Contract Farming and Rural Transformation: Evidence from a Field Experiment in Benin Aminou Arouna a Jeffrey D. Michler b Jourdain C. Lokossou c a Africa Rice Center (AfricaRice) b Department of Agricultural & Resource Economics, University of Arizona c International Crops Research Institute for the Semi-Arid Tropics (ICRISAT) CSAE Conference, Oxford, 17-9 March, 2019 Arouna, Michler, & Lokossou Contract Farming and Rural Transformation 1 / 17
Introduction Rural Transformation ◮ The process of development is defined by structural transformation (Lewis, 1954) ◮ Requires a transition from subsistence to commercial agriculture (Collier & Dercon, 2014) ◮ Commercialization entails: ◮ Increased farm size ◮ Increased productivity ◮ Increased market participation ◮ Should result in greater income for (surviving) farmers Arouna, Michler, & Lokossou Contract Farming and Rural Transformation 2 / 17
Introduction Production Constraints ◮ Farmers face numerous production constraints that limit their ability to commercialize ◮ Lack access to credit (Berg, 2013; Stephens & Barrett, 2011) ◮ Produce below the production frontier (Hanna et al., 2014; Sherlund et al., 2002) ◮ Face price uncertainty due to limited access to output markets (de Janvry et al., 1991; LaFave & Thomas, 2016) Arouna, Michler, & Lokossou Contract Farming and Rural Transformation 3 / 17
Introduction Contract Farming ◮ Vertical coordination between producer and processor ◮ Integrates farmer into the market ◮ Shifts price risk and capital expenses to processor ◮ Production contracts have become a popular mechanism ◮ Barrett et al. (2012), Bellemare & Bloem (2018), and Ton et al. (2018) review literature ◮ Maertens & Vande Velde (2017) in Benin ◮ To date, almost no RCT evidence on impact of contract farming ◮ Ashraf et al. (2009) - failed export crops scheme Arouna, Michler, & Lokossou Contract Farming and Rural Transformation 4 / 17
Introduction Research Questions 1. What is the impact of contract farming on... 1.1 Rice area planted (ha)? 1.2 Productivity (kg/ha)? 1.3 Market participation (% of rice sold)? 1.4 Rice income per capita (USD)? 2. Which contract attributes are most effective at reducing risk and easing constraints? 2.1 Price guarantee 2.2 Technical assistance 2.3 Input loans Arouna, Michler, & Lokossou Contract Farming and Rural Transformation 5 / 17
Study Design Study Location ◮ Entreprises de Services et Organisations de Producteurs de Bante (ESOP) ◮ Included 953 rice farmers in Benin ◮ Organized into 107 farmer groups ( ∼ 8 per group) ◮ Baseline in 2016 ◮ Randomization and contract signing mid 2016 ◮ Endline in early 2017 Arouna, Michler, & Lokossou Contract Farming and Rural Transformation 6 / 17
Study Design Contracts ◮ Treatment 1 (T1) ◮ Sets price ($0.27 kg) and ◮ Treatment 3 (T3) quantity ◮ Rice variety (IR841) ◮ Everything in T1 and T2 ◮ Date and location of collection ◮ Plus input loans ◮ ∼ 45 kg of IR841 ◮ Treatment 2 (T2) ◮ ∼ 150 kg of fertilizer ◮ Everything in T1 ◮ Control (C) ◮ Plus 3-5 visits from ESOP ◮ No contract extension agents ◮ Provide direction on production decisions Arouna, Michler, & Lokossou Contract Farming and Rural Transformation 7 / 17
Study Design Experimental Design Farm Households ( n = 855 ) ∼ 0 . 75 ∼ 0 . 25 Treatment Pure Control ( n = 635 ) ( n = 220 ) ∼ 0 . 2 ∼ 0 . 4 ∼ 0 . 4 T1 T2 T3 ( n = 114 ) ( n = 252 ) ( n = 269 ) Arouna, Michler, & Lokossou Contract Farming and Rural Transformation 8 / 17
Study Design Contracts Arouna, Michler, & Lokossou Contract Farming and Rural Transformation 9 / 17
Study Design Balance Test [T1-C] [T2-C] [T3-C] [T2-T1] [T3-T2] [T3-T1] Household size − 0 . 888 − 0 . 366 − 0 . 142 0 . 523 0 . 224 0 . 747 ( 0 . 689 ) ( 0 . 546 ) ( 0 . 613 ) ( 0 . 804 ) ( 0 . 722 ) ( 0 . 846 ) Age of HoH (years) 1 . 647 0 . 969 − 0 . 815 − 0 . 678 − 1 . 784 − 2 . 462 ( 1 . 394 ) ( 1 . 147 ) ( 1 . 110 ) ( 1 . 572 ) ( 1 . 290 ) ( 1 . 541 ) Male HoH (=1) 0 . 077 − 0 . 076 0 . 027 − 0 . 152 0 . 103 − 0 . 049 ( 0 . 080 ) ( 0 . 071 ) ( 0 . 063 ) ( 0 . 086 ) ( 0 . 067 ) ( 0 . 079 ) Experience (years) − 0 . 327 1 . 582 − 0 . 214 1 . 909 − 1 . 796 0 . 113 ( 1 . 250 ) ( 0 . 840 ) ( 0 . 657 ) ( 1 . 467 ) ( 0 . 997 ) ( 1 . 364 ) Primary Edu (=1) − 0 . 017 0 . 009 − 0 . 021 0 . 027 − 0 . 030 − 0 . 004 ( 0 . 032 ) ( 0 . 024 ) ( 0 . 025 ) ( 0 . 031 ) ( 0 . 024 ) ( 0 . 030 ) Farming household (=1) 0 . 029 − 0 . 005 0 . 000 − 0 . 035 0 . 006 − 0 . 029 ( 0 . 036 ) ( 0 . 031 ) ( 0 . 032 ) ( 0 . 040 ) ( 0 . 035 ) ( 0 . 040 ) Training in rice (=1) − 0 . 334*** 0 . 100 0 . 023 0 . 434*** − 0 . 077 0 . 357*** ( 0 . 063 ) ( 0 . 084 ) ( 0 . 090 ) ( 0 . 091 ) ( 0 . 111 ) ( 0 . 095 ) Member of assoc. (=1) − 0 . 047 0 . 008 0 . 006 0 . 055 − 0 . 002 0 . 053 ( 0 . 034 ) ( 0 . 020 ) ( 0 . 016 ) ( 0 . 036 ) ( 0 . 020 ) ( 0 . 031 ) Arouna, Michler, & Lokossou Contract Farming and Rural Transformation 10 / 17
Study Design Treatment Effect Simple Mean Difference (SMD) y ir = α + δ SMD T i + X ir β + ρ r + ǫ ir Analysis of Covariance (ANCOVA) y ir = α + δ ANCOVA T i + µ y ir , pre + X ir , pre β + ρ r + ǫ ir Arouna, Michler, & Lokossou Contract Farming and Rural Transformation 11 / 17
Study Design Hypothesis Testing ◮ Cluster standard errors at farmer group level ◮ Multiple Hypothesis Testing ◮ Possible that significant results emerge due to chance, not actual treatment effects ◮ Bonferroni (1935) adjustment ◮ Holm (1979) adjustment ◮ List et al. (2018) step-wise correction ◮ Anderson (2008) sharpened q -values Arouna, Michler, & Lokossou Contract Farming and Rural Transformation 12 / 17
Results Treatment Effects of Farming Contract Rice area (ha) Productivity (kg/ha) Market participation (%) Rice income pc (USD) SMD ANCOVA SMD ANCOVA SMD ANCOVA SMD ANCOVA (1) (2) (3) (4) (5) (6) (7) (8) Treatment effect 0 . 179*** 0 . 178*** 480 . 2*** 472 . 5*** 34 . 79*** 34 . 83*** 139 . 6*** 139 . 2*** ( 0 . 057 ) ( 0 . 057 ) ( 106 . 2 ) ( 105 . 8 ) ( 2 . 433 ) ( 2 . 440 ) ( 32 . 05 ) ( 32 . 64 ) Ctrl Mean Dep. Var 0.772 1,652 24.96 60.70 Observations 855 855 855 855 855 855 855 855 R 2 0.07 0.07 0.09 0.10 0.50 0.50 0.25 0.25 Arrondissement FE Yes Yes Yes Yes Yes Yes Yes Yes Household Covariates Yes Yes Yes Yes Yes Yes Yes Yes Arouna, Michler, & Lokossou Contract Farming and Rural Transformation 13 / 17
Results Treatment Effects of Each Contract Characteristics Bonferroni-adjusted p-values for Wald tests between coefficient estimates Rice area (ha) Productivity (kg/ha) Market participation (%) Rice income pc (USD) SMD ANCOVA SMD ANCOVA SMD ANCOVA SMD ANCOVA (1) (2) (3) (4) (5) (6) (7) (8) Difference between T2 & T1 0 . 105 0 . 102 1 . 000 1 . 000 0 . 001*** 0 . 001*** 0 . 628 0 . 617 Difference between T3 & T1 1 . 000 1 . 000 1 . 000 1 . 000 0 . 000*** 0 . 000*** 0 . 013** 0 . 013** Difference between T3 & T2 0 . 009*** 0 . 008*** 0 . 286 0 . 283 0 . 000*** 0 . 000*** 0 . 000*** 0 . 000*** Observations 855 855 855 855 855 855 855 855 R 2 0.08 0.08 0.10 0.10 0.55 0.55 0.29 0.29 Arrondissement FE Yes Yes Yes Yes Yes Yes Yes Yes Household Covariates Yes Yes Yes Yes Yes Yes Yes Yes Arouna, Michler, & Lokossou Contract Farming and Rural Transformation 14 / 17
Results Differences Between Contract Characteristics Price guarantee [T1-C] SMD ANCOVA Ctrl Mean Dep. Var Rice area 0.040 0.037 0.772 (0.095) (0.095) Productivity 455.2*** 447.3*** 1,652 (164.5) (161.6) Market participation 20.05*** 19.40*** 24.96 (2.844) (2.779) Rice income 55.92** 56.33** 60.70 (23.95) (24.74) Arouna, Michler, & Lokossou Contract Farming and Rural Transformation 15 / 17
Results Differences Between Contract Characteristics Extension training [T2-T1] SMD ANCOVA T1 Mean Dep. Var Rice area -0.115** -0.117** 1.008 (0.056) (0.056) Productivity -110.6 -99.35 2,133 (110.6) (110.2) Market participation 8.949*** 8.729*** 51.13 (2.122) (2.117) Rice income -22.21 -20.10 190.3 (22.32) (21.39) Arouna, Michler, & Lokossou Contract Farming and Rural Transformation 15 / 17
Results Differences Between Contract Characteristics Input loans [T3-T2] SMD ANCOVA T2 Mean Dep. Var Rice area 0.129*** 0.132*** 0.904 (0.040) (0.040) Productivity 160.1** 155.9** 2,036 (79.68) (76.92) Market participation 10.86*** 10.88*** 57.25 (1.637) (1.634) Rice income 106.1*** 106.1*** 146.5 (18.83) (18.85) Arouna, Michler, & Lokossou Contract Farming and Rural Transformation 15 / 17
Results Differences Between Contract Characteristics Extension & loans [T3-T1] SMD ANCOVA T1 Mean Dep. Var Rice area 0.023 0.023 1.008 (0.048) (0.049) Productivity 112.7 109.4 2,133 (108.9) (106.8) Market participation 16.75*** 16.40*** 51.13 (2.072) (1.994) Rice income 90.78*** 92.22*** 190.3 (24.85) (25.34) Arouna, Michler, & Lokossou Contract Farming and Rural Transformation 15 / 17
Recommend
More recommend
