An Evolving Paradigm of Agricultural Mechanization Development A n Evolving Paradigm of A gricultural M echanization D evelopm ent H ow M uch Can Africa Learn from A sia? Edited by Xinshen Diao, Hiroyuki Takeshima, and Xiaobo Zhang How Much Can Africa Learn from Asia? About IFPRI The International Food Policy Research Institute (IFPRI), a CGIAR Research Center established in 1975, provides research-based policy solutions to sustainably reduce poverty and end hunger and malnutrition. IFPRI’s stra- tegic research aims to foster a climate-resilient and sustainable food supply; promote healthy diets and nutrition for all; build inclusive and efficient mar- kets, trade systems, and food industries; transform agricultural and rural econ- omies; and strengthen institutions and governance. Gender is integrated in all the Institute’s work. Partnerships, communications, capacity strengthening, and data and knowledge management are essential components to translate IFPRI’s research from action to impact. The Institute’s regional and country programs play a critical role in responding to demand for food policy research and in delivering holistic support for country-led development. IFPRI collabo- rates with partners around the world. About IFPRI’s Peer Review Process IFPRI books are policy-relevant publications based on original and innova- tive research conducted at IFPRI. All manuscripts submitted for publica- tion as IFPRI books undergo an extensive review procedure that is managed by IFPRI’s Publications Review Committee (PRC). Upon submission to the PRC, the manuscript is reviewed by a PRC member. Once the manuscript is considered ready for external review, the PRC submits it to at least two exter- nal reviewers who are chosen for their familiarity with the subject matter and the country setting. Upon receipt of these blind external peer reviews, the PRC provides the author with an editorial decision and, when necessary, instructions for revision based on the external reviews. The PRC reassesses the revised manuscript and makes a recommendation regarding publication to the director general of IFPRI. With the director general’s approval, the manu- script enters the editorial and production phase to become an IFPRI book. An Evolving Paradigm of Agricultural Mechanization Development How Much Can Africa Learn from Asia? Edited by Xinshen Diao, Hiroyuki Takeshima, and Xiaobo Zhang A Peer-Reviewed Publication International Food Policy Research Institute Washington, DC Copyright © 2020 International Food Policy Research Institute (IFPRI). This publication is licensed for use under a Creative Commons Attribution 4.0 International License (CC BY 4.0). Subject to attribution, you are free to share (copy and redistribute the material in any medium or format), adapt (remix, transform, and build upon the material) for any purpose, even commercially. Third-party content: The International Food Policy Research Institute does not necessarily own each component of the content contained within the work. The International Food Policy Research Institute therefore does not warrant that the use of any third-party-owned individual component or part contained in the work will not infringe on the rights of those third parties. The risk of claims resulting from such infringement rests solely with you. If you wish to re-use a component of the work, it is your responsibility to determine whether permission is needed for that re-use and to obtain permission from the copyright owner. Examples of components can include, but are not limited to, tables, figures, or images. Recommended citation: Diao, X., H. Takeshima, and X. Zhang, eds. 2020. An Evolving Paradigm of Agricultural Mechanization Development: How Much Can Africa Learn from Asia? Washington, DC: International Food Policy Research Institute. https://doi.org/10.2499/9780896293809. This is a peer-reviewed publication. Any opinions expressed herein are those of the authors and are not necessarily representative of or endorsed by the International Food Policy Research Institute (IFPRI). The boundaries and names shown and the designations used on the maps do not imply official endorsement or acceptance by IFPRI. International Food Policy Research Institute, 1201 Eye Street, NW, 12th floor, Washington, DC 20005 USA, Telephone: +1-202-862-5600, www.ifpri.org ISBN: 978-0-89629-380-9 DOI: https://doi.org/10.2499/9780896293809 Library of Congress Cataloging-in-Publication Data may be found on page viii. CONTENTS List of Tables ix List of Figures xvii Abbreviations and Acronyms xxiii Foreword xxvii Abstract xxix Acknowledgments xxxi Part 1: Synthesis of the Lessons Chapter 1 An Evolving Paradigm for Africa and Synthesis of the Lessons from Asia 3 Xinshen Diao, Jed Silver, Hiroyuki Takeshima, and Xiaobo Zhang Part 2: Early-Adopter Asian Countries Chapter 2 Mechanization Outsourcing Clusters and Division of Labor in Chinese Agriculture 71 Xiaobo Zhang, Jin Yang, and Thomas Reardon Chapter 3 Farm Machinery Use and the Agricultural Machinery Industries in India : Status, Evolution, Implications, and Lessons Learned 97 Madhusudan Bhattarai, Gajendra Singh, Hiroyuki Takeshima, and Ravindra S. Shekhawat Chapter 4 The Evolution of Agricultural Mechanization in Sri Lanka 139 Fredrick Abeyratne and Hiroyuki Takeshima Chapter 5 Evolution of Agricultural Mechanization in Thailand 165 Rob Cramb and Viboon Thepent Chapter 6 Evolution of Agricultural Mechanization in Viet Nam 203 Hiroyuki Takeshima, Yanyan Liu, Nguyen Van Cuong, and Ian Masias Part 3: Late-Adopter Asian Countries Chapter 7 Evolution of Agricultural Mechanization in Bangladesh : The Case of Tractors for Land Preparation 235 Mansur Ahmed and Hiroyuki Takeshima Chapter 8 Myanmar’s Rapid Agricultural Mechanization: Demand and Supply Evidence 263 Myat Thida Win, Ben Belton, and Xiaobo Zhang Chapter 9 Evolution of Agricultural Mechanization in Nepal 285 Hiroyuki Takeshima and Scott E. Justice Part 4: African Countries Chapter 10 The Rapid—but from a Low Base—Uptake of Agricultural Mechanization in Ethiopia: Patterns, Implications, and Challenges 329 Guush Berhane, Mekdim Dereje, Bart Minten, and Seneshaw Tamru vi Contents Chapter 11 Agricultural Mechanization in Ghana: Alternative Supply Models for Tractor Hiring Services 377 Xinshen Diao and Hiroyuki Takeshima Chapter 12 Evolution of Agricultural Mechanization in Kenya 401 Hugo De Groote, Cliff Marangu, and Zachary M. Gitonga Chapter 13 Evolution of Agricultural Mechanization in Nigeria 423 Hiroyuki Takeshima and Akeem Lawal Chapter 14 Agricultural Mechanization in Tanzania 457 Geoffrey C. Mrema, David G. Kahan, and Andrew Agyei-Holmes Contributors 497 Index 501 Contents vii Library of Congress Cataloging-in-Publication Data Names: Diao, Xinshen, editor. | Takeshima, Hiroyuki, editor. | Zhang, Xiaobo, 1966- editor. Title: An evolving paradigm of agricultural mechanization development : how much can Africa learn from Asia? / edited by Xinshen Diao, Hiroyuki Takeshima, Xiaobo Zhang. Description: Washington, DC : International Food Policy Research Institute, 2020. | Includes bibliographical references and index. Identifiers: LCCN 2020025874 (print) | LCCN 2020025875 (ebook) | ISBN 9780896293809 (paperback) | ISBN 9780896293816 (epub) Subjects: LCSH: Farm mechanization—Government policy—Africa. | Farm mechanization—Government policy—Asia. Classification: LCC S760.A35 A54 2020 (print) | LCC S760.A35 (ebook) | DDC 338.1/61096—dc23 LC record available at https://lccn.loc.gov/2020025874 LC ebook record available at https://lccn.loc.gov/2020025875 ix LIST OF TABLES 1.1 Key aspects of theoretical framework covered in country chapters 24 1.2 Focus of empirical analyses in selected chapters 25 1.3 Share (percentage) of agricultural area with different levels of soil workability constraints 38 2.1 Average agricultural production and input at the household level, China, 2009–2012 79 2.2 Ordinary least squares estimation of production function for three crops, China, 2009–2012 80 2.3 Estimation of production function for three crops using Levinsohn and Petrin method, China, 2009–2012 81 2.4 Use of machinery in Chinese agricultural production, percentages, 2008 and 2012 85 2.5 Summary statistics of combine service enterprise survey in Peixian county, China, 2013 91 3.1 Change in use of farm machinery and related factors of Indian agriculture, 1960–2012/2013 99 3.2 Trends of farm holdings in India, 1983 to 2010/2011 101 3.3 Distribution of tractors across selected states in India, 1982 and 2012 101 3.4 Major farm machinery used and annual market size of farm machinery in India, 2014/2015 102 3.5 Percentage of mechanization by major crop and by operation in India, 2013 105 3.6 Correlation coefficients between tractor density and selected factors in India, 1982 and 2012 107 3.7 Leading tractor manufacturers in India, 2009/2010 to 2015/2016 116 3.8 Status of farm mechanization industry in India, 2014 116 3.9 Tractor investment and farm size dynamics in semi-arid villages in India, 2001–2014 121 3.10 Effects of tractor use for land preparation on the use of human labor and animal (bullock) traction for land preparation (per year, all production seasons combined) 123 3.11 Effects of combine harvester use on the use of human labor for harvesting and threshing (per year, all production seasons combined) 124 3.12 Effects of tractor use on overall labor and animal use at the household level 125 3.13 Effects of tractor use for land preparation on the use of chemical fertilizer per acre (per year, all production seasons combined) 126 3.14 Yield effects of tractor and combine harvester use on key crops 127 3.15 Wage rates for agricultural labor (by sex), bullocks, and tractors in a typical dryland village in Andhra Pradesh, India, 2011 128 4.1 Evolution of economic structure, employment structure, and mechanization in Sri Lanka, 1960s–2010s 142 4.2 Declining farm sizes in Sri Lanka, 1960–2002 147 4.3 Holding size and use of machinery for paddy land preparation and harvesting, selected districts, Sri Lanka, 2013/2014 148 4.4 Labor and machinery costs for selected crops, Sri Lanka, 1979/1980–2013 150 4.5 Power sources used for selected operations in maize production in Monaragala district 153 4.6 Awareness, affordability, and use of farm implements for land preparation in finger millet production in study locations 153 x List of tabLes 4.7 Annual sales of farm machinery, Sri Lanka, 2011/2012– 2013/2014 155 5.1 Number of holdings using farm machinery and equipment by source, Thailand, 2013 168 5.2 Holding size and land use by region, Thailand, 2013 172 5.3 Number of agricultural machinery businesses, Thailand, 2009 186 5.4 Production of agricultural machinery in Thailand, 2001 and 2012 187 6.1 Mechanization level and economic structure in Viet Nam, 1960–2020 206 6.2 Proportion of machine rental by region, Viet Nam, 1992–2008 208 6.3 Median total land cultivated per household (in square meters) by region, Viet Nam, 1992–2008 208 6.4 Tractor ownership by region, Viet Nam, 1992–2008 209 6.5 Tractors per 1,000 farm households, by region, Viet Nam, 2007 209 6.6 Real median daily wage of male agricultural labor, Viet Nam, 1992–2008 (’000 VND) 217 6.7 Regression results on machine use 218 7.1 Evolution of economic and employment structure and mechanization in Bangladesh, 1980s–2010s 239 7.2 Profile of farm machines of farm households, Bangladesh, 2000–2010 243 7.3 Characteristics of tractor/power tiller adopter and owner households, Bangladesh, 2000–2010 248 7.4 Determinants of adoption and ownership/service provision of tractors, Bangladesh, 2000–2010 250 7.5 Determinants of adoption of tractors and/or power tillers, Bangladesh, 2000–2008 254 7.6 Probit model and linear regression models (including correlated random effects variables) 256 8.1 Share of farm households using machinery for land preparation and harvesting, Myanmar, 2006–2016 (percentages) 270 8.2 Real purchase value of selected machinery and average annual growth rate, Myanmar, 2000–2015 271 List of tabLes xi 8.3 Share of households using owned versus rented machinery for land preparation and combine harvesting, by farm size category, Myanmar, 2015/2016 274 8.4 Cumulative number of agricultural machinery dealers selling different types of machine, Myanmar, 1995–2016 277 8.5 Annual sales of selected machinery by dealerships in Mingalar Than Myint compound, Myanmar, 2012–2016 277 8.6 Share of 2016 sales of selected machinery, by type of finance, Myanmar 280 9.1 Evolution of different scales of mechanization in Nepal, 1970– 2016 287 9.2 Land-to-labor ratio, terrain ruggedness, and draft animal holdings, Nepal, 2010 288 9.3 Share (percentage) of farm households using tractors, by year and agroecological belt, Nepal, 1995–2010 288 9.4 Mechanization and cropping patterns (percentage of farmers growing each crop), Nepal, 2010 289 9.5 Farm household characteristics, Nepal, 1995–2010 293 9.6 Daily agricultural wages (in kg of milled rice a day of labor purchases), Nepal, 1995 and 2010 296 9.7 Types of hiring service providers interviewed, Nepal, 2016 302 9.8 Size of landholdings by interviewed tractor service providers, Nepal, 2016 302 9.9 Extent of custom hiring service operations, in number of days used per year,a Nepal, 2016 303 9.10 Sources of financing for tractors and power tillers 303 9.11 Breakdown of key cost components for four-wheel tractors and power tiller operations (US dollarsa per year per machine), excluding costs for attachments, Nepal, 2016 304 9.12 Median owned and annually cultivated farm size (ha) of tractor owners, differentiated by hiring-out status, Nepal Terai, 2003 and 2010 305 9.13 Change in owned farm size and operational size, by tractor owners and non-owners (panel samples), Nepal, 1995, 2003, 2010 307 xii List of tabLes 9.14 Effects of adopting tractors on household income (total income and agricultural income), Nepal, 1995, 2003, 2010 307 9.15 Effects of tractor use on livestock revenue, land rental revenue, and off-farm income, Nepal Terai and Hills, 1995, 2003, 2010 310 9.16 Effects of tractor use on off-farm income-earning activities (person-hours within 12 months), Nepal Terai and Hills, 1995, 2003, 2010 310 9.17 Effects of mechanization on agricultural input use (land, fertilizer, labor), Nepal Terai and Hills, 1995, 2003, 2010 311 9.18 Effects of mechanization on real revenue per hectare, Nepal Terai and Hills, 1995, 2003, 2010 313 9.19 Effects of tractor custom hiring service on agricultural returns to scale in Nepal Terai, 2010 313 9A.1 Significance (p-values) of differences in averages between tractor adopters and non-adopters for each variable used in inverse probability–weighted regression 323 10.1 Ownership of machinery, Ethiopia, 2013 340 10.2 Use of machinery, percentage of plots, Ethiopia, 2013 342 10.3 Use of machinery by crop, percentage of plots, Ethiopia, 2013 343 10.4 Share of households that use machines, by zone, as reported by focus groups, Ethiopia, 2013 344 10.5 Comparison of some characteristics of mechanization- intensive zones with other zones, Ethiopia, 2013 351 10.6 Importance of machine rental services, by farm size, Ethiopia, 2013 355 10.7 Labor and land productivity in the wheat cluster in southeast Ethiopia, 2013 364 10.8 Regressions to examine the association of the use of machines for harvesting and threshing on wheat yield in southeast Ethiopia, 2013, quintals/ha 365 10.9 Labor and land productivity by tractor use in mechanization- intensive zones, Ethiopia, 2013 366 10.10 Regressions to examine the association of the use of tractors on wheat yield in mechanization-intensive zones, quintals/ha, Ethiopia, 2013 368 List of tabLes xiii 11.1 Crop area per rural person, by region, 2000 and 2010 382 11.2 Changes in the farm size distribution, 2005/2006 and 2012/2013 383 11.3 Agricultural wages and annual average growth rate in Ghana, 1991–2012, by type of work 385 12.1 Sampling design of the four surveys, Kenya, 1992–2012 409 12.2 Factors affecting the adoption of farm mechanization, Kenya, farm household survey of 2012 416 13.1 Farmland endowments, farm sizes, labor, and animal use by region, Nigeria, 2010 and 2012 424 13.2 Percentage of farmers using tractors or animal traction in 2010 and 2012 rainy seasons in Nigeria 425 13.3 Percentage of area using tractors or animal traction in 2010 and 2012 rainy seasons in Nigeria 425 13.4 Farm size distribution in Nigeria and corresponding mechanization levels, 2010–2012 426 13.5 Estimated area and area shares of tractor use in Nigeria, 2010–2012 427 13.6 Major characteristics of each type of farm household in Nigeria, 2010 428 13.7 Real male wages per day in Nigeria for land clearing and land preparation (in kilograms of local rice grain at local purchase price), 2010–2012 429 13.8 Determinants of the area cultivated by tractors (pseudo-panel double hurdle model; marginal effects evaluated at the mean of observations), Nigeria, 2010–2012 430 13.9 Sources of tractor services in 2010 rainy season (percentages), Nigeria 438 13.10 Area cultivated annually per tractor (hectares per year), Nigeria, 2013 441 13.11 Average annual tractor use by farming activity (all tractors combined), Nigeria, 2013 441 13.12 Profitability differences between types of tractor owner- operators (in US$1,000s), Nigeria, 2013 443 xiv List of tabLes 13.13 Effects of tractor adoption on animal traction use, Nigeria, 2010, 2012 445 13.14 Effects of tractor adoption on chemical fertilizer use, crop revenues, and household income, Nigeria, 2010, 2012 445 13.15 Effects of tractor adoption on household labor use and off- farm income, Nigeria, 2010, 2012 446 13.16 Soil types and tractor use in Nigeria, 2010, 2012 448 13B.1 Balancing properties of propensity score matching 455 13B.2 First-stage regression of propensity score matching method 456 14.1 Significant difference in utilization rates per price of machines between Siam Kubota and Amec power tillers in Mbarali district, Tanzania, 2012/2013 469 14.2 Number of tractors and implements in surveyed districts, Tanzania, 2015 476 List of tabLes xv xvii LIST OF FIGURES 1.1 Overview of sequential adoption of mechanization according to Pingali, Bigot, and Binswanger (1987) 15 2.1 Number of agricultural workers and amount of machinery power, China, 1978–2012 77 2.2 Average cost of cooperative ownership of combine harvesters 84 2.3 Demand for mechanization services 87 2.4 Cost per hectare and area harvested by combine service enterprises, Peixian county, China, 2013 92 3.1 Two-wheel and four-wheel tractor sales in India between 2005 and 2015 103 3.2 Breakdown by horsepower of tractors sold in India between 2000 and 2012 103 3.3 Map showing major agroclimatic zones of India 106 3.4 Tractor density and agricultural labor density across the states of India, 2012 107 4.1 Holding size and use of machinery for paddy land preparation and harvesting 148 5.1 Regions of Thailand 167 5.2 Monthly wage rates in Thailand by sector, 1980–1995 170 5.3 Agricultural land, agricultural employment, and land-to- worker ratio in Thailand, 1971–2014 174 5.4 Imports and exports of agricultural machinery and parts, Thailand, 2009–2014 187 5.5 Use of two-wheel tractor to power an axial flow pump to irrigate a rice field in Central region, Thailand 189 6.1 Trend of tractor ownership, machine renting, and labor hiring for all farming operations in Viet Nam, 1992–2008 210 6.2 Tractor ownership and machine rental for larger- and smaller- holders, based on rice planting area, Viet Nam, 1992–2008 211 6.3 Relationship between tractor ownership and rice planting area, Viet Nam, 1992 and 2008 211 6.4 Relationship between machine rental and rice planting area, Viet Nam, 1992 and 2008 213 6.5 Trend of total land cultivated per household and total annual cropland cultivated per household, Viet Nam, 1992–2008 213 6.6 Kernel density of total cultivated area, Viet Nam, 1992 and 2008 214 6.7 Kernel density of total annual cropland, Viet Nam, 1992 and 2008 214 6.8 Kernel density of rice planting area, Viet Nam, 1992 and 2008 215 6.9 Median daily real male and female agricultural wage, Viet Nam, 1992–2008 216 6.10 Proportions of samples providing tractor rental services, conditional on rice planting area, among tractor owners, Viet Nam, 2004–2008 223 7.1 Agricultural labor, cattle/buffalo, machinery, and real wages, Bangladesh, 1992–2013 238 7.2 Agricultural machinery use in South Asian countries 240 7.3 Kernel density of land distribution among adopters and owners of machinery, Bangladesh, 2000–2010 244 7.4 Landownership and adoption of agricultural machinery, Bangladesh, 2000–2010 244 7.5 Percentage of owners hiring out their tractors/power tillers, Bangladesh, 2000–2010 245 7.6 Mean adoption rate of tractors/power tillers by ecology, Bangladesh, 2000–2008 247 xviii List of figures 7.7 Mean adoption rate of tractors/power tillers by division, Bangladesh, 2000–2008 247 8.1 Location of village tracts surveyed, Myanmar, 2016 267 8.2 Machinery and draft animal use in paddy cultivation, Myanmar, 2015/2016 268 8.3 Cumulative purchases of selected machinery, Myanmar, 1990–2015 270 8.4 Share of households using machinery for land preparation and harvesting, by farm size group, Myanmar, 2015/2016 272 8.5 Use of owned or rented machinery for land preparation and for harvesting in paddy cultivation, Myanmar, 2006–2016 273 8.6 Number and location of machinery suppliers, Myanmar, 2010, 2013, and 2016 278 9.1 Agroecological belts in Nepal 286 9.2 Share of Terai farm households renting tractors or using only draft animals, by farm size (left = tractor renters; right = draft animals only), Nepal, 2003 and 2010 292 9.3 Correlation between village development committee–level share (percentage) of tractor-using farm households and share of households engaged in farming, Nepal, 1995–2010 308 9A.1 Balancing properties of inverse probability–weighted regressions 325 10.1 Real daily wages of unskilled laborers in rural areas of Ethiopia, 2003 to 2016 (left panel), and rural wage data from other developing countries (right panel) 335 10.2 Real prices of oxen in Ethiopia, 2004–2016 335 10.3 Evolution of farm sizes of smallholders (less than 25 ha), Ethiopia, 2004–2015 337 10.4 Commercial farms in Ethiopia, 2014/2015: Crops grown by percentage (left panel) and area under commercial farms by zone, ’000 ha (right panel) 338 10.5 Rainfall patterns in Ethiopia 339 10.6 Imports of agricultural machines into Ethiopia, three-year moving averages—value of main agricultural machines, 2001– 2014 (left panel); number of four-wheel tractors, 2005–2014 (right panel) 346 List of figures xix 10.7 Tractor sales by Adama Agricultural Machinery Industry (part of Metals and Engineering Corporation), Ethiopia, 2012/2013–2015/2016 348 10.8 Imports of combine harvesters into Ethiopia, three-year moving averages, by value, 2001–2014 (left panel); by number, 2007–2014 (right panel) 349 10.9 Farm size and grain area, by zone, Ethiopia, 2014/2015 350 10.10 Mechanization and daily wage rates, Ethiopian birr/day, 2015 351 10.11 Seasonal changes in use of combine harvesters (left panel) and in harvesting charges (right panel), Ethiopia, 2015/2016 358 10.12 Seasonal movement of tractors between the Arsi, Bale, Borena, and Harari zones, Ethiopia, 2015/2016 359 10.13(a) Seasonal movement of combine harvesters in Arsi, Bale, and neighboring areas, Ethiopia, November 2015 to February 2016—Route 1 360 10.13(b) Seasonal movement of combine harvesters in Arsi, Bale, and neighboring areas, Ethiopia, November and December 2015— Route 2 361 10.14 Density functions on association of combine harvesters with labor productivity (left) and with crop yields (right), Ethiopia, 2013 364 10.15 Density functions on the association of tractor use and labor productivity, Ethiopia, 2013 367 10.16 Seasonal movements in real rural wages in Ethiopia, July 2004–June 2014, monthly wage index (average yearly wage is 1.00) 370 11.1 R-value measure of farming system evolution in Ghana, 1961–2014 381 11.2 Tractor imports in Ghana, 2003 to 2012 388 12.1 Map with primary sampling units of different surveys, Kenya, 1992–2012 409 12.2 Farm implements used, by agroecological zone, farm households, Kenya, 2014 411 12.3 Evolution of farm mechanization in Kenya from 1992 to 2012 412 xx List of figures 12.4 Evolution of farm mechanization in Kenya from 1992 to 2012, by agroecological zone and year 413 12.5 Evolution of farm mechanization in Kenya from 1992 to 2012, by year and by agroecological zone 414 13.1 Seasonality of tractor use, hours per month, by tractor owners (left panel); average distance between home district and where hiring services are provided (right panel); Nigeria, 2014 442 14.1 Tractors in use in Tanzania in different years, 1950–2015 459 14.2 Draft animals in use in different regions, Tanzania, 2005 464 14.3 Four-wheel tractors in use in different regions in Tanzania, 2005 466 14.4 Four-wheel tractors in use in different regions in Tanzania, 2015 467 14.5 Age of four-wheel tractors in Tanzania, 2005 468 14.6 Number of two-wheel tractors imported by Tanzania, 2005–2014 468 14.7 Distribution of two-wheel tractors in different regions, 2014 469 14.8 Level of mechanization in the four case study districts, Tanzania, 2015 476 14.9 Size of farms operated by four-wheel tractor owners in the case study districts, Tanzania, 2015 477 14.10 Size of farms in Mbarali district, Tanzania, farmed by tractor owners in 2013 477 14.11 Distances from home to shops selling spare parts, Mbarali district, Tanzania, 2013 483 List of figures xxi xxiii ABBREVIATIONS AND ACRONYMS AAMI Adama Agricultural Machinery Industry (Ethiopia) ADB Asian Development Bank (Nepal) AED Agricultural Engineering Division (Thailand) AEHE Agricultural Equipment Hiring Enterprises (Nigeria) AERI Agricultural Engineering Research Institute (Thailand) AGITF Agriculture Inputs Trust Fund (Tanzania) AMD Agricultural Mechanization Department (Myanmar) AMSEC Agricultural Mechanization Services Enterprise Center (Ghana) ASDP Agricultural Sector Development Program (Tanzania) ASDS Agricultural Sector Development Strategy (Tanzania) ATA Agricultural Transformation Agency (Ethiopia) AUC African Union Commission BDT Banglashi taka (Bangladesh) BOI Thailand Board of Investment (Thailand) Br Ethiopian birr (Ethiopia) CAADP Comprehensive Africa Agriculture Development Programme CAMARTEC Centre for Agricultural Mechanisation and Rural Technology (Tanzania) CBU completely built-up units (Ethiopia) CHS custom hiring service CHSC custom hiring service center (India) CIF cost, insurance, and freight (Nepal) CIMMYT International Maize and Wheat Improvement Center CKD completely knocked-down CNC computer numerical control (Thailand) CRS constant returns to scale CSA Central Statistical Agency (Ethiopia) DAP draft animal power EAs enumeration areas (Myanmar) ESAP economic structural adjustment program (Tanzania) ESS Ethiopia Socioeconomic Survey (Ethiopia) FACASI Farm Mechanization and Conservation Agriculture for Sustainable Intensification project (Tanzania) FAO Food and Agriculture Organization of the United Nations FMRC Farm Mechanization Research Centre (Sri Lanka) FOB free on board (Nepal) 4WTs four-wheel tractors FtF Feed the Future (Ethiopia) GDP gross domestic product (Ethiopia) GGS Great Groundnut Scheme (Tanzania) GMM generalized method of moments (India) GS government sourced (Nigeria) ha hectare HIES Household Income and Expenditure Survey (Bangladesh) HIMIs high- and intermediate-mechanization implements (Ethiopia) hp horsepower HTT hand-tool technology (Tanzania) HYVs high-yielding varieties ICRISAT International Crops Research Institute for the Semi-Arid Tropics (India) IFPRI International Food Policy Research Institute IMF International Monetary Fund (Ghana) IPW inverse probability weighting (Nepal) IRRI International Rice Research Institute (Thailand) IV instrumental variables (India) KD knocked-down (Ethiopia) KES Kenya shillings (Kenya) LGA local government area (Nigeria) LMIs low-mechanization implements (Ethiopia) xxiv abbreviations and aCronyms LP Levinsohn and Petrin (2003) LPG liquefied petroleum gas (Thailand) LSFs large-scale farms (Tanzania) LSMS-ISA Living Standards Measurement Study–Integrated Surveys on Agriculture (Nigeria) LSP local service provider (India) MAAS Myanmar Aquaculture-Agriculture Survey (Myanmar) MASL meters above sea level (Ethiopia) METEC Metals and Engineering Corporation (Ethiopia) MIP mechanization implementation program (Nigeria) MMK Myanmar kyats (Myanmar) MNCs multinational corporations (Tanzania) MoANR Ministry of Agriculture and Natural Resources (Ethiopia) MS market sourced (Nigeria) MSFs medium-scale farms (Tanzania) NARI National Agricultural Research Institute (Sri Lanka) NC North Central region (Nigeria) NCAM National Center for Agricultural Mechanization (Nigeria) NE North East region (Nigeria) NGN Nigerian naira (Nigeria) NGO nongovernmental organization NGV natural gas for vehicles (Thailand) NLSS Nepal Living Standards Survey (Nepal) NTM Nigeria Truck Manufacturers (Nigeria) NW North West region (Nigeria) OFCs other field crops (Sri Lanka) OFY Operation Feed Yourself (Ghana) OLS ordinary least squares PAN Peugeot Automobile Nigeria Limited (Nigeria) PBAM Peixian Bureau of Agricultural Mechanization (China) PBB Pingali, Bigot, and Binswanger (1987) PIM Policies, Institutions, and Markets (CGIAR Research Program) PSFs peasant subsistence farms (Tanzania) PSM propensity score matching (Nigeria) QFP Quality Food Products Ltd (Tanzania) R&D research and development RCRE Research Center for the Rural Economy (China) Rs rupees abbreviations and aCronyms xxv SAMA Sustainable Agricultural Mechanization in Africa SAP Structural Adjustment Program SE standard error (Kenya) SKD semi-knocked-down SMAM Sub-Mission on Agricultural Mechanization (India) SMS short message service SSA Africa south of Sahara SSCFs small-scale commercial farms (Tanzania) SSFs small-scale farms (Tanzania) SUMA-JKT National Service Corporation Sole Agri-Machinery Project (Tanzania) SUR seemingly unrelated regressions (Nepal) SW South West region (Nigeria) TAMS Tanzania Agricultural Mechanization Strategy (Tanzania) THB Thai baht (Thailand) THS tractor hiring services THU tractor hiring unit (Nigeria) TRAMA Tanzania Tractor Manufacturing Company (Tanzania) 2WTs two-wheel tractors VAT value-added tax VDC village development committee (Nepal) VDSA Village Dynamics in South Asia (India) VEAM Viet Nam Engine and Agricultural Machinery Corporation (Viet Nam) VHLSS Vietnamese Household Living Standard Survey (Viet Nam) VLSS Vietnamese Living Standard Survey (Viet Nam) VND Vietnamese dong (Viet Nam) VWON Volkswagen of Nigeria Limited (Nigeria) ZOI zone of influence (Ethiopia) xxvi abbreviations and aCronyms xxvii FOREWORD Agricultural mechanization plays an important role in economic transforma- tion. In recent years, rapid urbanization, increased food demand, rising rural wages, and seasonal labor bottlenecks have led to a resurgence of policy sup- port to promote agricultural mechanization in Africa south of the Sahara. The African Union’s Agenda 2063 commits countries to banishing the hand hoe by 2025 as part of the goal of achieving overall productivity and food security improvements through a modern and environmentally sustainable agriculture sector. This book makes an important contribution to our understanding of the development of mechanization in Africa. The authors emphasize that appro- priate strategies in Africa should be guided by better knowledge of mechaniza- tion supply and demand, as seen through a lens that combines farming-system evolution and induced innovation theories, as well as knowledge of the para- digm shift that has been occurring in Africa over the past three decades. The book’s case studies show that this “new paradigm” raises new chal- lenges in Africa. As demand for mechanization has risen in Africa, including among smallholders, the supply-side gaps compared to Asia have become more pronounced. Significant opportunities exist for supply-side reform, including the identification of appropriate machinery for different locations, dissemina- tion of knowledge about such types of machines, mitigation of coordination failures, exploitation of the multifunctionality of machines, and facilitation of migratory service provision. A range of policies can be helpful, such as import policies and promotion policies, as well as institutional and capacity develop- ment policies. To be effective, such policies will need to strike the right bal- ance between direct and indirect interventions—adhering to key principles of prioritizing market-led hiring services, eliminating distortions, and providing relevant public goods. The policies and challenges highlighted in this book have implications not only for agricultural production but also, at a broader level, for agrifood sys- tem transformation and rural revitalization, which are essential for achieving inclusive, sustainable growth in Africa. The book’s analytical framework and empirical results promise to be valuable for both policymakers and researchers. Johan Swinnen Director General xxviii foreword xxix ABSTRACT Over the last few decades, urbanization, increased food demand, rising rural wages, and seasonal labor bottlenecks have led to the resurgence of policy- maker and development stakeholder interest in promoting agricultural mech- anization in Africa south of Sahara (SSA). Unlike the predictions based on farming system intensification and on the induced innovation frame- work used by Pingali, Bigot, and Binswanger in their 1987 book Agricultural Mechanization and the Evolution of Farming Systems in Sub-Saharan Africa, however, the supply of mechanization in SSA has failed to meet demand, due to both market failures associated with various characteristics that are unique to the continent (such as complementary technologies, infrastructure, and machine size) and improper government interventions. Meanwhile, lessons from Asia—where significant mechanization growth has occurred despite smallholder dominance—are becoming increasingly relevant to SSA. The potential for such South–South knowledge exchange has so far been limited due to the insufficient empirical information in both regions. In SSA, there is little knowledge regarding in-country variations of demand potential and adoption of mechanization. In Asia, despite the large overall body of knowl- edge of the continent’s experience, the literature has not coherently docu- mented how and when mechanization grew and what impacts it had, nor has it provided sufficient quantitative evidence on these issues. This book attempts to fill this knowledge gap. In so doing, the book revis- its the approach paradigm of Pingali, Bigot, and Binswanger (1987), empha- sizing the market failures that can cause the undersupply of mechanized services and assessing potential government failures based on evidence from some recent government interventions in mechanization in studied SSA countries. The book also emphasizes the divergent patterns of mechanization between Asian and African countries, pointing out the challenges specific to SSA and their policy implications. The book begins with a chapter describing the farming systems approach, the importance of functional mechanization hiring markets and supportive policies, and key differences between SSA and Asian mechanization, including the market failures specific to SSA. The rest of the book is structured with stand-alone case studies for 13 countries—8 Asian and 5 African—to document the evolution of mechanization in these countries. Two of the key underlying questions are what SSA policymakers can draw from the Asian mechanization experience and whether there is a need to develop a new paradigm that better accounts for the Africa-specific experience. This book makes at least four significant contributions to the lit- erature, with important policy implications. First, based on solid country evi- dence, it provides a detailed overview of the development and current state of mechanization in African and Asian developing countries. Second, using a similar structure for case studies in the country chapters, it allows the reader to easily draw comparisons between countries and between Asian and SSA mechanization in a clearer light. Third, it endows national policymakers and the development community with broader knowledge that can be adapted to local contexts, helping them to learn from successes and avoid replicating past failures. Finally, as a whole, the book serves as a precursor for further research and a call to create an updated paradigm of mechanization development. xxx abstraCt xxxi ACKNOWLEDGMENTS The editors wish to acknowledge the valuable contributions of research- ers and stakeholders at the International Conference on South–South Knowledge Sharing on Agricultural Mechanization, held in Ethiopia in 2017, and programs by the CGIAR Research Program on Policies, Institutions, and Markets (PIM), led by the International Food Policy Research Institute (IFPRI) and funded by CGIAR Fund donors. Editors and authors are also grateful for the constructive comments and suggestions from the external reviewers, Jerry Shively and other IFPRI Publications Review Committee members, which significantly improved the book. Special thanks go to Gershon Feder, Paul Dorosh, Steve Biggs, and the late Hans Binswanger- Mkhize for their feedback and comments on the conceptualization of issues and the development of the framework. Editors also benefited greatly from the valuable insights of Frédéric Baudron, Lidia Cabral, Josef Kienzle, M. A. Sattar Mandal, Keijiro Otsuka, and David Spielman, among others. Authors of various chapters benefited from the support of in-country collaborators, including IFPRI and International Maize and Wheat Improvement Center (CIMMYT) country offices and local field teams. Editors also thank Mia Ellis for editing various chapter manuscripts. China and Kenya chapters con- sist largely of materials that have already been published in the journals China Economic Review and Agricultural Mechanization in Asia, Africa and Latin America. Finally, the writing of this book would not have been possible with- out the financial support of PIM, the United States Agency for International Development, and the Regional Strategic Analysis and Knowledge Support System in Asia. Part 1 Synthesis of the Lessons AN EVOLVING PARADIGM FOR AFRICA AND SYNTHESIS OF THE LESSONS FROM ASIA Xinshen Diao, Jed Silver, Hiroyuki Takeshima, and Xiaobo Zhang Abstract: Africa has experienced a paradigm shift in mechanization in the past three decades. The “new paradigm” has also given rise to new challenges and pol- icy issues. By synthesizing the recent experiences in African and Asian countries, this chapter draws lessons from Asia and Africa under this new African para- digm. In doing so, the chapter first lays out the guiding theoretical framework used in 1987 by Pingali, Bigot, and Binswanger (PBB), based on the literature on farming systems evolution and induced technological change. The chapter then describes the “new paradigm,” which builds on PBB but also integrates the additional dimension of market failures associated, on the supply side, with custom hiring services, which have become the most common mode of mecha- nization among smallholders in developing countries. Applying this expanded framework, the chapter then reviews the Asian experience first. It highlights how mechanization has grown in the continent, having largely avoided supply-side market failures, thanks to several factors: smaller machine sizes;1 increased opportunities for multifunctional uses of machines; more secure land tenures that allow integration with formal credit markets; and the supportive, rather than distortive, nature of government subsidy policies. The chapter then turns to the experiences in Africa south of the Sahara (“Africa” hereafter) and highlights the emerging patterns of spatial variations in demand that are still largely con- sistent with the PBB framework. However, the chapter also stresses that market failures associated with custom hiring services on the supply side are substan- tial due to features unique to Africa, including the dominant types of large trac- tors, in addition to higher financial constraints on tractor ownership resulting from lack of secure land tenures and weak penetration of formal credit markets, as well as other barriers due to limited multifunctionality, lack of migratory 1 In this book, we use the term “size,” in relation to tractors, to refer loosely to horsepower. Our definition of size is not based on any clear-cut engineering threshold. Rather, we focus on size aspects when we highlight the differences in typical tractor horsepower between Africa and Asia that are relatively universal, as described in the later section, and their implications for the nature of constraints and market failures that Africa is facing. Chapter 1 3 services due to insufficient infrastructure and coordination failures, and insuf- ficient technologies complementary to mechanization. Based on country expe- riences in Asia and Africa, the chapter also highlights key government policies that have not always been successful, including import restrictions (or removal thereof), inefficient technology and skill promotion, and insufficient provision of public goods. Last, given the country experiences and the identified appro- priate roles of governments, the chapter concludes by describing the key lessons that are important for Africa’s mechanization pathway forward, including (1) understanding the emerging nature of demand, (2) actively promoting private hiring services, (3) eliminating or reducing distortions, and (4) prioritizing the mechanization technologies appropriate for African contexts. Introduction Agricultural transformation is imperative for growth and poverty reduction in Africa. Yet the desired progress has been elusive. The region is a net food importer despite the fact that agriculture accounts for 60 percent of employ- ment. Main food crop yields are estimated at about half the world average, and rural poverty, hunger, and malnutrition are persistent (AfDB 2016). Recently, increased (albeit still insufficient) attention has been paid to promoting a Green Revolution–style agricultural intensification, focusing on improved seed varieties, fertilizer, and agrochemicals that increase the land productiv- ity. In comparison, much less emphasis has been placed on addressing seasonal labor constraints and rising rural wages through mechanization to promote agricultural transformation. Mechanization is a labor-saving technology that enables farmers to expand cultivation area and free up labor for other agricultural functions or nonfarm income generation.2 Early efforts to promote mechanization in Africa often failed due to abundance of rural laborers within most rural farm households 2 In this book, the term “mechanization” is defined in a broad sense, including both technolo- gies themselves and processes that involve their use. The term “tractorization” is used where the focus is specifically on tractors, and mechanization is used if the focus is generally on increased mechanical power. Mechanization can sometimes encompass tractorization if, for example, the process happens to involve a switch from draft animals to tractors. The term “farm power” is used where the focus is more on motive energy inputs in farming, which are provided through either human labor, animal work, or machinery (these all take the form of motive energy, as opposed to energy embedded in other types of inputs, such as fertilizer). The term “agricultural machinery” is used where the focus is on the physical capital items that convert energy inputs into the desired form of energy outputs. The term “agricultural implements” is used in a similar way, but specifically for machine attachments, such as plows and harrows. Similarly, “labor-saving” is defined in a broad way, including both in economic terms (for example, saving on the cost of labor) and ergonomic terms (for example, reduced labor 4 PART 1: SYNTHESIS OF THE LESSONS that limited farmer incentives to intensify production (Pingali, Bigot, and Binswanger 1987, hereafter “PBB”). However, agricultural mechanization has gained renewed attention recently in Africa. Some indications suggest that farming systems have evolved sufficiently in many locations of Africa for farm- ers to demand mechanization (Mrema, Baker, and Kahan 2008; Diao, Silver, and Takeshima 2016; Binswanger-Mkhize 2017) and that increased mecha- nization adoption has occurred in new pockets across Africa recently (FAO 2016; Malabo Montpellier Panel 2018). Despite recent progress, however, the spread of mechanization in Africa has lagged far behind that of Asia, where mechanization has been widely adopted in most countries in recent years, including in many low-income and labor-abundant Asian countries. Moreover, research on mechanization in developing countries, and in Africa in partic- ular, is scarce, and knowledge and insight about mechanization in Africa, in terms of both collecting statistics about it and understanding its drivers and impacts, remains limited. Although different opinions exist in the literature for understanding fac- tors affecting mechanization in Africa, the framework developed by PBB in 1987, based on the farming systems evolution hypothesis, which emphasizes the demand side of mechanization, remains one of the most important guid- ing frameworks, together with other guiding strategies, for pursuing mecha- nization development. In their seminal volume, Agricultural Mechanization and the Evolution of Farming Systems in Sub-Saharan Africa, PBB argued that widespread public efforts to promote mechanization often failed because African farming systems had simply not intensified enough to generate suf- ficient demand for mechanization among farmers. Their theory also fits the patterns observed in Asia, where farming systems had already undergone widespread intensification and draft animal power (DAP) had been in use for a much longer period (Lawrence and Pearson 2002). Twenty years after publi- cation of the PBB book, Pingali further asserted that “where the demand side factors are in place, agricultural intensification and the adoption of mechan- ical power occurs in Africa in a similar pattern to Asia and Latin America” (2007, 2787). Broadly speaking, farming systems have intensified in many places in Africa, with a shortened fallow period and an increasing share of annual crop areas among total agricultural land. However, as the rest of this chapter and the African case studies in this book show, the characteristics of demand for mechanization remain complex. Moreover, supply does not appear requirement and reduced drudgery), both of which are becoming increasingly important moti- vations in Africa to address the mechanization challenge (for example, Kormawa et al. 2018). CHAPTER 1: AN EvOLvINg PARAdIgm FOR AFRICA ANd SYNTHESIS OF THE LESSONS FROm ASIA 5 to have responded at nearly the levels observed in Asia. As a result, the farm- ing systems hypothesis alone has been insufficient to explain mechanization in Africa (Binswanger-Mkhize 2017). Altogether, with the intensifying farming system and growing relevance of modern mechanization technologies, there is a need for a closer understanding of not only demand but also, most impor- tant, the increasing relevance of supply-side constraints on mechanization in Africa, which we describe as a “new paradigm” for mechanization in Africa. Mechanization could help farmers overcome the labor constraints pres- ent in agriculture, reduce drudgery in various farming operations, expand farm sizes where land is available, and permit higher levels of intensification in more labor-intensive farming activities. Although tractor plowing per se is not directly associated with yield growth, it enables key operations to be done on time, which is especially relevant for rainfed agriculture in areas with short planting windows. Evidence from Asian as well as some African coun- tries also suggests that tractor use is associated with higher cropping intensity and use of fertilizer. Combine harvesters have also become part of mechani- zation practice. As shown in the case studies of Asian countries and Ethiopia, use of combine harvesters has the potential to significantly reduce postharvest losses, thus increasing outputs per unit of land. In a highly optimistic scenario, these effects of mechanization can contribute to Africa’s forestalled agricul- tural transformation. However, for this to occur, substantial improvements in agricultural engineering research; varietal development; and market develop- ment for inputs, credit, and outputs would likely have to take place to comple- ment mechanization. In many African countries, supply elasticity is limited partly due to technological backwardness, leading to persistently high reliance on food imports despite relatively high food prices (for example, in the case of rice, Gyimah-Brempong, Johnson, and Takeshima 2016). It is essential to understand and address these complex issues hindering mechanization devel- opment in Africa. It has often been suggested that Africa can learn from Asian experiences of agricultural transformation, including mechanization. Asian experiences in mechanization are diverse, and distinct patterns of mechanization across Asian countries could offer many lessons to Africa. One thing that seems to be common among many Asian countries is that mechanization has often started with little direct intervention from the governments. Manufacturing of spare parts and simple tools often grew out of innovations by local entre- preneurs at the early stages of manufacturing-sector development (Diao et al. 2014). In some countries where governments did get involved, they primar- ily attempted to overcome market constraints for the private supply to meet 6 PART 1: SYNTHESIS OF THE LESSONS existing farmer demand, and provided key public goods to overcome market failures as well as education (Rijk 1986). As is described later in this chap- ter, despite the limited successes of the African government–led mechaniza- tion programs three decades ago (as diagnosed by PBB), recent efforts by a few African governments suggest that the past lesson is still relevant for today’s issues. These issues include limited machine utilization rates and the insuffi- cient provision of soil and machinery technological knowledge to the private sector, among others (Kormawa et al. 2018). The challenge to African poli- cymakers is how to appropriately address key market failures that prevent the private sector’s supply from meeting emerging demand, and how to better identify the characteristics of demand that help overcome supply bottlenecks. Later sections of the chapter describe market failures that lead to risks and uncertainty about making machine investments due to spatially and tempo- rally variable demand in hiring markets, uncertainty about and limited oppor- tunities for multifunctional use of machines, and insufficient machinery and oil information and knowledge that can be provided by the government as public goods (Diao et al. 2017).3 This book aims to update the PBB framework by integrating an addi- tional dimension—market failures on the supply side of mechanization asso- ciated with custom hiring services, the most common mode of mechanization among smallholders in developing countries—in order to account for the recent mechanization patterns observed in Africa alongside those in Asia. By doing so, we intend not to dispute the underpinnings of PBB’s hypothesis, but rather to capture the emerging challenges of mechanization, particularly those highlighted by recent experiences in Africa. An updated framework is not only important for further research but also a crucial tool for African policymakers to develop judicious approaches to supporting mechaniza- tion development. 3 PBB hypothesized that the supply-side constraints are rarely binding, and although there are many grounds to expect this hypothesis to hold in general (such as the private sector’s ability to innovate mechanical technologies compared with biological technologies), testing it formally has been challenging. Further, despite the improved understanding and recognition of market failures (including those in information, risk, and finance) and of the role public-sector institu- tions play in sustaining private sector–led growth (Rodrick 2007; Naudé 2011), these new per- spectives have not been applied sufficiently to understand how the public sector can “speed up” and “raise efficiency” of the private sector’s responses to meet demand. The potential roles of the public sector to assist the private sector in such ways in the short run is important because African governments and international communities are under pressure to meet development goals that are becoming increasingly time-sensitive (for example, 2025 goals for mechanization achievements envisaged under the African Union’s Agenda 2063, and 2030 goals under more general Sustainable Development Goals, among others). CHAPTER 1: AN EvOLvINg PARAdIgm FOR AFRICA ANd SYNTHESIS OF THE LESSONS FROm ASIA 7 This book presents evidence from 13 countries in Asia and Africa to lay the foundations for the new paradigm. These chapters largely avoid policy pre- scriptions, and instead aim to provide a thorough overview of where mecha- nization stands in each country and how it has developed to that point. They are intended as resources for policymakers, academics, and lay readers to draw upon when considering how to encourage the development of mechanization in specific contexts. This book’s approach of focusing on individual countries for case studies differs from that of PBB, which focused on the African con- tinent as one geographic region and drew collective lessons from various loca- tions within it. The book also contributes to the efforts in integrating mechanization into the mainstream Africa-wide agenda, including the African Union’s Agenda 2063 (a strategic framework for the socioeconomic transformation of the continent over the next 50 years), the Comprehensive Africa Agriculture Development Programme (CAADP), and the Malabo Declaration. Under Agenda 2063, Aspiration #1, “To achieve a prosperous Africa based on inclu- sive growth and sustainable development,” Goal #5 commits countries to ban- ish the hand hoe by 2025 as part of the goal of achieving overall productivity and food security enhancement through a modern and environmentally sus- tainable agriculture sector (Malabo Montpellier Panel 2018). Although this is a political goal and its economic rationale remains to be investigated, it sym- bolizes the growing interest in mechanization within the African community. The CAADP platform, which commits African countries to spend 10 percent of national budgets on the agriculture sector to achieve a 6 percent annual growth rate in the sector, recognizes the importance of agricultural mech- anization in promoting intensification (Diao, Silver, and Takeshima 2016; FAO 2016). Increasingly, agricultural mechanization has been integrated into CAADP’s Pillar #4, Integrated Research, Technology Dissemination and Adoption (Malabo Montpellier Panel 2018). The Malabo Declaration fur- ther recognizes the slow pace of mechanization along the agriculture value chain and emphasizes the importance of investments into suitable, reliable, and affordable mechanization and energy supplies in order to double pro- ductivity by 2025 (Malabo Montpellier Panel 2018). In October 2018, after intensive expert consultations with a broad range of stakeholders, the African Union Commission (AUC) and Food and Agriculture Organization of the United Nations (FAO) launched the Sustainable Agricultural Mechanization in Africa (SAMA) framework, which has been integrated into CAADP and the Malabo Declaration (Kormawa et al. 2018) and recognizes that agricul- tural mechanization in Africa is an indispensable pillar for attaining the 8 PART 1: SYNTHESIS OF THE LESSONS Zero Hunger vision by 2025, as stated in the Malabo Declaration of 2014 (AUC 2018). The book is written at a crucial time, and its recommendations are expected to become part of the policy dialogue and debate on how to develop and disseminate modern agricultural technologies in order to double agricul- tural productivity in Africa. The book emphasizes mechanization promotion as an important component of agricultural technology, together with agricul- tural research and development (R&D), irrigation, and so on, which requires the public sector to increase its investment while avoiding the creation of new market and trade distortions that can discourage private investment. We expect the insights of the book to be consistent with the spirit of CAADP to prioritize and coordinate investments (World Bank 2007, 24). We also expect the book to inform the implementation of SAMA, launched by the AUC and FAO in 2018 (Kormawa et al. 2018). SAMA consists of 10 elements, includ- ing (but not limited to) appropriate technologies, business models, financing mechanisms, manufacturing growth, technology development and transfer, and inclusiveness focusing on smallholders and their organizations—all of which are highly relevant for mechanization policy and promotion. Moreover, by providing concrete examples and experiences from what Asian countries have achieved, this book tries to promote South–South learning and Africa– Asia collaboration in searching for the proper pathways for African countries to adopt mechanization. This book is edited mostly from economists’ perspectives, although chap- ters are written by a mixture of agricultural economists and agricultural engineers. The book is intended for a range of stakeholders. For policymak- ers at higher levels, the book will be useful for any ministry of agriculture that oversees direct policies on agricultural engineering and the governance of agrifood systems that encompass the farm sector as well as upstream and downstream sectors (such as machine industries and agricultural machin- ery service providers on the one hand, and rural transportation of harvests on the other). However, our interpretations of PBB (and updating thereof) that highlight the roles of a broad set of economywide factors, including macro- economic factors, provide useful insights to other policymaking institutions, such as ministries of finance, trade and industry, or education, into how their policies can have profound effects on agricultural mechanization, and fur- ther offer a useful common base of knowledge on which they may coordinate their policies toward unified development strategies. The descriptions in this book, which distinguish the modern sector and the more traditional small businesses involved with mechanization, also provide better understanding CHAPTER 1: AN EvOLvINg PARAdIgm FOR AFRICA ANd SYNTHESIS OF THE LESSONS FROm ASIA 9 for policymakers at an operational level and establish more realistic expecta- tions of the impact of their programs and interventions; in this way, this book can serve as a reference for them to better communicate these expectations to stakeholders. Furthermore, the book offers useful historical perspectives to the newer and younger generation of policymakers, who are increasingly assuming political positions in Africa. This book also offers detailed descrip- tions of mechanization adoption at the subregional level and of the structure of the existing mechanization market sector. These can be useful for the pri- vate sector and specific industries that are interested in assessing the potential market opportunities of mechanization in Africa. The book also serves as an important document that communicates economic perspectives on agricul- tural mechanization in comprehensive and holistic ways to the agricultural engineering research community. Last, the book can be useful for researchers and students in the agricultural economics research community who are inter- ested in learning about and conducting empirical research on economic issues surrounding agricultural mechanization. The remainder of this chapter is structured as follows. First, we describe the analytical scope of the book. We then develop our updated theoretical frame- work for mechanization. This updated framework integrates PBB’s hypothesis, centered around the theory of farming systems development by Boserup (1965) and the induced innovation theory of Hayami and Ruttan (1970, 1985), and expands it to account for common market failures in agricultural machinery investment and mechanized service provision. We then describe the selection of case study countries and their linkages with the framework. Next, we apply this framework to help explain the divergence between mechanization trends in Asia and those in Africa. We then pay attention to the role of government policies in shaping mechanization, before concluding with some recommenda- tions that provide context for the remainder of the book. analytical Scope of the Book This book primarily focuses on tractors—both four-wheel tractors (4WTs) and power tillers, also known as walk-behind tractors—though some chapters also cover combine harvesters, another example of a motive, power-intensive mechanization technology, mainly because highly specialized hiring service providers have emerged in some countries, which allow smallholders to get access to the services of combine harvesters. We acknowledge that agricultural mechanization is not just tractor use and that it involves the use of many other types of equipment. Although future studies need to investigate these broader 10 PART 1: SYNTHESIS OF THE LESSONS categories of mechanization to provide a more holistic view of agricultural mechanization in Africa, many other types of commonly used agricultural machinery in Africa are less associated with the development of the hiring ser- vice markets that have eased smallholders’ access to mechanization, one reason why we did not cover them in this book. There are several other reasons that justify this book’s focus on tractors and the important role they play in agricultural mechanization. First, tractors have historically been considered one of the major mechanical innovations in agriculture that can replace animal and human power for the toughest part of farming operations—land preparation. Replacing animal and human power with a tractor makes it possible for farm size to expand and for more and more land to be brought under cultivation. Moreover, tractors as a substitution for animal and human power make it much easier to command more power per worker and significantly raise labor productivity in agriculture (Hayami and Ruttan 1970). Hayami and Ruttan (1970) went as far as to consider the trac- tor the single most important mechanical innovation. In an extreme case like that of the United States, tractors alone have historically raised gross domestic product by significant margins (Steckel and White 2012). Second, partly related to the first point, the adoption of tractors generally means a considerable leap from human or animal power because of the great dif- ference in horsepower, and thus is likely to have significant effects in reducing drudgery associated with manual farm work and in enhancing welfare, as well as a modernizing effect on the agricultural sector. Third, among major farm- ing operations, the use of tractors for land preparation often precedes significant mechanization of other operations. Land preparation is the most energy- demanding farming operation in rainfed agriculture (Lal 2004; Baudron et al. 2015), and primary tillage is one of the first operations to be mechanized when a new source of mobile power becomes available (Binswanger 1986). Fourth and finally, tractors are unique in various dimensions and face dis- tinct challenges not encountered by other machines. For example, because threshing is less time-bound than some other operations, the rental market for threshing machines developed in the United States in the 19th century and in Asia in the 20th, with fewer timeliness constraints than was the case for tractors (Binswanger and Donovan 1987, 15). Multifunctionality of trac- tors, intensively exploited in Asia, also makes tractors unique because tools for other operations, such as planters, weeders, sprayers, and carts for transporta- tion, are often attached to tractors when the latter are introduced. Also, trac- tors have served as an important source of power to run stationary equipment such as irrigation pumps and threshers (Diao, Silver, and Takeshima 2016), CHAPTER 1: AN EvOLvINg PARAdIgm FOR AFRICA ANd SYNTHESIS OF THE LESSONS FROm ASIA 11 especially before the widespread adoption of cheaper machines such as die- sel pumps (IRRI 1983) or the emergence of modern substitutes such as solar pumps. Because tractors are relatively unique, with few substitutes (compared with other types of equipment), information is relatively more available across countries about tractors than about other farm implements. Focusing on trac- tors therefore provides us one way to see more clearly the gaps between Asia and Africa. Furthermore, for many African countries, tractors may be one of the most important binding constraints on the current stage of mechaniza- tion, especially where intermediate technologies such as animal traction have spread relatively widely. Although agricultural mechanization encompasses not only crop pro- duction but also fishery and livestock production, the focus of this book is mechanization for crop production. This is not only because crop production accounts for the most significant part of agriculture in both Africa and Asia, but also because mechanization of crop production often represents the first stage of mechanization development. The share of crop production remains greater than 75 percent of the gross production value of agriculture in Africa, and 70 percent in Asia (FAO 2019a). The share of rural households engaged in crop production is also high and dominant in both Africa and Asia. Focusing primarily on crop production therefore still captures the important aspects of agricultural mechanization at both continents’ current stage of agricultural development. In fact, focusing on tractors as the core of crop-related mechanization can implicitly cover some power-intensive activities associated with fishery and livestock production, such as on-farm production of fodder (from maize and so on), transportation of fodders or water, and transportation of animals for slaughtering or sales. However, due to data availability on the use of machin- ery in livestock raising, we cannot do any empirical analysis explicitly on live- stock mechanization in this book. Similarly, as with crop production, more control-intensive activities, such as identification, medication, vaccination, evisceration, and processing and packaging, may be mechanized only after the mechanization of power-intensive activities has been widely adopted. Such adoption represents a more advanced stage of mechanization and is not covered by this book, although mechanization in these areas is likely to be increasingly important in the future in both Asia and Africa. Last, the book focuses primarily on production and does not directly address issues associated with postharvest and storage activities, or with pro- cessing and marketing along the value chain of agriculture (Breuer, Brenneis, and Fortenbacher 2015), except where we touch on the multifunctionality of 12 PART 1: SYNTHESIS OF THE LESSONS tractors (for example, their use for transport). We make this choice because most mobile operations, which constitute one type of power-intensive oper- ation, occur at the production stage, whereas most operations in postharvest stages are stationary (PBB). The mechanization of mobile operations typi- cally faces a different set of constraints than that of stationary operations. For example, activities such as milling, grinding, pounding, pressing, crushing, and threshing typically do not face the timeliness problems associated with plowing, and so efficient rental markets are relatively easily established (PBB; Binswanger and Donovan 1987). Similarly, these service providers often pro- vide other postharvest services. For example, in Ghana, many rice mills pro- vide drying and storage services (Takeshima, Agandin, and Kolavalli 2017). As long as there are still challenges associated with the mechanization of these activities, treating them together with mechanization for mobile operations, such as that provided by tractors, may be difficult in a single book. At the same time, mechanization of some activities, such as packaging and grading, may be adopted at later stages when demand for high-value processed foods rises substantially, at which point mechanization of power-intensive activities would have sufficiently spread. Nevertheless, as is shown in this chapter, some of the recommendations in the book may also apply to postproduction stages, and therefore our book provides a useful framework that can be adopted for analyses of postharvest mechanization issues in the future. a theoretical Framework for the Evolving Paradigm Demand for mechanization depends on the level of agricultural intensifica- tion, land–labor ratios, labor costs, and the development of a hiring market. We first summarize the theoretical framework around these components and describe how our framework builds on elements of the conventional frame- work, before examining how it explains trends in Asia and Africa. Farming Systems Evolution One of the elements of the conventional framework is Boserup’s theory of endogenous farming systems evolution, established in her seminal work, The Conditions of Agricultural Growth (1965). PBB’s hypothesis can be seen as an extension of this theory. Boserup argued that the evolution of farming systems is an interactive, endogenous process driven by increasing population pressure and rising demand for agricultural products through market development. Farmers respond to this process by shortening fallow periods and intensifying CHAPTER 1: AN EvOLvINg PARAdIgm FOR AFRICA ANd SYNTHESIS OF THE LESSONS FROm ASIA 13 production by adopting modern inputs, such as improved seeds, fertilizers, and agrochemicals, to increase land productivity (Boserup 1965). The hypoth- esis was further developed in Ruthenberg’s 1971 book, Farming Systems in the Tropics, which saw its third edition published in 1980. Ruthenberg specified that when farming systems move from long fallows to short fallows (at most two years of fallow per year of cultivation) or annual cultivation, plowing— whether with animals or tractors—becomes necessary to limit weed growth as well as to bring nutrients to the surface of the soil (Ruthenburg 1980). Before this stage, stumps and other field obstacles make plowing more challeng- ing, especially where the use of appropriate plows (such as disc plows) is not profitable (PBB). However, once this stage is reached, the grasses that emerge between seasons cannot be removed by burning. Consequently, labor require- ments become too high for manual hoeing alone (Boserup 1965; Ruthenburg 1980). The crux of PBB’s hypothesis focused on emphasizing that demand for plowing is insufficient if viewed from the evolution stage of such a farming system in significant parts of Africa, and that this principle had largely con- strained adoption of mechanization until recently. PBB further hypothesized that mechanization is a sequential process adopted at different stages of agricultural intensification, with a shift from human muscle to animal power to machinery. The most power-intensive oper- ations, such as plowing and threshing, are mechanized before harvesting is mechanized. Figure 1.1 illustrates these sequences according to intensifica- tion level, power source, and functions mechanized. In PBB’s view, in the areas where animal traction is feasible, bypassing it to move directly from hand hoes to tractors is not cost-effective due to the costs of destumping fields and forgo- ing the benefits of animal by-products. They demonstrate that although try- panosomiasis is a major constraint to developing draft animals, it becomes less of a problem as population density increases and more forests are converted to crop fields. This book’s online Appendix 1D discusses in greater detail how the patterns in Figure 1.1 have, in fact, been widely observed in Asia and elsewhere. PBB’s fundamental explanation of the low adoption of mechanization in Africa focused on African small farmers, for whom farming systems had not evolved sufficiently for them to demand plowing using tractors. However, some African countries did have more tractors than many Asian countries prior to the 1960s, because of their colonial history and related farmland dis- tribution in which large-scale farming was carried out mostly by white settlers who had a long history of mechanization (Acemoglu, Johnson, and Robinson 2002). Excluding such large-scale farms that still existed after independence, 14 PART 1: SYNTHESIS OF THE LESSONS the smaller African farmers continued to follow farming systems character- ized by long- to medium-fallow stages, and their agricultural products faced relatively inelastic demand due to low population density, lack of urbaniza- tion, and poor market access. Thus, there was limited market demand for mechanization among the majority of African farmers, and mechanized ser- vices were predominantly provided by the public tractor hiring scheme. In the few systems in which mechanization was concentrated, PBB found it to be highly correlated with intensification levels and to have developed with lim- ited government intervention. Induced Technological Change Equally important for understanding mechanization is Hayami and Ruttan’s induced technological innovation theory (1970, 1985). The simple intu- ition behind this theory is that the public and private sectors are driven to develop and adopt technologies that can help to overcome constraints caused by the scarcity of factor endowments—land or labor. Under this theory, in addition to technology innovation, institutional innovations, which include agricultural R&D as well as changes in property rights, tenancy, and labor arrangements, are expected to respond similarly to agricultural endowments. The induced innovation theory explains why mechanization, a labor-saving technology, was adopted much earlier in land-abundant North and South American countries than elsewhere; it also explains why Japan and some other land-constrained Asian countries first adopted land-saving technologies such as high-yielding varieties and intensive use of fertilizer, before machine power replaced animal power. FIGuRE 1.1 Overview of sequential adoption of mechanization according to Pingali, Bigot, and Binswanger (1987) Intensification level Low (forest and bush fallow) Medium (short fallow) High (annual cultivation) Industralization Source of power Human Human and animal Animal and machine Machine Functions mechanized None Plowing (animal) Plowing, threshing, harvesting, milling Seeding, weeding, winnowing, harvesting Source: Adapted from Pingali, Bigot, and Binswanger (1987). CHAPTER 1: AN EvOLvINg PARAdIgm FOR AFRICA ANd SYNTHESIS OF THE LESSONS FROm ASIA 15 The intensification process described by PBB requires greater labor input in the beginning, in response to a growing labor endowment resulting from the growing population density. At the later stage, the industrial sector’s pull of labor out of agriculture and emigrants’ remittances to rural areas cause a labor shortage in agriculture and a rising rural wage rate, which lead to more modern mechanization development, as has been experienced in the United States, Japan, and more recently, other Asian countries (Hayami and Ruttan 1970; Binswanger-Mkhize 2017). This outline shows the importance of an overall economic transformation process for agricultural mechanization. Importantly, however, the mechanization process also varies across coun- tries and regions. For example, many African countries have already expe- rienced rapid urbanization and a growing service sector that leads to labor movement out of the agricultural sector, but their domestic food produc- tion responds to such transformation less through agricultural intensification than through other means. That is, despite the fact that urbanization leads to growth in domestic demand for food, that demand is increasingly met by imports. Understanding the full mechanisms underpinning the relationship between overall economic transformation and agricultural mechanization, including why some Asian countries have seen greater mechanization com- pared with some in Africa as a response to structural transformation, is there- fore important. For mechanical technologies, market incentives have generally been consid- ered more effective than biological technology in inducing innovation, includ- ing in countries such as the United States (Hayami and Ruttan 1985). Unlike the farming systems theory, the induced innovation theory explicitly considers public institutions as part of the technology development and adoption pro- cess, recognizing that technological change is unlikely to originate solely and automatically from the evolution of farming systems, and instead is likely also to require institutional innovations within both the public and private sectors. DISTINCTION BETWEEN PBB FRAMEWORK AND INDUCED INNOVATION THEORY Micro-elements of the PBB framework are consistent with induced innovation theory, yet there are important distinctions between the two. For example, conventional induced innovation theory does not explain explicitly why mech- anization (or innovation toward mechanization) did not emerge in Africa before the farming system evolution, when land was more abundant than labor. Induced innovation theory, such as that described by Hayami and Ruttan (1970, 1985), is largely founded on the existence of a highly intensified 16 PART 1: SYNTHESIS OF THE LESSONS farming system, such as that of the United States and Japan. Whereas the premise of the theory is that relative land abundance induces innovation in land-complementary technologies, authors such as Hayami and Ruttan (1970, 1985) have shown that this pattern holds in already highly intensified farming systems. The theory does not imply that mechanization could emerge wher- ever land is abundant, regardless of farming system. At the preintensifica- tion stage, the traditional land-complementary strategy for farming is shifting cultivation and fallowing, rather than developing mechanical technologies. Mechanical technologies, which are labor-saving and land-complementary, require sufficient market demand for agricultural outputs, which comes from increased population density and urbanization, as described in PBB. To understand mechanization in Africa in the early days, it is impor tant to integrate induced innovation theory with the PBB framework because induced innovation alone cannot explain when and why mechanization occurred or did not occur in Africa, even though many countries in the con- tinent are relatively more land abundant than Asia. To explain the evolution of the land-complementarity of mechanical technologies, induced innovation theory has to rely on the PBB framework, which draws on the farming systems evolution hypothesis. Both induced innovation theory and the farming systems evolution hypothesis relate to the broader demand-side drivers of technological change, and the PBB framework focuses specifically on the demand-side factors appli- cable to mechanization. We now turn to supply-side factors, which are less prominently considered as binding factors in PBB’s framework. The Supply Side of Mechanization: Hiring Markets and Market Failures One of the important components of our updated framework does not trace back to any particular strand of literature; instead, it focuses on the supply- side issue of mechanization, addressing market failures relating to agricul- tural machinery investment and mechanization custom hiring services—the most common mode of mechanization among smallholders in developing countries. In developing countries dominated by smallholders with limited wealth, most farmers are unlikely to be able to afford a tractor or other large machinery. Hiring in services often becomes the only way for many farmers to access mechanization. At the same time, farm sizes are often not large enough for tractor owners to fully utilize their machinery. Thus, hiring out services becomes necessary for owners to be able to recoup their investments. In this mechanized service market, private owner-operators are almost invariably the CHAPTER 1: AN EvOLvINg PARAdIgm FOR AFRICA ANd SYNTHESIS OF THE LESSONS FROm ASIA 17 most efficient way of supplying hiring services. First, private owner-operators have incentives for maximizing tractor utilization, which may not be the case for government hiring schemes. Second, on-farm benefits of tractor ownership ensure that owners can conduct plowing and other field operations on time for their own land. Third, farmer owner-operators have low risk associated with machine damage caused by irresponsible behaviors of some hired operators. The ability of the supply of hiring services to meet the growing demand for mechanization among small- and medium-scale farmers depends on the many factors affecting the decision of a few would-be buyers to invest in a trac- tor. For investment in a tractor to be viable, the revenues from hiring out ser- vices plus the timeliness benefits from using a tractor on one’s own fields, less the costs of fuel, maintenance, repairs and spare parts, payment to operators (when they are hired), machinery depreciation, and loan interest payments, must be enough to offset the investment over the course of the tractor’s use- ful life (Houssou, Diao, and Kolavalli 2014). The opportunities for hiring out services are therefore key to maximizing utilization rates in a way that ensures profitable ownership of a tractor or combine harvester. Although tractors can theoretically operate for 800–1,200 hours per year, short plowing periods determined by rainfall and temperature conditions can reduce this capacity to 300 hours per year (Hunt 1983; Culpin 1988). In addition to the length of the plowing season, achieving a break-even rate for the investment depends on many economic and technical factors that affect the development of the hiring market. First, without sufficient demand among farmers for mechanization services in their home areas, ownership of a tractor is unlikely to be profitable for a medium- or even large-scale farmer. On the other hand, farmers’ demand for hiring services depends not only on farming system evolution and the relative scarcity of labor at the national level, but also on whether the expected benefit of mechanization services out- weighs the service charges, a payoff that requires a high enough productivity level. The market price of hiring services, in turn, depends on the competitive- ness of service markets with enough service providers and, further, the pro- viders’ operating costs. If agricultural returns are low due to low productivity, then demand for paid services may be low. Given that returns on investment in tractors are determined by the utilization rate, low or uncertain demand in the local hiring market negatively affects the decision of a would-be buyer to invest in a tractor, particularly when long-distance mobility in service provi- sion is limited. Second, the utilization rate of a tractor—and hence opportunities for prof- itable tractor ownership in rainfed systems—greatly depends on the length 18 PART 1: SYNTHESIS OF THE LESSONS of the planting window, which may be as short as 30 days in semi-arid areas (Mrema, Kienzle, and Mpagalile 2018). This makes it extremely difficult in such contexts to reach a break-even point in investment and magnifies the cost of a tractor breakdown or other delay. Therefore, opportunities for providing multifunctional hiring services with a tractor, beyond plowing, can be vital. This may be achieved by using the tractor for water pumping, maize shelling, or processing of other crops, or for other functions such as transport, although certain stationary power applications such as pumping may be less relevant in areas where general motorized pumps or solar pumps are emerging as alter- natives. Opportunities for multifunctional tractor use depend on farmers’ demand for additional hiring services, which may be low in the places where an irrigation system is beyond reach for most farmers or where small-scale irri- gation technologies are underdeveloped. Farmers in these areas may not have adopted a practice, such as harrowing or multiple plowing, that requires trac- tor use multiple times in land preparation. Migratory service provision in plowing increases utilization by allowing tractor owners to use their machines for a longer period of the year by exploiting geographical variation in seasons. However, these opportunities may not currently exist in many African coun- tries. Migratory services rely on better road infrastructure, which depends on public investment. Migratory services are also subject to coordination failures, and it is unlikely for individual tractor owners to gauge the service demand and connect with customers in locations beyond their home areas. These issues can be particularly serious in Africa, more so than in Asia, because road infrastructure is poorer and service market networks are underdeveloped at an early stage of mechanization in Africa. Transporting tractors over long distances can be prohibitively costly where physical infrastructure is poor. Alternatively, identifying medium-scale commercial farmers and encourag- ing their growth in a way that raises returns on machinery can also stimulate the growth of potential suppliers of custom hiring services (Mpanduji 2000; Agyei-Holmes 2014). Even within a locality, significant obstacles hinder the efficient utilization of tractors, especially where plots are small and fragmented. In an area with small farm sizes, especially if the timing of production among small farmers is not uniform, traveling between plots, as well as turning and other maneu- vers in a small plot, increases time and fuel consumption. Again, this can be especially serious in Africa, where the dominant types of tractors are larger and the road infrastructure is poorer than in Asia. These obstacles all cut into the margins of tractor operation. One way to offset this disadvantage of scale in farm size is for small farmers to coordinate planting and to jointly hire a CHAPTER 1: AN EvOLvINg PARAdIgm FOR AFRICA ANd SYNTHESIS OF THE LESSONS FROm ASIA 19 tractor for plowing their fields at once, but this would require coordination efforts beyond tractor owners’ capacity. Moreover, there is a steep learning curve for both the technical and business aspects of tractor ownership and operation in hiring markets, which implies additional risks for tractor invest- ment. The complex soil and field conditions across small farms require expe- rience, which takes time for owners or hired operators to acquire. Otherwise, stumps and other obstacles hidden in unfamiliar fields can easily damage tractors. Apart from these risks, traditional land tenure systems can not only limit the consolidation of farmland but also prevent the land from being used as collateral, making credit for tractor purchases unavailable to many would-be buyers who are farmers. This is especially prevalent in countries with partic- ularly weak land tenure security. Thus, the lump-sum investment required becomes unfeasible for many potential owners. Even where credit is available, interest rates are often too high to be attractive for would-be buyers. The availability of appropriate technology is also imperative for hiring- market development. All of the potential market failures described above are exacerbated when tractor sizes are large. Larger, higher-horsepower tractors are generally more expensive, require higher utilization rates for breakeven, and possess higher barriers to entry than smaller models. Thus, larger trac- tor size exposes owners to greater hiring-market risk because any major delay or coordination failure has greater consequences in terms of recouping the higher investment cost. Thus, it is important to strike a balance between a tractor powerful enough to effectively plow local soils and one small and cheap enough to be owned and operated cost-efficiently in areas where those smaller tractors would be in fact more suitable. However, in Africa, with hir- ing markets in the early stages, the tractors available are manufactured for and previously owned in other countries. With limited knowledge of the suitabil- ity of different tractors for different within-country soil conditions, achiev- ing the balance between size and efficiency of tractors is beyond the capacity of individual owners or the private sector. These challenges are more severe in Africa, where soil conditions and other production environments are generally more diverse than in Asia (World Bank 2007), yet the public information to address them is limited. The evidence is still insufficient as to what the opti- mal size of tractors is in Africa, although small tractors, including two-wheel tractors, have been promoted in Africa from time to time over the course of several centuries. Continuous research is needed to shed more light on this issue and provide information to policymakers and stakeholders about the size of tractors. 20 PART 1: SYNTHESIS OF THE LESSONS Familiarity with animal traction can facilitate the adoption of tractors beyond the sequential nature of farming systems evolution, because in this case tractors are adopted simply to substitute for animal traction. Although the transition from animal traction to tractors requires learning a new tech- nology, in the places where farmers go straight from hand hoes to tractors, they must learn not only the new technology but also new land prepara- tion practices. In a society with established animal traction, formal or infor- mal hiring markets already exist for draft animal services. Moving from a tradition of hiring animals for land preparation to hiring tractors for plow- ing is therefore a much faster process, in terms of both new technology adop- tion and tractor hiring market development. Having used animal traction also helps new tractor owner-operators shorten their learning period for ser- vice provision. Of course, the potential of draft animals must also be evalu- ated against the risks of owning the animals, such as disease and competition with the growing demand for livestock products. Where feasible, leapfrogging draft animal technology should remain as one of the options (Kormawa et al. 2018). However, as is described in this book, animal power often preceded the growth of mechanical power in developing countries in the 20th century and has spread considerably in parts of Africa in the last few decades, suggesting that lessons from these experiences can be applied to other parts of Africa in the future. To summarize, we integrate induced technological innovation theory and market failure challenges in the development of markets for hiring services with PBB’s farming systems hypothesis to better explain contemporary mech- anization trends under the new paradigm. PBB focused on the relative effi- ciency of the private sector compared with the public sector in overcoming some of the aforementioned challenges associated with hiring-service oper- ations. In contrast, we emphasize that the private sector continues to face the remaining challenges, and the public sector must still play an active role in mitigating these challenges, not through direct interventions in hiring- service schemes, but instead through other measures. Demand is still a nec- essary precondition for adoption. Demand depends not only on farming systems but also on the availability of labor relative to land in the context of broader economic transformation. The development of mechanization hir- ing markets is constrained by many factors that can slow down the sponta- neous supply response of mechanization services. Our framework recognizes that certain market failures associated with investment risk and hiring mar- ket development are significant, with some form of public support required to overcome them. One key difference from PBB is that we highlight more CHAPTER 1: AN EvOLvINg PARAdIgm FOR AFRICA ANd SYNTHESIS OF THE LESSONS FROm ASIA 21 the complex nature of the demand for mechanization of specific activities, such as land preparation. Whereas PBB illustrated the nature of demand in a broader scheme, including its sequential nature, they focused relatively less on, for example, its variations across farm households or at the intensive margins (for example, the number of times each plot is plowed). The key will be under- standing the extent and determinants of such variation. Case Study Countries in the Book The validity of the theoretical framework presented in the previous section is assessed drawing on collective evidence from a set of countries. Key aspects of this framework are further investigated through focused empirical analyses from a subset of countries. Theoretical Framework and Country Chapters Countries covered in this book are selected based on various criteria. First, we focus on countries in which substantial research has been done by the International Food Policy Research Institute and other CGIAR centers so that we can make the best use of existing research results. In addition, Asian countries are selected to represent diversity in mechanization experiences with different levels of economic development and manufacturing capacity, and various mechanization-sector development patterns. Selected Asian coun- tries have also developed various mechanization supply models (for example, Bangladesh, China, and India, as highlighted in Diao et al. 2014). Although some Asian countries are not included due to the unavailability of substantial research with relatively detailed empirical assessments, the information that is available for excluded Asian countries (for example, in FAO and CSAM 2014) is generally consistent with the key patterns described in this book. Although China and India are much larger than other Asian countries included, their lessons can potentially be useful for subregional approaches within Africa, which have been promoted in recent years (Kormawa et al. 2018). African countries are selected from East Africa (Ethiopia, Kenya, and Tanzania) and West Africa (Ghana and Nigeria). Four of these countries are among the largest countries in Africa in terms of economically active popu- lation in agriculture, and they together (plus Ghana) account for 40 percent of all of the economically active population in Africa’s agriculture sector and more than 30 percent of Africa’s arable land, including the nation of South Africa and some northern African countries (USDA ERS 2018; GGDC 22 PART 1: SYNTHESIS OF THE LESSONS 2019). South Africa and some countries in northern Africa (such as Egypt) are much more advanced in mechanization than other areas due to domination of either large-scale commercial farms (in the case of South Africa) or irrigated agriculture (in the case of Egypt). Therefore, they are excluded from this book, which focuses, in the case of African countries, on smallholders and rainfed agriculture. For similar reasons, Latin America is not covered by the book. Although a general framework is provided in the previous section, many chapters of the book adopt different approaches to addressing this framework. Table 1.1 summarizes how the case study countries covered in this book are collectively linked to each key aspect of the theoretical framework described in the previous section. In addition, an online appendix, Appendix 1C, that provides a brief review of other African countries’ experiences, particularly Francophone countries and Lusophone countries, as well as experiences of some Latin American countries, is included as a supplement to the book. As is described in the over- view in Appendix 1C, although the experiences in the countries have been by no means identical, they do have much in common with those of the countries explicitly covered in this book. Empirical Framework and Approaches The empirical approach also differs across chapters also because of the avail- able primary data. Table 1.2 summarizes the types of data covered in each chapter’s empirical analysis. Note that Table 1.2 omits five country chap- ters (Chapters 4, Sri Lanka; 5, Thailand; 8, Myanmar; 11, Ghana; and 14, Tanzania), which include no quantitative empirical analyses due to the unavailability of data. The Viet Nam, Bangladesh, Kenya, and Nigeria chapters focus on the determinants of mechanization adoption. The China chapter focuses on the importance of machine rental, assessing the effect of machine rental on agri- cultural production using a structural production function. Similarly, the Nepal chapter focuses on the effects of machine rental on production technol- ogy characteristics and provides further insights into the effects on returns to scale. The India, Nepal, and Nigeria chapters focus on detailed aspects of the associations between tractor adoption and agricultural input uses (land, labor, draft animals, and other inputs such as chemical fertilizer or irrigation). The Bangladesh, Nepal, and Nigeria chapters assess the associations between trac- tor adoption and off-farm labor supply or incomes. The Nepal and Nigeria chapters also assess th