Sovereign debt crises produce dramatic effects – both for the countries directly involved and for other economies via various contagion mechanisms. Proposals to deal with such problems have been advanced. Examples include the Sovereign Debt Restructuring Mechanism (Krueger, 2002), and creation of an International Lender of Last Resort (Fisher, 1999). But these are mechanisms to deal with the default after it occurs. The best possible solution would somehow prevent the problem arising.
If emerging economies’ debt could be made contingent on the variables that influence government decisions on default, the problem would be solved. However, to a large extent, that just can’t be done.
For example, sovereign default is influenced by the political attitude of the current government towards integration into the world economy. Therefore, elections of new politicians affect the amount of debt payments a country is willing to make. If the amount to be paid were contingent on such political variables, we would observe fluctuations in debt payments rather than defaults – something more like an equity contract. The payments would fluctuate in accordance to the willingness/ability of the country to honour its debt and there would never be a breach of contract. Instead of infrequent but dramatic crises with long and costly renegotiation, we would observe frequent and smooth changes in the value of debt and debt payments.
The problem is that it is basically impossible to write such contracts. Two reasons stand in the way. First, the contact would have to specify quantitative measures of such political variables, but these (as well as many other factors that influence sovereign default) cannot be measured. Payments cannot be contingent on things that cannot be objectively determined. Second, even if they could be measured or proxied, the classic problems of adverse selection and moral hazard would rule out a market for such debt. For example, if a a particular political attitude led to lower debt payments, all politicians would have the incentive to act accordingly. Anticipating this, lenders would shun such contracts.
But not every form of contingency is subject to this problem. It is possible to write contingent contracts, but they need to be based on variables that are not subject to the problems of moral hazard and adverse selection. The trick is to find variables that strongly influence the ability or willingness of countries to pay their debts, but which are beyond the control of the borrowers.
For example, some pundits have been arguing for GDP-indexed contracts (Borenztein and Mauro, 2004). However, it is not clear that fluctuations in GDP are an important source of default risk. Indeed, empirical evidence shows that the relation between recessions and default episodes is weak (Tomz and Wright, 2006). There are also practical problems related to GDP-indexed debt. Measuring GDP is not a simple task; such figures are usually produced by national governments and even in the best of cases, they are subject to statistical errors and corrections that are difficult for outsiders to evaluate. In short, GDP figures may be unreliable and subject to manipulation.
Debt and world real interest rates
The interest rate is another candidate – one with superior features. First, real interest rates in the rich nations have strong effects on debt repayments. Emerging countries are always borrowing in order to pay their debts – that is one of the reasons why changes in the international financial environment have such a strong impact on those economies. The rate at which they borrow is directly influenced by the opportunity cost of the investors (world real interest rates). Thus, a rise in the US real interest rates leads American investors to require similar increases in expected payments from investments in emerging economies, which drives up the cost faced by emerging nations in refinancing and thus servicing their debt.
To understand the effect of world real interest rates on debt sustainability, consider a country that finds it optimal to keep honouring its debt contracts provided payment is not above $1 million a year. In a world without any uncertainty, with real interest rates at 1% a year, that country would be able to service debt as high as $100 million – without any uncertainty, there would be no default premium, as the country would always service its debt. However, with interest rates at 2% a year, any debt higher than $50 million would imply debt payments higher than $1 million a year, which would be unsustainable.
It is exactly this effect that kicks in following interest rate hikes in developed economies. Following the above numerical example, in the extreme case of a permanent and totally unexpected interest rate increase from 1% to 2%, the sustainable level of debt decreases by 50%. The example is stylised but illustrates how large and long-lasting interest rate hikes in the developed world affect the sustainable level of debt in emerging economies. In fact, the interest rates faced by emerging countries increase even more than world real interest rates. The possibility of default acts as a multiplier on US interest rate hikes. The probability of default rises with the cost of servicing debt, so the premium on borrowing also tends to increase. All this makes it even more difficult for emerging economies to honour their debt obligations.
This is the story of the 1980s debt crises. The dramatic increases in US rates at the beginning of the 1980s (from about 0% in the 1970s to around 4% in the 1980s) raised the cost of servicing debt, and most Latin American countries ended up in a debt crises – despite their differences in policies and distinct economic conditions. Foreign interest rates matter a lot for emerging economies’ debt sustainability. A decade after the crisis, the so-called Brady agreements allowed the main Latin American countries to write off around 30% of their debts.
Debt contingent on world real interest rates
A solution would be to link the debt payments negatively to US real interest rates; fluctuations in those rates would thus have no effect on the probability of emerging economy default. With such contingent contracts, if US real interest rates were higher than initially expected, debt payments would be smaller, so the country would need to borrow a smaller amount to honour its obligations. Conversely, if borrowing was cheaper than expected, payments would be higher. Lenders would use their expectations of future US interest rate fluctuations to predict the repayment streams and discount this into the price of the debt.
Such contingent contracts would remove (or minimise) one important source of default risk. That would be beneficial not only for preventing crises due to interest rate hikes but also for reducing the volatility of emerging economies’ default risk and, therefore, decreasing the uncertainties they face in international capital markets.
It is important that debt payments are contingent on real and not nominal rates, as making debt payments a function of inflation in the US only adds extra noise and, therefore, extra risk to emerging economies’ debt.
Allowing for contingencies on world real interest rates is simpler than conditioning payments on domestic GDP. Obtaining good estimates for expected real interest rates may not be trivial, but it is simpler than measuring GDP. Moreover, there are no questions about misreporting of data, as all the necessary information comes from easily-observable financial variables.
A final plus is that under such contingent contracts, payments to investors would be higher when real interest rates were lower – which usually coincides with worse economic conditions in the developed nations, so for the lenders, those contracts would have the desirable features of an insurance.
Real interest rates have been low across the world, and major real interest rate increases are not expected to occur in the short run. Alas, the same could have been said in 1978. Conditioning debt payments on world real interest rates would not remove the risk of emerging markets debt default – shocks to many other important variables would still threaten debt sustainability – but it would help. It is only a partial solution – but a good one.
Borensztein, Eduardo and Mauro, Paulo, 2004, "The case for GDP-indexed bonds", Economic Policy 2004, 165-216.
Fischer, S., 1999, “On the Need for an International Lender of Last Resort”, Journal of Economic Perspectives 13:4, 85-104.
Krueger, Anne, 2002, "A New Approach to Sovereign Debt Restructuring", IMF.
Tomz, Michael and Wright, Mark, 2006, "Do countries default in ‘bad times’?", forthcoming, Journal of the European Economic Association.
For further details, see “Optimal External Debt and Default” by Bernardo Guimaraes, CEPR Discussion Paper 6035.