Improving the Transparency and Delivery of a Subsidized Rice Program in Indonesia

Jordan Kyle Cohen
Sudarno Sumarto
Lampung, South Sumatra, and Central Java provinces, Indonesia
572 villages
2012 - 2014
Outcome of interest:
Corruption Transparency and accountability
Intervention type:
AEA RCT registration number:


Governments around the world face the challenge of ensuring that the rules governing social assistance programs for the poor are implemented as intended. Within Indonesia’s national subsidized rice program, researchers tested two interventions designed to improve service delivery and reduce program leakage: increasing program transparency by sending households official identification cards with information about the benefits they were eligible to receive, and providing villages with the option to privatize rice distribution. Providing information about benefits on official ID cards substantially increased the total benefits eligible households received and led to a large reduction in leakage. Privatizing rice distribution also improved program performance and reduced the price households paid for rice.

Policy issue

Social assistance programs like cash transfers and food subsidies are one of theprimary channels through which people interact with their government, butservice delivery presents a complex challenge. Without a clear understandingof how a government program is supposed to work, people are less equippedto expect and demand better performance.

Transparency and accountability between a government and its citizens arecritical for effective public service delivery. In theory, providing people withinformation about the government benefits that they are entitled to can helpthem advocate for better service delivery and hold government officialsaccountable. It is not clear, however, whether this occurs in practice.

One potential way to ensure social programs reach their intended beneficiaries is to provide people with information the benefits that they are eligible to receive.

Another possible way to improve program delivery is to privatize implementation, as governments can often provide stronger performance incentives to private contractors than to their own employees. However, the impact of privatization is unclear because stronger incentives could also push contractors to lower program quality to cut costs.

Context of the evaluation

Rice comprises a large share of the household budget for many low-income households in Indonesia.In 2012, the Government of Indonesia was considering distributing identificationcards to low-income households informing them about their eligibility for anational subsidized rice program, Raskin, in an effort to improve theprogram’s service delivery. The government requested evidence on theeffectiveness of the cards before rolling them out nationwide.

In response, the study authorspartnered with the governmentto conduct a fast, large-scale randomized evaluation of the Raskin ID cards in order to inform the government’s decision about whether toscale them up.

Raskin is Indonesia’s largest social benefit program and targets the poorest 30 percent of the population. At the time of the evaluation in 2012, Raskin-eligible households could purchase 15 kilograms of rice per month—about half of a typical household’s monthly rice consumption—at a subsidized price of one-fifth the market value. In 2012, the budget for Raskin was US$1.5 billion and the government distributed 3.4 million tons of subsidized rice to 17.5 million people. The program launched in 1998; at the time of this summary publication in 2019, it continued to operate under the name Rastra.

Administration of such a large benefit program posedmany challenges:

  • Lost rice (leakage): A substantial amount of rice intended fordistribution disappeared before it reached people due tocorruption, weak oversight, and inefficiencies. In 2012, eligiblehouseholds received only about one-third of their entitled benefits.
  • Awareness: Raskin was a well-known program, butbeneficiaries were not always aware of program rules andeligibility requirements.1

  • Targeting: Local officials were supposed to use an officialroster of eligible households to determine who could accessthe program, but instead they sometimes used their owndiscretion. In this study, 63 percent of ineligible households inthe comparison group reported that they were able to purchaseRaskin rice recently.
  • Pricing: Local officials inflated prices. For example, eligiblehouseholds in the comparison group in this study paid onaverage 42 percent more than the official subsidy price.
  • Still from animated commercial on family with social protection identification card
    TNP2K commercial about the social protection identification card

    Details of the intervention

    Information intervention

    In collaboration with the Government of Indonesia, researchersconducted a randomized evaluation measuring the impact of sendingRaskin-eligible households identification cards featuring informationabout the program. They designed the study to generate preliminaryresults in less than a year to inform the government’s decisionabout whether to roll out the cards nationwide.

    The government mailed identification cards to Raskin-eligiblehouseholds in 378 villages, randomly selected from 572 villagesacross three provinces. To better identify the different waysin which information had impacts on program delivery, thegovernment randomly assigned villages to four variations of theintervention or a comparison group:

    1. Basic Cards: Eligible households received cards withinformation about their eligibility status and the quantity ofsubsidized rice that they were entitled to purchase.
    2. Basic + Price Cards: Eligible households received cards thatincluded the same information as Basic Cards, plus the officialsubsidized price for Raskin rice.
    3. Cards + Public Information: In a subset of villages in the BasicCards and Basic + Price Cards groups, lists of eligible householdswere posted in public gathering areas and information aboutthe cards was played on a village loudspeaker.
    4. Cards to Lowest-Income Only: In a subset of villages in theBasic Cards and Basic + Price Cards groups, the governmentonly sent cards to households in the lowest 10 percent of theincome distribution (32 percent of eligible households).
    5. Comparison Group: In comparison villages, the governmentcontinued to run the program under the status quo with noidentification cards.

    Researchers conducted surveys of eligible and ineligible householdstwo months, eight months, and eighteen months after the cardswere mailed to measure key outcomes important to the nationalgovernment, including the amount of rice received by eligible(and ineligible) households, the price they paid for the rice, andindividuals’ understanding of their benefits.

    Privatization of last-mile service delivery

    In a randomly assigned 191 villages out of the same sample of 572, the central government introduced an auction procedure where citizens could bid for the right to distribute Raskin rice.

    To facilitate bidding, a facilitator shared information about Raskin and the new distribution process with the community. A local committee was then organized to examine bids, choose the winner, and monitor distribution.A random subset of 96 villages was encouraged to have a minimum of three bids before the auction could proceed.

    To control for the fact that the bidding intervention could improve program performance just by providing citizens with better information about Raskin, another 96 villages received an “information only” intervention, which provided citizens with the same program information and organized a committee to monitor Raskin distribution, but did not include the actual bidding process. The remaining 285 villages served as a comparison group and maintained the status quo rice distribution process.

    Raskin commercial (with English subtitles)


    Results and policy lessons

    Information intervention

    Identification cards with information about Raskin benefits increasedthe total benefits that eligible households received and reducedprogram leakages, despite incomplete program implementation.2

    Information increased peoples’ knowledge of their eligibilitystatus. Thirty-nine percent of eligible households that were sentcards knew their official eligibility status, relative to 30 percentin the comparison group—a nearly 30 percent increase.

    Information helped eligible households receive more benefitson average, and effects persisted over time. Eligible householdsin card villages received about 26 percent more total benefitson average relative to the comparison group in the two monthsprior to each survey (Figure 1).3 This increase was due to acombination of two factors: a 24 percent increase in the averagequantity of rice received, and a 2.5 percent decrease in theaverage price paid. Increases in benefits persisted for up toeighteen months after the intervention.

    Chart showing impact of identification card intervention compared to comparison group.

    Adding the official price of rice to the cards doubled theirimpact. Eligible households in the Basic + Price Cards groupreceived 24 percent more in total program benefits on averagerelative to the comparison group (a gain of Rp. 6,802 or US$0.69per month)—113 percent more than the gains seen by the BasicCards group (Figure 2).This difference was largely driven by those in the Basic + PriceCards group receiving a higher quantity of rice, rather than payinga lower price. This suggests that cards with price informationwere effective not because they encouraged officials to complywith the program’s pricing rules, but rather because theyimproved households’ bargaining power with local officials.

    Publicizing information about program eligibility furtherincreased households’ access to cards and knowledge of theireligibility status.Eligible households in the Card + Public Information group were19 percent more likely to actually receive a card and 50 percentmore likely to present the card when picking up their ricecompared to those in the card-only groups. Public informationalso increased eligible and ineligible households’ awareness oftheir own eligibility status.

    Public information doubled the average increase in total benefitsthat eligible households received. Eligible households in the Cards+ Public Information group saw a 34 percent increase on averagein total benefits relative to the comparison group (a gain of Rp.9,666 or US$0.99 per month)—100 percent more than the gainsseen by card-only groups.

    Chart showing impact of price and public information treatments compared to card only treatments.

    When cards were sent only to the lowest-income households,they had no impact on benefits for those households. Sendingcards only to the poorest subset of eligible households led tofewer complaints to local leaders about program administration,but did not change the total benefits that these households receivedrelative to the comparison group. When cards were sent only to thelowest-income households, those households fared no differentlythan when cards were sent to all eligible households. Eligiblemiddle-income households only received more benefits when theywere in villages in which all eligible households were sent cards.

    Cards reduced program leakage by 33–58 percent. Eligiblehouseholds that were sent cards received more benefits on averagethan the comparison group, ineligible households in card villagesreceived no less, and government spending on the program didnot change. Taken together, this implies that overall leakagewithin the program declined significantly. Researchers estimatethat the cards reduced leakage by 1–1.6 kg of rice per eligiblehousehold, a 33–58 percent reduction in “lost” rice.

    Privatization intervention

    Offering villages the option to privatize rice distribution also improved program performance and reduced the price households paid for rice.

    Overall, privatization improved program performance, though by a smaller magnitude. The bidding process led about 17 percent of villages to switch distributors.Villages generally prioritized suppliers who had relevant distribution skills as traders (who manage sales and distribution of goods for a living), access to transportation for distribution, and a savings account that could be used for business.

    The bidding process reduced price markup by IDR 49 (US$0.01) relative to the information-only intervention, equivalent to an 8 percent reduction in markup. Requiring three bids further reduced markup by about IDR 74 (US$0.01).

    In contrast, simply providing information without the bidding process had no effect on the price or quality of rice distributed, indicating that it was the bidding process and not increased transparency that led to improved program delivery.

    Policy lessons

    Delivering information to people can be a powerful tool for improving government service delivery. Providing information to peopleon benefits that they are entitled to through official channels like government ID cards can help balance the power between citizens and localofficials, reducing opportunities for corruption and increasing access to social services. This is possible to achieve even for a well-establishedprogram like Raskin.

    Making public information easily accessible can be empowering. Transparent eligibility information, through increasing people’s awarenessof their own rights and the rights of others, can further equip people with the knowledge they need to access the benefits they are entitled to.

    This type of information delivery can be highly cost-effective. The estimated increase in benefits that households received over thecourse of eighteen months was more than seven times the cost of the card program. The benefits of the cards exceeded the costs within justtwo months. This suggests that providing information to people about their rights to benefit from social services can be a cost-effective wayto reduce corrruption and improve access to services.

    Policy influence

    This study was developed in partnership with the Government of Indonesia with the explicit purpose of generating rigorous evidence toinform a national policy decision. Surveys were carefully timed to accommodate government deadlines and researchers quickly relayedresults to policymakers in under a year.

    In part based on these results, the government decided to scale up social assistance identification cards to 15.5 million households across thecountry, reaching over 65 million people in 2013. The cards included information on two other social benefit programs in addition to Raskin.Drawing on the results of another evaluation by J-PAL affiliated researchers, the scale-up included a local engagement process that enabledcommunities to reallocate cards for people who moved or for whom the community deemed too wealthy to receive the program.4

    For more details, see the Evidence to Policy Case Study.

    Abhijit Banerjee, Rema Hanna, Jordan Kyle, Benjamin A. Olken, and Sudarno Sumarto, 2018. "Tangible Information and Citizen Empowerment: Identification Cards and Food Subsidy Programs in Indonesia," Journal of Political Economy 126, (no 2): 451-491.

    Banerjee, Abhijit, Rema Hanna, Jordan Kyle, Benjamin A. Olken, and Sudarno Sumarto. "The Role of Competition in Effective Outsourcing: Subsidized Food Distribution in Indonesia." March 2017.

    Smeru Research Institute. 2008. “The Effectiveness of the Raskin Program.”; andWorld Bank. 2012. “Raskin Subsidized Rice Delivery: Social Assistance Programand Public Expenditure Review.”
    The identification card program was not fully implemented. Eligible households in theprogram were just 30 percentage points more likely to receive the cards than those inthe comparison group. Incomplete implementation was due to multiple factors. Somecards may have gotten lost in the mail. In villages with no postal infrastructure, cardswere sent directly to local leaders to distribute, providing an opportunity for them toblock distribution.
    Researchers calculated total benefits by multiplying the quantity of rice that householdsreceived by the difference between the market price of rice—what households wouldhave paid without the subsidy benefit—and the subsidized price that households paid.
    Alatas, Vivi, Abhijit Banerjee, Rema Hanna, Benjamin A. Olken, and Julia Tobias. 2012.“Targeting the Poor: Evidence from a Field Experiment in Indonesia.” American EconomicReview 102 (4): 1206–1240.