SPDI In Indonesia – Delivering Food Assistance

SPDI In Indonesia – Delivering Food Assistance

Informing Transition from Food Delivery to Food Vouchers in Indonesia

What is the difference in impact between welfare programs that provide food vs. programs that provide food-vouchers? Researchers affiliated with Evidence for Policy Design (EPoD) at Harvard Kennedy School and the Abdul Latif Jameel Poverty Action Lab (J-PAL) sought to answer this question based on the Indonesian government’s largest social assistance program, Rastra (formerly called Raskin). The program provides 10 kg of rice to the bottom 25 percent of Indonesia’s income distribution, and the Government of Indonesia randomly phased the transition from in-kind delivery to equivalent electronic vouchers. Using the SPDI approach, researchers accompanied the transition and evaluated program effectiveness and also relative delivery costs. Read below about how this research informed the transition of what became a key part of Indonesia’s social safety net during the COVID-19 pandemic. 

 

Step One: Identify

The in-kind delivery of rice involved centralized procurement and distribution from district-level warehouses to villages by village heads. This process resulted in problems such as low quality rice (moldy and full of small stones, for instance), irregular deliveries and even rice theft. Thus, on average, eligible households were receiving only around one-third of the intended subsidy.

Step Two: Diagnose

Those problems were associated with 1) a poorly functioning market, given the monopoly of government-supplied rice, that did not create incentives for increasing product quality and 2) issues in the distribution of rice causing subsidized rice to be distributed to more households than initially targeted.

Step Three: Design

The Government of Indonesia designed a new system, called BPNT, to replace in-kind delivery with digital vouchers. This new system would allow beneficiaries to purchase food from any vendor - public or private - that can process digital payments for the program. Additionally, the electronically managed system would transfer the vouchers directly to an account in the name of the beneficiary - which would, in theory, reduce the risks of irregular delivery.

Step Four: Implement

The first-stage rollout of this transition, covering more than 1.4 million households, began in 2017. For the following year, the Government deemed that 105 districts were ready to transition, but budgetary and logistical resources could only cover a fraction of them. Therefore, 42 were randomly assigned to receive the program in 2018, while the remaining 63 districts were randomly assigned to receive the program in 2019.

Step Five: Test

EPoD and J-PAL researchers tested the impact of the overall reform to switch from in-kind transfers (Rastra) to digital food stamps (BPNT) using an experimental design, comparing results between randomized districts that were considered ready for the transition. They measured the effects on poverty, changes in food consumption and rice price and on the administrative costs of the program. The results indicate that vouchers led to greater impacts on poverty, whilst substantially decreasing the administrative costs of delivery.

Step Six: Refine

A national rollout of the electronic vouchers happened in 2019. Since then, BPNT - whose name was changed to Sembako - was made more generous and flexible: the amount of assistance and number of beneficiaries has increased, and the types of food eligible for purchase have expanded. Program Sembamako was a crucial part of the COVID-19 assistance framework. Looking forward, the research team and the Indonesian government are interested in how delivering social assistance through digital payments could have positive knock-on effects on how people engage with the digital financial system. Hence, there is ongoing work investigating the impacts of the BPNT transition on financial inclusion outcomes.