Digital Cash Transfers in Bangladesh
08 Mar 2018 by Daniel Winstanley,
UNDP’s Social Security Policy Support (SSPS) Programme is working with the Government of Bangladesh to re-configure the current social security system so that economic growth is achieved in a more inclusive manner, with economic opportunities reaching the rural and urban poor and the protection of vulnerable groups against shocks. Support is provided primarily in two areas: governance of social protection and strengthening of systems.
In collaboration with SSPS’s sister project, Strengthening Women’s Ability for Productive New Opportunities (SWAPNO), and in partnership with the Bangkok Regional Hub and Innovation Facility, which is funded by UNDP and the Government of Denmark, UNDP is testing the delivery of social security cash transfers to SWAPNO beneficiaries through bKash’s and Dutch Bangla Bank Limited’s (DBBL) Rocket mobile-money services. bKash and Rocket are pre-existing mobile-money service providers in Bangladesh, with hundreds of thousands of agents across every local region (Union Parishads) of the country.
The goal of this initiative, titled Social Security Digital Cash Transfers, is simple: test an alternative, digital method of transferring social security cash transfers to Bangladeshi citizens, and provide rigorous evidence that this method brings significant savings in terms of costs and time as opposed to the current traditional method. After an intensive few months of methodological design, partnership development, procurement, and implementation, a regular flow of high quality data from this initiative is about to begin flowing in.
We chose to test only mobile-money because of an earlier test pilot of 8 payments that we conducted and that showed, on this limited sample, that beneficiaries of mobile-money enjoyed significant savings in time and cost as opposed to other digital alternatives. But would we be able to see those savings at scale? Fast-forward and we are in a 12-18 month study with a much larger sample size.
The sample of the project consists of women living below the Lower Poverty Line in the districts of Kurigram and Satkhira. We randomized a population of 4464 into 3 groups: 648 bKash, 648 Rocket, and 648 control. The control group, like the majority of Bangladeshi citizens who receive a social security cash transfer, will receive their money by being physically present at the nearest local bank, which has a list of all the names, amounts, and frequency of transfer. The bKash and Rocket group beneficiaries were all provided with the same low-cost feature phone (USD 15) to ensure consistency throughout the intervention. Their cash transfers will be automatically deposited into their mobile-money wallet at regular intervals, of which they can withdraw physical cash of any amount at any time from one of the hundreds of thousands of agents across the country of their respective service provider.
We then identified our sources of data collection: surveys, focus group discussions, and key informant interviews, complemented by a Geographic Information System (GIS) mapping exercise. The survey was designed, and built into an app in order for our local workers to be able to easily capture survey data on a tablet device. The survey consists of 25 questions in a self-reported format and will be carried out at regular intervals throughout the initiative on both intervention and control group beneficiaries. We’ve broken our research components into two: efficiency (time, cost, number of visits) and financial inclusion (interaction with mechanics of service provider, changes in money management, and interactions with other services of service provider). Additionally, included in UNDP’s signed Memoranda of Understanding with bKash and DBBL is a clause for accessing hard backend data to complement the survey data; this clause was established based on a carefully crafted Data Governance Framework, which includes having beneficiary approval for accessing such account information.
Further, we brought on a behavioural insights specialist, focusing on digital financial services, to visit Kurigram and Satkhira, talk with SWAPNO beneficiaries, and provide us insight into what might be the risks involved in our study, and how to capture any behavioural data. Such insights include the difficulty of some beneficiaries in reading an English-language keypad on their low-cost feature phone, and the relationship of UNDP’s local workers with the beneficiaries which is mostly informal in nature and therefore most guidance is performed in methods such as storytelling or in pictures as opposed to academic style instruction.
Partly as a result of that work, we further subdivided our intervention groups into 2 smaller groups (now 4 groups of 324). The idea is to have 324 bKash and 324 Rocket participants in each district to receive only an initial orientation on utilizing their new payment method (plus support from our local workers when requested), while the second groups will receive that, plus regular reminders and support throughout the intervention, such as a 15 minute pictorial exercise.
We’ve also conducted a GIS Mapping exercise, through collecting GPS data points from several hundred key locations, including beneficiary homes, agent points, and bank collection sites. The purpose of this exercise is to visually demonstrate the changes in distance required to withdraw cash between the intervention and control groups. In most Union Parishads across the country, all beneficiaries of any Government social security cash transfer will congregate at 1 specific bank, whereas with mobile-money agents, the distance to a cash-out point is dramatically reduced and less congested.
At this point, the first social security cash transfer has been delivered and beneficiaries are withdrawing some or all of the money as they see fit. The initial orientation with beneficiaries has previously been completed. The first survey will be conducted over the next few weeks.
An important lesson that we’ve learned along the way though, is that sometimes our ambitious timelines get slowed down for a variety of factors, such as the disastrous flooding that took place late 2017 in Kurigram.
Now that all the heavy lifting has been done to set up this research, the next steps include the continuous delivery of the cash transfers to the SWAPNO beneficiaries and regular surveys to track changes in our agreed upon metrics. At the end, we’ll compile, analyze, and report on of all our data we collect which will be turned into a policy document, and formally presented to the Government of Bangladesh with the intent to provide a pathway for national scaleup.