do contractual relations incentivize farmers’ adoption of...
TRANSCRIPT
Do Contractual Relations Incentivize Farmers’ Adoption of Multiple Innovations?: Evidence from the Indonesian Dairy Sector
Risti Permani, Ph.D. Professor Wendy Umberger, Ph.D.Global Food Studies, University of Adelaide
The 60th AARES Conference, 2-5 February 2016
Background
University of Adelaide 2
Technology adoption
• Adoption of a single type of technology
• Farmers may adopt in a stepwise pattern
Contract farming
• An important platform to allow smallholders to gain access to both markets and technologies
• Literature on contract farming and technology adoption is extensive
• Little has been done to assess the effects of farmers’ involvement in contract farming on the sequential adoption of multiple innovations.
– Imposing quality-based bonus payment to contracts between farmers and processors to address information asymmetries would improve input use and therefore milk quality (Saenger et al. 2013; Saenger et al. 2014)
• It remains unclear whether and how this vertical coordination has impacted farmers’ stepwise adoption of multiple innovations
Objective
University of Adelaide 3
• To investigate whether farmers who engage in contract farming would be more incentivized to (sequentially) trial multiple innovations
Methodology
University of Adelaide 4
Data collection (December 2014-January 2015)o >200 dairy households.o 6 cooperatives o Stratified random samplingo 20-page structured questionnaireo 10 moduleso Collaborative effort between Global Food Studies at
University of Adelaide and Bogor Agricultural Universityo Team: Four key researchers, 3 fieldwork coordinators, 16
enumerators, one data entry programmer (and her staff).
Have you used/done since 2010?
University of Adelaide 5
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Automatic milking machines
Milk processing (make yogurt)
Cooling milk in water tanks
Conserving forages for the dry seasons (hay, silage)
Milk pasteurisation
Biogas units
High protein concentrates (16% or higher)
Teat dipping after milking
Nutrient feed blocks
Record keeping
Feed legume forages (e.g. Leucaena)
Mastitis test
Grow new improved grasses (high yield)
Rubber/Plastic floor for the barn/cage
Stainless steel milking equipment
Improved milking hygiene to reduce TPC
Improving drinking water availability 24/7
Use of any fertilisers for the grass
Using detergents for milking equipment
Artificial Insemination (AI)
Adoption, 0 to 1
Innovations – adoption and timeline
Type of innovation Are you familiar with or have you heard of […]?
Have you used/done […] since 2010
Are you still using/doing [...]?
What year did you used/do [...] for the first time?
Artificial Insemination (AI) 0.99 0.98 0.97 2000
Using detergents for milking equipment 0.94 0.86 0.80 2001
Rubber/Plastic floor for the barn/cage 0.92 0.68 0.62 2008
Stainless steel milking equipment 0.88 0.68 0.58 2003
Use of any fertilisers for the grass 0.80 0.79 0.62 2002
Biogas units 0.80 0.25 0.09 2010
Mastitis test 0.79 0.58 0.33 2003
Record keeping 0.70 0.41 0.27 2003
Milk processing (make yogurt) 0.66 0.06 0.02 2010
Improving drinking water availability 24/7 0.66 0.71 0.46 2001
Automatic milking machines 0.65 0.03 0.02 2011
Cooling milk in water tanks 0.61 0.12 0.06 2009
Feed legume forages (e.g. Leucaena) 0.54 0.49 0.22 2000
Conserving forages for the dry seasons (hay, silage) 0.54 0.15 0.04 2011
Grow new improved grasses (high yield) 0.51 0.58 0.29 2004
Teat dipping after milking 0.49 0.32 0.10 2006
Improved milking hygiene to reduce TPC 0.44 0.69 0.30 2002
Milk pasteurisation 0.43 0.17 0.05 2006
High protein concentrates (16% or higher) 0.37 0.29 0.03 2005
Nutrient feed blocks 0.28 0.33 0.02 2002
University of Adelaide 6
Mastitis and record keeping
• Mastitits – requiring milk not being sold
• Record keeping is negatively associated with the incidence rate of clinical mastitis (Kivaria et al. 2007)
• First-time mastitis cases can be compared against farmer’s record book to see what changes preceded the development of mastitis.
• Mastitis adversely affected the quality of pasteurized fluid milk (Ma et al.2000). – It is recommended that the fluid milk industry
consider implementation of premium quality payment programs for low Somatic Cell Count (SCC) milks.
University of Adelaide 7
Empirical model
To look at the impacts of contract farming on farmer 𝑖′s decision to adopt multiple innovations• First stage: 𝐶𝑜𝑛𝑡𝑟𝑎𝑐𝑡𝑖 = 𝛼1 +
𝛼2𝑋𝑖 + 𝛼3𝑍𝑖 + 𝑢𝑖• Second stage: 𝐼𝑛𝑛𝑜𝑣𝑖 = 𝛽1 +
𝛽2𝑋𝑖 + 𝛽3𝐶𝑜𝑛𝑡𝑟𝑎𝑐𝑡𝑖 + 𝑣𝑖
Where 𝐶𝑜𝑛𝑡𝑟𝑎𝑐𝑡 is contract farming participation (Yes/No); 𝐼𝑛𝑛𝑜𝑣 is the innovation index.
Two innovations:• Mastitis test• Record keeping
Method:1) Logit – contract farming2) Multinomial logit - Innovation
index3) Multinomial logit with an
endogeneous regressor
University of Adelaide 8
Initial node
Mastitis test
Record keeping (3)
– 9.62%
No record keeping (1) – 32.69%
Adopt both (5)
Adopt none (0) – 38.46%
Record keeping
Mastitis test (4) – 4.81%
No mastitis test (2) –14.42%
Endogeneity
• 𝐶𝑜𝑟𝑟 𝐶𝑜𝑛𝑡𝑟𝑎𝑐𝑡𝑖 , 𝑣𝑖 ≠ 0
• Similarly, contract farming participation could be affected by adoption decisions where more ‘technology savvy’ farmers might be more likely to meet standards and participate in contract farming.
• As summarised by Bolwig et al. (2009): to deal with endogenous selection, three approaches can be used:
1. Matching estimators which require that selection into program is based only on the observed variables. However, if the assumption of participation on observed variables is doubted, then matching methods will be biased;
2. Instrumental variable (IV);
3. Heckman selection models
Title and Author Method Instrument
The Economics of Smallholder
Organic Contract Farming in Tropical
Africa (Bolwig, et al., 2009)
OLS, Two-stage IV, LIML and
FIML Heckman Selection models
Treatment: Participation in the
organic scheme
Outcome: Gross crop revenue
(log.)
(i) A dummy constructed from the ratio of
non-farm revenue to total revenue, taking
the value of one for those falling in the top
tercile and zero otherwise; (ii) A dummy to
proxy welfare, one if the walls of the
household are made of brick and zero
otherwise.
Impact of Contract Farming on
Income: Linking Small Farmers,
Packers, and Supermarkets in China
(Miyata, et al., 2009)
Heckman selection model
Treatment: Participation in
contract farming
Outcome:
Household income
The distance between the farm of a
household and the farm of the village
leader.
As You Sow, So Shall You Reap: The
Welfare Impacts of Contract Farming
(Bellemare, 2012)
IV model
Treatment: Household
participation in contract farming
Outcome:
Household income
A variable derived by interacting: (i) Dummy 1=’Yes’ to a hypothetical contract
farming question i.e. (randomly generated)
contingent valuation question ( Arrow et
al., 1993 and Mitchell and Carson, 1989);
and
(ii) The initial (hypothetical) value of
investment to enter a contract farming as
part of the same question (i.e. “Would you
be willing to enter a contract farming
agreement that would necessitate an initial
investment of [___]?”; the value ranges
from 25,000 - 150,000 Ariary or US$ 12.5 -
75).
An analysis of contract farming in
East Java, Bali, and Lombok,
Indonesia (Simmons, et al., 2005)
OLS and 2SLS
Treatment:
Contract participation
Outcome: gross margins and
labour use
N/A
Production Contracts and Farm
Business Growth and Survival (Key,
2013)
OLS and IV-2SLS
Treatment: Participation in
production contract
Outcome: Farm size growth
The local (county-level) availability of
production contracts.
University of Adelaide 9
Instrument
• The ‘availability of contracts’ (Key, 2013) as proxied by the average number of local buyers as an instrument.
• In our survey, farmers reported that two most important clauses in their contract are price certainty and supply exclusivity.
University of Adelaide 10
Descriptive statistics
University of Adelaide 11
Variable Contract farming DifferenceNo
(59.05%)Yes
(40.95%)HH gender (Male=1; Female=0) 0.984 0.988 -0.005
HH age (years) 45.984 44.267 1.716
Education (years of schooling) 2231.855 2128.337 103.518
Household size (age>10) 6.976 6.942 0.034
Car ownership (Yes=0) 62.895 57.919 4.977
Truck ownership (Yes=1) 4.411 4.744 -0.333
Motorbike ownership (Number) 0.339 0.116 0.222*
Mobile phone ownership (Yes=1) 0.032 0.023 0.009
Dairy as the main business (Yes=1) 1.355 1.395 -0.041
Herd size (number of cows) 0.944 0.942 0.002
Experience in dairy (years) 0.903 0.837 0.066
Distance to the nearest processor (km) 9.226 6.105 3.121**
Number of buyers 14.780 14.791 -0.011
Quality-based pricing (Yes=1) 6.018 4.647 1.371
Total plate count testing (Yes=1) 3.753 8.032 -4.278**
Standard errors in parentheses*** p<0.01, ** p<0.05, * p<0.1
#1 - Multinomial logit: Adoption of innovations(contract farming assumed to exogenous)
University of Adelaide 12
Dependent variable: Innovation index (between 0 and 4)
Innovation index=1Mastitis test
Innovation index=2Record keeping
Innovation index=3Mastitis testthen record keeping
Innovation index=4Record keeping then mastitis test
Ordered logit (0=none; 1=1innov; 2=2innov)
HH gender (Male=1; Female=0) -1.126*** 1.490*** 0.507** 0.253 -0.297
HH a -0.022 -0.004 0.017 0.002 0.008
HH age, square 0.000* 0 0 0 0
Education (years of schooling) -0.055 -0.02 0.008 0.062 -0.055
Education, square 0.004 0.002 0 -0.003 0.007Household size (age>10) 0.022 0.041* -0.032* -0.007 0.014Car ownership (Yes=0) 0.103 -0.076 -0.127 0.049 -0.201Truck ownership (Yes=1) 0.073 0.106 0.042 -0.417** 0.161Motorbike ownership (Yes=1) -0.037 -0.025 0.001 0.02 -0.078Mobile phone ownership (Yes=1) -0.214 -0.111 0.932*** -0.108 0.625Dairy as the main business (Yes=1) 0.103 -0.098 0.029 -0.026 0.06
Herd size (number of cows) -0.01 0.002 0.009** -0.006 0.03
Experience in dairy (years) 0.004 0.002 0.002 0.001 0.029
Distance to processor (km) 0.001 -0.008 0.004 0.003 0.019
Participation in contract farming (Yes=1) -0.051 0.012 0.006 0.062* 0.386
Quality-based pricing (Yes=1) 0.005 -0.002 -0.099 -0.032 -0.943**
Total plate count testing (Yes=1) 0.056 -0.007 0.114* 0.049 1.195*
Robust standard errors in parentheses*** p<0.01, ** p<0.05, * p<0.1; Base category is innovation index=0. All coefficients are
marginal effects.
#2 Probit: Participation in contract farming
University of Adelaide 13
Dependent variable:
Participation in contract farming
Probit
(1)
Probit
(2)HH gender (Male=1; Female=0) -0.165 -0.0410HH age (years) -0.103** -0.113**HH age (years), square 0.000998** 0.00109**Education (years of schooling) 0.168* 0.210**Education (years of schooling), square -0.00729 -0.00888*Household size (age>10) 0.174*** 0.164**Car ownership (Yes=0) -0.205 -0.198Truck ownership (Yes=1) 0.468 0.469Motorbike ownership (Yes=1) 0.0372 0.0385Mobile phone ownership (Yes=1) -0.125 -0.0777Dairy as the main business (Yes=1) -0.323 -0.267Herd size (number of cows) -0.0614** -0.0659**Experience in dairy (years) 0.00894 0.00838Distance to processor (km) -0.0226 -0.0427**Quality-based pricing (Yes=1) 0.591***Total Plate Count testing (Yes=1) -0.211Number of buyers 0.0288*** 0.0289***Constant 1.575 1.402Observations 210 210
Standard errors in parentheses*** p<0.01, ** p<0.05, * p<0.1. All coefficients are
marginal effects.
#3 - Multinomial logit: Adoption of innovations(contract farming assumed to endogenous)
University of Adelaide 14
Dependent variable: Innovation index (between 0 and 4)
Innovation index=1Mastitis test
Innovation index=2Record keeping
Innovation index=3Mastitis testthen record keeping
Innovation index=4Record keeping then mastitis test
Ordered logit (0=none; 1=1innov; 2=2innov)
HH gender (Male=1; Female=0) 0.424 -14.62 -10.43 -9.681 -0.0623
HH age -0.0909 -0.0497 0.229 0.0303 0.0109
HH age, square 0.00106 0.000183 -0.00231 -0.000224 -0.000127
Education (years of schooling) -0.249 -0.207 0.0415 1.549 -0.0143
Education, square 0.0189* 0.0215 0.000948 -0.0805 0.00294Household size (age>10) 0.140 0.389** -0.359 -0.105 0.0178Car ownership (Yes=0) 0.254 -0.694 -1.899* 1.109* -0.0435Truck ownership (Yes=1) -0.203 0.317 0.0246 -10.92 0.121Motorbike ownership (Yes=1) -0.249 -0.322 -0.130 0.401 -0.0284Mobile phone ownership (Yes=1) 0.737 0.585 14.36 -1.216 0.239Dairy as the main business (Yes=1) 0.404 -0.703 0.479 -0.640 -0.0963
Herd size (number of cows) -0.0581 -0.00380 0.107* -0.166 0.00803
Experience in dairy (years) 0.0315 0.0346 0.0426 0.0348 0.0150
Distance to processor (km) 0.0106 -0.0568 0.0663 0.0832 0.0101
Participation in contract farming (Yes=1) -0.115 0.161 0.170 1.674* 0.281*
Quality-based pricing (Yes=1) -0.428 -0.440 -1.834** -1.235 -0.550***
Total plate count testing (Yes=1) 1.608 1.325 3.008** 2.548 0.689**
Constant 0.452 14.21 -11.26 0.736 1.402
Number of observations 210 210 210 210 210
Robust standard errors in parentheses*** p<0.01, ** p<0.05, * p<0.1; Base category is innovation index=0. All
coefficients are marginal effects. STATA command gsem is used to run multinomial logit with an endogenous regressor
#3 - Multinomial logit: Adoption of innovations(contract farming assumed to endogenous)
University of Adelaide 15
Dependent variable: Innovation index (between 0 and 4)
Innovation index=1Mastitis test
Innovation index=2Record keeping
Innovation index=3Mastitis testthen record keeping
Innovation index=4Record keeping then mastitis test
Ordered logit (0=none; 1=1innov; 2=2innov)
HH gender (Male=1; Female=0) -0.588 14.76 10.42 9.797 -0.00370
HH age -0.0969 -0.0546 0.217 0.0265 0.00875
HH age, square 0.00109 0.000211 -0.00221 -0.000204 -0.000114
Education (years of schooling) -0.208 -0.177 0.0988 1.588 0.000429
Education, square 0.0173 0.0203 -0.00170 -0.0823 0.00233Household size (age>10) 0.142 0.388** -0.379 -0.109 0.0205Car ownership (Yes=0) 0.198 -0.732 -1.932* 1.080* -0.0578Truck ownership (Yes=1) -0.258 0.285 -0.00347 -11.22 0.134Motorbike ownership (Yes=1) -0.241 -0.314 -0.0915 0.408 -0.0232Mobile phone ownership (Yes=1) 0.681 0.558 14.57 -1.233 0.220Dairy as the main business (Yes=1) 0.362 -0.738 0.425 -0.661 -0.116
Herd size (number of cows) -0.0481 0.00339 0.117* -0.161 0.00955
Experience in dairy (years) 0.0353 0.0372 0.0471 0.0367 0.0167*
Distance to processor (km) 0.00527 -0.0604 0.0611 0.0796 0.00908
Participation in contract farming (Yes=1) -0.0941 0.174 0.181 1.679* 0.289*
Quality-based pricing (Yes=1) -0.335 -0.377 -1.763** -1.180 -0.523***
Total plate count testing (Yes=1) 0.772 0.546 2.207** 1.854 0.553*
Constant 1.406 -15.19 -32.29 -18.81
Number of observations 210 210 210 210 210
Robust standard errors in parentheses*** p<0.01, ** p<0.05, * p<0.1; Base category is innovation index=0. All
coefficients are marginal effects. STATA command gsem is used to run multinomial logit with an endogenous regressor
Concluding remarks
University of Adelaide 16
• There is a need to better understand the determinants of adoption of innovations given the declining rates of adoption.
• Endogeneity of contract farming
– Younger; more educated; smaller farms located close to a processor; quality-based pricing – more likely to engage in contract farming
– ‘Availability of contracts’ as an instrument
• Determinants of adoption of innovations vary between different innovation modes taking into account the sequence of adoption.
• Policy recommendations on the design and timing of programs to introduce new innovations to farmers and future ‘packaging of technologies’ by emphasizing ‘stepwise dissemination’.
References
University of Adelaide 17
• Barrett, C. B., Bachke, M. E., Bellemare, M. F., Michelson, H. C., Narayanan, S., Walker, T. F., 2012. Smallholder Participation in Contract Farming: Comparative Evidence from Five Countries, World Development. 40, 715-730.
• Bellemare, M. F., 2012. As You Sow, So Shall You Reap: The Welfare Impacts of Contract Farming, World Development. 40, 1418-1434.
• Bolwig, S., Gibbon, P., Jones, S., 2009. The Economics of Smallholder Organic Contract Farming in Tropical Africa, World Development. 37, 1094-1104.
• Ersado, L., Amacher, G., Alwang, J., 2004. Productivity and Land Enhancing Technologies in Northern Ethiopia: Health, Public Investments, and Sequential Adoption, American Journal of Agricultural Economics. 86, 321-331.
• Key, N., 2013. Production Contracts and Farm Business Growth and Survival, Journal of Agricultural and Applied Economics. 45, 277-293.
• Miyata, S., Minot, N., Hu, D., 2009. Impact of Contract Farming on Income: Linking Small Farmers, Packers, and Supermarkets in China, World Development. 37, 1781-1790.
• Saenger, C., Qaim, M., Torero, M., Viceisza, A., 2013. Contract farming and smallholder incentives to produce high quality: experimental evidence from the Vietnamese dairy sector, Agricultural Economics. 44, 297-308.
• Saenger, C., Torero, M., Qaim, M., 2014. Impact of Third-party Contract Enforcement in Agricultural Markets—A Field Experiment in Vietnam, American Journal of Agricultural Economics.
• Simmons, P., Winters, P., Patrick, I., 2005. An analysis of contract farming in East Java, Bali, and Lombok, Indonesia, Agricultural Economics. 33, 513-525.
• Wang, H. H., Wang, Y., Delgado, M. S., 2014. The Transition to Modern Agriculture: Contract Farming in Developing Economies, American Journal of Agricultural Economics. 96, 1257-1271.
Thank you!http://www.adelaide.edu.au/global-food
http://www.adelaide.edu.au/global-food/blog/