The Unified Database for Social Protection Programmes (UDB)

"The Unified Database for Social Protection Programmes (UDB) contains detailed socio-economic information about households  classified as the poorest 40 per cent of the Indonesian population. With data on more than 25 million households, the UDB is the largest database of its kind in the world.

Household registration

The data source for the UDB was the 2011 data collection for social protection programmes (PPLS 2011). The PPLS 2011 was designed specifically to address the issues which compromised the accuracy and effectiveness of the PSE 2005 and PPLS 2008. Households to be included in the UDB were identified using an innovative approach that combined the poverty mapping methodology from Elbers et al. (2003) with community suggestions. First, enumeration quotas—the number of households to be surveyed in each area—were estimated for all administrative levels, from the province to the village level. These quotas were based on the geographic distribution of households comprising the poorest 40 per cent. Second, pre-lists (names and addresses) of households to be surveyed for the UDB were established using the 2010 population census. Household welfare was estimated using proxy means testing (PMT), and households whose predicted welfare was below the enumeration quota threshold were selected to be surveyed by Statistics Indonesia. Third, in addition to households on the pre-lists, field enumerators were required to survey households that were not on the pre-lists but that ‘appeared poor’ to them (survey sweeping) or were recommended for inclusion by the three poorest households on the pre-lists (consultation with poor households).

In practice, the PPLS 2011 fielding exercise was not without its challenges. According to SMERU (2012), official implementation guidelines were often not followed in the field. For instance, in many districts, households deemed non-poor were removed from the pre-lists, either by enumerators or by community leaders, while consultation with poor households seldom occurred as envisioned. Despite these challenges, the PPLS 2011 covered 45–50 per cent of the population, close to the number originally targeted nationally.

Household classification

Indicators in the PPLS 2011 questionnaire were selected based on analysis of data from the national socio-economic  survey (Susenas). In order to limit potential manipulation by respondents with the intention of increasing their chances of receiving social assistance benefits, indicators were also determined based on the ease with which they could be observed by enumerators. Using these indicators, households registered in the UDB are classified according to their predicted welfare, estimated using 497 district-specific PMT formulas, which were derived from Susenas data (TNP2K, 2014).

The data collection and PMT ranking processes were districtspecific, but nationally comparable programme eligibility thresholds were required to ensure that similar households in different locations had an equal chance of receiving programme benefits. Susenas was used to determine the number and percentage of households in each decile for each district. Households in the UDB are classified into the four poorest deciles, which can further be divided into percentiles, ranging from the 5th to the 40th, in order to accommodate the needs of all stakeholders.

The nationally comparable decile or percentile classification in the UDB gives social protection programme implementers (UDB users) the flexibility to design and implement programmes within available budgets and for the coverage level desired.

Institutional arrangement

A permanent national targeting unit (UPSPK) was established to manage the UDB and to facilitate its use by providing information and technical assistance to users. UPSPK aims to optimise the use of the UDB and contribute to improving the effectiveness of social assistance programmes at the national and local levels. UPSPK is organised around three main functions:

• supporting users to access and use the UDB, including through the development of a simple set of data sharing procedures;

• developing, implementing and maintaining the information technology infrastructure to ensure the availability, security and integrity of the UDB data; and

• monitoring the use and evaluating the performance of the UDB for targeting, as well as conducting relevant research to inform continuous improvement of the system.

 

Unification of targeting using the UDB

The concept

The UDB was established to unify approaches to targeting social programmes in Indonesia. Adopting a unified approach to targeting, both to allocate geographic quotas and to select beneficiaries, offers several benefits to social assistance programmes. Using a single source to identify beneficiaries reduces overall administrative costs, since individual programmes do not have to allocate their own resources to accomplish this task. Such resources can instead be channelled to address other outstanding implementation issues. Programmes at both national and local levels are more likely to achieve complementarity, and the government has a tool to monitor their progress towards convergence and integration.

 Initial results: programme targeting accuracy

Since 2012, the main national social assistance programmes—Rice for Poor Households (Raskin), Public Health Insurance (Jamkesmas), the Conditional Cash Transfer Programme for Poor Families (PKH) and Cash Transfers for Poor Students (BSM)—have used the UDB to identify beneficiaries.

The incidence of benefits for Raskin, Jamkesmas and BSM have improved following the establishment of the UDB, in line with the findings of Bah, Bazzi et al. (2014). Raskin shows the least improvement, which may relate to the longheld practice of communities often sharing the subsidised rice, as opposed to it being used only for targeted beneficiaries.

The lack of change in coverage between 2011 and 2013 might also indicate that there was no room for improvement in Raskin, since it already covered nearly all of the poorest families. In the case of Jamkesmas and BSM, total coverage increased noticeably between 2011 and 2013, showing that using the UDB led to better targeting outcomes.

Furthermore, the poorest 10 to 30 per cent of households, rather than those less poor, appear to have benefited more from this increased coverage.

 Initial results: programme complementarity

In addition to improving the targeting accuracy of individual programmes, the single registry also intends to ensure households receive complementary benefits from multiple programmes. The UDB classifies households by decile, which allows its users to identify beneficiaries who need access to a combination of programmes. This is particularly beneficial for the poorest households, since the UDB can list them to receive all social assistance programmes. Data from Susenas in September 2013 showed that of the reported PKH beneficiary households, 76 per cent, 92 per cent and 28 per cent, respectively, also reported receiving Raskin, Jamkesmas and BSM. Moreover, 71 per cent received both Raskin and Jamkesmas. Local governments have also shown a strong interest in using data from the UDB. Between 2012 and September 2014, over 500 local government institutions had requested UDB data to plan and implement more than 1500 locally funded social assistance and poverty reduction programmes. These institutions reported using the data mostly to plan their programmes and identify potential programme beneficiaries (Bah, Mardiananingsih and Wijaya, 2014). Data provided to local governments are tailored to their needs as well as their programme criteria, and a number of these institutions have requested such data more than once.

Establishing the UDB allowed for the introduction of Social Protection Cards (KPS), which help increase programme complementarity at the national level. This was the first time Indonesia had launched a mechanism that would allow recipients to benefit from several programmes simultaneously. The UDB is an important step towards a more unified and integrated social protection system. It also provides the opportunity to improve complementarity between national- and local-level programmes, especially the local versions of national programmes, such as Jamkesda (locally funded regional health insurance) or Raskin-da (locally funded rice subsidies). The UDB, on request, provides local governments with beneficiary lists that are tailored to ensure complementarity either in coverage or in benefit packages.

Affordability

The UDB was established at a total cost of about IDR600 billion for data collection (the PPLS 2011) and with an average annual operating cost of IDR16.3 billion between 2012 and 2014. That is equivalent to about IDR26,000 per household registered, and the data have been used for three years, suggesting an annual cost per registered household of about IDR8700 or USD0.70. In comparison, Castaneda et al. (2005) showed that the annual cost per person registered varies between USD0.20 and USD1.20 in Latin American countries, placing the UDB among the most cost-effective targeting systems. In addition, the main social assistance programmes now share the same targeting system rather than each having to design and fund their own targeting mechanisms. As a result, the UDB reduces duplication of costs for beneficiary selection and targeting. The costs of implementing the UDB amount to about 0.5 per cent of the central government’s expenditures on the four main permanent national social assistance programmes (Jamkesmas/JKN, Raskin, BSM and PKH) between 2012 and 2014.8 It could be argued that the introduction of KPS further increases the cost-effectiveness of the UDB by reducing costs associated with issuing cards for beneficiaries of each programme.

Implementation of the UDB: remaining challenges

Despite early successes, continuous efforts are needed to address the remaining challenges and further improve the ability of the UDB to identify intended beneficiaries, enhance programme effectiveness and, ultimately, accelerate poverty reduction. Most outstanding implementation challenges relate to maintaining and updating the system, in terms of both household registration and classification.

Programme-specific updating

If the UDB is the only source used to select programme recipients, not being registered in the database would mean being excluded from most programmes. In order to mitigate this risk, the National Team for the Acceleration of Poverty Reduction (TNP2K) proposed the Replacements Summary Form

(FRP) mechanism. These forms are used to record outputs from community meetings where UDB beneficiary lists are amended and are subsequently used to update the database. The mechanism aims to provide a transparent means of updating information recorded in the UDB, as well as correcting inclusion errors in household classifications, an issue that can lead to social unrest. However, TNP2K monitoring activities revealed the need to improve the mechanism to ensure its effectiveness.

For instance, the implementation of the FRP mechanism is now supported by a regulation issued by the Ministry of Home Affairs, which makes heads of district governments responsible for its implementation by all villages in their area.

Nationwide re-certification

The planned update of the UDB in 2015 provides an opportunity to improve the registry’s performance in the short run. A significantly higher number of households were surveyed for the UDB than in previous censuses of poor households, and this issue of adequate coverage is a key factor in improving targeting accuracy. The next large-scale update will achieve greater accuracy by registering even more households, using a more transparent data collection process. Transparency will be increased through public consultations to identify households that should continue to be registered in the UDB and add potentially eligible households to survey pre-lists in order to make them more complete. The data collection should also be accompanied by a substantial and extensive information  campaign involving local governments, to ensure that the general public is well informed about the purposes, conditions and criteria for the UDB. This is key to improving public understanding of the single registry, as well as to facilitating the work of enumerators and relieving community leaders of the potential pressure of beneficiary selection.

Public awareness

Raising awareness among local-level implementers of national programmes, about the UDB and the beneficiary selection system it uses, is key to increasing acceptance by both UDB users and beneficiaries and to making the programmes more effective. Generally, more and better awareness-raising activities are needed at the local level. However, these efforts need to be differentiated to cater to the differing needs of local implementers, communities and beneficiaries. Local implementers require detailed information on programme goals and implementation procedures. Communities and beneficiaries require information on the selection mechanisms and criteria. If users and beneficiaries do not understand how the UDB works, they will reject the selection process and its results. This may lead to benefits being shared equally instead of being allocated to target beneficiaries.

Grievance redress system

Developing a credible and functioning grievance redress system (GRS) is also key to improving the UDB. Households and communities will have negative perceptions and will not accept the single registry if they have no way to voice their complaints about targeting or to request an assessment whenever they feel they are eligible for benefits they have not received. During interviews, officers from both central and local government institutions pointed out numerous errors in the UDB. However, looking at the actual figures, a relatively low share of households seems to have been misclassified. For example, not many changes result from either the FRP mechanism or from the local validation exercises which check PKH beneficiary lists for inclusion and exclusion errors.  An effective grievance redress mechanism will increase the credibility of the UDB, given its objective of including all households eligible for social assistance programmes".

Extracted from TNP2K Research Brief No.49/July 2015 (International policy center for inclusive growth)