Impact of the SGP

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 

In the initial period of the SGP implementation, the coverage of the

program was lower than had been targeted. Drawing on data from the

1999 SUSENAS Special Module, Sumarto et al. (2001) found that the

program only covered 4 per cent of students at the primary school level,

8 per cent at lower secondary, and 4 per cent at upper secondary level,

compared with the targeted 6, 17 and 10 per cents.36 Moreover, despite an

evident bias towards the poorest segments, there was also clear evidence of

leakage, since a relatively large proportion of the recipients also came from

better-off households. While the program reached 6 per cent of primary

school students at the lowest quintile of per capita consumption, 4 per cent

of students at the higher quintiles (and 2 per cent in the top quintile) were

also allocated scholarships. At the lower secondary school level, 12 per cent

of students at the poorest quintile received scholarships, but 7.5 per cent

of students at the higher quintiles (and 5 per cent at the top quintile) also

benefi ted. Meanwhile at the upper secondary level, 5 per cent of students

at the poorest quintile received the scholarship, as well as around 4 per cent

of students at the higher quintiles, (including 2 per cent of students at the

top quintile). Given these fi ndings, the authors argued that the targeting of

SGP was near to random (see Table 3.9).

According to CIMU (2000a: 16), apart from the data problem, one

possible explanation for the evident mis-targeting among upper secondary

students was the fact that at this level students who received scholarships

might have been among the poorest at their school, but they are not always

from the poorest segments of society. Although a high percentage of primary

students come from poor families, this proportion drops dramatically at the

lower and upper secondary levels. Thus, while 72 per cent of scholarship

recipients at the primary school level come from the two lowest expenditure

quintiles, the fi gure falls to 58 and 42 per cent at subsequent levels. Another

problem mentioned in the CIMU report was that by directing scholarships

to those who were already in school, a substantial number of the poorest

young people of school age who had already dropped out, who were not

attending school or attending ‘open junior high school’ (SMP Terbuka),

were not eligible for the scholarships. Hence, such poor students were never

likely to receive support from the program.

We should also again note that the SMERU study was based on the

SUSENAS data that were collected in February 1999, covering only the

previous six-month period. Hence, it only provides limited information on

the implementation of the program, and only focuses on the issue of the

targeting coverage of the SGP. It does not consider some of the other factors

that may be important in making an overall evaluation of the program.

Other studies of the scholarship program have suggested fi ndings that

are more positive. While the SGP had initially contributed to preventing

enrollment rates from declining sharply between the 1997/98 to 1998/99

academic years, it may have played a major role in increasing enrollment

rates in subsequent years. In the 1999/2000 and 2000/01 academic years, the

enrollment rate for all age groups increased, with the largest increase enjoyed

by students from the poorest expenditure quintile (CIMU, 2001b: 3).

Another study by Cameron (2002), applying regression analysis of

the probability of students becoming school dropouts using the 100

Village Survey data, argued that the scholarships signifi cantly reduced

the probability of dropout at the lower secondary level, but did not affect

dropout rates in primary and upper secondary schools, at least during the

fi rst few months of the program’s operations. However, some care should be

taken when interpreting her fi ndings or comparing these with other studies.

Firstly, the study used the 100 Village Survey data, which is more limited

and focused on poor villages in comparison with the SUSENAS Special

Module.37 Secondly, as Daley and Fane (2002) have argued, Cameron’s

results seemed to underestimate the effects of the scholarships in reducing

dropouts since her results relate only to dropouts in the course of a school

year. Using her data set, it is impossible to estimate the presumably larger

impact of the offer of a scholarship on the probability of re-enrollment at

the start of a school year.

Some modifications to the allocation rules, including a better

representation of poor private and religious schools, improved the coverage

and targeting of the SGP in later years of the program. This was assisted

by better performance of the allocation committees, increased community

participation and local awareness. In a nationwide survey, CIMU found that

7 per cent of all primary school students, 20 per cent of all lower secondary

school students and 11 per cent of all upper secondary school students

received SGP scholarships in the 1999/2000 academic year (CIMU, 2001a:

8). In terms of targeting, using data obtained from the 2002 SUSENAS,

Sparrow (2003a: 20) has calculated that in the academic year 2001/02, 70

per cent of the scholarships went to the poorest two expenditure quintiles,

an increase from 60 per cent in the fi rst year of the program as revealed by

the 1999 SUSENAS data. Meanwhile, the percentage of scholarships going

to the richest quintile decreased from 6 to 3 per cent.