Impact of Narrow Targeting Programs

К оглавлению
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 
17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 
34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 
51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 
68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 
85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 
102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 
119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 
136 137 138 139 140 141 142 

Relatively few detailed independent evaluations are available for the narrow

targeting schemes discussed above. The subsidized credit component of

the Tulong sa Tao program was examined to see how far the aim that

benefi ciaries should come from low-income groups in rural areas was met.

It appears that in practice detailed inclusion and exclusion criteria were

not enumerated and therefore there was no effort to screen prospective

benefi ciaries. Estimates suggest that only about one-third of benefi ciaries

were really from the low-income groups. Also there was no mechanism to

disqualify borrowers who were no longer poor. On the contrary, the project

allowed individual micro-entrepreneurs a higher ceiling on the second loan,

and an even higher ceiling for the third and each subsequent loan (Balisacan

et al., 2000).

The offi cial impact evaluation of the CIDSS was favorable, although

vague. It stated that ‘unmet basic needs were reduced by an average of 57

per cent’. An external evaluation of the project (Bautista, 1999) revealed

that the incidence of poor families in the sample of households covered

by CIDSS actually increased, although in the non-CIDSS communities,

the incidence also increased and by more than the increase in the CIDSS

sample. A more rigorous evaluation was conducted using data from the 1997

and 2000 Family Income and Expenditure Survey. Here the real per capita

incomes in the original set of priority provinces benefi ting from the CIDSS

were compared against an equivalent set of provinces.5 The analysis shows

that real per capita incomes decreased, but in comparison with the change

in the control group, this reduction is not statistically signifi cant. Poverty

incidence also increased between 1997 and 2000, but again, in comparison

with the control, this negative impact is not statistically different from that

in the other provinces. This simple evaluation suggests that the CIDSS did

not result in a signifi cant improvement in incomes or in poverty reduction,

at the provincial level. The alternative interpretation is that it will take some

time for community empowerment to translate into poverty reduction.

To compare alternative poverty targeting measures Balisacan et al. (2000)

simulated several experiments on geographical targeting. Given a fi xed

program cost of Pesos 10 billion, they simulated ten experiments, and for

each, the rates of leakage and undercoverage were computed. The assumed

objective was reduction of national poverty incidence, thus effectiveness

is measured as the impact on national poverty incidence for a given cost.

Of note is the fact that the strategy of the Lingap para sa Mahihirap (Care

for the Poor) program of the Estrada administration, which involved

directing assistance at an equal number of poorest households in each

province and city (regardless of differences in regional income), is by far

the least effective. The authors test for the infl uence of administrative or

screening costs of targeting and fi nd that once these are introduced, the

preferred ranking of alternatives changes. Once the administrative cost per

applicant rises to a modest Pesos 135 then a simple form of geographic

targeting, allocating funds only to the poorest provinces, comes to dominate

the ‘perfect targeting’ solution where each province and city receives the

same funding but to be distributed only among the genuinely poor. They use

this as evidence that if the screening cost is substantial it pays to implement

geographical targeting.