Impact of employment 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 

How effective were the targeting outcomes in these labor-intensive

employment safety net programs? Several preliminary studies drew on the

results of the 100 Village Survey,24 but the most complete analysis has

been by Sumarto et al. (2001), drawing on data provided by the SUSENAS

Special Module. This reveals that only 8.3 per cent of poor households (as

defi ned by those households in the lowest quintile of per capita expenditure)

were covered by these programs in the six months before February 1999.

Meanwhile, 70 per cent of those who took part in padat karya activities

were from non-poor households, even though participation dropped off

among the better-off sections of the community.

The authors use a simple targeting ratio to judge the effectiveness of the

schemes. This is defi ned as ‘the ratio of participation of the non-poor in a

program compared to the fraction of non-poor in the sample.’ It is derived

by dividing the proportion of non-poor benefi ciaries for each program by

0.8, which is the proportion of non-poor in the sample (as the poor are

the bottom quintile); see Sumarto et al. (2001: 14–18). If all recipients are

non-poor, the targeting ratio will be 1.25, whilst perfect targeting, where

all benefi ciaries are poor, will mean a targeting ratio of zero. Random

targeting, where the non-poor and poor are equally likely to be benefi ciaries,

will produce a targeting ratio equal to unity. For the employment creation

schemes the result was a targeting ratio of 0.88, leading to the conclusion

that targeting, far from being effective, was near to random. Table 3.9 gives

their estimates of targeting ratios for a number of anti-poverty programs.

Overall, ratios are high and relatively close to unity for all schemes, with

the lowest ratio of 0.83 found for the Health Card program (see below).

Although these results refer to a relatively brief and early period for these

schemes, the weak coverage of the poor relative to the better-off suggests

that these Social Safety Net programs were relatively ineffective.

Some other valuable insights gleaned from the SUSENAS data included

the low participation rates for women (19 per cent) compared with men (81

per cent), almost certainly a refl ection of the overwhelmingly heavy physical

labor that was offered by these programs. The wage rates on offer appear

to have not been very far below average wages for agricultural laborers,

perhaps helping to explain why so many non-poor were attracted to the

programs (giving a high leakage or type two error).

Table 3.9 Coverage and targeting of anti-poverty programs, August 1998–February 1999

Program coverage (%) Program targeting

Programs Total number Non-poor Coverage Proportion Targeting

of potential Poorest Richest (upper of all of total ratiob

recipientsa 20% 20% 80%) potential recipients

recipients from non-poor

(1) (2) (3) (4) (5) (6) (7)

Subsidized rice 50 385 444 52.64 24.33 36.90 40.09 0.74 0.92

Employment creation 50 385 444 8.31 2.53 4.94 5.61 0.70 0.88

Primary scholarships 29 745 369 5.80 2.04 3.60 4.03 0.71 0.89

Lower secondary scholarships 10 394 621 12.15 4.85 7.53 8.42 0.71 0.89

Upper secondary scholarships 6 430 146 5.40 1.96 3.32 3.71 0.71 0.90

Health Card 27 567 138 10.60 3.09 5.28 6.33 0.67 0.83

Nutrition 19 970 948 16.54 14.24 15.79 15.94 0.79 0.99

Note:

a The total number of potential recipients for each program:

• Subsidized rice and job creation programs: all 50.4 million Indonesian households.

• Scholarships: the total number of individual students enrolled at each school level: primary, lower secondary and upper secondary.

• Health care: all those individuals estimated to have visited a health care provider in the three months prior to the SUSENAS survey.

• Nutrition: the total number of individuals in the ‘pregnant women and children under three years old’ category.

b Targeting ratio is column (6) divided by the fraction of the non-poor in the population. By defi nition, non-poor households are 80 per cent of the

population, hence the targeting ratio is column(6) is divided by 0.8.

Source: Sumarto et al. (2001: Table Appendix 2), based on the 1999 SUSENAS.