DOES INCREASING REGISTRATION COSTS IN CASH TRANSFER PROGRAMS DISCOURAGE RICH FAMILIES REGISTRATION?

Responsible researcher: Eduarda Miller Figueiredo

Intervention Location: Indonesia

Sample Size: 400 villages

Sector: Income inequality

Variable of Main Interest: Probability of enrolling in the program

Type of Intervention: Income Transfer Program

Methodology: Experimental Design

By increasing the cost of applying for selection into cash transfer programs, will wealthy families be deterred from applying? This article aimed to answer this question through randomization of enrollment methods in the PKH Program in Indonesia. The results demonstrated that there is greater participation of poor families in the program after the increase in the registration cost. And the improvement in the segmentation of the program's beneficiaries was due to the fact that the rich predicted that they had a low probability of success and chose not to apply. However, they point out that increasing the cost may not be the best way to carry out such screening.

1. Policy Problem

Assistance programs need to separate the rich from the poor so that targeted aid is received by the individuals the program was designed to assist. To achieve this, different programs have different ways of incurring such separation: requirements that require manual labor, or even providing low-quality foods that cause the rich to choose to purchase higher-quality foods and not apply for the program. In other words, to dissuade the rich from participating in the program, the poor are forced to incur utility costs to receive the transfers.

Therefore, the article aims to determine whether, by reducing such costs for the poor to enter welfare programs, it is still possible to achieve substantial self-selection.

2. Implementation and Evaluation Context

The Keluarga Harapan Program (PKH) is a conditional cash transfer project administered by the Ministry of Social Affairs (DepSos) in Indonesia. It is a program that explores self-selection mechanisms that aim to assist families with per capita consumption below 80% of the poverty line, which is equivalent to the poorest 5% of the population studied. This program requires that the family has a pregnant woman and a child between 0-5 years old or a child under 18 who has not completed the 9th year of compulsory education. Beneficiaries will receive from 67 to 250 dollars per year, which is equivalent to about 3.5%-13% of the average annual consumption of poor families in the sample. Such payments can last for up to 6 years, being paid quarterly.

The determination of families below the minimum consumption required by the program occurs through tests with automatic screening for families that meet the demographic requirements. Therefore, every 3 years, a survey of households that are potentially eligible for all anti-poverty programs is conducted. The selected families go through two stages of questions, the first consisting of 5 questions for an initial filter and, after passing this filter, they receive around 30 questions about their home, property, education and occupation. The government therefore estimates the relationship between these variables and per capita household consumption to generate a district-level formula to predict the level of consumption. Therefore, individuals below this consumption level are eligible for the program.

3. Policy/Program Details

The study was carried out during the 2011 expansion of PKH into areas previously discovered by the project. 6 districts were chosen, which allowed the inclusion of a variety of different cultural and economic environments. Thus, 400 villages were randomly selected, 30% urban and 70% rural. Within these villages, one was selected. The villages contained around 150 families, each of which had its own administrative head. The authors experimentally varied the program enrollment process in these 400 villages, in which some enrolled in the program through the procedure that the government itself implements in other areas and in others the government's statistical system conducts consumption testing of potential beneficiaries in their houses and automatically enroll those who have been approved. Therefore, the main difference studied was whether the family had to actively apply to be selected for possible eligibility or whether they were automatically selected based on government survey results.

From December 2010 to March 2011, baseline data were collected from one randomly selected village from each village. After this, the government conducted targeting treatments. Another survey was carried out in August 2011, the month in which the distribution of funds took place. And finally, the final survey took place in January to March 2012.

4. Method

To reexamine self-selection into a welfare program based on the expected benefits and costs of enrollment, the authors assumed that families live for two periods and that their consumption equals their income.

Households' beliefs about eligibility for the program take Probit . There are two types of families present in the sample:

1. poor families who do not know what the government observes, knowing the real empirical probability that someone with their level of consumption will receive the program.
2. Rich families, who know more about the real rule the government uses.

Thus, if the family benefits from the program, they will receive additional income in the future; if they are not selected, they will receive nothing.

The turnout rate ratio is always greater than 1 because the wealthy are less likely to enroll given the higher costs and lower likelihood of receiving the benefit. Furthermore, the higher the proportion, the greater the fraction of poor people in the requesting population.

Making obtaining the benefit more difficult reduces the number of poor applicants and imposes excessive costs on all applicants, which is undesirable. Therefore, the only motivation for doing this is that it improves the poor-to-risk ratio, reducing government program costs per eligible beneficiary.

• Main Results

From a government perspective, self-selection can affect in two ways:

• Selection for characteristics that are observable by the government: Wealthy families who are less likely to be selected for the PKH may be less likely to enroll. Thus, this type of selection could potentially reduce the government's administrative costs, as it would reduce the number of interviews for the program.
• Selection on the unobservable component of consumption: Households with higher unobservable consumption may be less likely to enroll, this could arise if there is self-selection based on the opportunity cost of time or if households do not fully understand the score.

Therefore, if selection is occurring on unobservables (ii), the introduction of self-selection has the potential to lead to a poorer distribution of beneficiaries than automatic selection.

When testing whether the types of individuals selected in self-selection and automatic screening – which is the usual procedure adopted by the Indonesian government – ​​differ, they observed that the distribution of beneficiaries was poorer families in self-selection. However, this result does not demonstrate whether this is due to the inclusion of poor families, the exclusion of rich families or some combination of both factors. To answer this question, the authors used a non-parametric Fan regression of the probability of obtaining the benefit as a function of log per capita consumption. The results demonstrated that it is a combination of both factors: poorer families are more likely to receive the benefit while richer families are less likely to receive it.

After estimating several results, controlling for several factors to resolve doubts about how the introduction of application costs could impact the selection of families for the PKH, the authors demonstrated that the improvement in the segmentation of program beneficiaries, after the imposition of the cost for applying, was due to the fact that the rich predicted that they had a low probability of success and chose not to apply.

• Public Policy Lessons

The study demonstrates that administrative costs can be an interesting tool to improve selection compared to automatic screening. However, increasing the cost may not be the best way, as such an administrative barrier is seen as another obstacle to be overcome. Therefore, there is still a need to discover screening mechanisms to increase acceptance among the poor and discourage rich families from enrolling in such programs.