Published on January 7, 2008
Can GM Crops Help the Poor? Bt Cotton in Makhathini Flats, KwaZulu-Natal: Can GM Crops Help the Poor? Bt Cotton in Makhathini Flats, KwaZulu-Natal Lindie Beyers and Colin Thirtle, Imperial College of Science,Technology and Medicine & University of Pretoria Yousouf Ismaël, University of Reading Jenifer Piesse, Birkbeck College, University of London & University of Stellenbosch Slide2: Makhathini Flats Slide3: Land ownership Agronomic constraints: Agronomic constraints Pest problems: Pest problems Non-agronomic constraints: Non-agronomic constraints Adoption vs. Non-adoption: Yield, input use and profitability: Adoption vs. Non-adoption: Yield, input use and profitability Yield comparison: Yield comparison Gross margin comparisons: Gross margin comparisons Adopters vs. Non-adopters: Insecticide use: Adopters vs. Non-adopters: Insecticide use Production Efficiency of Makhathini Flats cotton producers: Production Efficiency of Makhathini Flats cotton producers The basic farm accounting measures are useful, but tell one very little about the reasons for any observed differences between farms. Yield is a very partial measure of productivity, which is of limited use when the amounts of non-land inputs used, such as labour and fertiliser, differ between farms. Stochastic Production Frontiers – Econometric Provides Farm Level Efficiencies and Tests Data Envelopment Analysis – Programming Information on the Effects of Farm Size Data, Estimation and Tests: Data, Estimation and Tests Output bales of cotton [a physical measure of output] Four inputs land (in hectares); insecticides (a value because of aggregation over different types); seed (25kg-bags, except in one case where the seed costs worked better); and labour (number of days of family and hired labour used for spraying, weeding and harvesting) Planting is excluded because all the farms used the mechanised planting services, which were available All the variables are in natural logarithms, so that the coefficients can be interpreted as elasticities, which must have values of between zero and unity to conform to production theory Since the values cannot be negative, a one-tailed significance test is appropriate. Stochastic frontier 1998/1999: Stochastic frontier 1998/1999 Stochastic Frontier 1999/2000: Stochastic Frontier 1999/2000 Test results: Test results Deterministic Frontier Programming Models: Deterministic Frontier Programming Models Whilst the stochastic frontier model results are entirely acceptable, deterministic frontier efficiency models are perhaps more reliable and easier to follow The data envelopment analysis (DEA) model has been widely applied to efficiency measurement problems DEA provides both a check on the stochastic frontier results and further information, especially on the farm size issue. DEA analysis confirms the stochastic frontier results DEA results: DEA results Income distribution and inequality: Income distribution and inequality The introduction of new technologies can have adverse effects on the distribution of income, as the voluminous literature on the green revolution showed The farmers who have resources to adopt may become richer increasing inequality (even if non-adopters do not suffer a reduction in income) If they are disadvantaged and actually lose land to the better off, this situation is exacerbated and their income levels may actually fall. In this study, we are fortunate to have data from the first year of adoption, when very few farmers use the new technology Thus, 1998/1999 can serve as a benchmark for tracking the changes in the distribution of land and incomes that result from the introduction of the Bt variety. Slide19: The measures of inequality used are the Gini-coefficient and the Lorenz curve The Gini ratio of the area between a Lorenz curve and the diagonal and the total area under the diagonal Lorenz curve is defined by the cumulative shares of income/wealth attributable to proportions of the population Consequently, 0 <Gini < 1 0 absolute equality of income distribution 1 absolute inequality of income distribution Per capita income inequality: Per capita income inequality Conclusions: Conclusions The farmers who adopted the Bt-cotton variety benefited from the new technology, according to all the measures used. 1999/2000 was a bad year, due to unusually heavy rainfall and the Bt adopters suffered far less of a fall in yields than those who did not adopt. Efficiency frontiers consider the efficiency with which all inputs are converted into outputs, using only the more reliable input and output quantity data and avoiding prices. Both deterministic and stochastic frontiers were used. In either case, the results confirm the farm accounting results, showing that the Bt cotton adopters were considerably more efficient than those who used the non-Bt varieties were. Slide22: There did not appear to be any agronomic or other technical impediments to adoption. All smallholders could benefit, provided that credit is made available. So there is no reason to expect that the ceiling has been reached. But the suppliers have considerable monopoly power, which they may exploit to appropriate a greater share of the benefits by raising their prices. The next stage is to survey the same farmers this season and to investigate the structure of the industry with a view to understanding the upstream and downstream distributional consequences of the introduction of GM crops.