For the last 15 years, I have been focused on (some may say obsessed with) the delivery of basic services such as health and education. In the last two years, as a member of the Pathways for Prosperity Commission, I’ve been interested in how technology can promote inclusive growth. Imagine my delight therefore when the Commission issued a report on technology and service delivery!
The report, “Positive disruption: health and education in a digital age,” starts with the right questions: What are the reasons why health and education services are failing in many developing countries, and can technology alleviate those problems? The first question is similar to the one asked in the 2004 World Development Report, “Making Services Work for Poor People,” which concluded that the reason was a failure of accountability—of service providers to policymakers, and of policymakers or politicians to citizens (Figure 1).
Figure 1: The framework of accountability relationships
Source: World Development Report 2004
The framework in “Positive disruption” has the same actors (with the addition of intermediate organizations), although the relationships are two-way (Figure 2).
Figure 2: A simplified system framework for health and education
Source: Pathways for Prosperity Commission
Both frameworks lead to the conclusion that interventions that improve only one link in the chain are insufficient to improve service delivery. For instance, a program to introduce time-stamp machines to monitor nurses’ presence (and thereby reduce absenteeism) in Udaipur, India led to the machines’ being vandalized by the nurses and no change in absence rates. “Positive disruption” documents several examples of technologies, such as the One Laptop Per Child initiative in Peru, that had little effect on children’s math and reading scores, partly because it didn’t relieve other constraints in the education system, such as teachers’ knowledge.
But the report goes further to show how technology can play a role in strengthening the system. For instance, Mindspark—digital personalized learning software that facilitates tailoring learning to each student’s level—showed 38 percent improvement in math scores after just four and a half months in a pilot, after-school program in Delhi. Digital technologies can capture and share data to make administrative systems work better. The data can be used to provide incentives for better performance (digital feedback to health workers in Mali led to a 10 percent increase in the number of houses visited); fill knowledge gaps among frontline workers (such as online courses in HIV for Zambian clinicians); and induce better uptake of services by citizens (SMS reminders to Kenyan patients to take their medicines).
“Positive disruption” is appropriately cautious about the limits, and possible unintended consequences, of using digital technology for improving service delivery. The monitoring of teachers and nurses may reduce absenteeism initially, but the effects sometimes dissipate over time, as in the example from Udaipur above, or on teachers in India more generally. And the sharing of data, which makes the administrative systems operate more efficiently, involves a host of privacy issues especially when it comes to individuals’ health data. Unless the appropriate norms and rules of data privacy, protection, and governance are in place, these innovations could backfire or elicit a backlash from the population (or both).
Despite its hopeful, wide-ranging, and balanced treatment, “Positive disruption” missed an opportunity to highlight the role of technology in strengthening another link in the service-delivery chain: that between citizens and politicians. However much we help politicians hold service providers accountable (through digital monitoring, incentives, etc.), if the politician does not find it in his or her interest to improve service delivery, they will not implement the programs. And if people do not vote along service delivery lines, and vote instead along caste, ethnicity, or religious lines for example, then there is no incentive for the politician to hold teachers and doctors accountable for performance. Furthermore, if teachers are part of a patronage system where no-show jobs are part of the currency, then the incentives for reform are even weaker.
One reason why people, especially poor people, don’t vote along service delivery lines is that they may not know enough about the politician’s performance. Digital technology can play a role in getting this information out to poor people. The information has to be accessible and salient. The recent effort in India to produce data on child health by political constituency, and publish it ahead of the elections, is a step in that direction. Sometimes, politicians find it more effective to promise their constituents private goods, including cash, rather than public goods such as health and education, if they vote for them. There is evidence that “vote-buying” leads to poorer health and education outcomes. But even if they have received cash, people have a choice when they enter the voting booth—if it is genuinely a secret ballot. Electronic voting—another technological development—can reinforce the anonymity of the ballot. As Thomas Fujiwara shows, the Brazilian districts that replaced paper ballots with electronic voting had better public health outcomes.
In short, while “Positive disruption” presents a realistic, well-grounded, and yet ambitious vision for harnessing technology to improve service delivery, I think it should be even more ambitious, and explore ways of using digital technology to strengthen the weakest link in the service delivery chain, namely the ability of citizens, especially poor citizens, to hold politicians accountable.
Africa is the world's breadbasket—or should be. It has vast arable land, grows a wide variety of crops and has vast irrigation potential with seven major rivers. Yet, Africa imported $43 billion worth of food items in 2019. Digital technologies ... are eliminating the traditional inefficiencies of smallholder food production and helping to close the yield gap.