From time to time I come across development practitioners who believe that it’s a bad idea for poor countries to borrow money, particularly non-concessional loans, to finance health or education projects. Moreover, recent research by the Human Development Practice at the World Bank shows that as countries graduate from softer IDA financing to the somewhat more expensive terms of IBRD loans, there is a disproportionate decline in the share of social sectors in the programs the World Bank supports in those countries. So, it seems that not only do some analysts believe that borrowing for social sectors (especially on less concessional terms) is a bad idea, country policymakers also appear to reflect this in their actual borrowing behavior.
What is the rationale for this belief? Does it make economic sense? In this blog, I set out why I think decisions to borrow for the social sectors should be made in the same way as decisions to borrow for so called ‘hard sectors’ like infrastructure or industry. I also invite alternative views to see where my own thinking has missed some important consideration, or is just plain wrong!
I’ve been given two kinds of arguments in support of not borrowing for social sector projects. The first is about their ability to repay the borrowing by generating enough foreign exchange. And the second is skepticism about the productivity of government spending in these areas. Let me take each in turn.
Ability to pay
The repayment capacity argument essentially says that health and education have a tenuous and very long-term impact on growth and exports and any foreign loans invested in these sectors would have to be repaid from ‘productive investments’ in other sectors. There are two problems with this argument. First, there is now a huge body of research and evidence that without a healthy and educated population countries cannot compete in today’s global marketplace, and that this is going to be even more true in the technology-driven world of tomorrow. This thinking is reflected, among other initiatives, in the recent Human Capital project of the World Bank.
More fundamentally, the idea of linking repayments of loans to the individual investment projects they finance ignores the concept of fungibility and the need to focus on debt sustainability at the level of the country, not simply the project. Of course, countries should carefully assess how they will repay the international loans they are taking on, based on realistic projections of growth, exports, fiscal and external imbalances, and allowing for unanticipated shocks. They should also be realistic about the time it takes for many investments to yield economic, financial, and social results.
And before embarking on any investment project, whether financed from borrowing or national resources, it is important to assure oneself that financing will be available to meet the longer term recurrent costs associated with the investment (it is legitimate to note here that it is harder to stop paying teachers salaries or health care recurrent costs than to let an infrastructure project fall into disrepair because of inadequate maintenance). Nevertheless, calculations of debt sustainability are best done on the volume and terms of overall borrowing. It matters little as to which particular loan finances which specific project. More generally, the concept of fungibility means that assigning specific bits of funding to specific projects is at best a convenience and has little economic significance.
Let’s illustrate with a simplified hypothetical exercise. Suppose a country has an investment budget financed by a billion dollars of its own savings and a hundred million dollars borrowed internationally at 10 percent. Let’s also suppose that the money is invested equally between social sector projects and an offshore energy project whose output of $50 million per year is exported entirely. Let’s further accept for the moment (the flawed argument) that the social sector spending contributes nothing to the running of the offshore project. Now, this country as a whole is comfortably able to service its debt from its export earnings but this is completely unaffected by whether the loan was taken for the project itself or for some other $100 million project in the social sectors. I accept that this is a simplified illustration and one can make the case that for very large projects, the decision on financing and debt sustainability cannot be delinked from the performance of the project being financed. However, more often than not, these cases concern large infrastructure projects – not investments in health or education.
Productivity of government spending
The other line of reasoning, often from finance ministry officials and their counterparts, is a general skepticism about waste and inefficiency in government spending for (particularly) education. No doubt there are many examples of such waste, although it’s only fair to point out that spectacular examples of corruption and waste can also be found in large infrastructure projects across the globe.
A related concern is that government spending on (again primarily) education doesn’t show compelling results in terms of learning outcomes. Before borrowing more money for expanding education programs, the argument goes, governments need to focus on improving the productivity of the existing system. How productive it is to invest in education systems as they are, and how to link expanding educational opportunities with improving learning outcomes, is the subject of much ongoing international debate, including as part of the research being done by colleagues at CGD. Indeed, it seems a priority to me to find more effective ways of sharing international examples of what works in delivering better learning and under what conditions. There is also value in understanding how technology can be effectively and affordably deployed at scale to improve learning outcomes. But these questions relate to how – and how much – governments should spend on education not on how that spending is financed. Whether you continue to finance the social sector spending from taxpayer money or borrowing doesn’t seem to me to enter the equation. Indeed, turning down an international education project loan on these grounds could result in foregoing the technical support and international experience that might have led to some marginal improvement in the performance of the sector.
The bottom line for me is that it’s a false distinction to borrow internationally for ‘hard sectors’ but not for education or health or other social sector investments. In both cases, the investments need to be the right ones, well prepared, and the debt burden should be sustainable at the aggregate country level, but the choice of sectors should not be the determining factor in the decision to borrow.
I look forward to comments and reactions and plan to summarize them in the comments below.
I am grateful to Amanda Glassman, Charles Kenny, Justin Sandefur, Sanjeev Gupta, and Liesbet Steer for valuable comments on this blog, and to Kelsey Ross for research support.