CONCLUSIONS

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 

It is apparent from the above account that the targeting of poverty

alleviation programs in Indonesia has been a diffi cult and frustrating process

for central government planners attempting to allocate scarce budgetary

resources as effi ciently and effectively as possible. Although poor families

did benefi t to a certain degree, all the programs that we have considered have

suffered from two common problems: the type one and type two errors of

undercoverage and leakage, respectively. Undercoverage occurred as many

poor households were not reached by the program and have therefore not

received the assistance that was actually intended for those who were in the

disadvantaged section of the community, especially during the immediate

Crisis period. At the same time, there has been a signifi cant amount of

leakage, with far too many non-poor households – in some instances

a disturbingly sizable number from the higher quintiles of income and

expenditure, those who are clearly not poor – able to access program benefi ts.

Admittedly, some programs performed far better than others, and there is

some evidence, although not conclusive, that the targeting effectiveness of

some of the key social safety net programs may have improved after the

initial implementation period.

In most cases, the targeting measures that have been adopted in Indonesia

have been either geographic targeting or individual and household targeting.

Sometimes, these two approaches have been used in tandem for a particular

program (see Table 3.5). Yet the effectiveness of both these methods has been

limited by the scope and accuracy of the data that have been available.

As far as a regional or spatial perspective on poverty is concerned, reliable

and authoritative poverty statistics from BPS have only been accessible as

far as the province level. Although useful for revealing the macro picture

across the entire archipelago, this is far too broad a level to be really useful

for the purposes of budgetary allocation and poverty program targeting. As

a result, planners were forced to rely on far more unreliable sources of data

to distinguish need in different parts of the country, since detailed, up to

date, and accurate information about poverty levels throughout Indonesia

– especially in the wake of the economic crisis – was simply unavailable.

We can report, however, on the early results of a major collaborative

research program now underway that has been attempting to address this

problem. A group of researchers from an independent research institute

have been working with BPS staff to test the feasibility of developing

detailed ‘poverty maps’ for the entire country that will provide program

planners with a targeting tool to identify poverty incidence disaggregated

down to lower levels of administrative authority. The results of a recently

published pilot study in three provinces suggest that the methodology will

be invaluable for calculating poverty and inequality indicators at least as

far as the sub-district (kecamatan) level. Work is now underway to extend

the methodology to the remaining provinces. Despite the promising nature

of this research, some limitations remain. Extending the analysis down

to the village level has not proved to be reliable. Nor would these poverty

maps necessarily remain useful in the event of a sudden shock – either

from environmental or economic causes – as the particular nature of any

crisis may render them out of date. Hence, there will be a need for a regular

process of revision (Suryahadi et al., 2003b).

Individual or household targeting has presented a separate problem. As

we have noted, the annual Core SUSENAS (200 000 households) and the

SUSENAS consumption module (65 000 households), which is conducted

every three years, only cover a representative sample of the population and

so cannot be used for household targeting of a national poverty alleviation

program that is delivered throughout the entire country. A complete

household survey covering all of Indonesia’s urban areas and rural villages,

although theoretically feasible, would be extremely expensive and hence an

overwhelming additional cost burden on these programs. Consequently,

program planners have been forced to rely on the BKKBN registration of

family welfare status as the only available source of data covering the entire

country. Despite efforts to respond to the criticisms of using these data as

a targeting tool (principally the concentration on fi xed assets that do not

capture transitory shocks and the initial inclusion of non-economic criteria

of no relevance for assessing poverty status), problems still remained. At

the local level, those charged with the responsibility of allocating program

benefi ts frequently pointed out that the BKKBN registration did not

include any reference to factors such as household income levels, source

and status of employment, family size and number of dependents. As a

result, additional ‘local criteria’ were sometimes included in the targeting

process that was fi nalized at the grass-roots level.

There has also been some limited use of self-targeting, in particular in

the short-lived employment safety net programs. This was achieved by

attempting to hold wage rates well below minimum or average rates so

that only those who were really desperate for work would be attracted to

the program. This would also actively discourage those who are already in

paid employment from participation. This was sound in theory, and has

reportedly worked in other cases to ensure that the poorest individuals

benefi ted from the program, but as we have noted above, the administration

and implementation of the employment programs was poor and rules were

often fl outed so that wages were not always held to these low levels, thus

subverting the original intentions.

From time to time, some have advocated adopting this self-targeting

principle as a solution to the problems of excessive leakage in the food

security program. It has been argued that a lowering of the quality of

the subsidized rice that Bulog distributes would insure that only poor and

needy families applied to purchase an allocation. However, this has not been

taken up as policy – thankfully in our view – as there are many practical

political, administrative and indeed moral objections to attempting such a

risky strategy with an essential food staple and a perishable commodity.

In addition to the actual targeting methodology, another factor – or more

correctly, a range of factors – exercising a considerable effect on targeting

outcomes is the administrative capacity of government agencies to design,

plan and implement programs according to a consistent set of objectives.

There are many aspects to this problem but it should be stressed that the

implementation of the social safety net programs was taking place during

a period of immense social and political fl ux throughout Indonesia. The

capacity of central government agencies to deliver programs effectively

was under closer scrutiny than ever before. Furthermore, in 1999 Indonesia

embarked upon a radical and far-reaching decentralization process, with

most major tasks of government rapidly devolved down to the district

(kabupaten and kota) levels. As a result, increased levels of cooperation

between the layers of government were required to achieve satisfactory

outcomes. This was not always in evidence, especially with tension emerging

between the provincial and the district levels of authority.

Nevertheless, the ability of government agencies to overcome some of

the logistical problems and organize complex administrative arrangements

for some of the social safety nets was an impressive achievement. A prime

example was the system put in place so that students who were receiving

school scholarships were able to withdraw the funds on a monthly basis

through the local post offi ce system. Another was the regular delivery of a

signifi cant tonnage of subsidized rice to over 44 000 separate distribution

points throughout the archipelago. Delays and administrative problems

sometimes occurred but by and large these arrangements made a major

contribution to ensuring the relative success of these programs.

The relationship between central levels of administrative authority and

the local level, especially in a country as large and complex as Indonesia,

nevertheless suggested that a certain amount of fl exibility would be required

as programs were being implemented. As the social safety net programs got

under way, it soon became apparent to central government offi cials that

‘local voice’ also had to be taken into account, and that targeting directives

that were regarded as unacceptable (for example, program benefi ts should

only be delivered to those on the BKKBN list of KPS families) were likely

to be simply ignored or signifi cantly altered at the grass-roots level. Attempts

to design programs according to administrative targeting criteria determined

in Jakarta were soon revealed to be incapable of being implemented in the

fi eld. In some cases, the importance of incorporating some local decision

making into the program design had been well understood from the outset

(for example, in the targeting approach adopted for the school scholarships

program). In other cases, it required a process of trial and error as central

government offi cials came to terms with what was actually occurring during

the initial phases of certain programs. As a result, offi cials then made the

required modifi cations to the offi cial program guidelines. This was certainly

what happened with the OPK/Raskin subsidized rice program.

For this reason, it seems quite misleading to draw a distinction between

de facto versus de jure targeting as one recent study has characterized the

process, since this implies that what was occurring at the local level was in

breach of program rules that had been determined by the central government

agencies (Pritchett et al., 2002: 31–32). As we have noted above, in the most

recent version of the subsidized rice program’s offi cial guidelines the fi nal

decision about which families were to become benefi ciaries was passed down

to village-level decision makers for fi nal determination.

Of course, as a number of authors have pointed out, too much local

fl exibility can lead to undesirable targeting consequences: there is always the

risk that program benefi ts will be shared out among so large a number of

recipients that the essential purpose of the program as a poverty alleviation

measure is lost. This certainly occurred in many locations in the case of the

subsidized rice program. There is also the risk that local elites can divert the

benefi ts of the program to those for whom it was not intended. Corruption

remains a widely acknowledged problem in Indonesia and clearly occurs

at all levels. The social safety net programs were not free from instances of

theft and misappropriation of material benefi ts, sometimes involving local

offi cials. Dealing with this problem effectively requires a commitment on

the part of government that has not always been evident.

Several observers have noted that effective targeting and successful

implementation of poverty alleviation programs can be assisted – especially

in the special circumstances of crisis-related programs – if there is a relatively

simple message that can be easily communicated to the wider public. This was

certainly the case with the cheap rice program and the scholarships program,

as the essential elements of these measures were easy to comprehend. Some

limited publicity through the mass media was attempted, although this can

add to the implementation costs of such programs.

Other programs (for example, PDM-DKE and the entirety of the health

sector safety net, JPS-BK) were exceedingly complex, and hence the essential

purpose of these programs could not be communicated to the general public

in a way that made it clear who the recipients really were and what they

should expect by way of program benefi ts.

Admittedly, the importance of the need to persuade and inform – rather

than issue directives – is a relatively new concept in the post-New Order

environment in Indonesia. Government offi cials have so far failed to come to

grips with the value of effective communication about the essential purposes

and aims of government programs. Most of the effort that was spent in this

area leading up to the implementation of the poverty alleviation programs

was directed at informing local offi cials at the various layers of government

about the details of the administrative procedures that they were required

to follow. A far more concerted effort needs to be directed at informing the

wider community.

Another element of the implementation of poverty alleviation programs

that has not yet received suffi cient attention is an effective monitoring and

reporting process. In the early stages of the implementation of the Social

Safety Net program, monitoring activities were carried out in specifi c areas

by a range of community groups and non-government organizations. Some

of these activities were supported by foreign donor agencies through small

grants mainly to promote the idea of transparency and public accountability.

However, the monitoring that resulted was somewhat ad hoc and often

adversarial, with the most controversial social safety net (PDM-DKE)

absorbing most public attention. Subsequently, monitoring of many other

programs has been quite limited. Certain districts and government agencies

have recruited local university students or NGOs to monitor the delivery

of subsidized rice to the delivery points, but their responsibilities did not

extend to the actual distribution of rice to recipients.

The two programs where a major effort was made to establish a rigorous

and effective monitoring process were the education and health social safety

nets. In both cases, a central independent monitoring unit was established

in Jakarta, with branches also set up at the provincial level throughout the

country. In addition to regular monitoring through fi eld visits by staff and an

active publications program of reports and newsletters to inform interested

parties of program results, investigations were conducted into any reported

irregularities in the operation of the program. The education program

monitoring and reporting initiative seems to have been especially useful in

keeping track of the program and ensuring that targeting effectiveness was

maintained at a satisfactory level. Yet such an ambitious operation with

specialist staff recruited for lengthy periods was an expensive operation, and

was only possible through the fi nancial backing of those donor agencies who

wished to ensure that the program that they were supporting was achieving

its objectives. It remains unlikely that the Indonesian government would

contemplate replicating such an ambitious operation in other programs, no

matter how desirable and useful this may have been.