Financial Exchange Rates and International Currency Exposures

This paper builds a database of international currency exposures for a large panel of countries over 1990-2004 and uses it to construct financially-weighted exchange rate indices, showing that trade-weighted indices are an inadequate guide to the balance-sheet consequences of currency movements and that most countries held short foreign-currency positions in the mid-1990s before substantially reducing them over the following decade. The method is to estimate, asset class by asset class, the currency composition of each country’s foreign assets and liabilities – combining BIS international banking and securities statistics, the IMF’s Coordinated Portfolio Investment Survey, UNCTAD bilateral FDI data, World Bank Global Development Finance, national and central bank sources, and the Lane-Milesi-Ferretti External Wealth of Nations dataset – and then to weight the asset classes by their shares in the international balance sheet. Three findings follow. First, financially-weighted and trade-weighted exchange rates move quite differently: the mean and median within-country correlation between the net financial index and the trade index is negative in the full sample and in the developing sample, and even for industrial countries, where it is positive, it averages 0.41 with a median of 0.70 against pairwise correlations above 0.85 between any other pair of indices (Table 1, Section 5.1.1). Second, in 1994, 70 percent of countries had a net negative position in foreign currencies with an average exposure weight of minus 27 percent, and over 20 percent were below minus 50 percent – but by 2004 the mean and median had moved to minus 7 percent, only about 10 percent of countries remained at minus 50 percent or worse, and 86 percent of industrial countries had positive exposure. The decomposition attributes this shift to improving net foreign asset positions and a move toward equity and FDI liabilities (which are denominated in local currency) plus reserve accumulation, rather than to greater domestic-currency denomination of international debt: beyond the euro area there is “effectively no change in the foreign-currency share of debt liabilities,” and for the top quartile of improvers over 50 percent of the increase in total assets came from reserves (Tables 5-6, Section 5.2.2). Third, exchange rate valuation shocks are large and not quickly reversed: the 75th percentile of the absolute currency valuation effect is 2.8 percent of GDP for advanced countries, 3.8 percent for emerging and 5.3 percent for developing countries, and regressing the total valuation term on the currency valuation term gives roughly one-for-one pass-through with R-squared of 0.4 to 0.6 for developing and emerging countries, against a coefficient near 0.6 and R-squared of only 0.06 to 0.09 for advanced countries, whose larger equity positions leave more room for price-driven valuation effects (Tables 9-10, Section 5.3). The authors are explicit about scope. The analysis is “partial equilibrium in nature, since we effectively treat exchange rate movements as exogenous.” The currency positions are estimated, not observed: “we have made many assumptions in constructing our estimated international currency exposures,” and in some cases missing data are imputed by modelling the relation between country characteristics and international financial holdings, so “estimated data will not be perfectly accurate, nor will every assumption made fit every country perfectly.” Cross-border hedging via derivatives is unobserved, though the paper argues the omission is modest, citing an estimate that only about 10 percent of foreign equity positions are hedged and noting that hedging between two domestic residents leaves the country’s aggregate exposure unchanged.

Jay C Shambaugh | Macro Paper Warehouse

Financial Exchange Rates and International Currency Exposures