Africa’s Changing Farmland Ownership: The Rise of the Emergent Investor Farmer T.S. Jayne, Jordan Chamberlin, Lulama Ndibongo Traub, N. Sitko, Milu Muyanga, Felix K. Yeboah, Chewe Nkonde, Ward anseeuw, A. Chapoto, and Richard Kachule Presentation at the Annual World Bank Conference on Land and Poverty Washington D.C March 14 – 18, 2016 1. Need IAPRI’s logo

Presentation Outline Africa’s Structural Transformation: Which Way? Rapid Changes in Africa’s Farmland Structure Question: Do we have the right structural transformation model for Africa? Conclusion: Guiding policy-makers Data source: Medium-scale farm survey in Mchinji, Kasungu and Lilongwe Districts undertaken by LUANAR, UP, and MSU (2014-2015). The rise of medium-scale farms seems to have been especially rapid in recent years, since 2005. For example, Anseeuw et al (2015) shows that the land controlled by medium-scale holdings (between 5 and 100 ha) in three districts of Malawi has almost doubled between 2000 and 2015, from 2,544 ha in 2000 to 4,726 ha in 2015 (Figure 4). In the 2005-2015 period, the land under medium-scale holdings has increased by 49.1 percent. If the trends documented in these three districts of Malawi provide a general indication of changes in landholdings in the country, Anseeuw et al. (2015) conclude that roughly 300,000 hectares has been newly acquired by medium/large-scale holders since 2005, slightly more than 10% of the total area under cultivation in Malawi. The apparent rapid rise in medium-scale farms over the past decade in Malawi is truly remarkable in a country where the majority of rural people face acute land scarcity and where household poverty is highly correlated with very small farm size.

Africa’s Structural Transformation: Which Way? North-South Divide: What structural transformation process is Africa following? Is it; Asian-style structural transformation process with its “unimodal” small-scale farming system? Or, Latin America’s Bimodal farming system with its highly concentrated farm structure?

Brazilian-style Structural Transformation Bimodal Agricultural System Skewed land distribution 65% Source: OECD-FAO Agricultural Outlook 2015

Asian-style Structural Transformation Unimodal agricultural system Low Gini coefficient on land ownership distribution Notes: 1. Starts with the Green Revolution  Rising agricultural productivity through intensification Surplus production and rising on-farm income  Rising demand for goods & services from off-farm sectors Rising labor demand in non-ag sectors Rural-urban migration and gradual urbanization  Agriculture’s share of total GDP declines. 2. Development thinkers assume the standard structural transformation models to be the best fit for the African-context 3. Conventional wisdom would tell us, yes, this form of transformation would be the best fit

Will Africa’s Structural Transformation unfold as it did in Asia? North-South Divide: What structural transformation process is Africa following? Is it; Asian-style structural transformation process with its “unimodal” small-scale farming system? Or, Latin America’s Bimodal farming system with its highly concentrated farm structure?

Rapid Changes in Africa’s Farm Structure Rise of the ‘emergent’ medium-scale Farmer: Three sub-categories of medium-scale farmers (5 – 100 ha) “emergent” urban-based investor farmers ~ 60% “elite” rural-based farmers ~ 35% Successful smallholder expansion farmers ~ 5% Note: Rough estimates across three countries – Kenya, Zambia and Southern Ghana Sharolyn: Move the “Categories of medium-scale farmers up…and earlier”

Evidence of medium-scale farmers   Number of farms % growth in number of farms % of total cultivated area Ghana 1992 2005 0-2 ha 1,458,540 1,725,024 18.3 25.1 12.5 2-5 ha 578,890 957,722 65.4 35.6 24.1 5-10 ha 116,800 256,620 119.7 17.2 14.6 10-20 ha 38,690 110,076 184.5 11 12 20-100 ha 18,980 46,143 143.1 11.1 11.7 >100 ha -- 6,958 388.6* 25 Total 2,211,900 3,102,543 100 Sources: Ghana Living Standards Surveys 1992/3, 1999/2000 and 2005/2006. * Percentage change from 1999 to 2005. Source: Ministry of Agriculture Crop Forecast Surveys, 2009, 2012. *2001 figures are land under cultivation. ** Growth rate computed from 2009-2012 only. “na” means not available. Source: Kenya Central Bureau of Statistics, Welfare Monitoring Survey II, 1994: Basic Report (Kenya: Central Bureau of Statistics, Office of the Vice-President and Ministry of Planning and National Development, 1996). Kenya National Bureau of Statistics, Kenya Integrated Household Budget Survey 2005-2006 (Nairobi, Kenya: Kenya National Bureau of Statistics - Ministry of Planning and National Development, 2006). *** For reasons explained in Section 3, we believe that the official Kenya data sets underreport farms over 10 hectares. Overall: the absolute number of farms of all sizes is increasing over time, both the number of small-scale farms and the area under cultivation by small-scale farms are growing at a much lower rate than farms in the 5-10 and 10-100 hectare range. In other words, medium-scale farms are accounting for a rising proportion of total farmland. Consequently, the share of small-scale farms in total area under cultivation is falling in some countries even as the numbers of small farms continue to rise. According to official data, farms over 10 hectares account for 48.7% of total area under cultivation in Ghana, 28% in Kenya, and 27% in Zambia (Table 1). Kenya: In highly land-constrained Kenya, rural population growth and land subdivision has led to an alarming rise in the proportion of very small farms (Table 1). Between 1994 and 2006, the proportion of Kenya’s farms smaller than one hectare rose from 44.8 to 67.2 percent. Over 95 percent of Kenya’s farms were below three hectares in 2006, up from 83.3 percent in 1994. We also see a major decline in the number of farms over 10 hectares, again likely due to subdivision. However, within this largest farm size category, we see a curious 230 percent increase in average landholding size over this twelve-year period, from 13.2 to 31.1 hectares, and the percentage of total farmland accounted for by this group actually rose from 24.5% to 28%. While we cannot conclusively identify the reason for this increase, it is consistent with evidence that shows rapid new entries of relatively large landowners, and/or consolidation of farms in this size category, even as the overall number of farms in this size category declines In Kenya’s case, farmland held by large-scale domestic owners is possibly grossly under-reported. Namwaya (2004) reports that over 600,000 ha, or roughly one-seventh of Kenya’s total land under cultivation, are held by the families of the country's three former presidents, and that most of this land is in relatively high-potential areas. Source: Ghana Living Standards Surveys

% growth in number of farms % of total cultivated area Evidence of medium-scale farmers   Number of farms % growth in number of farms % of total cultivated area Zambia 2001 2012 0 – 2 ha 638,118 748,771 17.3 34.1 16.2 2 – 5 ha 159,039 418,544 163.2 45 31.7 5 – 10 ha 20,832 165,129 692.6 14.3 25 10 – 20 ha 2,352 53,454 2272.7 6.6 15 20 – 100 ha -- 13,839 53.3** 12 Total 820,341 1,399,737 100 Source: Ministry of Agriculture Crop Forecast Surveys, 2009, 2012. *2001 figures are land under cultivation. ** Growth rate computed from 2009-2012 only. “na” means not available. Source: Kenya Central Bureau of Statistics, Welfare Monitoring Survey II, 1994: Basic Report (Kenya: Central Bureau of Statistics, Office of the Vice-President and Ministry of Planning and National Development, 1996). Kenya National Bureau of Statistics, Kenya Integrated Household Budget Survey 2005-2006 (Nairobi, Kenya: Kenya National Bureau of Statistics - Ministry of Planning and National Development, 2006). *** For reasons explained in Section 3, we believe that the official Kenya data sets underreport farms over 10 hectares. Overall: the absolute number of farms of all sizes is increasing over time, both the number of small-scale farms and the area under cultivation by small-scale farms are growing at a much lower rate than farms in the 5-10 and 10-100 hectare range. In other words, medium-scale farms are accounting for a rising proportion of total farmland. Consequently, the share of small-scale farms in total area under cultivation is falling in some countries even as the numbers of small farms continue to rise. According to official data, farms over 10 hectares account for 48.7% of total area under cultivation in Ghana, 28% in Kenya, and 27% in Zambia (Table 1). Kenya: In highly land-constrained Kenya, rural population growth and land subdivision has led to an alarming rise in the proportion of very small farms (Table 1). Between 1994 and 2006, the proportion of Kenya’s farms smaller than one hectare rose from 44.8 to 67.2 percent. Over 95 percent of Kenya’s farms were below three hectares in 2006, up from 83.3 percent in 1994. We also see a major decline in the number of farms over 10 hectares, again likely due to subdivision. However, within this largest farm size category, we see a curious 230 percent increase in average landholding size over this twelve-year period, from 13.2 to 31.1 hectares, and the percentage of total farmland accounted for by this group actually rose from 24.5% to 28%. While we cannot conclusively identify the reason for this increase, it is consistent with evidence that shows rapid new entries of relatively large landowners, and/or consolidation of farms in this size category, even as the overall number of farms in this size category declines In Kenya’s case, farmland held by large-scale domestic owners is possibly grossly under-reported. Namwaya (2004) reports that over 600,000 ha, or roughly one-seventh of Kenya’s total land under cultivation, are held by the families of the country's three former presidents, and that most of this land is in relatively high-potential areas. Source: Zambia MoA Crop Forecast

% growth in number of farms % of total cultivated area Evidence of medium-scale farmers   Number of farms % growth in number of farms % of total cultivated area Kenya 1994 2006 0 – 2 ha 1,692,343 2,640,020 56 29.2 46.4 2 – 5 ha 525,363 332,011 -36.8 32.3 23.5 5 – 10 ha 93,871 17,451 -81.4 21.4 2.1 > 10 ha 92,498 19,493 -78.9*** 24.5 28 Total 2,404,075 3,008,975 100 Average Landholding Size: 1994: 13 ha 2006: 31 ha Source: Kenya Central Bureau of Statistics, Welfare Monitoring Survey II, 1994: Basic Report (Kenya: Central Bureau of Statistics, Office of the Vice-President and Ministry of Planning and National Development, 1996). Kenya National Bureau of Statistics, Kenya Integrated Household Budget Survey 2005-2006 (Nairobi, Kenya: Kenya National Bureau of Statistics - Ministry of Planning and National Development, 2006). *** For reasons explained in Section 3, we believe that the official Kenya data sets underreport farms over 10 hectares. Kenya: In highly land-constrained Kenya, rural population growth and land subdivision has led to an alarming rise in the proportion of very small farms (Table 1). Between 1994 and 2006, the proportion of Kenya’s farms smaller than one hectare rose from 44.8 to 67.2 percent. Over 95 percent of Kenya’s farms were below three hectares in 2006, up from 83.3 percent in 1994. We also see a major decline in the number of farms over 10 hectares, again likely due to subdivision. However, within this largest farm size category, we see a curious 230 percent increase in average landholding size over this twelve-year period, from 13.2 to 31.1 hectares, and the percentage of total farmland accounted for by this group actually rose from 24.5% to 28%. While we cannot conclusively identify the reason for this increase, it is consistent with evidence that shows rapid new entries of relatively large landowners, and/or consolidation of farms in this size category, even as the overall number of farms in this size category declines In Kenya’s case, farmland held by large-scale domestic owners is possibly grossly under-reported. Namwaya (2004) reports that over 600,000 ha, or roughly one-seventh of Kenya’s total land under cultivation, are held by the families of the country's three former presidents, and that most of this land is in relatively high-potential areas. Overall: the absolute number of farms of all sizes is increasing over time, both the number of small-scale farms and the area under cultivation by small-scale farms are growing at a much lower rate than farms in the 5-10 and 10-100 hectare range. In other words, medium-scale farms are accounting for a rising proportion of total farmland. Consequently, the share of small-scale farms in total area under cultivation is falling in some countries even as the numbers of small farms continue to rise. According to official data, farms over 10 hectares account for 48.7% of total area under cultivation in Ghana, 28% in Kenya, and 27% in Zambia (Table 1). Source: Kenya National Bureau of Statistics

Rapid Changes in Africa’s Farm Structure Rise of the ‘emergent’ medium-scale Farmer Medium-scale farmers land acquisition > foreign & domestic large-scale investors

African Medium-scale Land Acquisition This we changed from a table to a bar graph. Source: Jayne, et. al., 2014 (JIA)

Rapid Changes in Africa’s Farm Structure Rise of the ‘emergent’ medium-scale Farmer Medium-scale farmers land acquisition > foreign & domestic large-scale investors Largely urban-based and/or politically connected

% of National Landholdings held by Urban Households While many other regions of the world started to experience negative rural population growth soon after their economic transformation phase started, SSA is the only region of the world that is continuing to experience rural population growth, currently at 1.8% annually. There are projected to be 48% more people in rural Africa in 2050 than in 2015. This implies that farm consolidation in much of Africa will occur more slowly than in regions that have experienced rural depopulation for some time. Source: DHS

Categories of “emergent” farmers Type 1A: Investor Farmer Mode of entry into medium-scale farming status ---------- Zambia ---------- ---------- Kenya ----------  Acquisition of land from non-farm employment Acquisition of land from non-farm employment (n=164) (n=180) % of cases 58 60 % men 91.4 80 Year of birth 1960 1947 Years of education of head 11 12.7 Have held a job other than as a farmer (%) 100 83.3 Formerly or currently employed by the public sector (%) 59.6 56.7 Initial landholding size when started farming (ha) 106.6 22.6 Current landholding size (ha) 74.9 50.1 % of land currently under cultivation 24.7 46.6 Decade when land was acquired 1969 or earlier 1.1 6 1970-79 5.1 18 1980-89 7.4 20 1990-99 23.8 32 2000 or later 63.4 25 1. Move earlier & add source Source: MSU, UP, and ReNAPRI Retrospective Life History Surveys, 2015

Categories of “emergent” farmers Type 1A: Investor Farmer Mode of entry into medium-scale farming status ---------- Zambia ---------- ---------- Kenya ----------  Acquisition of land from non-farm employment Acquisition of land from non-farm employment (n=164) (n=180) % of cases 58 60 % men 91.4 80 Year of birth 1960 1947 Years of education of head 11 12.7 Have held a job other than as a farmer (%) 100 83.3 Formerly or currently employed by the public sector (%) 59.6 56.7 Initial landholding size when started farming (ha) 106.6 22.6 Current landholding size (ha) 74.9 50.1 % of land currently under cultivation 24.7 46.6 Decade when land was acquired 1969 or earlier 1.1 6 1970-79 5.1 18 1980-89 7.4 20 1990-99 23.8 32 2000 or later 63.4 25 1. Move earlier & add source Source: MSU, UP, and ReNAPRI Retrospective Life History Surveys, 2015

Categories of “emergent” farmers Type 1A: Investor Farmer Mode of entry into medium-scale farming status ---------- Zambia ---------- ---------- Kenya ----------  Acquisition of land from non-farm employment Acquisition of land from non-farm employment (n=164) (n=180) % of cases 58 60 % men 91.4 80 Year of birth 1960 1947 Years of education of head 11 12.7 Have held a job other than as a farmer (%) 100 83.3 Formerly or currently employed by the public sector (%) 59.6 56.7 Initial landholding size when started farming (ha) 106.6 22.6 Current landholding size (ha) 74.9 50.1 % of land currently under cultivation 24.7 46.6 Decade when land was acquired 1969 or earlier 1.1 6 1970-79 5.1 18 1980-89 7.4 20 1990-99 23.8 32 2000 or later 63.4 25 1. Move earlier & add source Source: MSU, UP, and ReNAPRI Retrospective Life History Surveys, 2015

Categories of “emergent” farmers Type 1A: Investor Farmer Mode of entry into medium-scale farming status ---------- Zambia ---------- ---------- Kenya ----------  Acquisition of land from non-farm employment Acquisition of land from non-farm employment (n=164) (n=180) % of cases 58 60 % men 91.4 80 Year of birth 1960 1947 Years of education of head 11 12.7 Have held a job other than as a farmer (%) 100 83.3 Formerly or currently employed by the public sector (%) 59.6 56.7 Initial landholding size when started farming (ha) 106.6 22.6 Current landholding size (ha) 74.9 50.1 % of land currently under cultivation 24.7 46.6 Decade when land was acquired 1969 or earlier 1.1 6 1970-79 5.1 18 1980-89 7.4 20 1990-99 23.8 32 2000 or later 63.4 25 1. Move earlier & add source Source: MSU, UP, and ReNAPRI Retrospective Life History Surveys, 2015

Categories of “emergent” farmers Type 1B: Rural-based Elite Mode of entry into medium-scale farming status -------- Zambia -------- -------- Kenya -------- Expansion of rural elite (n=118) (n=120) % of cases 42 40 % men 92.9 82.5 Year of birth 1966 1945 Years of education of head 8.2 7.5 Have held a job other than as a farmer (%) 32.9 17.5 Formerly or currently employed by the public sector (%) 5.8 12.5 Initial landholding size when started farming (ha) 28.8 14 Current landholding size (ha) 38.2 32.7 % of land currently under cultivation 46.9 54.1 Decade when land was acquired 1969 or earlier 3.9 29 1970-79 6.7 24 1980-89 14.8 20 1990-99 32.2 18 2000 or later 9 1. Move earlier & add source Source: MSU, UP, and ReNAPRI Retrospective Life History Surveys, 2015

Categories of “emergent” farmers Type 1B: Rural-based Elite Mode of entry into medium-scale farming status -------- Zambia -------- -------- Kenya -------- Expansion of rural elite (n=118) (n=120) % of cases 42 40 % men 92.9 82.5 Year of birth 1966 1945 Years of education of head 8.2 7.5 Have held a job other than as a farmer (%) 32.9 17.5 Formerly or currently employed by the public sector (%) 5.8 12.5 Initial landholding size when started farming (ha) 28.8 14 Current landholding size (ha) 38.2 32.7 % of land currently under cultivation 46.9 54.1 Decade when land was acquired 1969 or earlier 3.9 29 1970-79 6.7 24 1980-89 14.8 20 1990-99 32.2 18 2000 or later 9 1. Move earlier & add source Source: MSU, UP, and ReNAPRI Retrospective Life History Surveys, 2015

Categories of “emergent” farmers Type 1B: Rural-based Elite Mode of entry into medium-scale farming status -------- Zambia -------- -------- Kenya -------- Expansion of rural elite (n=118) (n=120) % of cases 42 40 % men 92.9 82.5 Year of birth 1966 1945 Years of education of head 8.2 7.5 Have held a job other than as a farmer (%) 32.9 17.5 Formerly or currently employed by the public sector (%) 5.8 12.5 Initial landholding size when started farming (ha) 28.8 14 Current landholding size (ha) 38.2 32.7 % of land currently under cultivation 46.9 54.1 Decade when land was acquired 1969 or earlier 3.9 29 1970-79 6.7 24 1980-89 14.8 20 1990-99 32.2 18 2000 or later 9 1. Move earlier & add source Source: MSU, UP, and ReNAPRI Retrospective Life History Surveys, 2015

Categories of “emergent” farmers Type 1B: Rural-based Elite Mode of entry into medium-scale farming status -------- Zambia -------- -------- Kenya -------- Expansion of rural elite (n=118) (n=120) % of cases 42 40 % men 92.9 82.5 Year of birth 1966 1945 Years of education of head 8.2 7.5 Have held a job other than as a farmer (%) 32.9 17.5 Formerly or currently employed by the public sector (%) 5.8 12.5 Initial landholding size when started farming (ha) 28.8 14 Current landholding size (ha) 38.2 32.7 % of land currently under cultivation 46.9 54.1 Decade when land was acquired 1969 or earlier 3.9 29 1970-79 6.7 24 1980-89 14.8 20 1990-99 32.2 18 2000 or later 9 1. Move earlier & add source Source: MSU, UP, and ReNAPRI Retrospective Life History Surveys, 2015

Rapid Changes in Africa’s Farm Structure Rise of the ‘emergent’ medium-scale Farmer: 2000 – 2015: emergent farmers land acquisition > foreign investors Largely urban-based and/or politically connected Rising landholding concentration

Movement in Gini coefficient: Rising Landholding Concentration Period Movement in Gini coefficient: Ghana (cult. area) 1992  2013 0.54  0.70 Kenya (cult. area) 1994  2006 0.51  0.55 Zambia (landholding) 2001  2012 0.42  0.49 The rise of medium-scale farms has led to a concentration of landholdings. In the study of Ghana, Kenya and Zambia by Jayne et al (2014), the Gini coefficients of landholdings rose in all three countries substantially, e.g., in Ghana from 0.52 in 1992 to 0.65 in 2005. While landholdings in most of Africa are not as concentrated as in Latin America, where Gini coefficients can be as high as 0.90, the Gini coefficients in the three African case studies are substantially higher than most Asian countries and appear to be rising over time (Jayne et al., 2003, Jayne et al., 2014). Clearly, in all three countries, the idea of a “unimodal” and egalitarian farm structure within Africa’s indigenous farming population has become extraordinarily outdated. Source: Jayne et al. 2014 (JIA)

Question: Do we have the right structural transformation model for Africa? Impact of emerging farmland ownership patterns is poorly understood…

Areas of Impact Changing land markets A shift from customary to statutory tenure system for farmland Question: How is the “modernization” of land institutions affecting access to land by local rural communities?

Rising Land Scarcity Source: UN 2013 While many other regions of the world started to experience negative rural population growth soon after their economic transformation phase started, SSA is the only region of the world that is continuing to experience rural population growth, currently at 1.8% annually. There are projected to be 48% more people in rural Africa in 2050 than in 2015. This implies that farm consolidation in much of Africa will occur more slowly than in regions that have experienced rural depopulation for some time. Source: UN 2013

Area of Impact Changing land markets Changing farmland distribution and the multiplier-effect Question: The strength and location of the multiplier-effect from agricultural growth is currently unknown: Egalitarian land distribution  broad-based growth Urban-based investor farmers  will we see rural poverty reduction?

% of National Landholdings held by Urban Households While many other regions of the world started to experience negative rural population growth soon after their economic transformation phase started, SSA is the only region of the world that is continuing to experience rural population growth, currently at 1.8% annually. There are projected to be 48% more people in rural Africa in 2050 than in 2015. This implies that farm consolidation in much of Africa will occur more slowly than in regions that have experienced rural depopulation for some time. Source: DHS

Question: do we have the right structural transformation model for Africa? Impact of emerging farmland ownership patterns is poorly understood…because… Data issues…

The Data Issues To date available national databases: Sample proportional to population  Systematically under-sample large farms And/or exclude non-smallholder farming sectors by design Urban households  not prompted about farmland owned or cultivated away from main urban residence 1. Fix the size of the font…make it larger.

Farmland distribution in Panama – Ag. Census 1. Include rectangle on click and show both Census and LSM data in the same frame. Source: Lowder et al (2015)

Farmland distribution in Panama – LSMS Source: Lowder et al (2015)

The Data Issues This Study: Integrated Public Use Micro-level Surveys (IPUMS) Demographics and Health Survey (DHS) for urban household farmland ownership Large-scale Crop Forecast Surveys in Zambia MSU/UP/ReNAPRI Retrospective Life History Survey on Medium-scale farmers 1. Change the last bullet point

Conclusions Medium-scale farms: the fastest growing segment of the farm sector in Sub-Saharan Africa. Current institutional systems and methodological approaches for collection of data - systematically missing the most dynamic portion of the sector. Africa’s farmland ownership structures and the degree of land inequality more heterogeneous across space. Implications of rising urban-based investors on agricultural growth spillover effects on rural non-farm economy and on rural poverty need to be examined What does this all mean for African Governments in their attempt to move forward to promote economic development

Thank You