The returns to pre-primary education in developing countries

Despite critiques of the 'Heckman curve' (see here), it is reasonable to believe that the returns to education are not only positive, but somewhat higher for the first years of education than for later years. That would be a simple application of what economists refer to as diminishing marginal returns (to education). So, I was interested to read this post on the Development Impact blog by Lelys Dinarte-Diaz and Alaka Holla, which summarises the results of a meta-analysis of the benefit-cost ratio for pre-primary education in developing countries. They write that:

Using robust-variance meta-analysis to aggregate study-level estimated effects, we find significant average effect sizes of preschool interventions on outcomes measured during preschool – around 0.15 standard deviations for cognitive skills and around 0.12 standard deviations for social-emotional skills and behavior.  

Next, to see whether these estimated benefits are worth their cost, we then take our sample of studies from low- and middle-income countries and restrict our attention to those that provided sufficient information for us to back out the cost per child of the pre-primary intervention. To convert these average effects into a monetary value, we looked to a few longitudinal studies that could estimate how much earnings in adulthood increase when cognitive skills increase during the early childhood period. With this conversion factor, along with other empirically informed assumptions, our most conservative calculations suggest benefit-to-cost ratios ranging from 1.7 to 14.2. That is, in all cases where we had data, benefits outweighed the costs. Returns were certainly positive.

You can find the working paper by Holla et al. here. The results are based on 50 studies conducted across 19 developing countries. The benefit-cost ratios certainly suggest that there is a lot of value in pre-primary education. This in itself doesn't tell us anything about the Heckman curve, because we would also need to know corresponding benefit-cost ratios for higher levels of education.

In their blog post, Dinarte-Diaz and Holla go on to discuss why there is under-investment in pre-primary education in developing countries:

Despite these high returns, it isn’t too difficult to understand why there is under-investment in a service like preschool. Though a government would have to front the costs today of expanding preschool or improving its quality, it is only in the future that the government would realize most of its monetized return once children enter the labor market as adults. That is, the returns could lag the costs by about 15-20 years. With political cycles and attention spans much shorter than this, getting preschool on any agenda would be a challenge.

As I discuss in my ECONS102 class, credit constraints are a serious impediment to investment in education for poorer households. The issue is that poorer households find it more difficult than richer households to save in order to pay for education costs, and more difficult (or impossible) to borrow to pay for education costs. And, if a poorer household can borrow, they tend to pay higher interest costs - because they are higher risk to lend to, lenders will require a higher interest rate in order to compensate for the risk.

These points about credit constraints extend to poorer countries in a very similar way. Dinarte-Diaz and Holla offer bonds as a solution, noting that:

The World Bank’s Treasury also raises funds to support the Bank’s work through Sustainable Development Bonds - $43 billion worth in 20 different currencies last year – which have had a AAA rating since 1959. They also have cause-specific bonds such as green bonds and bonds focused on women’s empowerment and on health and well-being.  

Why shouldn’t we add a preschool bond to the list?

This makes sense. In many developed countries, we deal with household credit constraints in education either by the government paying the costs of education (as is the case for primary and secondary education), or through student loans (as is often the case for tertiary education). The proposal by Dinarte-Diaz and Holla is a good suggestion, but with a potentially serious flaw. If it is the World Bank that is issuing and paying back the bond, then the suggestion is good. The World Bank is seen as being relatively low risk to lend to. However, the question becomes whether investment in pre-primary education is the best use of funds offered to the World Bank.

However, if Dinarte-Diaz and Holla are suggesting that developing countries themselves raise the money for pre-primary education through issuing bonds, then I think that runs into the same problems that poorer households do - bonds issued by developing countries tend to have high interest rates, to compensate for the risk of default. And issuing bonds for pre-primary education would compete with all of the other things that a country may wish to issue bonds for.

The takeaway message for me is that pre-primary education is very much worth the investment, but the real challenge is how that investment would be funded.