Last week, the Center for Global Development hosted a presentation and conversation about new evidence on the long-term earnings gains resulting from improvements in early grade literacy and numeracy. We heard Lee Crawfurd present new results showing sizeable earnings gains in Indonesia, Jishnu Das demonstrate striking gains in Pakistan, and perspectives from Jo Bourne of the Global Partnership for Education and Jennifer Swift-Morgan of the Prevail Fund. It was an informative, engaging, and lively hour. Here’s a quick rundown.
Why is this so important?
If there’s one thing it seems hard to argue with, it’s the value of making sure that kids can read and do arithmetic. Many education systems don’t manage this: recent surveys in rural India and in Uganda showed one in four and one in ten third-grade students can read a second-grade text. Those countries are not unique. But policymakers don’t seem convinced: they consistently overestimate how well children in their countries master these basic skills. Budgets likewise haven’t shifted: a recent analysis of education projects from a multilateral development bank showed no increased focus on quality or learning in recent years. Strikingly, that same analysis showed a boost in early childhood projects, potentially because of compelling evidence for the long-term economic returns to investments in early childhood. So, do investments in literacy and numeracy in the first years of primary school yield long-term gains in earnings?
Lee Crawfurd helped answer this question with recent evidence from Indonesia. While studies on the impact of schooling on earnings are plentiful, studies on foundational literacy and numeracy are scarce. Drawing on a dataset that followed the same children from ages 7-12 (when they were tested for reading and math skills) through their late twenties (once most were out of school), Crawfurd showed that children who achieved early literacy and numeracy had at least 4 percent higher earnings in adulthood. (The number is even higher under certain assumptions.)
Jishnu Das went on to show even larger gains in villages in Pakistan. A randomly assigned intervention—specifically, providing report cards on the quality of schools—led to gains in test scores, between 0.1 and 0.15 standard deviations. (If you’re wondering how big that is, it’s just above the median if you look at the effect of many, many education interventions.) What kind of impact did that have 15 years later? After taking all the costs into account, wages rose enough to yield a rate of return each year of more than 20 percent! The government makes more on taxes from the increased earnings than it takes to run the program in the first place. Suggestive evidence from correlations in other countries (India, Peru, Ethiopia) suggests that if anything, wages may rise even more in those settings. (These results are all forthcoming!)
What are some of the implications of these findings for donors?
Jo Bourne highlighted that nearly half of the countries working with the Global Partnership for Education are signaling that foundational learning is a key priority. This is an opportunity for national governments and international partners to work together, as governments lead reforms at scale. Providing those governments with evidence on the earnings gains helps education policymakers in their national planning discussions. But we can’t measure these long-term returns without robust data systems in place. Jennifer Swift-Morgan highlighted that aid for basic education is dropping right now, even as the evidence we’re seeing from Crawfurd and Das shows us it should be going way up. Donors can focus on what works and stick around until an education system has reached a "new normal": if a child goes to school, they come out knowing how to read and do math.
What questions remain?
We need better evidence on how to reinforce early gains, carrying improvements throughout a child’s educational career, as well as an understanding of what a healthy education ecosystem requires: for example, a violent school environment is bad for learning. We need to look at equity: which children are being left behind even as schools improve? We also see wide variation in the quality of programming for foundational literacy and numeracy: we need to figure out how to get the design and the implementation right for consistent success.
A point that came up repeatedly is that if we’re seeing sizeable earnings gains from these modest early grade learning improvements, what could we see with big improvements? Das showed that, in his sample from Pakistan, a 0.15 standard deviation improvement translates to going from 17.4 percent of children being able to read with comprehension to 18.7 percent. We should be setting our sights much, much higher. To paraphrase Swift-Morgan, “How much good could come from foundational literacy and numeracy programs that are actually good?”
Each of these panelists shared much more than I can do justice to here. I recommend you watch the full event here. If you want to jump to a particular spot, here’s your cheat sheet with direct links:
