.0272) 0.136*** (0.0357) 0.154*** (0.0427) 0.135*** (0.0285) 0.174*** (0.0370) 0.0909* (0.0457) Interest in public affairs 0.178*** (0.0490) 0.216*** (0.0628) 0.116 (0.0789) 0.289*** (0.0534) 0.290*** (0.0709) 0.298*** (0.0821) Contacted government authority 0.338*** (0.0342) 0.426*** (0.0453) 0.243*** (0.0526) 0.722*** (0.0339) 0.772*** (0.0470) 0.676*** (0.0497) Joined others to raise an issue 0.536*** (0.0207) 0.528*** (0.0265) 0.558*** (0.0339) 0.242*** (0.0214) 0.250*** (0.0295) 0.239*** (0.0314) Close to incumbent party 0.0688 (0.0545) 0.114 (0.0699) –0.00934 (0.0882) 0.0543 (0.0560) 0.0156 (0.0780) 0.0811 (0.0816) Continued OUP CORRECTED PROOF – FINAL, 29/10/19, SPi 104 Danielle Resnick Pew Research Center. 2010. Millennials: Confident. Connected. Open to change. Washington, DC, U.S.A.: Pew Research Center. Putnam, R. 1993. Making democracy work: Civic traditions in modern Italy. Princeton, NJ, U.S.A.: Princeton University Press. Raleigh, C. 2015. Urban violence patterns across African states. International Studies Review 17: 90–106. Resnick, D., and J. Thurlow. 2015. Introduction: African youth at a crossroads. In African youth and the persistence of marginalization: Employment, politics, and prospects for change, ed. D.Resnick, and J.Thurlow. New York, N.Y, U.S.A.: Routledge. Resnick, D., and J. Thurlow. 2016. The political economy of Zambia’s recovery: Structural change without transformation? In Structural change, fundamentals, and growth, ed. M. McMillan, D. Rodrik, and C. Sepúlveda. Washington, DC, U.S.A.: IFPRI. Rosenstone, S.J., and J.M.Hansen. 1993. Mobilization, participation, and democracy in America. New York, N.Y, U.S.A.: Palgrave Macmillan. Rüdig, W., and G.Karyotis. 2013. Who protests in Greece? Mass opposition to austerity. British Journal of Political Science 44 (3): 487–513. Ryder, N.B. 1965. The cohort as a concept in the study of social change. American Sociological Review 30 (6): 843–61. Sall, I. 2015. Trust in political institutions in Senegal: Why did it drop? Afrobarometer Policy Paper No. 24. http://afrobarometer.org/sites/default/files/publications/Policy%20 papers/ab_r6_policypaperno24_trust_in_political_institutions.pdf. Sall, M. 2012. The REVA Plan in Senegal: Does modern farming change minds of young people about agriculture? Presented at the Young People, Farming and Food Conference, March 19–21, Accra, Ghana. Scholz, J., and M.Lubell. 1998. Trust and taxpaying: Testing the heuristic approach tocollective action. American Journal of Political Science 42 (2): 398–417. Smets, K., and A. Neundorf. 2014. The hierarchies of age-period-cohort research: Political context and the development of generational turnout patterns. Electoral Studies 33: 41–51. Stasavage, D. 2005. Democracy and education spending in Africa. American Journal of Political Science 49 (2): 343–58. Sumberg, J., N. A. Anyidoho, M. Chasukwa, B. Chinsinga, J. Leavy, G. Tadele, S.Whitfield, and J.Yaro. 2015. Young people, agriculture and employment in rural Africa. In African youth and the persistence of marginalization: Employment, politics, and prospects for change, ed. D.Resnick, and J.Thurlow. London, U.K.: Routledge. Tufekci, Z., and C.Wilson. 2012. Social media and the decision to participate in pol it ical protest: Observations from Tahrir Square. Journal of Communication 62 (2): 363–79. Urdal, H. 2006. A clash of generations? Youth bulges and political violence. International Studies Quarterly 50 (3): 607–29. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Troublemakers, Bystanders, and Pathbreakers 105 Valenzuela, S., A.Arriagada, and A.Scherman. 2012. The social media basis of youth protest behavior: The case of Chile. Journal of Communication 62 (2): 299–314. Valenzuela, S., A.Arriagada, and A.Scherman. 2014. Facebook, twitter, and youth engagement: A quasi-experimental study of social media use and protest behavior using propensity score matching. International Journal of Communication 8: 2046–70. Verba, S., K.L.Schlozman, and H.Brady. 1995. Voice and equality: Civic voluntarism in American politics. Cambridge, MA, U.S.A.: Harvard University Press. Wattenberg, M. 2008. Is voting for young people? New York, N.Y, U.S.A.: Pearson Longman. Watts, M. 2009. Oil, development, and the politics of the bottom billion. Macalester International 24 (1): 1. Wonacott, P. 2012. Youth protests shake politics across Africa—Angola’s long-serving president, though seen as secure in re-election, is one of many facing pressure in Arab Spring’s wake. Wall Street Journal 30 August: A16. World Bank. 2016. Africa’s pulse: An analysis of issues shaping Africa’s economic future. Washington, DC, U.S.A.: World Bank. Yang, Y., and K.C.Land. 2013. Age—period—cohort analysis: New models, methods, and empirical applications. Boca Raton, FL, U.S.A.: CRC Press, Taylor & Francis Group. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi OUP CORRECTED PROOF– FINAL, 29/10/19, SPi PART II CASE STUDIES ON THE ROLE OF YOUTH EMPLOYMENT IN STRUCTURAL TRANSFORMATION OUP CORRECTED PROOF– FINAL, 29/10/19, SPi OUP CORRECTED PROOF– FINAL, 29/10/19, SPi Emily Schmidt and Firew Bekele Woldeyes, Rural Youth and Employment in Ethiopia In: Youth and Jobs in Rural Africa: Beyond Stylized Facts. Edited by: Valerie Mueller and James Thurlow, Oxford University Press (2019). ©International Food Policy Research Institute. DOI: 10.1093/oso/9780198848059.003.0005 5 Rural Youth and Employment in Ethiopia Emily Schmidt and Firew Bekele Woldeyes 5.1 Introduction The economic growth literature argues that as an economy grows, the location and structure of labour transitions from primarily rural, agriculture-focused activities to more urbanized activities in the industry and service sectors. This structural transformation improves the livelihood of those who earn higher wages outside of agriculture, as well as increases land to labour ratios of those who remain in agriculture. Increases in household income (via diversified labour portfolios) often provide capital to spur innovation and agricultural productivity growth in rural areas. Over the last few decades, Ethiopia’s economic development strategy, the Agricultural Development Led Industrialization (ADLI) strategy, aimed to increase agricultural productivity, and in doing so, encouraged labour diversification via the development of rural nonfarm activities. This mode of development is supported by a large body of research literature which suggests that growth in the rural nonfarm sector is driven by agricultural productivity growth (Haggblade, Hazell, and Reardon2002, Haggblade, Hazell, and Dorosh2006, Mellor1976). Given Ethiopia’s focus on ADLI, agricultural production has increased substantially and the country has experienced impressive economic growth over the last decade of approximately 11 per cent per year. However, macroeconomic trends suggest that Ethiopia’s economy remains at a very early stage in its structural transformation. A puzzle presents itself as to how such growth can be maintained given that Ethiopia is one of the least urbanized countries in Africa south of the Sahara (84 per cent of the total population lives in rural areas) with ap proxi mate ly three quarters of the population engaged in agricultural activities (FDRE Population Census Commission2008, Central Statistical Agency2013). From a policy point of view, understanding how youth can take advantage of employment opportunities, both in the agriculture and non-agriculture sectors, will inform future economic growth potential in years to come. Slow urbanization paired with vibrant economic growth suggests that rural youth will remain an important component of the agricultural labour force, while OUP CORRECTED PROOF– FINAL, 29/10/19, SPi 110 Emily Schmidt and Firew Bekele Woldeyes also seeking to diversify into non-agricultural, higher-value labour opportunities. Within the agricultural sector, transformation includes moving from low-value cereal production, which is characteristic of current Ethiopian agricultural production patterns, to high-value crops, such as fruit and vegetables. Rural youth may seek to modernize agricultural practices and utilize new technologies to enhance agricultural growth in the medium term. Regarding the overall economic landscape, as structural transformation progresses in Ethiopia, youth may drive labour diversification trends from predominantly rural agricultural ac tiv ities to more urban focused manufacturing and service sector activities. This chapter examines current trends in labour diversification in Ethiopia, focusing on youth employment activities, and explores the structure of livelihood decisions given underlying agricultural endowments. Although the majority of rural youth work exclusively on their own family farm, it is not clear that focusing on agriculture is a strategy that will provide a sufficient livelihood for future generations. Recent data collected by the Ethiopia Socioeconomic Survey (ESS) suggest that youth in particular may have less access to important agricultural assets than do their elders. In response to these constraints, one might expect youth to implement more intensive farming. However, the same data show that households headed by youth are not more likely to use agricultural production en han cing technologies than are mature-headed farming households. Given that youth face constraints in the agricultural sector, we examine youth nonfarm labour engagement in rural and small town areas. We find that youth (those aged 25 to 34 years) have a greater probability of working in nonfarm enterprises (NFE) compared to mature individuals (ages 35–64). The majority ofindividuals working in nonfarm employment are engaged in small-scale trade activities, such as street and market vending, while there exists limited demand for more skilled labour in the construction and manufacturing sectors. Our analysis suggests that push factors are at play with regards to nonfarm diversification, whereby those that live in areas with less favourable agricultural potential, who possess few assets, like livestock, and have less access to agricultural credit are more likely to seek off-farm work. This chapter provides evidence that youth are currently driving the limited structural changes observed in employment patterns in Ethiopia’s economy via employment diversification into nonfarm enterprises. However, low demand for higher-skilled labour, including in the rural nonfarm sector, remains a major obstacle to achieving structural transformation in the near to medium term. The remainder of the chapter is organized as follows: the second section reports employment trends in Ethiopia with a focus on youth activities in rural, small town, and urban areas. The third explores the difference in agricultural production practices between mature-headed households and youth-headed households. Thefourth section focuses on youth nonfarm activities using a multinomial logit OUP CORRECTED PROOF– FINAL, 29/10/19, SPi Rural Youth and Employment in Ethiopia 111 model to explore correlates of youth decisions to work in the nonfarm sector. The fifth discusses results of the multinomial logit, and then the chapter concludes. 5.2 Employment in Ethiopia 5.2.1 Employment Trends We utilize two nationally representative survey datasets to explore overall labour activity in Ethiopia: the National Labour Force Surveys (NLFS) of Ethiopia and the Ethiopia Socioeconomic Survey (ESS). Although the NLFS data provide nationally representative data on labour trends in the country, it restricts data collection to ‘main occupation’. Thus, we are unable to assess the portfolio of economic activities individuals pursue, in particular rural nonfarm work by members of farming households. In order to provide a more comprehensive analysis of labour participation, we complement the NLFS evaluation with a detailed labour decomposition in rural, small town, and large cities using the ESS. The ESS requests that each individual household records the amount of time worked on agriculture (own-farm), wage, and nonfarm enterprises over a 12-month period. Given that 82 per cent of Ethiopia’s population reside in rural areas, where a majority of individuals define their primary occupation as agriculture, the ESS supports a more diversified analysis of individual work portfolios. 5.2.2 National Labour Force Surveys, 2005 and 2013 When analysing the NLFS data, we adopt the Central Statistical Agency (CSA) definition of the active labour force. Labour force participants include individuals who are at least 10 years old (Ethiopia does not limit the labour force to retirement at 64 years old). Within the working age population, individuals who are not engaged in work and would not be available to take up work if it was offered, as well as individuals who are students, handicapped, or have long-term illnesses, are not considered a part of the active labour force.1 Although we follow CSA definitions for active labour force, we adjust the 2013 definition of economically active to provide a more accurate comparison with the 2005 NLFS data. In 2005, data collected on occupation and industry concentrated on individuals who worked at least four hours per day, while in 2013, individuals 1 The CSA does not include ‘seeking work’ as a criterion for being considered economically active. This is due to local conditions of inadequate labour absorption, a large share of labour force being self-employed, and inconsistencies in time accounting of individuals who work in the informal labour market. OUP CORRECTED PROOF– FINAL, 29/10/19, SPi 112 Emily Schmidt and Firew Bekele Woldeyes who reported working at least one hour were provided an industry classification. This modification in the questionnaire resulted in a large share of household unpaid family labour (firewood and water collectors) being classified as working in the household services sector in 2013. We adjust for this discrepancy by reclassifying individuals who stated their main occupation as ‘wood and water collection’ into ‘not in the labour force’ in order to provide comparable estimates of employment shares within sectors between the two survey years. We report labour shares using NLFS 2013 official definitions (including wood and water collectors in the economically active population) as well as our adjusted 2013 statistics.2 Between 2005 and 2013, while official definitions suggest a greater transition out of agriculture from 80 to 73 per cent of economically active population, we find that, after adjusting the data for water and firewood collectors, the share of people working in agriculture decreased by only 3 percentage points from 80 to 77 per cent over this period (Table5.1). Overall employment shares in the services sector also reflect the reallocation of water and firewood collectors. Official statistics reported the overall services sector to encompass 20 per cent of the eco nomic al ly active population in 2013, of which private household work increased from 6 to 36 per cent of service employment. Adjusting for the discrepancies between the 2005 and 2013 surveys in the definition of those employed, we find that the service sector employed 16 per cent of the economically active population in 2013. Finally, the industry sector has not experienced significant growth over the last decade in terms of job creation, with employment shares increasing by only about one percentage point between 2005 and 2013. Although the government of Ethiopia has made significant investments in education with an emphasis on increasing access to secondary education 2 In 2013, 88 per cent of individuals who reported water and wood collecting as their primary occupation reported that this activity was classified as unpaid family worker. The majority of these workers were female (89 per cent) and rural (94 per cent). This category was not present or accounted for in 2005. For more information on these individuals see Appendix5.A1. Table 5.1. Employment shares by industry of the economically active population (ages 10 years and older), per cent Industry 2005 2013 unadjusted 2013 adjusted Agriculture 80.2 72.7 76.6 Industry 6.7 7.3 7.7 Services 13.2 20.0 15.7 Private households 6.0 36.4 10.2 Source: National Labour Force Survey (2005, 2013). ‘2013 unadjusted’ represents the national CSA-based definition of sectors. ‘2013 adjusted’ reclassifies water and firewood collectors as ‘not economically active’—this sort of work is not taken into account in the figures in this column. OUP CORRECTED PROOF– FINAL, 29/10/19, SPi Rural Youth and Employment in Ethiopia 113 op por tun ities, non-agricultural workers are predominantly engaged in low-skill sectors. Sales workers make up 29 per cent of non-agricultural work, of which street and local market vendors comprise 42 per cent (Table5.2). Formal shopkeepers and informal home-brewed alcohol sellers comprise almost equivalent shares of 22 and 21 per cent of sales workers, respectively. These employment trends suggest a mode of development that is moving, albeit slowly, towards a service sector focused economy. However, the specific service activities that individuals are engaged in reflect a low level of development with limited labour demand. In order to better understand employment activities within the Ethiopian economy, we disaggregate employment numbers by geographic area (rural, small town, and urban areas) and by age group. Focusing on youth, the data suggest that rural youth are primarily engaged in agriculture, while a greater share of youth living in ‘other urban’ locations and in large cities are engaged in non-agricultural work. As per the CSA definition of ‘other urban’, we can assume that these centres represent secondary cities that are urban centres with populations of less than 100,000 people and are not considered regional capitals. When comparing the percentage share of individuals working in agriculture between rural areas and these secondary cities, diversification is primarily occurring in the secondary cities whereby 22 and 12 per cent of youth aged 15 to 24 years and 25 to 34 years, respectively, report their primary occupation is in agriculture (Table5.3). However, it is important to note that ‘other urban’ represents only 12 per cent of the economically active population, both overall and for youth. Evaluating employment transitions between 2005 and 2013 disaggregated by age and spatial domain, the NLFS results presented in Table5.3 suggest very little Table 5.2. Share of non-agricultural employment in Ethiopia by occupational group, 2013, present Occupational group Share of non-agricultural employment Sales workers 30.2 Street and market salespersons 43.6 Shop salespersons 22.1 Alcohol sales 20.6 Other sales 13.7 Construction and mining 10.6 Food processing, wood and garment craft 7.6 Refuse workers 7.0 Teacher 6.5 Personal service worker 5.9 Other 32.2 Source: National Labour Force Survey (2013). OUP CORRECTED PROOF– FINAL, 29/10/19, SPi 120 Emily Schmidt and Firew Bekele Woldeyes individuals from reaching their full working potential. Those who are solely engaged in own-farm activities, 78 per cent of the overall working population, report being economically active for only about half of the year. Given limited nonfarm labour demand, as well as the large share of rural youth that work exclusively in agriculture, we now assess to what degree youth are leading any agricultural transformation processes in Ethiopia, particularly those that involve specialization in high-value crops or the utilization of modern technologies. The goal of education policy in Ethiopia has been, in part, to produce educated farmers who would then be able to effectively adopt new agricultural technologies (MOE2005). These objectives continue to underpin the national education policy. This would suggest that as rural youth create their own, independent households and acquire their own agricultural land, they may seek solutions to increase agricultural productivity and overall welfare via agricultural intensification, diversification, and modernization. Table5.7 compares the characteristics of agricultural households located in rural and small towns (less than 10,000 people), disaggregated by the age of the household head.7 Several differences stand out. First, youth-headed households have access to significantly less agricultural land compared to mature-headed households. Youth-headed households own and operate approximately 0.8 and 1.4 hectares, respectively, compared to mature-headed households that own and operate 1.5 and 1.7 hectares, respectively. Limited data suggest that the young youth-headed households (aged 15–24 years) have greater difficulty accessing land than the experienced youth-headed households (between the ages of 25 and 34 years), however this result should be read with caution given the small sample size of young youth-headed households. Landlessness is also greater among youth-headed households. For example, 7 per cent of mature-headed households living in rural and small town areas are landless, compared to 14 per cent of youth-headed households. Similarly, the share of landlessness among the youngest households (15–24 years old) reaches 21 per cent, while 13 per cent of experienced youth-headed households are landless (Table5.7). This follows recent research by Bezu and Holden (2014) who found that youth in the rural south of Ethiopia have limited access to agricultural land due to land scarcity and land market restrictions. Headey, Dereje, and Taffesse (2014) also report declining farm sizes over time, with younger rural households facing larger constraints in obtaining agricultural land. Finally, youth are not more likely to implement agricultural enhancing technologies (improved seed, cash crop production, and row planting) compared to 7 Research shows that a variety of factors affect household uptake of agricultural technologies. Extensive literature has analysed specific issues including: physical and human capital endowments (Pender and Fafchamps2006); access to agricultural extension (Abrar, Morrissey, and Rayner2004); supply of seeds (Dercon and Hill2009); heterogeneity of fertilizer success (Suri2011); risks of negative shocks (Dercon and Christiaensen2011); and access to credit (Duflo, Kremer, and Robinson 2011). OUP CORRECTED PROOF– FINAL, 29/10/19, SPi Rural Youth and Employment in Ethiopia 121 mature-headed households. However, compared to mature households, youth use more technologies that are labour-reducing, such as herbicides and tractors. This is in line with recent work by Bachewe et al. (2015) and Minten et al. (2013) who found that substitution of labour with labour-saving modern inputs, in particular herbicides, is increasing in lieu of time spent on weeding. This may be due to the smaller household size of youth-headed households which creates a labourconstrained environment in which such households will seek technologies to decrease labour demands in agricultural work. Overall, these figures suggest that agriculture may not be the optimal or first choice of employment among youthheaded households, given the current environment. Table 5.7. Agricultural household-level characteristics in rural and small town areas, by age cohort of household head, means  MatureheadedHHs (35–64) YouthheadedHHs (15–34) Young youthheaded HHs (15–24) Experienced youth-headed HHs (25–34) Land characteristics     Operated area, ha 1.74 1.38 * 0.75 ** 1.46 Owned area, ha 1.49 0.82 *** 0.59 * 0.85 *** Landless, % 7.2 14.0 *** 20.6 *** 13.2 *** Good agricultural potential, % 26.4 24.1 16.5 ** 25.0 Agricultural inputs    Inorganic fertilizer, % 59.9 57.7 44.5 *** 59.3 Organic fertilizer, % 67.7 59.8 *** 53.0 *** 60.7 *** Irrigation, % 09.8 10.8 14.5 10.4 Herbicide, % 33.9 39.3 *** 33.9 40.0 *** Tractor, % 4.2 6.6 *** 08.9 ** 06.3 ** Improved seed, % 30.1 29.0 26.5 29.3 Row planting, % 50.0 45.1 *** 51.2 44.4 *** Grow cash crop, % a82.3 78.3 *** 75.6 ** 78.7 ** Receive agricultural credit,% 23.8 20.5 ** 08.4 *** 21.9 Receive agricultural extension, % 45.8 40.2 *** 26.2 *** 41.8 ** Household characteristics     Household size, number 6.18 4.85 *** 3.82 *** 4.97 *** Number of observations 2,752 1,024 135 889 Note: t-tests are relative to mature households; *** p < 0.01, ** p < 0.05, * p < 0.1. a Cash crops include beans, nuts, sesame and other seeds, spices, fruit, vegetables, coffee, chat, cotton, sugar cane, and tobacco. Source: Ethiopia Socioeconomic Survey (2013/14). OUP CORRECTED PROOF– FINAL, 29/10/19, SPi 122 Emily Schmidt and Firew Bekele Woldeyes 5.4 Correlates of Youth Engagement in Nonfarm Employment A rich literature has evaluated the determinants of nonfarm labour engagement including disaggregated analysis of individuals’ decisions to seek out skilled versus unskilled nonfarm labour opportunities (Reardon, Berdegue, and Escobar 2001, Winters et al.2009, Mduma and Wobst2005, Bezu, Holden, and Barrett 2009); market access and nonfarm participation (Fujita, Krugman, and Venables 1999, Renkow 2006, Henderson, Shalizi, and Venables 2001, Fafchamps and Shilpi 2003, Deichmann, Shilpi, and Vakis2009); and effects of income or wealth on nonfarm labour choices (Bezu, Barrett, and Holden 2012, de Janvry and Sadoulet 2001, Woldehanna and Oskam 2001, Dercon and Krishnan 1996). However, largely missing from the literature on Ethiopia is an in-depth evaluation of the transition of youth from employment on-farm into the nonfarm sectors. A recent report by the World Bank outlines the opportunities and challenges for youth employment in Africa and provides a comprehensive overview of potential growth sectors, including agriculture. However, the discussion in this overview is limited to country and regional levels (Filmer and Fox2014). Bezu and Holden (2014) evaluated the determinants of youth aspirations to pursue nonfarm employment in Ethiopia. However, they did not examine the experience of youth that already are in the nonfarm work force. This section addresses some of these knowledge gaps by evaluating the determinants of youth employment in the nonfarm sector. Appendix 5.A4 provides the average values for key variables used in the empirical analysis. The profiles of rural and small town workers in Ethiopia differ in terms of individual, household, and location characteristics. Youth between the ages of 25 and 34 years are generally more active in the nonfarm sector (wage or nonfarm enterprise), while youth between the ages of 15 and 24 tend to work more on own-farm labour. Those that diversify into wage labour activities have completed more schooling (36 per cent completed primary school). Primary school completion rates of about 10 per cent are approximately the same for in di vid uals that diversify into a nonfarm enterprise activity and those that work solely on own-farm activities, suggesting that nonfarm enterprise activities do not require a significantly different skill set or greater experience level than does own-farm work. Compared to own-farm workers, wage and nonfarm enterprise workers report higher annual expenditure per capita, which may be associated with higher potential profitability of off-farm work. In this analysis, nonfarm labour activities refer to any labour that is conducted off the own-family farm. We limit our sample to individuals living in rural or small town areas. We are interested in assessing workers that choose to diversify into nonfarm labour activities, meaning wage or nonfarm enterprise activities, in addition to working on their own-family farm in either planting or harvesting. OUP CORRECTED PROOF– FINAL, 29/10/19, SPi Rural Youth and Employment in Ethiopia 123 The sample is split into three categories: individuals who work solely on their own family farm (omitted); individuals who report working in a mix of own-family farm and NFE; and individuals who report working in a mix of own-family farm and wage work.8 Correlates with diversification are estimated using a multi nomial logit model: l og , π π αβ wi ni wr wi X      =+ where π wi is the odds of seeking off-farm work w, π ni is the odds of remaining on the family farm and working solely in agriculture, and parameter α wr is the baseline hazard of work in region r for the specific work type w. β w is a vector of parameter estimates. X is a vector that denotes the factors that influence labour choice. In order to take into account unobserved variables within districts that affect employment, such as access to infrastructure, information, or agroecological zone, standard errors are clustered at woreda level. Finally, the coefficients in a multinomial logit model are calculated in relation to a base outcome and thus are difficult to interpret directly. However, average marginal effects can be predicted, so we focus the discussion on the reported marginal effects. We set our base outcome as individuals who work exclusively on their own-farm. We estimate three models to assess correlations between youth and livelihood choice. The first model pools the rural and small town working samples ages 15 to 64 years old to test if youth (aged 15–34) are more likely to enter off-farm labour opportunities compared to mature individuals (aged 35–64). The second model is limited to youth aged 15–24 to evaluate how individual, household and location variables are correlated with nonfarm labour. The third model is limited to youth aged 25–34 and follows the same methodology of the second model to evaluate how more established youth are engaged in the labour market. We split youth categories assuming that young youth and experienced youth differ by the amount of work experience they have attained and the social network they have built, which may affect an individual’s ability to secure off-farm work. 5.4.1 Potential Determinants of Engaging in Off-farm Employment in Rural Ethiopia Diversification into nonfarm employment is shaped by a variety of conditioning factors. Individuals may be pushed from agricultural work into nonfarm activities 8 Estimation of the multinomial logit model assumes that probabilities of alternative choices are independent of each other, referred to as the Independence of Irrelevant Alternatives (IIA). In order to test for this independence, we use the Small-Hsiao test, and find we are unable to reject the IIA assumption for the multinomial logit model presented in this analysis (Small and Hsaio1985). OUP CORRECTED PROOF– FINAL, 29/10/19, SPi 124 Emily Schmidt and Firew Bekele Woldeyes in order to seek out sufficient sources of income, alternatively individuals may be pulled into nonfarm activities given higher returns to labour and capital compared to agriculture. 9 Lucas (2015) focuses on rural–urban migration issues and argues that these decisions are driven by differentials in opportunities across locations. Focusing specifically on the choice of diversifying labour portfolios, in di vid uals differ in their ability to take advantage of nonfarm opportunities based on their human, physical, and financial capital. For example, some individuals are more educated, with better access to savings for start-up capital and greater options for nonfarm work due to proximity to a market or transportation network. We include a variety of explanatory variables in our multinomial logit analysis in order to account for differences across individuals and households. At the individual level, we include a variable that disaggregates odds of employment by age in the regression that pools all age groups: youth aged 15–24, youth aged 25–34, and mature aged 35–64 (omitted category). Including these variables in a multivariation regression allows us to adjust for inherent differences across age groups which may mask the interpretation of the descriptive statistics discussed above. These age variables attempt to capture experience level and potential life-cycle effects, as well as to explicitly evaluate youth participation in nonfarm activities. We also include whether or not the individual is a household head, female, or married. If an individual is the household head, she or he may be more inclined to stay working on the farm in order to insure sufficient agricultural output. In addition, Ethiopia’s land tenure system requires residency on the farm to maintain usufruct rights to farmland which may create greater disincentives for household heads to seek alternative employment. Education—measured by whether an individual completed primary school—is also an important factor, given that it improves the value of labour, raises the opportunity costs for an educated individual to stay at home and engage in lower paying agricultural work, and potentially enhances the individual’s social network to facilitate access to nonfarm jobs. At the household level, we include a variety of variables that take into account household assets. For example, owning a relatively large agricultural land area may indicate better farming potential and food self-sufficiency, which may incentivize individuals to remain in agriculture. Alternatively, larger land holdings may be associated with higher crop incomes, which could provide start-up capital for work in the nonfarm sector.10 Due to restrictions on land ownership in Ethiopia, land rental markets are very active, thus we include both total agricultural land owned and total agricultural land operated to account for these factors. Given that 9 Moretti (2004) and Ciccone and Peri (2006) examine pull factors of migration and their links to agglomeration economies. 10 See Reardon et al. (2007) and Bezu and Barrett (2012) for a greater discussion on land holdings and nonfarm labour diversification. OUP CORRECTED PROOF– FINAL, 29/10/19, SPi Rural Youth and Employment in Ethiopia 125 we are limited to cross-sectional data, it is possible that we introduce simultaneity bias in our regression framework by including household assets. For example, not only does land holding size potentially affect diversification but diversification could affect land holding size. This is particularly accute when using contemporraneous explanatory variables. Thus, we discuss the results in terms of correlates of diversification rather than addressing causation. We also include livestock ownership in the form of Tropical Livestock Units owned by the household, per capita expenditure of the household, and whether a household is located in an area with good agricultural potential. We hypothesize that youth who have access to land with good agricultural potential are less likely to seek nonfarm employment. In addition, we include whether a household has experienced a flood or drought during the last year in order to take into account potential fluctuations in agricultural productivity. Such fluctuations may incentivize individuals to seek other forms of employment as a means of insurance against agricultural uncertainty. In addition to physical endowments at the household level, individuals coming from larger households with greater potential labour resources may exhibit a greater probability of working in the nonfarm sector because their labour would not be as critical for agricultural production within such households. In order to account for differences in female and male labour roles in rural Ethiopia (for example, it is rare for females to cultivate land in Ethiopia—see Deininger, Ali, and Tekie2008), we include the number of working age (ages 15–64) females and the number working age males within the household, as well as total household size. Household variables are included to differentiate between households that have received agricultural credit or agricultural extension. These variables represent incentive factors to stay working on-farm because they are targeted to augment agricultural productivity.11 Finally, distance to a market or trafficked road captures a household’s locational potential for nonfarm labour opportunities, as well as assessing the effect on transaction costs, and thus, an individual’s willingness to seek nonfarm labour. Job search costs would be lower for those that live closer to markets or key transportation corridors, while at the same time they may be better informed of potential job opportunities. 11 Recent work evaluated credit via microfinance programmes aimed at nonfarm activities and found mixed results with regards to such credit inducing greater engagement in non-agricultural income earning opportunities. Hagos (2003) found a positive effect of microfinance credit programmes on income level changes derived from self-employment. However, no effect was found on participation in wage employment. Bezu and Holden (2014) reported that access to savings and credit are significant factors for transitioning into high-return rural nonfarm activities. Tarozzi, Desai, and Johnson (2015) evaluated access to microfinance credit in Amhara and Oromiya on a variety of outcomes and found no significant effects on nonfarm enterprise creation. OUP CORRECTED PROOF– FINAL, 29/10/19, SPi 126 Emily Schmidt and Firew Bekele Woldeyes 5.4.2 Results and Discussion The coefficients in a multinomial logit model are calculated in relation to a base outcome and are difficult to interpret directly. However, average marginal effects can be predicted, so we focus the discussion on the reported marginal effects. Model 1 evaluates whether youth are more likely to diversify into wage or nonfarm enterprise opportunities in addition to working on own-family farm, by testing whether the coefficients on the age indicators are statistically different than zero. Analysis suggests that older youth (age 25–34) have a greater probability of diversifying into nonfarm enterprise activities compared to mature in di viduals, however this does not hold true for youth ages 15–24 (Table5.8). It may be that the younger cohort of youth have not built up sufficient work experience or developed an appropriate social network to successfully engage in a nonfarm enterprise. Although older youth are more engaged in nonfarm enterprise labour, wage labour is less accessible to Ethiopia’s youth. According to Model 1, youth (regardless of their age) are no more active in the wage labour market than are mature individuals (aged 35–64). Focusing specifically on youth (Models 2 and 3), those who are located in areas with good agricultural potential have a greater probability of diversifying into nonfarm enterprises, especially older youth aged 25–34 years old (Table 5.8, Model 3). This supports the findings of previous research that contended that local nonfarm income is greater in better agroclimatic areas, whereas migration is a more common strategy in unfavourable climatic areas (Reardon1997, Reardon et al.2007). Woldehanna and Oskam (2001) reported that households in Tigray during good production seasons prefer nonfarm enterprise work over wage employment, suggesting that a good production season gives farmers the financial capacity to start a nonfarm enterprise. Youth aged 15–24 that are located in good agricultural productivity areas are also more likely to mix farm and nonfarm enterprise work (Table 5.8, Model 2). Research conducted by Bezu and Holden (2014) found that a lack of access to land was driving youth migration from agriculture. Our analysis suggests that youth (age 25–34) with greater land ownership have a 3 per cent lower probability of diversifying into wage labour compared to working exclusively on their own family farm. Overall, assets and capital are associated with mature youth employment decisions. Greater ownership of land, livestock, and access to agricultural credit decrease the probability that mature youth diversify out of farming into wage labour activities (Table5.8, Model 3). We do not witness the effects of asset ownership in the young youth sample due to lack of variation among variables; young youth have less asset accumulation across all categories. Similar to agricultural endowments, agricultural shocks can have an effect on an individual’s choice to seek alternative income sources outside of agriculture. OUP CORRECTED PROOF– FINAL, 29/10/19, SPi Rural Youth and Employment in Ethiopia 127 Continued Table 5.8. Multinomial models of determinants of type of labour engagement for rural workers in Ethiopia, by age cohort Explanatory variables Model 1 Model 2 Model 3 Working population: age 15–64 Young youth working population: age 15–24 Experienced youth working population: age 25–34 Mix of own-farm and wage work Mix of own-farm and nonfarm enterprise Mix of own-farm and wage work Mix of own-farm and nonfarm enterprise Mix of own-farm and wage work Mix of own-farm and nonfarm enterprise Age 15–24, 0/1 –0.016 0.013 – – – –  (0.010) (0.019)     Age 25–34, 0/1 0.000 0.049 *** – – – –  (0.006) (0.015)     Household head, 0/1 0.002 0.037 ** 0.019 ** 0.030 –0.014 0.049 (0.009) (0.017) (0.009) (0.035) (0.016) (0.033) Female, 0/1 –0.036 *** 0.037 ** –0.004 0.034 –0.061 ** 0.079 *** (0.011) (0.016) (0.009) (0.021) (0.025) (0.028) Married, 0/1 0.005 0.011 –0.001 0.039 –0.006 0.016 (0.007) (0.016) (0.010) (0.024) (0.012) (0.034) Completed primary school, 0/1 0.045 *** 0.001 0.030 *** 0.018 0.055 *** –0.032 (0.008) (0.018) (0.010) (0.020) (0.011) (0.035) Adult (age 15 to 64) males in household, number –0.006 ** –0.022 *** 0.000 –0.009 –0.015 –0.003 (0.003) (0.007) (0.004) (0.010) (0.008) (0.021) Adult (age 15 to 64) females in household, number 0.004 0.002 0.002 0.001 0.008 0.018 (0.003) (0.008) (0.004) (0.011) (0.005) (0.019) Expenditure, ‘000 birr/ capita/year 0.005 *** 0.025 *** 0.003 *** 0.016 *** 0.006 *** 0.031 *** (0.001) (0.005) (0.001) (0.006) (0.002) (0.011) Agricultural area owned, ha –0.002 –0.006 0.007 –0.003 –0.026 ** 0.001 (0.002) (0.006) (0.011) (0.009) (0.011) (0.022) Agricultural area operated, ha 0.002 0.004 –0.017 0.006 0.004 –0.010 (0.002) (0.007) (0.014) (0.009) (0.002) (0.014) Receive agricultural extension, 0/1 –0.016 ** –0.003 –0.007 –0.012 –0.018 0.013 (0.006) (0.017) (0.010) (0.025) (0.011) (0.030) Receive agricultural credit, 0/1 –0.006 –0.021 0.001 –0.004 –0.068 ** –0.051 (0.009) (0.019) (0.009) (0.024) (0.027) (0.033) OUP CORRECTED PROOF– FINAL, 29/10/19, SPi 128 Emily Schmidt and Firew Bekele Woldeyes Table 5.8. Continued Explanatory variables Model 1 Model 2 Model 3 Working population: age 15–64 Young youth working population: age 15–24 Experienced youth working population: age 25–34 Mix of own-farm and wage work Mix of own-farm and nonfarm enterprise Mix of own-farm and wage work Mix of own-farm and nonfarm enterprise Mix of own-farm and wage work Mix of own-farm and nonfarm enterprise Livestock ownership, Tropical Livestock Units –0.002 * –0.004 0.000 –0.005 –0.006 * 0.000 (0.001) (0.003) (0.000) (0.003) (0.003) (0.006) Experienced drought, 0/1 0.005 ** –0.079 ** 0.005 –0.054 * 0.013 –0.095 (0.007) (0.036) (0.009) (0.050) (0.017) (0.070) Experienced flood, 0/1 –0.031 -0.005 –0.007 0.028 –0.366 *** –0.028 (0.020) (0.050) (0.012) (0.044) (0.069) (0.102) Good agricultural potential land, 0/1 –0.010 0.052 * –0.020 * 0.042 * –0.011 0.086 *** (0.009) (0.020) (0.011) (0.025) (0.017) (0.031) Distance to nearest market, km –0.001 0.001 –0.001 –0.001 –0.003 0.002 (0.001) (0.002) (0.001) (0.003) (0.002) (0.003) Distance to nearest major road, km 0.001 0.002 –0.004 0.005 0.004 0.000 (0.003) (0.006) (0.005) (0.007) (0.003) (0.010) Observations 7,567 2,526 1,754 Note: The base outcome for all three models are individuals that work exclusively on their own-farm. Standard errors in parentheses; *** p < 0.01, ** p < 0.05, * p < 0.1. Source: Ethiopia Socioeconomic Survey (2013/14). We find this particularly true for youth aged 25–34 years. In this case, those who experienced a flood during the last year had a 37 per cent less probability of expanding into a wage labour job, however this effect may be temporary and reflect a post-shock necessity of rehabilitating own agricultural land rather than an overall trend of off-farm labour activity. Alternatively, this analysis does not capture wage or nonfarm employment at alternative locations, thus a reduction in wage labour may coincide with an increase in wage labour outside of the sample woreda. Research on the impact of shocks on labour diversification suggests a greater propensity to diversify. Bezu and Barrett (2012) assessed employment transitions OUP CORRECTED PROOF– FINAL, 29/10/19, SPi Rural Youth and Employment in Ethiopia 129 out of agriculture between 2004 and 2009 and found that shocks that reduced agricultural income motivated individuals to seek out high-return rural nonfarm employment. Similar results of agricultural shocks increasing longer term nonagricultural earnings were reported by Porter (2012) using data from 1994–2004. While good agricultural potential and greater access to capital and assets (agricultural credit and livestock) decrease the likelihood of diversifying into nonfarm labour, distance to a market or road does not affect the probability of youth finding nonfarm employment. This may be due to several reasons. First, a large share of the rural population in this sample live relatively far from a market (on average about 55 km). Second, thin labour markets in small towns and rural areas may limit youth’s ability to take advantage of off-farm wage opportunities simply because there is not enough off-farm labour demand. These relationships suggest that in Ethiopia the rural and small town nonfarm sector is influenced primarily by push factors (lack of land, agricultural services, and assets) rather than driven by urban or small town labour demand. When comparing youth labour decisions to diversify into nonfarm employment, it becomes apparent that experienced youth (ages 25–34) have a greater likelihood of engaging in nonfarm work (in addition to own-family farm labour) compared to mature individuals (age 35–64). Although difficult to determine from cross-sectional data, the analysis presented in this chapter suggests that youth—in particular, experienced youth—may be driving the small share of labour diversification in rural and small towns. Moving forward, understanding if labour diversification occurs step-wise—meaning, individuals move from working exclusively on their own-farm activities to diversifying into nonfarm in add ition to own-farm activities, and then finally transitioning fully into nonfarm labour activities—will provide greater insight into Ethiopia’s likely economic trajectory over the next few decades. If this is the mode of labour transition within Ethiopia, we may be witnessing the initial transition of an economy moving towards greater structural transformation. 5.5 Conclusion Over the last several decades, Ethiopia has focused its public investments in economic growth according to its ADLI strategy. This led to large increases in agricultural output. Simultaneously, the country has experienced impressive economic growth at approximately 11 per cent per year during the last decade. Although these trends point to structural transformation as a major driver of economic growth, labour force survey data suggest that Ethiopia remains at a very early stage in its structural transformation. Whereas one would expect to see a transition out of agriculture into higher value nonfarm employment, we find OUP CORRECTED PROOF – FINAL, 29/10/19, SPi 232 Xinshen Diao, Eduardo Magalhaes, and Margaret McMillan Table 8.8. Continued Rural Other urban Dar es Salaam Total Farming 59.8 23.0 1.0 41.0 Others 5.8 4.0 2.2 4.8 Note: This table is prepared based on the question ‘what was your main occupation before you started this business?’ in the MSME survey, and a unique answer is provided by individual MSME owners. Number of the full sample obs. is 4,163, and 1,853 for the MSMEs with youth as owners. The sum of each column in the first panel for all MSMEs is 100. Source: Authors calculations using the MSME Survey 2010. other adult owners of MSMEs in the country’s largest city seem to give up a paid government job to open their own enterprises deserves more analysis to fully understand the motivation of small business entrepreneurs. The second question is: for what reason did you choose your line of business? Responses to this question are reported in Table8.9 by three broad sectors: manufac tur ing, trade services, and other services as well as by rural, other urban, and Dar es Salaam. Again, the responses that are statistically different between young and other adult owners are reported for these two subgroups separately in the table. In rural areas, half of all business owners say that the reason they chose their line of business is because they saw a market opportunity. This response is similar for manufacturing and trade services. The shares for the similar response are lower in other urban areas and in Dar es Salaam than in rural areas. The second most common reason for operating in a line of business in rural areas is that the owners’ capital could only finance that line of business, which is more common in urban areas. This seems to indicate that capital constraint for small business development is more binding in urban than in rural areas. The third most common reason for picking a line of business in rural areas was prior experience in that line of business, but apportionment for this reason are much lower than the two previous ones. Among young and other owners, the order of these three reasons are similar, which tells us that when coming to decide what kind of business, the reasons are similar regardless of young and other adult entrepreneurs in either rural or urban areas. The third question is: if you were offered a full-time salary-paying job, would you take it? Responses to this question are reported in Table8.10 and indicate that only 46.6 per cent of all small business owners would leave their current business for a full-time salaried position, but the share is higher in rural areas and other urban areas (47.8 and 48.6 per cent respectively) than in Dar es Salaam (37.7 per cent). Young adults show a stronger preference for a full-time salary job (instead of being a business owner). In rural areas 54 per cent of young owners would prefer a fulltime paying job, compared to 42 per cent for other adults. This is similar in urban OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Table 8.9. Reasons for business choice by broad sector in MSME survey (weighted, percentage) Rural enterprise Other urban enterprises Enterprise in Dar es Salaam Manufacturing Trade services Rural total Manufacturing Trade Services Other urban total Manufacturing Trade services Dar Es Salaam total All MSMEs I had previous experience inthis line 25.0 15.2 18.3 39.6 15.5 19.5 37.0 9.7 18.5 Friends/relatives are in this line 20.6 13.4 14.8 21.0 19.4 17.8 13.2 16.4 12.8 I saw a market opportunity 48.2 51.6 50.0 36.3 43.1 41.6 14.6 46.4 39.2 My capital could only finance this business 36.1 42.1 41.8 26.2 47.4 43.3 46.8 46.8 47.7 No apparent reason 2.9 6.0 4.5 4.3 2.9 4.2 3.2 5.2 4.7 I could start business gradually 0.0 0.1 0.1 0.0 0.1 0.5 0.0 1.0 0.6 Goods are easy to manufacture and sell 1.4 2.0 2.0 0.2 1.5 1.6 0.0 0.9 1.0 I just wanted to be near myhouse 0.8 1.0 0.9 0.0 1.8 1.2 0.0 0.0 0.4 I have been trained in it, Iam an expert 1.6 0.4 0.6 6.2 0.1 1.1 9.4 0.4 0.9 Goods are available 0.3 0.4 0.5 0.0 0.4 0.2 0.0 2.7 1.4 I perceived it to be profitable 1.3 1.6 1.7 0.0 2.6 1.8 0.0 0.2 0.1 I liked it 0.7 1.0 1.3 3.5 1.5 1.6 1.8 1.1 1.1 Business does not have many problems 1.4 0.6 0.7 0.0 0.3 0.3 0.0 1.6 1.2 Other 1.0 2.5 2.2 5.7 2.4 3.1 1.3 3.2 2.4 None 0.3 1.3 0.8 2.0 0.9 1.2 1.3 0.0 1.1 Continued OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Youth owners I had previous experience in this line 22.9 13.9 17.6 42.7 12.5 18.5 30.8 11.1 19.2 Friends/relatives are in this line 22.3 14.9 15.0 26.8 25.2 22.1 2.2 21.4 15.4 I saw a market opportunity 52.1 51.7 50.9 37.7 43.3 40.4 2.2 43.0 36.2 My capital could only finance this business 37.7 44.5 43.6 15.4 45.3 39.9 74.3 48.6 50.0 No apparent reason 1.9 6.0 4.8 8.7 4.8 6.8 0.0 2.6 4.0 I just wanted to be near my house 0.0 1.2 0.7 0.0 0.5 0.6 0.0 0.0 0.0 I liked it 0.7 1.4 2.0 2.5 3.0 2.1 4.0 0.0 0.7 Other adult owners I had previous experience inthis line 26.2 16.5 18.9 37.3 18.2 20.5 42.4 7.8 17.7 Friends/relatives are in this line 19.7 11.9 14.6 16.6 14.0 13.6 22.7 10.0 9.8 I saw a market opportunity 46.1 51.6 49.2 35.2 42.9 42.8 25.4 50.7 42.7 My capital could only finance this business 35.2 39.8 40.3 34.3 49.4 46.6 22.9 44.5 45.0 No apparent reason 3.5 6.0 4.3 0.9 1.2 1.7 5.9 8.5 5.4 I just wanted to be near my house 1.2 0.8 1.0 0.0 3.0 1.8 0.0 0.0 0.9 I liked it 0.8 0.7 0.7 4.3 0.2 1.2 0.0 2.4 1.6 Note: Number of the full sample obs. is 4,163, and 1,853 for the MSMEs with youth as owners. Multiple answers are allowed for individual MSME owners. Source: Authors calculations using the MSME Survey 2010. Table 8.9. Continued Rural enterprise Other urban enterprises Enterprise in Dar es Salaam Manufacturing Trade services Rural total Manufacturing Trade Services Other urban total Manufacturing Trade services Dar Es Salaam total OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Rural Nonfarm Enterprises in Tanzania 235 Table 8.10. Job satisfaction in MSME survey (weighted, percentage) Rural Other urban Dar es Salaam Total All MSMEs If you were offered a full-time salary paying job, would you take it? 47.8 48.6 37.7 46.6 Who would you rather work for? A large private company 17.9 27.4 43.1 24.0 Government 68.6 62.9 44.6 63.9 Someone else’s business 10.5 6.2 10.8 9.1 Anywhere 3.0 3.4 1.5 3.0 And why do you say that? Better security of income 81.7 83.8 81.4 82.3 Shorter hours 5.1 5.7 3.5 5.1 Less risk 1.8 1.8 3.0 1.9 To get pension 1.5 1.6 – 1.4 I am less educated 2.2 0.9 1.9 1.8 They listen to the opinions of the employees 1.1 1.0 – 1.0 As long as I get a living 0.6 0.4 – 0.5 Job security 2.1 0.6 – 1.4 Others 2.2 2.3 9.9 3.1 None/Nothing 1.7 2.1 0.2 1.6 Youth owners If you were offered a full-time salary paying job, would you take it? 54.0 52.9 38.2 51.1 Who would you rather work for? A large private company 20.0 36.7 33.8 27.2 Government 65.4 54.8 49.6 59.9 Someone else’s business 11.0 6.0 14.0 9.7 Anywhere 3.7 2.6 2.7 3.2 And why do you say that? Better security of income 81.5 86.2 80.4 82.9 Shorter hours 4.0 4.6 3.5 4.1 Less risk 2.2 2.7 5.5 2.8 To get pension 0.9 1.4 – 1.0 I am less educated 2.7 0.7 3.6 2.2 They listen to the opinions of the employees 0.9 0.4 – 0.6 As long as I get a living 0.8 0.2 – 0.5 Job security 2.5 0.4 – 1.5 Others 2.8 1.4 6.6 2.8 None/Nothing 1.6 1.9 0.4 1.6 Other adult owners If you were offered a full-time salary paying job, would you take it? 42.6 44.4 37.1 42.5 Who would you rather work for? A large private company 15.7 16.9 54.3 20.4 Government 72.0 72.2 38.7 68.3 Continued OUP CORRECTED PROOF – FINAL, 29/10/19, SPi 236 Xinshen Diao, Eduardo Magalhaes, and Margaret McMillan areas where 52.9 per cent of young adults prefer a full-time paying job compared to 44 per cent for other adults. Approximately, 64 per cent all respondents who would prefer a full-time sal aried job say they would like to work for the government, and the proportions are 68.6 and 62.9 per cent in rural and other urban areas but only 44.6 per cent in Dar es Salaam, where more government jobs are concentrated. The responses from rural and other urban MSME owners are consistent with results reported in Duflo and Banerjee’s analysis of the economic lives of the poor (2007). Between youth and other adults, other adults seem to prefer a government job than youth in rural and other urban areas, but the opposite occurs in Dar es Salaam where youth prefer government jobs than other adults. Large private companies are more attractive to small business owners in Dar es Salaam than in other places particularly among other adults. The predominant reason for preferring a fulltime salaried position is better security of income, which is consistent between youth and other adults and across different locations. Table 8.10. Continued Rural Other urban Dar es Salaam Total Someone else’s business 9.9 6.5 7.0 8.5 Anywhere 2.4 4.3 – 2.7 And why do you say that? Better security of income 81.8 80.9 82.6 81.6 Shorter hours 6.2 6.9 3.5 6.1 Less risk 1.4 0.6 – 1.0 To get pension 2.1 1.9 – 1.8 I am less educated 1.7 1.1 – 1.4 They listen to the opinions of the employees 1.3 1.7 – 1.3 As long as I get a living 0.4 0.6 – 0.4 Job security 1.7 0.7 – 1.2 Others 1.5 3.3 13.8 3.5 None/Nothing 1.7 2.3 – 1.7 Note: This table is prepared based on three questions: (1) ‘If you were offered a full-time salary paying job, would you take it?’ (2) ‘Who would you rather work for?’ and (3) ‘Why do you say that?’ A unique answer is provided by individual MSME owners to each of the last two questions. Rural, urban, and national total MSMEs, MSMEs owned by youth, and MSMEs owned by other adults for the sum of these two questions are 100 respectively. Number of the full sample obs. is 4,163, and 1,853 for the MSMEs with youth as owners. Source: Authors calculations using the MSME Survey 2010. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Rural Nonfarm Enterprises in Tanzania 237 8.5.3 The Productive Heterogeneity of Rural Enterprises We use kernel densities of the log of value added to examine the productive heterogeneity of MSMEs. Value added is computed as the firm’s average monthly sales minus the firms’ average monthly costs of production and is in nominal units of local currency. Firms in the MSME database report sales monthly and thus we can take seasonality into account. Our analysis of the productive heterogeneity of firms in the MSME sector reveals two important features of these firms. First, there is a significant degree of productive heterogeneity among both rural and urban enterprises. This can be seen by examining the density of the log of value added per worker in Figure8.2. Value added per worker is reported for both rural enterprises and urban enterprises, and urban enterprises are further disaggregated into other urban and Dar es Salaam. We also create a figure (Figure8.3) for value added per worker for small businesses owned by youth and by other adults. In both Figures8.2 and8.3 the vertical lines represent average labour productivity in Tanzania’s economy in 2010 in the agricultural sector (light gray), the trade services sector (middle gray) and the manufacturing sector (dark gray). Surprisingly, the distribution of the log of value added per worker or labour productivity for rural firms is almost identical to the distribution for 0 .1 .2 .3 Kernel density of the log value added per worker –10 –5 0 5 Rural Other urban Dar Es Salaam Figure 8.2. The distribution of the log of value added per worker among MSMEs in 2010 by location OUP CORRECTED PROOF – FINAL, 29/10/19, SPi 238 Xinshen Diao, Eduardo Magalhaes, and Margaret McMillan urban firms. In fact, a test of stochastic dominance13 rejects the hypothesis that the rural and urban distributions are not identical. Also, there is little difference between small businesses owned by youth and owned by other adults. One reason for this may be the fact that medium-sized enterprises that are mainly in urban areas appear to be under-sampled in the MSME survey due to the reliance on households for the sampling framework. Figures8.2 and8.3 also reveal that a little over half of the firms in the MSME sector have labour productivity levels higher than the economy-wide average in agriculture and that this is true in all the three locations. This is not surprising and is consistent with evidence presented by Diao, Kweka, and McMillan (2017) who show that average productivity in the sectors dominated by small firms is consistently higher than average productivity in agriculture. We can also see from Figure 8.2 that around 15 per cent of rural MSMEs have labour productivity higher than economy-wide manufacturing labour productivity. This too is consistent with Diao, Kweka, and McMillan (2017) who find that 15 per cent of total MSMEs in Tanzania account for 70 per cent of the total value added generated by the MSME sector. By contrast, the remaining 85 per cent of the MSMEs account 13 A stochastic dominance test compares the cumulative density function of the log of the value added in rural and urban areas. To establish whether the two curves originate from the same distribution, we have used the Komolgorov-Smirnov test of equality of distribution. We were not able to reject the null hypothesis that the distributions of urban and rural value added were the same (p-value=0.9). 0 .1 .2 .3 Kernel density of the log value added per worker –10 –5 0 5 youth other adults Figure 8.3. The distribution of the log of value added per worker among MSMEs in 2010 by young and other adult owners OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Rural Nonfarm Enterprises in Tanzania 239 for only 30per cent of the value added generated by the MSME sector. These results underscore the productive heterogeneity of the MSMEs in both rural and urban areas. 8.6 Using the MSME Survey to Identify ‘High Potential’ RuralEnterprises The main takeaway from the previous section is that some rural MSMEs have the potential to contribute significantly to rural transformation in Tanzania. To identify the characteristics of these firms, we separate the ‘high potential’ MSMEs from the rest using both qualitative information and performance-based measures of labour productivity. Following Lewis (1979), we call these groups of firms the ‘in-between’ firms. This terminology is meant to capture the idea that the characteristics of these firms place them somewhere in-between Tanzania’s modern (most productive) and informal (least productive) firms. For example, the in-between firms may keep written accounts and be quite productive, but not be registered. For the purposes of this chapter, we include in the in-between group of firms only firms whose owners report that they would not quit their job for a sal ar ied position, and in which labour productivity is greater than economy-wide labour productivity in trade service sector, represented by a vertical line in the middle of Figures8.2 and8.3 between the line for the economy-wide labour prod uct iv ity of agriculture on the left and the line for the economywide labour prod uct iv ity of manufacturing on the right. Using these criteria, we find that out of 5,609 sampled firms 944 firms that can be classified as belonging in the in-between sector, 45 per cent of which are rural firms, accounting for 10per cent of total sampled rural firms. To understand what distinguishes businesses with high potential, we have identified several explanatory variables that have been commonly used and tested in the relevant literature. These variables cover a host of business characteristics and are likely to provide a complete picture of the determinants of high potential businesses. Four categories of variables were identified: owner’s personal characteristics, business characteristics, infrastructure and technology, and financial and other services. Using a probit model, we explored the characteristics of the most successful businesses considering these four categories of variables as explanatory variables in the regression. The dependent variable is 1 if a firm belongs to the in-between14 category, the one we indicated in the fourth section as with high potential, and 0 otherwise. The regression shows that several 14 As a reminder, for a firm to be considered in-between it must fulfil two conditions: First, its owner would not leave the business for a full-time salaried job, and second, labour productivity must be higher than the economy-wide trade productivity. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi 240 Xinshen Diao, Eduardo Magalhaes, and Margaret McMillan vari ables that we expected to be statistically significant were not. The absence of significance does not mean that these variables are irrelevant to the analysis. Rather, it is possible that this is a consequence of the cross-sectional nature of the data, which constrains to provide comparison points for both business and owner’s characteristics. Nevertheless, we highlight a few main findings from the probit regressions focusing on the marginal effects in Tables 8.A4(a)–8.A4(c) in the Appendix. Results for the owner’s personal characteristics suggest that in-between firms are less likely to be headed by females on the order of 5 to 8 per cent depending on the location of the business. The two most important determinants of inbetween firms are whether the owner is not poor and whether the owner views the business as growing. Both variables are highly significant and lead to non-negligible increases in the probability of being in the in-between category. Common sense confirms the feasibility of these results. Non-poor business owners are less likely to require loans and to have to sacrifice on labour or the quality of services provided due to lack of resources. Similarly, business owners who are optimistic about their firm’s future and potential are also more likely to be driven to achieve success and to use resources in ways that are productive. Thus, none of the results described showed differences in the signs and presence of significance across the three locations (rural, other urban, and Dar es Salaam). The variable of ‘seeing business as a market opportunity’ shows significant results for national total and urban enterprises, a 3 and 4 per cent increase in the probability of being in the in-between category, respectively, but not for rural enterprises. ‘Membership to a business association’ also increases the probability of a firm falling in the in-between category. The absence of significance for young business owners is not surprising given that most business owners are young. Education, marital status, and whether the owner has taken expert advice are not significant. Education illustrates this point well. Evidence abounds about the positive effects of education on the performance of both businesses and individuals, but it needs panel data to verify it, as gains from education often are not immediately reflected in the performance of businesses. Similar logic can be applied to the owner receiving expert advice. Without more information about when individuals completed their education or how long has the owner received expert advice, it becomes difficult to gauge the true importance of these variables. The lack of significance of these variables may also reflect the lack of variability in the data. A considerably more varied pattern of significance across geographies is found for business characteristics. Results for the variables under the business characteristic are not particularly surprising. For example, a one-unit increase in the number of employees reduces the probability of being an in-between firm by slightly over 2 per cent across locations. Businesses that run full-time, on the other hand, increase the probability of being in the in-between category by 6–7 per cent depending on location. Both signs fall within our expectations for two reasons. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Rural Nonfarm Enterprises in Tanzania 241 First, the survey is designed to capture small businesses. Second, running a fulltime business indicates that the business probably experiences some degree of success, or else owners would be inclined to find other sources of income. Less obvious, however, is the fact that holding written accounts is only significant in rural areas. Moreover, the coefficients for firms that have regional customers and whose number of daily customers exceeds 20 a day are positive and significant both nationally and in rural areas, which suggest that the national results were likely driven by rural areas. Changes in the predicted probabilities associated with having regional customers range from 3–5 per cent for all firms and for rural firms respectively. Firms with a daily number of customers higher than 20 are also more likely to be in the in-between category; the predicted probabilities range from 5.8 to 6.8 per cent nationally and in rural areas. The lack of significance of written accounts, regional customers, and number of daily customers in urban areas could be the result of lower variability among urban firms relative to rural areas, particularly because the survey has a much greater number of rural enterprises than urban. Non-significant variables such as market access, the nature of firms’ suppliers (whether small traders or nationwide), and whether the firm has a licence may be a result of the fact that the average firm in the dataset is relatively young (about four years old). This, in turn, might suggest that not enough time has gone by for firms to be able to develop a consistent and systematic source of supply. A similar case could be argued for market access. As for whether firms have licences, the lack of significance might be a result of delays from regulatory agencies to provide licences without much red tape. In all cases, additional data about when firms started resorting to suppliers or gained access to markets (or time to receive a licence) would shed light on additional potential reasons for the absence of significance of these three variables. The variables related to infrastructure and technology are also considered in the regressions. The gains in predicted probability obtained by using a mobile phone to conduct business range from 2.5 to nearly 8 per cent nationally and in urban areas respectively. Curiously, the fact that the owner owns a calculator is only significant at the national level with an associated change in the predicted probability of being in the in-between category of 4.2 per cent. The most im port ant factor in the set of infrastructure and technology variables is ‘whether the business uses electricity to light their businesses’. Increased predicted probabilities are observed across geographies, with gains ranging from nearly 7 per cent in rural areas to slightly over 8 per cent in urban areas. The absence of significance for variables such as ‘whether the business has office equipment or a cooling facility’ might also be related to the age of the firm and its ability to set up a complete and fully functional office structure. Here too, panel data would provide insights about the importance of these variables. Variables related to the access of financial and other services do not show a systematic pattern of significance. Significant results are observed for firms that OUP CORRECTED PROOF – FINAL, 29/10/19, SPi 248 Xinshen Diao, Eduardo Magalhaes, and Margaret McMillan Table 8.A4(c). Probit results for probability of being in-between rural and urban enterprises, marginal effect: infrastructure, technology, and financial services All Urban Rural Infrastructure and technology Owner uses a mobile to conduct business 0.0450** 0.0779*0.0143 –0.0216 –0.0407 –0.0163 Firm owner has a calculator 0.0424** 0.0464 0.0372 –0.021 –0.0327 –0.0258 Business has office equipment 0.0285 0.0407 0.0219 –0.0265 –0.0435 –0.0205 Business owns a cooling facility 0.0202 –0.0106 0.11 –0.0317 –0.0374 –0.072 Business uses electricity to light business 0.0802*** 0.0837** 0.0692* –0.0252 –0.0326 –0.0392 Financial and other services Owner regularly sends and receives moneyfor business 0.0279 –0.0187 0.0249 –0.0339 0.0232 –0.0212 Firm has received legal services 0.0343 0.159 –0.0824** –0.0883 –0.133 –0.033 Firm has received technical services –0.00217 –0.0268 0.0134 –0.0493 –0.0668 –0.0493 Owner uses profit to expand business 0.0329 0.0892** –0.00386 –0.0228 –0.0387 –0.0207 Owner uses profits to buy stocks in advance 0.0149 0.0158 0.0114 –0.0143 –0.0286 –0.014 Owner uses profits to invest in buildings andland 0.0458 –0.0314 0.0357 –0.0482 0.0566* –0.0339 Observations 5,551 5,590 5,570 Nsub 5,551 1,427 4,124 F 10.38 4.553 27.28 P-value 0 3.16E-08 0 Note: Number of the full sample obs. is 4,163, and 2,310 for rural. *** p < 0.01, ** p < 0.05, * p < 0.1. Source: Authors estimation using MSME dataset. References Banerjee, A.V., and E.Duflo. 2007. The economic lives of the poor. The Journal of Economic Perspectives 21 (1): 141–67. Bezu, S., and C.Barrett. 2012. Employment dynamics in the rural nonfarm sector in Ethiopia: Do the poor have time on their side? Journal of Development Studies 48(9): 1223–40. Davis, B., S.Di Giuseppe, and A.Zezza. 2014. Income diversification patterns in rural sub-Saharan Africa: Reassessing the evidence. Policy Research Working Paper 7108, the World Bank. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Rural Nonfarm Enterprises in Tanzania 249 Diao, X., J.Kweka, and M.McMillan. 2017. Economic transformation from the bottom up: Evidence from Tanzania. IFPRI Discussion Paper 1603. Washington DC, U.S.A.: International Food Policy Research Institute. De Brauw, A., V.Mueller, and H.L.Lee. 2014. The role of rural–urban migration in the structural transformation of sub-Saharan Africa. World Development 63 (2014): 33–42. Gollin, D., and R.Rogerson. 2009. The greatest of all improvements: Roads, agriculture, and economic development in Africa. Department of Economics, Williams College, A mimeo. Haggblade, S., P.Hazell, and J.Brown. 1989. Farm–nonfarm linkages in rural subSaharan Africa. World Development 1 (8): 1173–201. Haggblade, S., P.Hazell, and T.Reardon. 2010. The rural nonfarm economy: Prospects for growth and poverty reduction. World Development 38 (1): 1429–41. Jin, S., and K.Deininger. 2008. Key constraints for rural nonfarm activity in Tanzania: Combining investment climate and household surveys. Journal of African Economies 18 (2): 319–61. Lewis, W.A. 1979. The dual economy revisited. The Manchester School 47 (3): 211–29. Nagler, P., and W.Naudé. 2016. Non-farm enterprises in rural Africa: New empirical evidence. Policy Research Working Paper 7066. Washington, DC: U.S.A.: The World Bank Group. Tanzania, NBS (National Bureau of Statistics). 2006. Tanzania 2002 Census: Analytic Report, Volume X. Dar es Salaam: National Bureau of Statistics, Ministry of Planning, Economy, and Empowerment. Tanzania, NBS (National Bureau of Statistics). 2007. Analytic Report for Employment and Earnings Survey 2002. Dar es Salaam: National Bureau of Statistics, Ministry of Planning, Economy, and Empowerment. Tanzania, NBS (National Bureau of Statistics). 2011a. National Household Budget Survey 2000–2001. Dar es Salaam: National Bureau of Statistics, Ministry of Finance. Tanzania, NBS (National Bureau of Statistics). 2011b. Integrated Labour Force Survey 2014, Analytical Report. Dar es Salaam: National Bureau of Statistics, Ministry of Finance. Tanzania, NBS (National Bureau of Statistics). 2014a. Household Budget Survey: Main Report, 2011/12. Dar es Salaam: National Bureau of Statistics, Ministry of Finance. Tanzania, NBS (National Bureau of Statistics). 2014. Tanzania 2012 Census: Basic demographic and socio-economic profile. Dar es Salaam and Zanzibar: National Bureau of Statistics, Ministry of Finance, Office of Chief Government Statistician, Ministry of State, President’s Office, State House and Good Governance. Tanzania, NBS (National Bureau of Statistics). 2014c. National Accounts of Tanzania Mainland 2001–2013. Dar es Salaam: National Bureau of Statistics, Ministry of Finance. Tanzania, NBS (National Bureau of Statistics). 2014d. Revised National Accounts Estimates for Tanzania Mainland, Base Year, 2007. Dar es Salaam: National Bureau of Statistics, Ministry of Finance. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi 250 Xinshen Diao, Eduardo Magalhaes, and Margaret McMillan Tanzania, NBS (National Bureau of Statistics). 2014e. Formal Sector Employment and Earnings Survey: Analytical Report 2013. Dar es Salaam National Bureau of Statistics, Ministry of Finance. Tanzania, NBS (National Bureau of Statistics). 2015. Integrated Labour Force Survey 2014, Analytical Report. Dar es Salaam: National Bureau of Statistics, Ministry of Finance. Tybout, J.R.2000. Manufacturing firms in developing countries: How well do they do, and why? Journal of Economic Literature 38 (1): 11–44. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Elisenda Estruch, Lisa Van Dijck, David Schwebel, and Josee Randriamamonjy, Youth Mobility and its Role in Structural Transformation in Senegal In: Youth and Jobs in Rural Africa: Beyond Stylized Facts. Edited by: Valerie Mueller and James Thurlow, Oxford University Press (2019). © International Food Policy Research Institute. DOI: 10.1093/oso/9780198848059.003.0009 9 Youth Mobility and its Role in Structural Transformation in Senegal Elisenda Estruch, Lisa Van Dijck, David Schwebel, and Josee Randriamamonjy 9.1 Introduction Senegal is a youthful country, with over 60 per cent of people below the age of 24years and up to 77 per cent of the population below the age of 35 years. Over 100,000 new young job seekers (between 15 and 34 years old) join the labour market every year in Senegal (World Bank2016). Although the agricultural sector employs nationally almost half of the labour force, the sector accounts only for about 16 per cent of GDP (World Bank2017). Low productivity growth in this sector combined with an unbalanced labour market has led to a wider economic stagnation of the rural economy. This creates major challenges for the rural youth to access productive and decent jobs and reach their work and life aspirations (Hathie et al.2015). Whilst there is a growing body of empirical analysis on youth employment in Sub-Saharan Africa, im port ant gaps remain and especially in relation to youth in rural areas in the Francophone countries of West Africa, such as Senegal. This chapter aims to complement the existing literature on employment dynamics in Senegal, focusing in depth on rural youth in the context of structural transformation. First, a stocktaking of the state of the agricultural sector and the wider rural economy is provided and how this relates to youth employment dynamics in rural areas of Senegal. In Section 9.2, through a literature review, the context for rural youth employment is discussed by looking in detail at the agricultural sector, the rural-non farm economy and the challenges and opportunities characterizing the rural labour market. Section 9.3 analyses the transitions of youth, either to the rural non-farm economy (RNFE) or to urban areas and abroad and attempts to provide a better understanding of patterns, motivations, and constraints youth face when engaging in the RNFE or in migrant labour. To do so, we use descriptive statistics from the latest Household Data Surveys of 2001 (ESAM-II) and 2011 OUP CORRECTED PROOF – FINAL, 29/10/19, SPi 252 Elisenda Estruch et al. (ESPS II) and the Migration and Remittances Household Survey.1 The findings suggest that there are limited rural employment op por tun ities for youth, leading to a slow pace of rural poverty reduction. Rural youth still work mainly in poor quality jobs in agriculture, although they increasingly try: (i) to diversify their and their family’s income by engaging in rural nonfarm employment, or (ii) to look for options outside rural areas by migrating to urban areas or abroad. Section 9.4 follows with a brief review of the main policies and programmes that have been implemented in the same period as the analysed data in Senegal in order to identify their strengths, weaknesses, and areas for improvement. We conclude with a discussion around the key findings. 9.2 State of Agriculture and the Rural Labour Market Senegal is a lower-middle income country with the fourth largest economy in West Africa. In 2015, Senegal had a GDP growth rate of 6.5 per cent, becoming the second fastest growing economy in West Africa (World Bank2016). However, between 2000 and 2015, Senegal experienced a lower average economic growth, of 4.1 per cent, than the rest of Sub-Saharan Africa (SSA) with 6 per cent. This economic growth has not been large enough to match the high population growth, leading to an average GDP per capita growth of only 2.3 per cent in that period, half of that experienced in SSA (ibid). In economic terms, the Senegalese economy is mainly dominated by the service sector, accounting for 60 per cent of GDP, followed by the industrial sector with 24 per cent of GDP (Figure9.1). Although the agricultural sector employs nationally around half of the labour force, the sector accounts only for about 16 per cent 1 See description of the data in the annex. Agriculture 16% Industry 24% Services, etc. 60% Figure 9.1. GDP contributions, Senegal, 2017 Source: World development indicators, World Bank. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Youth Mobility and its Role in Senegal 253 of GDP. With an annual growth rate of 3.5 per cent, the agricultural sector is growing slower than the service and industrial sectors. Compared to the rest of Sub-Saharan Africa, the agricultural sector is also growing at a slower pace (Figure9.2). Growth levels fluctuate heavily as Senegal’s agriculture is highly vulnerable to unpredictable weather; severe floods and droughts are common and only 1.3 per cent of agricultural land is equipped for irrigation. 9.2.1 Agricultural Sector On a macroeconomic level, the agricultural sector in Senegal seems to be stagnant, as labourand land-productivity have remained constant or have grown slowly over the past decade (Figure9.3 and Figure9.4). Senegal’s low labour prod uct iv ity is caused by an unchanged employment rate in agriculture coupled with low techno logic al improvement (Seck 2016, Shaw 2014). The agricultural sector employs 46 per cent of the total employed population, composed by a male and female employment rate in the sector of respectively 49 and 44 per cent. Contrary to the Sub-Saharan African average, the employment rate has decreased only slightly, leading to a lower and stabilized agricultural value added per worker (ibid). Similarly, land productivity, here projected by cereal yield, is around 20 per cent lower than the Sub-Saharan Africa average. Land and labour productivity are influenced by a low uptake of technology. For example, in 2014, per hectare of arable land only 7 kilogram of fertilizer was used in Senegal, compared to 16 kilogram on average in Sub-Saharan Africa (World Bank2017). 0% 1% 2% 3% 4% 5% 6% 7% Agriculture Industry Services GDP GDP per capita Senegal SSA Figure 9.2. Drivers of economic growth, average growth, 2000–2015 Source: World development indicators, World Bank. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi 254 Elisenda Estruch et al. According to the World Bank, agriculture is at fourth place in government spending after education, health, and nutrition, which represent the bulk of the state budget. The percentage of public spending on agriculture evolved in terms of GDP, rising from 9.8 to 10.9 per cent between 2005 and 2009 (World Bank2014). The country would have thus achieved the Maputo objective, which requires that a minimum of 10 per cent of total public expenditure be directed at agriculture. However, at a time when the authorities declare their intention to prioritize agriculture in their strategy of accelerated growth, the growth in investment in the sector appears insufficient (Oya and Ba2013). Out of the total budget for agriculture, subsidies for the distribution of agricultural inputs are the main component of expenditure with 46 per cent of resources. Much needed infrastructure works follow with 11 per cent. By contrast, the resources devoted to agricultural 300 500 700 900 1100 1300 2000 2002 2004 2006 2008 2010 2012 2014 SSA SEN Figure 9.3. Labour productivity in agriculture (agricultural value added per worker, constant 2010 US$) Source: World development indicators, World Bank. 600 800 1000 1200 1400 1600 2000 2002 2004 2006 2008 2010 2012 2014 SSA SEN Figure 9.4. Land productivity (cereal yield, kg per ha) Source: Food and Agriculture Organization (FAO). OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Youth Mobility and its Role in Senegal 255 research (3.1 per cent) and training (0.5 per cent) are negligible. Intra-sector allocation of expenditure shows that agricultural public expenditure is concentrated on crop farming, receiving nearly 64 per cent of the total budget. Other sectors, such as livestock and fisheries, are hence left with little funds and support (World Bank2014). 9.2.2 Rural Nonfarm Economy The Senegalese rural nonfarm economy (RNFE) is dominated by manufacturing, and wholesale and retail trade, and is limited in size and employs only 23 per cent of the employed population (ANSD2013). The RNFE is mainly composed of informal small-scale businesses and is characterized by a high level of selfemployment. The main hindrances to formalization of businesses are the lack of entrepreneurial and technical education or skills as well as underinvestment in infrastructure and the weakness of regulation and taxation institutions. Youth have special difficulties in finding the resources to create or upgrade their own businesses (Ndione2015). Increasingly, more rural households are diversifying their incomes and en gaging both in farm and nonfarm activities. In nearly a quarter of all rural households, both farm and nonfarm activities coexist (ANSD2013). While nonfarm activities play an important role in household livelihood strategies in rural Senegal, they typic al ly complement farm activities rather than constitute a viable farm exit strategy (Ndione2015). In order to reduce rural poverty effectively and develop a wider rural economy, it is important to enhance the RNFE and its linkages with farm economy. The effects emerge through backward and forward production linkages from agriculture to rural input suppliers and agro-processors, and through expenditure linkages as farm incomes are spent on locally produced goods and services or invested in nonfarm activities (Maertens2009). Spillover effects from the nonfarm economy to the farm economy increase productivity and profitability in agriculture (Anríquez and Stamoulis2007). In Senegal, linkages between farm and nonfarm economic sectors, which are important in creating multiplier effects for growth and rural development, are weak. Small farmers are only to a limited extent involved in the value chain leading to low profitability and to a lack of employment opportunities and attractiveness for youth (Davis et al.2002). Moreover, the agribusiness sector, which in many cases is more labour intensive and where added value along the agrifood value chain can be generated more easily, is small in Senegal compared to the Sub-Saharan Africa average (Schaffnit-Chatterjee2014). OUP CORRECTED PROOF – FINAL, 29/10/19, SPi 256 Elisenda Estruch et al. 9.2.3 Rural Labour Market The stagnant productivity in the agricultural sector is indicative of a wider, crosssectoral problem of an unbalanced rural labour market, with a labour supply characterized by a large young workforce with limited education and access to productive resources, and a labour demand characterized by low rural investments, access to markets, and rural job creation (Hathie et al.2015). Moreover, lack of access, especially for youth, to productive resources such as land, credit, and social capital, are restraining overall rural productivity levels in both agricultural and non-agricultural activities (ibid). Data from ANSD 2004 and 2013 shows that indeed the Senegalese labour supply is characterized by a large share of poorly educated and low-skilled labour force. There are significant education limitations in the country, as the percentage of people over 25 years old having completed primary and secondary education is respectively 22 per cent and 6 per cent (ANSD2013). Low levels of education characterize especially the rural and female population. The situation has however improved in the last decade as shown by the increase in the gross enrolment rate in secondary schooling between 2001 and 2011 from 16 to 40 per cent (ANSD 2004, 2013). National alphabetization rates have also grown from 38 to 52 per cent, and from 23 to 33 per cent in rural areas (ibid). Rural areas continue to lag behind in terms of educational attainment as shown by the fact that only 30 per cent of the rural population has benefited from any form of education compared to 51 per cent in urban areas. There also remains a 10 per cent gap between rural young men and women (ANSD2013). Moreover, in rural areas, few youth transition from primary to secondary education, due to high dropout rates and poor quality of education, which leads to low levels of literacy. Furthermore, the actual skills of the workforce are often not aligned with the labour market demand needs (Guarcello2007). Skills acquired during higher education often do not match those required in the rural labour market. This indicates a deep structural shift that needs to take place in order to create more high-level jobs and utilize the increased human capital (ibid). At the same time, in line with the process of structural transformation, as agriculture moves away from subsistence farming and the RNFE increases in relevance, improving the skills of the rural workforce is key to facilitate the mobility of surplus labour within agriculture and towards other activities (ibid). Labour supply is strongly influenced by the high population growth and subsequent youth bulge. The Senegalese population amounts to 14.3 million people and grew at an annual rate of 3.1 per cent between 2001 and 2011 (ANSD 2004,2013). The country is demographically quite young with an average age of 22 years and with one-third of the population aged between 15 and 34 years. This means that a growing number of young people are entering the labour market every year, OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Youth Mobility and its Role in Senegal 257 resulting in an excess of labour supply and subsequent pressures on the labour market to integrate them all. Between 2001 and 2011, the strongest growth in population was observed in older age groups, indicating the youth bulge is slowing down (ANSD2013). In rural labour markets, often the most decisive factors in improving the pay and working conditions of rural workers are a dynamic agricultural sector, increased public and private investment, and a tighter labour market (Oya2010). In other words, besides improving the skills development and other supply-side policy interventions, demand-side interventions, such as investments targeting job creation, are critical (ILO2008). The formal labour demand is limited, as Senegal’s rapid population growth was not matched by job creation in the formal sector at an equal rate and thus led to a booming of the informal sector. The latter is dominated by subsistence agriculture and small-scale informal (family-based) firms (Golub and Hayat2014). Although wages in the formal sector are three times higher than those in the informal sector, the formal private sector employs just 6 per cent of the total population (ibid). Conversely, in the informal sector, registration and taxation are absent, formal contracts and social protection rare, wages inefficiently low, underemployment the norm, and labour rights weak (Roubaud and Torelli2013). Furthermore, the small size of the operations of many rural enterprises poses also constraints in terms of job creation as they lack economies of scale. Ultimately, agricultural labour demand is limited and characterized by intra and inter-year variations, on account of weather conditions and the seasonal nature of agricultural labour, limiting the number of long-term formal contracts even further (Leavy and White1999). There are still more people living in rural areas than in urban areas, re spective ly 57 and 43 per cent of the population, although urbanization is occurring at remarkable pace: while the urban population increases at an annual rate of 3.9 per cent, the rural population grows annually by 2.6 per cent (World Bank 2017). Urbanization is also driven by significant internal migration: by 2009, almost 2 million Senegalese, or 14 per cent of the total population, had migrated within the country (IOM2009). In the past, migration was essentially from rural to urban areas, generally from the semi-arid regions towards Dakar; or from rural to rural areas following a seasonal pattern (that is, those who migrate during the rainy season to provide additional support) (ibid). Today, the Dakar region, but more generally the urban axis or Dakar–Thiès–Touba (the second largest city in the country), polarizes 60 per cent of migration and represents 47 per cent of the country’s population. The attractiveness of this urban axis largely determines the structure and dynamism of economic activities, including service provision to surrounding areas. It also determines the opportunity space for many rural youth (in particular from the Groundnut Basin). Nonetheless, there are also other secondary cities and towns to consider, such as M’bour, which is growing due to OUP CORRECTED PROOF – FINAL, 29/10/19, SPi 264 Elisenda Estruch et al. relied more on remittances to meet their needs (Ndiaye et al.2015). Reduced labour participation can also entail a positive outcome of migration when the flow of income increases educational uptake or reduces hazardous work. According to the stocktaking of migrants in 2009 (World Bank2009b), Senegal counts over 1.2 million migrants, of which 75 per cent are internal migrants. Over half of all migrants come from rural areas and 60 per cent are young, aged 15 to 34 (ibid). Looking into migration flows, MRHS 2009 captures information from households that had internal or international migrants about the pre and post residence of former member (N=1278 for youth 15-34-year-old, and N=929 for non-youth 35-year-old and above). Figure9.7 indicates that more people are leaving rural areas than arriving. Rural youth are more likely to migrate to the urban areas (5.5 per cent of all rural youth) while their older counterparts are more likely to move abroad (4.3 per cent of all rural non-youth). Youth are leaving rural areas more often than their older counterparts do (8.5 per cent of all rural youth versus 7.6 per cent of all rural non-youth), indicating that the rural exodus has a young face. Remarkably, people who move internationally are more likely to come from rural areas, for all ages. The fact that many people, particularly older people, are moving from urban to rural areas might indicate they stay and work for a while in urban areas but return to their place of origin at a certain time. The main reasons for internal migration in Senegal are family-related (63per cent), for work (17 per cent), and for education (14 per cent). When only looking at the rural population, the importance of work increases considerably (to73 per cent for internal migration and to 83 per cent for international). Work 5.5%5.5% 0.9% 0.9% 2.1% 2.1% 2.7% 2.7% 2.4% 2.4% 0.4% 0.4% 2.8% 2.8% 0.5% 0.5% 4.3% 4.3% 2.5% 2.5% 4.1% 4.1% 0.3% 0.3% 0.0% 1.0% 2.0% 3.0% 4.0% 5.0% 6.0% Rural to urban Rural to rural Rural to international Urban to urban Urban to rural Urban to international Non-youth (>34 years) Youth (15–34 years) Figure 9.7 Migration by origin, destination and age (as a percentage of total population) Source: Authors’ calculations based on MRHS 2009 from the World Bank. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Youth Mobility and its Role in Senegal 265 isequally the main reason to migrate for rural youth (58 per cent for those moving in tern al ly and 69 per cent for those moving abroad). The main reasons for people to move out of the rural space are the attraction of employment opportunities outside rural areas or the lack of employment opportunities in the rural space itself. Gender is also a key factor that distinguishes the reasons rural individuals migrate: women’s number one reason to migrate is family-related (63 per cent) (ESPS-II). The sector in which rural migrants were employed before moving varies by destination. Table9.2 describes the migration patterns of youth active in agriculture before moving. Senegalese youth migrants leaving rural agriculture are most likely to move abroad, followed by moving to an urban area, whereas a low percentage migrate to another rural area. Those moving to an urban area are mainly self-employed outside the agricultural sector. Looking at the characteristics of migrant youth and their households, based on labour transition data, it emerges that the youth who move out of agriculture but stay in rural areas to work in the RNFE are on average more frequently males, older, and less educated than youth undertaking a migration-induced labour transition (see Table9.A3 in the Annex for details). Especially in terms of education, the difference is substantial (1.65 years versus 4.34 years). The households of the sub-group of youth who leave agriculture but remain rural are on average smaller, located in smaller communes, and poorer than households of youth undertaking a migration-induced labour transition. This indicates that the lesseducated workers from poorer households are staying in rural areas while more educated workers from richer households are leaving the rural space altogether. The potential employment opportunities and the distance of migration influences the chosen destination. Less-educated workers might have more employment possibilities in rural areas, and poorer or disadvantaged households cannot always afford the elevated cost associated with long-distance migration. Table 9.2. Migration patterns of youth leaving agriculture Exit rural agriculture to Rural Urban Abroad Agriculture 3.40% 10.26% Self-employment 38.23% 17.14% Other (paid or unpaid family worker) 17.07% 7.01% Non-agriculture 2.75% 32.23% Self-employment 24.11% 65.56% Other (paid or unpaid family worker) 20.58% 10.30% Total N=665 6% 42% 51% Source: Authors’ calculations based on MRHS 2009 from the World Bank. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi 266 Elisenda Estruch et al. 9.3.3 Are Youth Driving Rural Transformation? A gradual rural transformation is taking place in Senegal as shown by the decrease in agricultural employment and the increase in the RNFE, although the latter is still small and involves few people. The young population nevertheless leads this process of rural transformation. However, when looking at the type of employment that this process entails, the increase in employment in the RNFE has been entirely associated with more unpaid family work, indicating that youth are being pushed more out of agriculture than attracted towards productive activities in the RNFE. The RNFE is dominated by wholesale and retail trade, by repair of motor ve hicles, motorcycles, and household goods, followed by manufacturing, real estate, and construction. Between 2001 and 2011 wholesale and retail trade increased from 7.8 per cent to 11.4 of rural youth total employment and manufacturing from 2.1 per cent to 4.5 per cent, increasing employment opportunities for youth in rural areas. Comparing households receiving and not receiving remittances, data show that households with youth and receiving remittances are more likely to combine farm and nonfarm employment than households with youth not receiving remittances, indicating that the presence of youth influences the decision to use remittances to diversify the households’ income. These income flows allow the recipient households to diversify into the non-agricultural sector by dedicating productive members, most often youth, to these activities (see Table9.A4 in the Annex). Most of the households with youth receiving remittances are nonfarm households. This logic is twofold: on the one hand, nonfarm households are eco nom ic al ly wealthier than other types of households so they have more resources to engage in migration; on the other hand, remittances sent back to these households strengthen their economic position which facilitates the transitions from agriculture to the RNFE. 9.4 Agricultural Policies Influencing Youth and Rural Transformation in Senegal The employment outcomes of rural youth in Senegal are in part a consequence of the agricultural, employment, and rural development policies of the Senegalese government over the years. Policy interventions have an impact on rural labour markets as well as on agricultural productivity, shaping the challenges and op portun ities rural youth face. During the French colonization, the Senegalese agricultural sector was strongly state-guided. After the country’s declaration of independence in 1960, the agricultural sector struggled to adapt, resulting in a deterioration of the sector and an OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Youth Mobility and its Role in Senegal 267 elevated dependency on imported food (Oya and Ba2013). After decades of weak performance in the agricultural sector, as of the year 2000, the sector regained national attention and a more active agricultural policy was put in place in order to economically revive the sector. As such, in the following years a set of large-scale policies were adopted, namely: • the Agro-Sylvo-Pastoral Orientation Act (LOASP), which is the legal framework for the development of agriculture and the reduction of poverty in rural areas (2004); • the Return to Agriculture plan (REVA) to promote youth employment in agriculture and avoid distress migration (2006); • annual programs of structural adjustment in an attempt to reduce the domin ance of the groundnut production and diversify agricultural production (2003–7); and • the Great Push Forward for Agriculture, Food and Abundance programme (GOANA), aimed to increase domestic production of Senegal’s main food and export crops (especially rice) and achieve selfsufficiency and food security by 2015 (2008). In December 2013, the government launched the Emerging Senegal Plan (PSE) which has been the reference for economic and social policy in the medium and long term, with the aim of making Senegal an emerging economy by 2035. Job creation is a key priority for the PSE and the plan envisages increasing the decent work opportunities at the rate of 100,000 to 150,000 new jobs per year. In line with the PSE, new programmes and projects were initiated, including • the Accelerated Programme for Agriculture in Senegal (PRACAS) which is the agricultural component of the PSE and the most important programme for agriculture (2014); • the Programme to Promote Youth and Female Employment (PAPEJF), which aims to contribute to the creation of decent jobs for youth and women (2013); and • the Project to Support the Promotion of Entrepreneurship of Youth in Rural Areas (PAJER) specifically targeting youth, entrepreneurship, and value chains (2015). Acknowledging the specific lack of employment opportunities for youth, the National Agency for the Promotion of Youth Employment (ANPEJ) was created in 2014 to coordinate all youth-employment-related policies, envisaging policy coherence and efficient action. Moreover, a specific policy to promote youth employment in rural areas (PPEJMR, Politique de Promotion de l’Emploi des Jeunes en Milieu Rural) was under formulation in 2016 with the aim of addressing OUP CORRECTED PROOF – FINAL, 29/10/19, SPi 268 Elisenda Estruch et al. rural youth’s lack of access to labour market information, productive resources, and entrepreneurial skills. Between 2000 and 2012, many of the agricultural policies were ambitious and lacked policy coherence, which hampered the implementation. Moreover, there was more attention on agribusiness development, neglecting the needs of smallholders and rural youth (who often encounter obstacles in accessing land, credit, and so on). Also, migration policy was strongly focused on preventing rural migration instead of efficiently managing the flows of migrants and the remittances to rural areas (Oya and Ba2013, Antil2010, Banque Mondiale2006). As shown in previous sections, the employment situation of rural youth did not improve significantly between 2000 and 2012. On the one hand, employment in agriculture decreased but this decrease was not offset by more productive jobs in the RNFE. The policies and programmes in place would have neither led to a diversification of agricultural production, nor did the country become self-sufficient in rice and other staple crops. On the other hand, poverty decreased in rural areas but farmers and households dependent on agriculture have less secure and more vulnerable employment and have a higher incidence of poverty than the rest of the population. Since 2012, the new government has been implementing what appears to be a more coherent policy framework for the agricultural sector and rural development. The needs of rural youth and their employment challenges are specifically addressed in many programmes and projects such as the ones described above. As they are relatively recent, it will take some time to assess their actual implementation and impacts. 9.5 Conclusions In Senegal, low agricultural productivity growth, fuelled by underinvestment in the sector and an unbalanced rural labour market, has led to a stagnant agricultural sector which has hampered the performance of the wider rural economy. Accordingly, rural economic development, in combination with a high population growth, has been too limited to enable significant results in reducing poverty in rural areas. This situation is creating major challenges for the rural population, inparticular rural youth, to access productive and decent jobs. The large majority of the rural population in Senegal is employed in agriculture, albeit there is a general downward trend in agricultural employment, particularly among rural youth. Data analysis shows that there has been a considerable re alloca tion of agricultural labour into urban areas and abroad, and to a lesser extent into the rural nonfarm sector. These patterns would indicate that a process of rural transformation is taking place in the country. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Youth Mobility and its Role in Senegal 269 This gradual rural transformation is mainly brought forward by youth, who are entering the labour market, as their share of employment in the RNFE is growing faster than that of their older counterparts. However, some concerns emerge as the type of employment opportunities are being found in rural areas, as they are increasingly working as non-paid family workers in the RNFE, indicating that they are pushed out of agriculture rather than pulled by remunerative and decent employment opportunities in the nonfarm economy. Rural youth and their families are increasingly diversifying incomes by en gaging in the agricultural and the nonfarm sector. Education and wealth of the family are key factors driving such diversification at household level: more educated youth within wealthier families are more prone to engage in the RNFE than less educated youth from less wealthy families. Hence, complementing or substituting income with rural nonfarm employment is a livelihood strategy providing a pathway out of poverty. Many Senegalese youth are leaving the rural areas altogether and migrating to urban areas or abroad. Once again, higher education and household wealth play akey role in determining who undertakes this pathway. Those who are more educated and wealthier are moving to urban areas or abroad, while the least educated youth from the poorest households are more likely to migrate to another rural area to find employment in the nonfarm economy. Youth who exit agriculture by migrating internally are most likely to find a job as self-employed in the nonfarm economy in urban areas. In the past, the Senegalese government’s actions to tackle structural problems and lift agricultural productivity in the rural economy have had insufficient results in terms of agricultural productivity and job creation, especially of youth. Conversely, a new generation of policies and programmes seems to be now in place, targeting rural youth’s needs associated with financial support and a more supportive and stable policy environment. Developing the human capital of rural youth has become a priority for the government, although it is still necessary to further focus on enhancing the labour demand in rural areas. To conclude, there remains a large untapped potential in the rural economy to boost opportunities for rural youth. Efforts to promote farmers’ productivity and incomes would also be helpful, as the skills of the agricultural workforce, especially youth, are crucial to tap into this potential. Moreover, higher labour prod uct iv ity reduces vulnerability and vice versa, and thus contributes to rural poverty reduction. At the same time, as young people may also exit agriculture, more jobs need to be created in off-farm agriculture-related activities. Development of midstream and downstream value chains promotes off-farm employment, providing opportunities for inclusive rural transformation. The participation of young entrepreneurs should be ensured in this process of transformation, as well as the support for the development of agro-industry and OUP CORRECTED PROOF – FINAL, 29/10/19, SPi 270 Elisenda Estruch et al. infrastructure to better connect rural and urban areas. Improved connections between rural areas and nearby small urban centres can play a catalytic role in mediating the rural–urban nexus and providing small-scale producers with greater opportunities, as well as serving as hubs for a thriving nonfarm sector. It is therefore essential to support the dynamic development of employment opportunities. Inclusive policies and more targeted investments in agricultural and rural development, leading to rural transformation, should be further strengthened to promote more and better opportunities for the younger generations, while at the same time helping to end rural poverty in Senegal. Annex: Data Description For the data analysis of youth employment in the rural nonfarm economy (RNFE), two nationally representative cross-sectional surveys were used: the second ‘Senegal Household Survey’ (ESAM-II), conducted in February–April 2001, and the ‘Senegal Poverty Monitoring Survey’ (ESPS-II), conducted in August–December 2011. Both surveys were implemented by the Senegalese National Agency for Statistics and Demography (ANSD) in cooperation with the World Bank and other UN agencies (United Nations Development Programme, World Food Programme, and ILO). The surveys include questions on household composition, education, household welfare, labour characteristics, sources of income, and more specific farm features such as land and technology. ESAM-II pooled 6,624 households of which 3,240 were rural. ESPS-II surveyed in total 17,891 households of which 7,560 were rural. To account for seasonal variation in agriculture, a 12-month recall period was used. Only a subsample (one-third) of the sampled households were administered the household consumption expenditure module that ESPS-II featured. Therefore, a separate household asset index that covered the entire sample was created to measure household wealth (following Sahn and Stifel2003). The data used to create the ‘commune size’ variable was retracted from ANSD. For the migration analysis, the ‘Migration and Remittances Household Survey’ (MRHS) implemented by the World Bank in 2009 was used (World Bank2009a). The survey was conducted both at household and individual level and addresses labour market status, expenditure, motivation for migration, remittances, and so on. The MRHS surveyed a total of 17,878 individuals and 1,983 households, ofwhich 36 per cent had no migrants, 30per cent had internal migrants, and 34per cent had international migrants. For the purpose of this analysis, youth are defined as people between 15 and 34 years old and the working population is defined at 15–65 years old. All data allows controlling for rural/urban areas and sectors, as well as for main household and individual characteristics. All structure and growth figures reported are derived from weighted data. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Youth Mobility and its Role in Senegal 271 Table 9.A1. Individual characteristics of rural youth employed according to sector ofemployment 2011 Rural areas (12-month reference) Youth employed inagriculture (N = 10,021) Youth employed in rural nonfarm economy(N = 2233) T-test Mean SE Mean SE Individual characteristics Age 23.05 0.06 25.39 0.11 Female (%) 0.43 0.00 0.35 0.01 *** Married (%) 0.54 0.00 0.49 0.01 *** Years of schooling 1.70 0.03 3.11 0.09 Household characteristics Age of HH head 52.13 0.15 50.39 0.33 *** Female HH head (%) 0.07 0.00 0.20 0.01 Years of schooling of HH head 0.81 0.02 1.94 0.08 HH size 14.38 0.08 12.54 0.14 *** Age dependency ratio 1.11 0.01 0.96 0.01 *** Wealth index (%) – Poorest tercile 0.65 0.00 0.36 0.01 *** – Middle tercile 0.25 0.00 0.35 0.01 – Richest tercile 0.11 0.00 0.29 0.01 Farm size (ha) 6.50 0.08 4.07 0.11 *** Land tenure (%) HH land owned 0.89 0.00 0.68 0.01 *** HH land rented in 0.07 0.00 0.04 0.00 *** HH land used for free 0.02 0.00 0.25 0.01 * Distance to market (in km) 61.45 0.67 28.70 0.95 *** Community characteristics Population commune 6,2303 592.28 9,9709 3561.00 Note: (a) Paired student t-test of youth employed in agriculture versus without youth employed in the rural nonfarm economy; (b) *** means statistical significance at 1% level, * means statistical significance at 10% level. Source: Authors’ calculations based on ESPS-II,2011. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Table 9.A2. Households with youth characteristics according to sector of employment Households with youth Farming only HH (N = 3636) Mixed farm—nonfarm HH (N = 1961) Nonfarm only HH(N = 1602) Paired t-test Mean SE Mean SE Mean SE Farming-only versus mixed Mixed versus. nonfarm Age of HHhead 50.43 0.24 52.64 0.32 48.76 0.37 *** Female HHhead 0.10 0.00 0.11 0.01 0.24 0.01 Married HHhead 0.92 0.00 0.93 0.01 0.86 0.01 *** Years of schooling ofHHhead 0.81 0.04 1.29 0.07 2.41 0.11 Household size 10.47 0.09 12.93 0.15 9.40 0.13 *** Age dependency ratio 1.27 0.01 1.15 0.01 1.14 0.02 *** Wealth index – Poorest tercile 0.71 0.01 0.50 0.01 0.36 0.01 *** *** – Middle tercile 0.21 0.01 0.34 0.01 0.36 0.01 – Richest tercile 0.07 0.00 0.16 0.01 0.29 0.01 Farm size (ha) 4.81 0.09 4.81 0.13 3.28 0.10 *** Land tenure – HH land owned 0.86 0.01 0.83 0.01 0.58 0.02 *** *** – HH land rented in 0.05 0.00 0.06 0.01 0.02 0.00 *** – HH land sharecropped 0.01 0.00 0.02 0.00 0.01 0.00 – HH land used for free 0.07 0.01 0.09 0.01 0.38 0.01 ** HH uses fertilizer 0.36 0.01 0.38 0.01 HH uses agricultural equipment 0.49 0.01 0.45 0.01 *** Distance to market (inkm) 62 1.13 39 1.08 24 0.90 *** *** Population of commune 61,550 1,164 70,397 1,709 110,356 4,816 Note: (a) *** means statistical significance at 1% level, ** means statistical significance at 5% level. Source: Authors’ calculations based on ESPS-II,2011. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Youth Mobility and its Role in Senegal 273 Table 9.A4. Sector of employment of rural households by remittances status HH with youth Receiving remittances N = 1187 Not receiving remittances N = 651 T-test Mean SE Mean SE Farming only HH 0.20 0.01 0.26 0.02 ** Mixed farm HH 0.19 0.01 0.16 0.01 Nonfarm HH 0.51 0.02 0.56 0.02 *** Notes: (a) Paired student t-test of HH with 15–34 receiving remittances versus HH with 15–34 not receiving remittances; (b) ***means statistical significance at 1% level, ** means statistical significance at 5% level, * means statistical significance at 10% level. Source: Authors’ calculations based on MRHS 2009 from the World Bank. Table 9.A3. Characteristics of migrated youth and their households according to labour transition Youth (15–34 years old) migrants from rural areas Agriculture to rural nonfarm economy N = 70 Other transition from rural areas N = 258 T-test Individual characteristics of migrants before migration Mean SE Mean SE Female 0.03 0.02 0.23 0.03 *** Age of departure 22.27 0.65 20.33 0.39 ** Years of schooling before migration 1.65 0.35 4.34 0.38 *** Characteristics of former household of migrants Average household size 11.31 0.77 12.72 0.43 Average monthly per capitaexpenditure 8,348 540 11,967 1,178 Commune population 46,257 14,664 187,143 300,692 *** Notes: (a) Paired student t-test of youth labour transition agriculture to rural nonfarm economy versus youth other labour transition; (b) *** means statistical significance at 1% level, ** means statistical significance at 5% level. Source: Authors’ calculations based on MRHS 2009 from the World Bank. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi 280 Valerie Mueller, Gracie Rosenbach, and James Thurlow surprising, since younger farmers are often expected to be more open to new technologies and knowledge than older famers are, even when they have similar levels of schooling. Overall, the new evidence presented in the country chapters suggests that it does not necessarily follow that having a younger population leads to greater adoption of improved farm technologies. Moreover, while education is strongly associated with the use of improved technologies, at least in our case study countries, the effect of input use on farm productivity may still be relatively small. This could explain why agricultural productivity growth in Africa has remained sluggish, despite substantial improvements in educational attainment. 10.2.3 Youth are More Likely to Engage in Rural Nonfarm Activities, but the Level of Off Farm Employment Remains Low and Most Youth Continue to Work in Agriculture History suggests that higher agricultural productivity and commercialization leads to an expansion of the rural nonfarm economy, including the processing and trading of agriculture-related products, as well as other kinds of occupations that arise to serve nonfarm workers as they concentrate around rural markets. Thus, as agricultural transformation progresses, we expect to see more farm households diversify into nonfarm activities or even specialize in off-farm work. This process should create new job opportunities for rural youth, especially those without access to farm land. The prevailing view is that youth themselves, by being better educated and less inclined to work on the farm, may be wellpositioned to establish rural businesses and drive this stage of agricultural transformation. New evidence from our country case studies confirms that youth are generally more likely than adults to be employed in off-farm jobs. However, the extent to which this is true varies across countries, and its implications for rural trans - forma tion are unclear. Chapter 5 on Ethiopia, for example, found that youth aged 25–34 years have a greater probability of working in nonfarm enterprises, but that the size of the nonfarm sector remains extremely small. While youth are driving growth in off-farm employment, the nonfarm sector itself is not a significant driver of rural transformation in Ethiopia. Similarly, Chapter 6 on Malawi found little evidence of any significant process of rural transformation or of youth being in the vanguard of any changes in employment patterns. Malawian youth are more likely to extend their schooling than start new businesses in the rural nonfarm economy, and once Malawian youth leave school, they are still more likely to work in agriculture. Chapter 7 on Ghana is more supportive of the prevailing view. It found that youth are far more likely than adults to run nonfarm enterprises, especially in the OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Conclusion 281 less developed northern region of the country. However, Ghana is transforming rapidly and many households, irrespective of the age of their members, are leaving agriculture to work in the rural nonfarm economy. What is concerning, however, is that young Ghanaians are less likely than adults to find work in the formal sector, and instead engage in informal trading and other low-productivity occupations. Ghana’s youth are participating in the country’s structural change, but it is adults who are benefiting more from this process. As in Ethiopia and Malawi, the Senegalese case study in Chapter 9 found that youth who leave agriculture are often pushed into unpaid family work in rural nonfarm enterprises, rather than being pulled into more remunerative and decent off-farm employment op por tunities. Perhaps for this reason, most Senegalese youth who leave agriculture go in search of employment outside of the country. Chapter 8 on Tanzania used firm-level data to examine the factors that determine the success of rural nonfarm enterprises. As in the other case studies, the authors found that youth are more likely than adults to engage, and even specialize, in rural nonfarm activities. However, amongst rural firms, those businesses that are run by adults are generally more productive than those run by youth. Overall, the five case study chapters reveal considerable differences across countries, but they all conclude that, even when youth are more likely to work in the rural nonfarm economy, they are also more likely to have low productivity jobs in the informal sector or run less successful nonfarm businesses. This limits the contributions of youth to rural transformation in these countries. 10.2.4 Education is Important for Rural Nonfarm Employment, but on Its Own it is not Enough to Ensure Success A common finding across all five case study chapters is that people working in nonfarm jobs tend to have more years of schooling. This is often seen as one of the reasons why youth are more likely to work off the farm. Chapter 8 on Tanzania, for example, found that households whose heads have at least secondary education are as much as 38 per cent more likely to engage in nonfarm ac tivities. However, the returns to education may be lower for youth than for adults. As mentioned above, adults in Tanzania are far more likely to work in more product ive sectors or run more successful nonfarm businesses. These households are also less likely to be poor. Moreover, all the case studies identify factors other than education that are also important determinants for participation and success in off-farm work. Chapter 7 on Ghana, for example, found that better access to markets, public transportation and electricity increase the likelihood of a rural household working in the nonagricultural sector, regardless of whether the household head is young or educated. Chapter 6 on Malawi found that education alone is insufcient to en able youth to OUP CORRECTED PROOF – FINAL, 29/10/19, SPi 282 Valerie Mueller, Gracie Rosenbach, and James Thurlow obtain nonfarm employment. The authors found that older Malawians are more likely to work in the nonfarm sector, and from this they concluded that capital accumulation, work experience, and the development of social and economic networks are more important than education in enabling in di vid uals to find work outside of agriculture. Chapter 5 on Ethiopia also found that younger youth (aged 15–24) are no more likely to engage in nonfarm work. The authors concluded from this that work experience and social networks may be important, beyond just education levels, which are higher for this age group. Thus, while the higher education levels of youth are undoubtedly an asset, there is no guarantee that youth will lead or benefit from the rise in the rural nonfarm economy. Investing in education is therefore necessary but insufcient. Investments in infrastructure and market development are also important to ensure that youth (and adults) can participate in the process of agricultural transformation. 10.2.5 Participation in the Rural Nonfarm Economy Differs for Young Men and Women Evidence on the roles of young men and women in the agricultural trans for mation process is mixed. On the one hand, Chapter 6 found that males dominate off-farm employment in Malawi, whereas women are more likely to remain in agriculture. Interestingly, the authors find that women who recently gave birth are less likely to engage in nonfarm employment of any sort, whereas infant care does not appear to affect the extent to which women engage in farm work. Chapter 9 also found that young men in Senegal are more likely than young women to be employed in the rural nonfarm economy, possibly reflecting the large education gap between men and women in this country. As in Malawi, Senegalese men were also found to be more likely than women to emigrate in search of work. In contrast, Chapter 7 found that being a female-headed rural household in Ghana increased the probability that the household engages in nonfarm ac tivities, although this relationship weakens over time. Although Chapter 8 found that men and women in Tanzania are almost equally likely to work in the rural nonfarm economy, the authors also found that women are less likely to run more productive or successful nonfarm enterprises. Female-run enterprises in Tanzania tend to be in lower-value manufacturing, such as food processing, rather than higher-value services. It should be noted, however, that these gender differences are less important than other factors, such as education, in explaining why some businesses are successful. One implication from this is that investing in women’s education in Tanzania, and possibly elsewhere, should help close the gender gap by allowing young women to participate in and benefit more from agricultural transformation. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Conclusion 283 10.2.6 Youth are More Likely to Migrate, but not Always for Work Reasons or to Urban Areas Agricultural transformation is initially driven by increases in agricultural product iv ity and then by deepening farm-nonfarm linkages as farmers commercialize and rural markets become more important. Eventually, however, the focus shifts to supplying urban consumers in fast-growing cities and towns. At this stage, workers in rural areas may decide to migrate in search of better job op por tun ities in urban areas. Agriculture continues to be the core of the rural economy, but the national economy is increasingly driven by urban development. The prevailing view is that, given land shortages in rural areas and higher returns to education in urban areas, youth are more likely than adults to migrate, and may therefore contribute more to economic growth and structural change. Chapter 2 used household panel survey data to examine the pattern of migration in four African countries, and to evaluate what determines youth decisions to migrate in two of these countries. The author found that younger people are more likely to migrate. However, most migration is between rural areas, rather than to cities and towns, and the main reasons for migrating are often not work-related. Rural–rural migration, for example, is higher amongst women and is often mo tivated by marriage. Migration distances are also quite short. In contrast, workers who move to urban areas are not only more likely to claim moving for employment reasons, but they also travel far greater distances. That said, there is considerable variation across countries. In Ethiopia, for example, many young migrants to urban areas move to attend secondary schools that are unavailable in rural areas. Migration generally leads to greater income diversification, but new migrants often work in agriculture before finding employment in the nonfarm sector. Chapter2 found that, in Malawi and Tanzania, migration to urban destinations offered more employment opportunities for youth in the non-agricultural sector. However, while the probability of finding a job in a high-return activity is higher for youth who move to urban areas, this does come with greater risk of becoming unemployed. In contrast, intra-rural migration also promotes income diversification, but migrants are less likely to work exclusively in rural nonfarm jobs, and the income gains from migration are smaller. The author of Chapter2 concludes that rural–urban migration in Malawi and Tanzania facilitates the movement out of agriculture, usually into higher-return activities. However, only a small share of rural youth become rural–urban migrants. In contrast, far more youth are likely to migrate between rural areas and so this is the more formative mobility pattern in the transformation process. It is the main driver of income diversification amongst youth, although it rarely involves a shift into exclusive non-agricultural employment. This means that, while youth are more likely to urbanize than adults, the importance of this for youth and for structural change should not be overstated. OUP CORRECTED PROOF – FINAL, 29/10/19, SPi 284 Valerie Mueller, Gracie Rosenbach, and James Thurlow 10.2.7 Rural Nonfarm Job Opportunities are Better Closer to Bigger Cities, but Jobs in the Food System may be More Important Closer to Smaller Towns As mentioned above, the case studies found that access to urban markets is a major factor in determining the likelihood of youth working in the rural nonfarm sector. However, two of the case studies also found that it is important to differentiate between urban centres of different sizes. Chapter 6 on Malawi, for example, found that better access to large urban centres of 50,000 people or more is strongly associated with nonfarm employment, but that proximity to smaller urban centres had little influence on the employment choices of youth in surrounding rural areas. The authors concluded that smaller towns have less of a role to play in changing labour patterns and contributing to structural change in Malawi. Chapter 7 on Ghana conducted more detailed analysis of how urbanization affects employment outcomes for youth and adults. The authors found that proximity to large cities in the south of the country greatly increased the likelihood of a household engaging exclusively in non-agricultural work. The authors also found that manufacturing is more dominant in areas that are less urbanized (i.e. closer to smaller towns than bigger cities). This is because informal manufacturing in these rural areas primarily consists of food processing for local markets, which can take place at the household level. In contrast, households living closer to big cities aremore likely to have members employed in the service sector, including jobs in the formal sector and outside of the agriculture-food system. Unlike in Malawi, the authors conclude that smaller towns in Ghana are im port ant for promoting youth employment in rural areas. The difference between the two countries may be that economic growth and urbanization are slower in Malawi and less of the country’s rural population live in peri-urban areas. In fact, Malawi’s rural population density is one of the highest in Africa. For these reasons, the linkages between small towns and rural areas in Malawi may be weaker, or less important, than they are in Ghana. 10.2.8 Youth are Only Slightly More Likely to Protest than Adults, but They are More Likely to be Driven by Concerns about Unemployment Some of the concerns about youth employment in Africa stem from the view that underemployed youth are especially prone to anti-government behaviour, including public protests and violence. A contrasting view is youth are better educated today and so may place more demands on their governments to enact policy reforms that address employment issues. Chapter 4 examined the political participation of youth using historical data on local protests and household surveys from 16 African countries. The author asked whether youth are more OUP CORRECTED PROOF – FINAL, 29/10/19, SPi Conclusion 285 likely to protest than adults, and if the issues that motivate youth have changed over time? She found that youth are more likely to protest than adults, but that the gap is extremely small, suggesting that concerns about youth protest may be overstated. Protest activity is a form of mobilization used in almost equal measure by both age groups. For young and old, being better educated and/or poor are strong motivators for protest. However, the author found that young people are also more likely to protest if they are unemployed and if they lack trust in political institutions. If governments in Africa wish to avoid protests, then youth employment needs to be a high priority, and job creation projects need to match young people’s skills and aspirations. Governments need to generate greater trust that youth policies and initiatives are aimed at enhancing youth’s long-term economic prospects rather than simply mobilizing their short-term political support. 10.2.9 Youth Employment is Now a Major Policy Goal, but the Means of Achieving this Goal are not Well Represented in Current Policies Creating more and better jobs for youth is a major policy priority for most African countries today. This differs from the early-2000s, when policies often focused on poverty reduction rather than job creation (i.e. on ‘pro-poor’ rather than ‘inclusive’ growth). However, making youth employment a policy goal does not necessarily mean that national policies include the kinds of interventions needed to promote youth employment. Chapter 3 developed a framework for systematically classifying policies based on whether they adequately address key constraints to youth employment in rural areas. The authors applied the framework to 47 national, rural and agricultural policies in 13 African countries. They found that policies tend to be strongest on labour supply issues, such as self-employment and skills development. Most national policies, for example, emphasize rural education as a means of improving the prospects of young job seekers. Policies are much weaker on labour demand issues, such as how to stimulate private sector job creation in the agriculture-food system beyond the farm—an area that the country case study chapters identified as being particularly important for rural youth. Industrial policies, for example, rarely identify concrete interventions for private sector development or discuss how demand for young rural workers will be incentivized. As the case study chapters found, rural youth are far less likely than adults to be employed in the private sector, and so national policies should explicitly support informal businesses in rural and peri-urban areas. The authors found that social and policy dialogue is the weakest area in the design of national policies. A lack of participation by rural youth in the policy process means that their specific needs are given insufcient attention. For [Document text truncated for crawler view.]