METHODOLOGICAL FRAMEWORK 21
because we used only two periods (before and after project start), we are not concerned about serial correlation among multiple periods. Another reason for the possible dependence of the errors is clustering of the sample. The propensity scores were computed using binary probit regression mod- els. We estimated three probit models for three comparisons: (1) Fadama II beneficiaries compared with all nonbeneficiaries, (2) Fadama II beneficiaries compared with nonbeneficiaries in Fadama II communities, and (3) Fadama II beneficiaries compared with nonbeneficiaries outside Fadama II communi- ties. The dependent variable in each model is a binary variable indicating whether the household was a beneficiary of the project. The explanatory variables used in computing the propensity scores are those expected to jointly determine the probability to participate in the project and the outcome. We focused on the determinants of income and productive assets when selecting the independent variables for computing the PSM. We assumed that rural infrastructure should be included in produc- tive assets. These variables are summarized in Table 4.2. Consistent with the CDD approach, Fadama II supported economic groups only. Hence to better understand participation in Fadama II, we analyzed the determinants of membership in EIGs. This analysis adds more information to the PSM analysis, because the PSM model assessed the determinants of mem- bership in Fadama II only, whereas the EIG analysis involved any economic group—even those that did not qualify or were not covered by Fadama II. We used the same covariates as those used for the PSM model (Table 4.2). Elite capture is one of the potential problems occurring in CDD projects. If the program benefits accrue more to the well-off than to poor beneficiaries, income distribution will be more skewed, leading to increased inequity. We analyzed the impact of the program across asset terciles and agroecological zones. We divided the beneficiaries into three groups of poverty terciles using the value of productive assets prior to the program as an indicator of wealth.
To understand the impact of Fadama II on income distribution, we com- puted the Gini coefficient and the coefficient of variation for beneficiaries and nonbeneficiaries before and after the project.10 We used both household consumption expenditure as well as income to measure the Gini coefficient. Each of these measures has its advantages and drawbacks. The main dis- advantage of the consumption data is measurement error. Consumption expen- diture was collected using the household survey, and farmers were asked
10 The Gini coefficient is a measure of inequality, ranging from zero if income (or any other sta- tistic) is equal across all members of a society to one if income (or any other statistic) belongs to or characterizes only one person in the society.
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60 |
Page 61 |
Page 62 |
Page 63 |
Page 64 |
Page 65 |
Page 66 |
Page 67 |
Page 68 |
Page 69 |
Page 70 |
Page 71 |
Page 72 |
Page 73 |
Page 74 |
Page 75 |
Page 76 |
Page 77 |
Page 78 |
Page 79 |
Page 80 |
Page 81 |
Page 82 |
Page 83 |
Page 84 |
Page 85 |
Page 86 |
Page 87 |
Page 88 |
Page 89 |
Page 90 |
Page 91 |
Page 92 |
Page 93