Borrowers normally seek to minimize the variability of debt service obligations and the overall cost of borrowing. In making a decision they may decide to hedge. This implies choosing currencies which will insulate the economy from currency volatility and changes in terms of trade because their real values rise and fall with the borrower's real income. Or they may decide to speculate on expectations of changes in exchange rates and interest rates. For any given loan, the "hedging" component (choosing a currency to minimize the expected cost of borrowing) would normally be adjusted by considering the speculative component (expected exchange rate changes). However, predicting movements in exchange rates and interest rates is exceptionally difficult. In addition, the factors on which a decision depends - such as the relation between the borrower's real income and the exchange rate of another currency - may be very volatile, so that past figures may be misleading. One way of mitigating these difficulties is to base the currency composition of a country's debt on its pattern of trade. A country then borrows in the currencies it earns from its exports while maintaining its reserves in the currencies in which it makes its imports. Since the borrower's terms of trade are likely to improve if the currency of an export market appreciates, the higher costs of servicing debt in that currency will be at least partially offset. However, even this strategy may be deceptively simple and practical. It is not necessarily consistent with the hedging and speculative strategies and might lead to inappropriate borrowing. There seems to be no foolproof alternative to making pragmatic judgements on variations in trade and in interest and exchange rates.
Some countries have informal rules on the currency composition of their debt and may, for example, insist on dollar-denominated loans, despite much lower interest rates on other currencies, on the grounds that the dollar is likely to depreciate over time. More explicit guidelines have been adopted by some industrial countries. They may, like Sweden, diversify their debt portfolios by increasing the number of currencies in which they borrow. Or they may seek to minimize exchange risk by borrowing in currencies linked to their own. At one time Ireland had almost half of its debt in deutsche marks since this is the dominant currency in the European Monetary System, to which Ireland belongs. In the early 1980s, by far the major proportion of developing country borrowing was in US dollars, whether because such loans were more readily available than those in other currencies or because such were actively chosen by debt managers. The appreciation of the dollar at that time often harmed the borrowing country. By contrast, some developing countries benefited from their non-dollar currency loans at that time - some loans made by the World Bank between 1978 and 1982 subsequently had effective interest rates of less than 1 percent a year when expressed in dollars.