Rural media, agricultural technology adoption and productivity: Evidence from smallholder rice farmers in Burkina Faso Alia D., Nakelse T., Diagne A., Ouedraogo M., Guissou R. Presenter :Tebila, Nakelse Africa Rice Congress, 2013 Africa Rice Center (AfricaRice) October, 2013 Alia D., Nakelse T., Diagne A., Ouedraogo M., Guissou R. Presenter :Tebila, Nakelse (Africa Rice Cen Rural media, agricultural technology adoption and productivity: Evidence from smallhold October, 2013 1 / 18
22
Embed
Th5_Rural media, agricultural technology adoption and productivity: Evidence from smallholder rice farmers in Burkina Faso
3rd Africa Rice Congress Theme 5: Innovation systems and ICT tools for rice value chain Mini symposium 1: Rice knowledge management using traditional media Author: Nakelse
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Rural media, agricultural technology adoption andproductivity: Evidence from smallholder rice farmers in
Burkina Faso
Alia D., Nakelse T., Diagne A., Ouedraogo M., Guissou R.
Alia D., Nakelse T., Diagne A., Ouedraogo M., Guissou R. Presenter :Tebila, Nakelse Africa Rice Congress, 2013 (Africa Rice Center (AfricaRice))Rural media, agricultural technology adoption and productivity: Evidence from smallholder rice farmers in Burkina FasoOctober, 2013 1 / 18
Introduction
Introduction
Agriculture accounts up to 40 percent to the total GPD and up to 60percent to export revenue and remains a major source of income
Productivity increase rely mainly on area expansion which is gettingbecause of
I The demographic pressure due to population and urbanizationI Climate change
Thus, raising productivity through adoption of improved varietiesalong with best crop management practices and mechanization couldbe a sustainable option
However, the diffusion and the adoption of most of these technologiesare still incomplete and even low in Africa (Abdulai and Huffman,2007)
Alia D., Nakelse T., Diagne A., Ouedraogo M., Guissou R. Presenter :Tebila, Nakelse Africa Rice Congress, 2013 (Africa Rice Center (AfricaRice))Rural media, agricultural technology adoption and productivity: Evidence from smallholder rice farmers in Burkina FasoOctober, 2013 2 / 18
Introduction
Introduction
Among the ICTs, radio is one the fastest for communicatingagricultural information. It has been used as extension tool sinceseveral years (Bereh, 2002).
Radio have been quite successful in spreading agricultural informationin several countries such as Brazil and Cote d’Ivoire (Dimelu andAnyawu, 2004 ; Wele 1991)
This paper proposes to contribute to the growing literature andempirically examines the effect of ICT use on the adoption ofimproved rice varieties and productivity
Alia D., Nakelse T., Diagne A., Ouedraogo M., Guissou R. Presenter :Tebila, Nakelse Africa Rice Congress, 2013 (Africa Rice Center (AfricaRice))Rural media, agricultural technology adoption and productivity: Evidence from smallholder rice farmers in Burkina FasoOctober, 2013 3 / 18
Theoretical framework and methods Rural media and the Rubin Causal Model
Rural media and the Rubin Causal Model
Let denote decision variable by D and by D = 1 if the farmer adoptand D = 0 if he do not adopt with following selection rule (SeeHeckman , 2010 ; Heckman et al 2006) : D = 1(µD(Z ) > V )
I µD(Z ) = Utility gain with adoptionI V a random scalar variable with cumulative distribution function FV
Farmer also faces two hypothetical or potential outcomesI Y1 = µ1(X ) + U1 under adoptionI Y0 = µ0(X ) + U0 without adoption with
X a subset of observed covariates andUD an error term.
Alia D., Nakelse T., Diagne A., Ouedraogo M., Guissou R. Presenter :Tebila, Nakelse Africa Rice Congress, 2013 (Africa Rice Center (AfricaRice))Rural media, agricultural technology adoption and productivity: Evidence from smallholder rice farmers in Burkina FasoOctober, 2013 4 / 18
Theoretical framework and methods Rural media and the Rubin Causal Model
Rural media and the Rubin Causal Model
How do Radio rice programs listening (W ) affect adoption ?First rural media affects adoption and hence the outcome bymodifying the information set of the decision maker
Second, it could also increase the knowledge of an adopting or even anon-adopting farmer about how to correctly use an existing improvedrice variety which has a complex input use requirement.
Alia D., Nakelse T., Diagne A., Ouedraogo M., Guissou R. Presenter :Tebila, Nakelse Africa Rice Congress, 2013 (Africa Rice Center (AfricaRice))Rural media, agricultural technology adoption and productivity: Evidence from smallholder rice farmers in Burkina FasoOctober, 2013 5 / 18
Theoretical framework and methods Rural media and the Rubin Causal Model
Average Treatmnet Effect (ATE )
Individual level effect of adoption on outcome : Y1 − Y0 = β
Population level effect : ATE (x) = β̄(x) = E (Y1 − Y0|X = x), ATTand ATU
BUTFarmer decision to adopt or not to adopt is likely to be endogenous ;farmer with expected high gain may sort into adoption (Diagne et al,2012 ; Ravaillon, 2012)
Alia D., Nakelse T., Diagne A., Ouedraogo M., Guissou R. Presenter :Tebila, Nakelse Africa Rice Congress, 2013 (Africa Rice Center (AfricaRice))Rural media, agricultural technology adoption and productivity: Evidence from smallholder rice farmers in Burkina FasoOctober, 2013 6 / 18
Theoretical framework and methods Rural media and the Rubin Causal Model
Local Average Treatment Effect (LATE )
To consider the unobserved characteristics effect we use MarginalTreatment Effect (MTE) version of LATE with listening rice radioprogram (W ) as instrument .
I MTE (u) = E (Y1 − Y0|X = x ,U = u)
I LATE (p1, p0,X = x) =∫ p
p MTE (u)du/(p1 − p0)
p1 is the probability of adoption when farmer listened rice radioprogram(W = 1)
p0 is the probability of adoption when farmer not listened rice radioprogram(W = 0)
Thus LATE is the mean effect on outcome (productivity) for farmerwho choose to adopt because of listening rice radio program.
Alia D., Nakelse T., Diagne A., Ouedraogo M., Guissou R. Presenter :Tebila, Nakelse Africa Rice Congress, 2013 (Africa Rice Center (AfricaRice))Rural media, agricultural technology adoption and productivity: Evidence from smallholder rice farmers in Burkina FasoOctober, 2013 7 / 18
Data and descriptive statistics Data
Data
The data set for this analysis is extracted from the rice statisticssurvey for Burkina Faso collected in 2009-2010 by NARS incollaboration with NASS and AfricaRice
Stratification at province level with villages randomly selected in eachstratum proportionally to the importance of the stratum and a fixednumber of rice farming households selected in each village
The size of sample drawn was originally 760 . The number ofhouseholds effectively surveyed was 650 in Burkina Faso.
Thus the final data used have 284 households.
Alia D., Nakelse T., Diagne A., Ouedraogo M., Guissou R. Presenter :Tebila, Nakelse Africa Rice Congress, 2013 (Africa Rice Center (AfricaRice))Rural media, agricultural technology adoption and productivity: Evidence from smallholder rice farmers in Burkina FasoOctober, 2013 8 / 18
Data and descriptive statistics Data
Descriptive Statistics
Descriptive Statistics by adoption statusNo Yes Diff tstata
Quantitative variablesLogarithm of Rice yield in 2008 (Kg/Ha) 0.69 0.88 -0.19 -2.4**Household size 9.4 9.13 0.28 0.3Number of radio receptors 1.88 1.52 0.36 2.0**Categorical variablesHave listened rice programs on radio 0.1 0.42 -0.33 -4.0***Have acquired some knowledge 0.02 0.4 -0.37 -4.7**Have applied knowledge acquired 0.02 0.36 -0.34 -4.3***Gender of Head of household 0.76 0.8 -0.04 -0.5Has secondary Activity 0.13 0.06 0.06 1.4Has agriculture as main activity 0.93 0.95 -0.02 -0.5Has at least some primary education 0.33 0.42 -0.09 -1Sample size 242 42
Alia D., Nakelse T., Diagne A., Ouedraogo M., Guissou R. Presenter :Tebila, Nakelse Africa Rice Congress, 2013 (Africa Rice Center (AfricaRice))Rural media, agricultural technology adoption and productivity: Evidence from smallholder rice farmers in Burkina FasoOctober, 2013 9 / 18
Data and descriptive statistics Data
Descriptive Statistics
Descriptive Statistics by listening statusNo Yes Diff tstata
Quantitative variablesLogarithm of Rice yield in 2008 (Kg/Ha) 0.83 0.88 -0.05 -0.9Household size 9.45 8.7 0.75 1.2Number of radio receptors 1.46 1.72 -0.26 -2.1**Categorical variablesHave listened rice programs on radioHave acquired some knowledge 0.01 0.91 -0.9 -15.5***Have applied knowledge acquired 0.01 0.82 -0.82 -14.4***Gender of Head of household 0.77 0.83 -0.06 -1.2Has secondary Activity 0.07 0.08 -0.01 -0.3Has agriculture as main activity 0.95 0.93 0.02 0.5Has at least some primary education 0.34 0.51 -0.16 -2.7***Sample size 106 178
Alia D., Nakelse T., Diagne A., Ouedraogo M., Guissou R. Presenter :Tebila, Nakelse Africa Rice Congress, 2013 (Africa Rice Center (AfricaRice))Rural media, agricultural technology adoption and productivity: Evidence from smallholder rice farmers in Burkina FasoOctober, 2013 10 / 18
Data and descriptive statistics Data
Descriptive Statistics
On average the yield is 19% higher for adopters
The percentage of household that have listened rice program on radiois significantly higher in the group of improved varieties adopter
Actual adoption rate is significantly higher in the group of farmer thathave listened rice program on radio
This might suggest a positive association between having listened riceprogram on radio and improved varieties adoption
In fact there is no significant difference between the yield of farmerthat have listened rice program on radio and the farmers that have
Alia D., Nakelse T., Diagne A., Ouedraogo M., Guissou R. Presenter :Tebila, Nakelse Africa Rice Congress, 2013 (Africa Rice Center (AfricaRice))Rural media, agricultural technology adoption and productivity: Evidence from smallholder rice farmers in Burkina FasoOctober, 2013 11 / 18
Data and descriptive statistics Data
Descriptive Statistics
On average the yield is 19% higher for adopters
The percentage of household that have listened rice program on radiois significantly higher in the group of improved varieties adopter
Actual adoption rate is significantly higher in the group of farmer thathave listened rice program on radio
This might suggest a positive association between having listened riceprogram on radio and improved varieties adoption
In fact there is no significant difference between the yield of farmerthat have listened rice program on radio and the farmers that have
Alia D., Nakelse T., Diagne A., Ouedraogo M., Guissou R. Presenter :Tebila, Nakelse Africa Rice Congress, 2013 (Africa Rice Center (AfricaRice))Rural media, agricultural technology adoption and productivity: Evidence from smallholder rice farmers in Burkina FasoOctober, 2013 11 / 18
Data and descriptive statistics Data
Descriptive Statistics
On average the yield is 19% higher for adopters
The percentage of household that have listened rice program on radiois significantly higher in the group of improved varieties adopter
Actual adoption rate is significantly higher in the group of farmer thathave listened rice program on radio
This might suggest a positive association between having listened riceprogram on radio and improved varieties adoption
In fact there is no significant difference between the yield of farmerthat have listened rice program on radio and the farmers that have
Alia D., Nakelse T., Diagne A., Ouedraogo M., Guissou R. Presenter :Tebila, Nakelse Africa Rice Congress, 2013 (Africa Rice Center (AfricaRice))Rural media, agricultural technology adoption and productivity: Evidence from smallholder rice farmers in Burkina FasoOctober, 2013 11 / 18
Data and descriptive statistics Data
Descriptive Statistics
On average the yield is 19% higher for adopters
The percentage of household that have listened rice program on radiois significantly higher in the group of improved varieties adopter
Actual adoption rate is significantly higher in the group of farmer thathave listened rice program on radio
This might suggest a positive association between having listened riceprogram on radio and improved varieties adoption
In fact there is no significant difference between the yield of farmerthat have listened rice program on radio and the farmers that have
Alia D., Nakelse T., Diagne A., Ouedraogo M., Guissou R. Presenter :Tebila, Nakelse Africa Rice Congress, 2013 (Africa Rice Center (AfricaRice))Rural media, agricultural technology adoption and productivity: Evidence from smallholder rice farmers in Burkina FasoOctober, 2013 11 / 18
Data and descriptive statistics Data
Descriptive Statistics
On average the yield is 19% higher for adopters
The percentage of household that have listened rice program on radiois significantly higher in the group of improved varieties adopter
Actual adoption rate is significantly higher in the group of farmer thathave listened rice program on radio
This might suggest a positive association between having listened riceprogram on radio and improved varieties adoption
In fact there is no significant difference between the yield of farmerthat have listened rice program on radio and the farmers that have
Alia D., Nakelse T., Diagne A., Ouedraogo M., Guissou R. Presenter :Tebila, Nakelse Africa Rice Congress, 2013 (Africa Rice Center (AfricaRice))Rural media, agricultural technology adoption and productivity: Evidence from smallholder rice farmers in Burkina FasoOctober, 2013 11 / 18
Results and discussion Impact of media use on improved variety adoption
Impact of media use on improved variety adoption
The propensity to adopt improved varieties is higher for a farmer thathave listen rice program on radio than for a farmer that have never
Most of the other control variables have the expected sign but werefound to be not significant
Alia D., Nakelse T., Diagne A., Ouedraogo M., Guissou R. Presenter :Tebila, Nakelse Africa Rice Congress, 2013 (Africa Rice Center (AfricaRice))Rural media, agricultural technology adoption and productivity: Evidence from smallholder rice farmers in Burkina FasoOctober, 2013 12 / 18
Results and discussion Impact of media use on improved variety adoption
Marginal Treatment Effect
Figure : MTEAlia D., Nakelse T., Diagne A., Ouedraogo M., Guissou R. Presenter :Tebila, Nakelse Africa Rice Congress, 2013 (Africa Rice Center (AfricaRice))Rural media, agricultural technology adoption and productivity: Evidence from smallholder rice farmers in Burkina FasoOctober, 2013 13 / 18
Results and discussion Impact of media use on rice productivity
Impact of media use on rice productivity
The graph shows that the MTE is decreasing in the unobserved utilityand consequently in the propensity score
This is an evidence of diminishing marginal return with the likelihoodof a farmer to adopt a modern variety
Farmer with the highest return are more likely to adopt ( low ud )while those that are less likely to adopt ( high ud ) have a muchsmaller gross return
This arises the fact that farmers almost certainly base their choice toadopt or not adopt on things they know but the analyst does notknow
This gives rise to what Heckman et al. (2006) call essentialheterogeneity
Alia D., Nakelse T., Diagne A., Ouedraogo M., Guissou R. Presenter :Tebila, Nakelse Africa Rice Congress, 2013 (Africa Rice Center (AfricaRice))Rural media, agricultural technology adoption and productivity: Evidence from smallholder rice farmers in Burkina FasoOctober, 2013 14 / 18
Results and discussion Impact of media use on rice productivity
Local Average Treatment Effect
Alia D., Nakelse T., Diagne A., Ouedraogo M., Guissou R. Presenter :Tebila, Nakelse Africa Rice Congress, 2013 (Africa Rice Center (AfricaRice))Rural media, agricultural technology adoption and productivity: Evidence from smallholder rice farmers in Burkina FasoOctober, 2013 15 / 18
Results and discussion Impact of media use on rice productivity
Impact of media use on rice productivity
The total gross return of adoption is positive and relatively high forall potential adoption farmers
In the sub-group of farmers who listened rice program the averageimpact of adoption (the LATE ) is statistically significant andestimated to be an increase of about 39.2% of yield
This is the mean effect on yield for farmer who choose to adoptbecause of listening rice radio program.
The LATE for the actual non-adopters is significantly positive whichsuggests that if the actual non adopter were all exposed to riceprograms on radio and come to adopter improved varieties, their yieldgain will be positive and much higher than the average yield gainacross all adopting farmers
Alia D., Nakelse T., Diagne A., Ouedraogo M., Guissou R. Presenter :Tebila, Nakelse Africa Rice Congress, 2013 (Africa Rice Center (AfricaRice))Rural media, agricultural technology adoption and productivity: Evidence from smallholder rice farmers in Burkina FasoOctober, 2013 16 / 18
Results and discussion Conclusion
Conclusion
We use household-level survey data and apply the Rubin CausalModel that has emerged as the standard approach for evaluatingpolicy/program effect using observational data.
Adoption of modern varieties in 2008 was significantly higher forfarmers that had listened to the radio program on rice before 2008than those who had not
Using of rural radio appears to significantly increase the propensity ofadopting modern varieties
We also estimate that the LATE of adoption of improved varietiesinduced by listening to the rice radio program on rice yield wassignificant and positive
Alia D., Nakelse T., Diagne A., Ouedraogo M., Guissou R. Presenter :Tebila, Nakelse Africa Rice Congress, 2013 (Africa Rice Center (AfricaRice))Rural media, agricultural technology adoption and productivity: Evidence from smallholder rice farmers in Burkina FasoOctober, 2013 17 / 18
Results and discussion Conclusion
Thank you !
Alia D., Nakelse T., Diagne A., Ouedraogo M., Guissou R. Presenter :Tebila, Nakelse Africa Rice Congress, 2013 (Africa Rice Center (AfricaRice))Rural media, agricultural technology adoption and productivity: Evidence from smallholder rice farmers in Burkina FasoOctober, 2013 18 / 18