Should government payments be digitized?

Especially in the era of COVID-19, it is hard to imagine a world without WhatsApp or Zoom. Yet, these tech platforms did not even exist two decades ago. These and many other innovations are proof of the ability of digital technology to transform our lives at rapid speed. 

These changes extend far beyond just the individual level of messaging apps, reaching a much larger scale, with entire economies becoming more digital. Digitization of economies is being accelerated by the COVID-19 pandemic and provides an opportunity for countries to make large strides towards greater economic development. 

Policymakers across sub-Saharan Africa are increasingly investing in a particular digital technology with transformative potential—biometric identification (ID) systems. Biometric IDs are a form of identification that uses biometric information of individuals to record, identify, and verify their identity. They use traits such as fingerprints, eye retina and iris scanning, voice recognition, facial patterns, and body movement as biometric measures of verification. Traditional IDs usually use other documents or people to verify information about someone or are prepared at birth. This step of digitizing ID systems is partly to increase the provision of a legal ID, which is often essential for citizens to access basic government services and lead a dignified life. It will also enable broader use of digital payment systems and promote digitization of services more generally. 

Countries like Malawi, Senegal, and Uganda have rolled out national biometric ID systems, while others including Ghana, Kenya, Nigeria, and Tanzania are in the process of implementing mass registration campaigns for their new biometric IDs. More countries still, including Ethiopia, have announced plans to start enrollment for such IDs in the coming months and years. As this digital ID revolution sweeps the continent, what do we really know about its benefits and drawbacks? How does this rapid digitization and proliferation of biometric IDs affect the lives of citizens? And how can they best be structured to result in the most benefit? In this blog, we explore the benefits and drawbacks of digital IDs, as well as the various questions that remain unanswered.

The Good

Overcoming the problems associated with a lack of identification through digital IDs might have powerful implications. A digital repository may make it easier to implement and target government programs, while also enabling citizens to participate in the digital economy. Since these IDs can function as a universally valid proof of identity, they can easily be linked to existing functional systems—like bank accounts and payroll, asset registries, insurance, and school records—to avoid corruption and siphoning of resources, as well as double counting and targeting errors in beneficiaries. By easing processes to obtain voter IDs, drivers’ licenses, telecom services, and banking facilities, IDs can mobilize citizens’ political and economic participation. Additionally, IDs can catalyze the growth of a country’s fiscal capacity by not only minimizing tax evasion but also by being instrumental in promoting financial inclusion and growth of the formal economy. Evidence points to many potential benefits of digital ID systems such as:

1. Improved data: It is difficult to allocate funding or identify problem areas without access to quality, granular data. Digital ID systems may alleviate this issue by facilitating improved record-keeping and generating administrative data (routine information collected, used, and stored by governments and other organizations primarily for operational, rather than research purposes) on individuals in the country.  
2. Improved public service delivery: Many countries plan to link ID systems to digital payment and welfare systems, which should help make public services more efficient and accessible. These policy changes aim to eliminate leakages and enhance the accuracy of transactions as well as the ability of the government to monitor them. When linked with monitoring systems, digital IDs have reduced absenteeism among health workers, reduced data forgery, and improved patient adherence to treatment in India. 
3. Improved targeting: Following from the previous point, digital ID systems may make it easier to implement and target government welfare programs. Evidence from India shows that biometric IDs cost-effectively increased efficiency of welfare payments by reducing leakages and delays while not reducing program access, particularly to vulnerable households. Further, in India, beneficiaries overwhelmingly preferred the smartcard systems.
4. Better credit market efficiency: Digital IDs can help address barriers to accessing formal financial products. Evidence from Malawi shows that fingerprinting for credit can improve loan repayment rates and lead to a better functioning credit market. By using biometric identification, lenders were better able to identify borrowers. Introducing fingerprinting identification for microloans caused high-risk borrowers to take out smaller loans and to improve their repayment behavior.

The Bad

While there are many potential benefits to using digital IDs, they often have their drawbacks. Digital ID systems have seen large pushback in many countries due to fear of exclusion and erosion of privacy. Further, the implementation of digital ID systems could cause more pain than gain in specific contexts such as remote areas with limited connectivity. Governments need to carefully evaluate where digital ID systems will maximize welfare without inconveniencing or excluding the already marginalized. Some of the recorded challenges with digital ID systems are:

1. Exclusion errors: Stringent ID requirements can generate non-trivial costs in terms of exclusion and inconvenience to genuine beneficiaries. Evidence from India shows that making biometrics compulsory in the public food distribution system created exclusion problems and increased transaction costs, especially for vulnerable groups, without reducing corruption. These findings were corroborated by a second study, which found that biometric IDs led to limited improvements, such as more timely and reliable recording of PDS transactions, and resulted in the exclusion of households (usually the most vulnerable ones such as widows, the elderly, and manual workers) that were unable to pass the biometric authentication test. The authors argue that the imposition of the digital system here led to “pain without gain” as it did not address the type of leakages that were prevalent in the given context. 
2. Implementation concerns: Another study in India found that while a biometric attendance-monitoring intervention led to health improvements, its imperfect enforcement illustrated the limitation of technological monitoring solutions when they are not combined with changes in the broader rules governing health workers. While there were some health gains from this, the biometric devices led to lower work satisfaction among staff and increased difficulty hiring new nurses, lab technicians, and pharmacists. As a result of this and other factors, there was limited appetite at all levels of government to use the better quality attendance data to enforce the government’s human resource policies. 
3. Privacy and data misuse concerns: Biometric authentication raises important privacy and data misuse concerns. There has been major pushback in India and Kenya to biometric national ID systems due to concerns regarding safety and misuse of collected data. Large data sets have been shown to have great power in surveillance, as well as predicting and shaping people’s decisions. In light of this, it is crucial to ensure that strong data privacy and security legal frameworks and systems are in place.

The Unknown

While we are beginning to glean insights on some of the benefits and drawbacks of digital ID systems, many questions remain unanswered, particularly in the context of sub-Saharan Africa. With the broad impacts that digital IDs can have, there is little research on how they can improve the implementation of welfare programs, how these systems and linked programs affect citizens, and what unintended consequences may result from having access to IDs and their associated technology. As government reforms in both IDs and payments advance, rigorous research can help them answer these questions, including providing insights into design questions as reforms scale up.

The Digital Identification and Finance Initiative in Africa (DigiFI)—hosted by J-PAL Africa at the University of Cape Town—aims to fill this evidence gap by generating cutting-edge policy research projects focussed on the study of innovative government payment and ID systems. We lay out our research agenda in our framing paper. Some examples of questions we hope to answers in the digital ID space are:

• What are the best ways to design digital ID linked services for the greatest impact at the lowest cost?
• To what extent and when can digital ID systems improve targeting and efficiency in public programs?
• How can digital ID systems assist in building incentive systems to motivate public servants?
• How do digital IDs affect voter participation, the fairness of elections, and electoral outcomes? 
• Can digital IDs promote further digitization in financial systems and thus enhance financial inclusion? 
• Can digital ID systems encourage businesses to enter the formal sector? Do these reforms reduce entry costs to entrepreneurship and enable productive investment?
• Can digital ID systems help strengthen law-enforcing institutions and, in turn, affect private investment?

If you are a policymaker interested in jointly understanding how to design digital ID and payment systems in your country or a researcher wanting to add to this body of literature, please reach out to us at [email protected]. We are very interested in understanding digital systems across the continent and are keen to help co-generate analysis that can inform policy decisions on digital ID and payment systems.

Author’s Note: This is the first blog in the DigiFI series on the various aspects of their research and policy priorities. The next blog will focus on government to person digital payments. 

As of 2019, 469 million people across sub-Saharan Africa used mobile money. In 2019 alone, 50 million sub-Saharan Africans created a new mobile-money account, a 12 percent increase from 2018.  Across Africa, governments are exploring new ways of digitizing financial services and identification to reform policies. While there is a big push to go digital, our knowledge of its potential impacts are limited, particularly in the African context. These reforms may have transformative impacts for citizens through improved governance and public service delivery. But they also have the power to exclude marginalized groups or violate privacy rights. 

The Digital Identification and Finance Initiative (DigiFI) in Africa aims to generate rigorous evidence on the impacts of these technologies for both governments and citizens in sub-Saharan Africa. For example: How should digital identification (ID) systems be designed to maximize benefits while minimizing costs in a specific context? When is it appropriate to link a social protection program to a digital ID system? To what extent can digital ID systems and digital payments reduce leakages and improve targeting of social protection programs? Can digital ID systems and digital payments assist in building incentive systems to motivate public servants? For more information on DigiFI Africa’s research agenda, please see our framing paper. 

We are excited to launch a further exploration of these questions in DigiFI Africa’s new blog series. Over two months, this series will unpack key policy questions on digital ID and payments systems while also exploring a subset of the academic literature provided in our framing paper. This series includes posts on: 

• The benefits, challenges, and unknown impacts of digital IDs,
• Digitization of government-to-person payments, 
• Mobile money and person-to-person payments, and
• Possible barriers to effective public service delivery (e.g. targeting, leakages, incentives, and take-up) and opportunities for digitization to improve these processes, including high frequency process monitoring.

You can read about DigiFI’s ongoing studies. These include, but are not limited to, research on the impacts of linking the national biometric ID system in Kenya to social protection schemes, the relationship between digital IDs and poverty alleviation in Malawi, how digital tax systems can aid revenue collection in Uganda, and the role of digital cash transfers in responding to the COVID-19 pandemic in Ghana and Kenya. 

If you’re a policymaker or researcher thinking about the design of a digital financial or ID system or evaluating new reforms in the DigiFI research agenda, we encourage you to get in touch! We can be found on [email protected] and would love to hear from you.

COVID-19 and the resulting economic recession have disproportionately affected already vulnerable individuals. Governments across the world have responded with an unprecedented expansion in their social protection programming. 

There is also a push towards digital transfers to be able to safely transfer cash while curbing the spread of the disease. While there is evidence on the efficacy of grants during non-pandemic times, we know little about the effect of transfers during large shocks. 

For example, during COVID-19, cash grants might prove inadequate in stimulating the supply of goods and services as it is a demand-side intervention. So, are cash transfers helpful in responding to the pandemic? More specifically, to what extent could a pre-existing universal basic income build resilience to future shocks?

J-PAL Africa’s Digital Identification and Finance Initiative (DigiFI) was launched in 2019 to fund innovative research to help answer questions like this. In 2020, we funded follow-up surveys to a 2017 study on universal basic income in Kenya to identify whether the program helped recipients adjust to shocks resulting from the COVID-19 pandemic.

The study: The effects of a universal basic income (UBI) in Kenya

In 2017, Abhijit Banerjee, Michael Faye, Alan Krueger, Paul Niehaus, and Tavneet Suri, in collaboration with Innovation for Poverty Action (IPA) and GiveDirectly, launched a randomized controlled trial in Kenya to test the effectiveness of a UBI in eradicating extreme poverty. 

A UBI is a specific form of social protection: an unconditional cash transfer large enough to meet basic needs and delivered to everyone within a community. One of the arguments put forth by advocates of UBI is that the transfers provide a form of insurance against uninsurable or unpredictable risks (read a more extensive discussion of the potential benefits of UBI here). 

Responding to the need for rigorous large-scale experimental evidence to support the debate, researchers designed and launched a field experiment evaluating the impacts of UBI in 2017 in Bomet and Siaya counties of Kenya. 

Participants in the study were randomly assigned into four groups, three of which received digital cash transfers with varying frequency and size: 

1. One group received a large lump-sum transfer at the beginning of 2018, 
2. One group was assigned a smaller long-term transfer scheduled to be received regularly for twelve years, 
3. One group was assigned a short-term transfer for two years that had largely been concluded prior to the COVID survey, and 
4. A final group was assigned not to receive any transfer and is the comparison group.

Assessing UBI in light of COVID-19

The COVID-19 pandemic unfortunately served as a unique opportunity to examine to what extent a UBI can mitigate the impact of a large shock. 

Against this backdrop, the researchers conducted (and will continue to conduct) phone surveys in May 2020 to measure the effects of the pandemic and how they were mitigated by a UBI. To assess how outcomes in the COVID survey have changed since the onset of the pandemic, they compared this to data from the standard endline survey of households conducted between August and December of 2019. 

It is unclear if the positive effects of social protection transfers found during “normal times” (Bastagli et al., 2019) will hold during the pandemic. Concerns around supply chain disruptions may limit the effectiveness of demand-side interventions such as a cash grant. In addition, when examining non-material outcomes such as depression, the negative psychological impacts may have been too great for the cash transfer to have had any meaningful effect. 

Results from this project provide some of the first evidence of the impacts of social protection programs during the pandemic. 

The UBI led to modest but statistically significant improvements in well-being

Reduced hunger and physical and mental illness

There were modest (though significant) impacts of the UBI on recipients’ personal well-being in all three groups that received cash transfers (lump-sum, long-term transfer, and short-term transfer). 

Roughly two-thirds of households (68 percent) that did not receive the UBI experienced hunger. Recipients of the UBI were 4.9–10.8 percentage points less likely to report experiencing hunger during the last thirty days prior to the COVID survey. 

Recipients of the UBI were also 8–13 percent less likely to have had a household member sick during the thirty days prior to the May 2020 survey. Given the very low prevalence of the coronavirus in the areas studied at the time of the surveys (twelve cases total), these illnesses were almost certainly not COVID-19 cases. 

Finally, recipients were significantly less depressed in the short-term and long-term arm, though not the lump-sum arm. 

No harmful impact of UBI on public health and some evidence that it was helpful

Recipients of the UBI were significantly less (10–16 percent) likely to seek medical attention in the last thirty days, as they were less sick. This would have resulted in a decreased burden on hospitals and could have been helpful in freeing up public health system capacity which is vital during the pandemic. 

In addition, there is some evidence that the UBI reduced social interactions—such as visiting friends or relatives—which could reduce COVID-19 infections in the future. 

Mixed effects on resilience to large aggregate shocks

Regular and long-term transfers are likely to affect investment decisions and hence resilience to large shocks. On one hand, a long-term transfer could allow the individual to buy assets to buffer against any income shocks. On the other hand, the long-term transfer may encourage recipients to increase their exposure to riskier choices, such as starting a business. 

It is particularly interesting to compare the impacts of the UBI before and during COVID times to see how beneficiaries coped with the pandemic. An important caveat is that these results could be driven by seasonality and not the pandemic. As the researchers learn more, we will update the results. 

• No effect on incomes: Long-term, regular transfers led to an increase in risk-taking commercial activities. Pre-COVID, beneficiaries had diversified their income streams by creating new non-agricultural enterprises, which resulted in an increase in profits without substantial changes in labor or agricultural earnings. 

  However, when the pandemic hit these enterprises were not spared and the income gains they witnessed pre-COVID were wiped out but they did not close down.

• Higher resilience to hunger in the agriculture lean season: The UBI provides households with a safety net and hence a reduced risk of hunger. This allows them to take on more income risk. As discussed above, beneficiaries of the UBI were more likely to not go hungry during the pandemic. 

  This was also seen prior to the pandemic during the lean season (the season between planting and harvesting), which saw a significant reduction in the number of households who experienced hunger in the UBI transfer group. 

  However, during the harvest season (August–December 2019), there were no significant differences in hunger between the groups that did and did not receive the transfers. 

UBI: Unlikely tool of choice to respond to unanticipated shocks

While access to a UBI prior to and during a large shock improves well-being in the wake of said shock, it is an unlikely choice of tool to respond in such contexts. This is because the income gains from the UBI pre-pandemic are wiped out during the pandemic. This fits well with the theory that a UBI encourages risk-taking by providing a safety net. Hence, it is likely that recipients of a UBI have increased exposure to shocks. 

Further, it would likely be more effective to use a targeted response tailored to protect those worst affected by a crisis. This is by no means a criticism of UBI but rather a remark on its appropriateness in different contexts. 

Another key message from these results is the importance of supplemental income during large shocks such as the COVID-19 pandemic. These results highlight the need for infrastructure that allows cash grants to be provided universally or to a large proportion of the population in response to unanticipated crises.

Learn more about DigiFI Africa.

Read the full paper.