آشنایی با نوسانات نرخ ارز دو جانبه

This paper develops an empirical model of bilateral exchange rate volatility. We conjecture that for developing economies, external financial liabilities have an important effect on desired bilateral exchange rate volatility, above and beyond the standard optimal currency area (OCA) factors. By contrast, industrial countries do not face the same set of constraints in international financial markets. In our theoretical model, external debt tightens financial constraints and reduces the efficiency of the exchange rate in responding to external shocks. We go on to explore the determinants of bilateral exchange rate volatility in a broad cross section of countries. For developing economies, bilateral exchange rate volatility (relative to creditor countries) is strongly negatively affected by the stock of external debt. For industrial countries however, OCA variables appear more important and external debt is generally not significant in explaining bilateral exchange rate volatility.

What are the principal determinants of exchange rate volatility? This has been perhaps the biggest research question in international finance over the last three decades. Despite hundreds of follow-up papers, the results of Meese and Rogoff (1983) suggesting that movements in exchange rates are largely unpredictable remain largely intact. Our paper also focuses on exchange rate volatility. But we take an alternative perspective to the literature that has directly followed in the tradition of Meese and Rogoff. Rather than focusing exclusively on the time series properties of exchange rates relative to a single large currency such as the dollar or the euro, we are concerned with understanding what drives bilateral exchange rate volatility across countries. Looking at a large cross section of both developing and developed countries, we seek to identify the main determinants of bilateral exchange rate volatility between country pairs. Our starting point is the optimal currency area (OCA) hypothesis of Mundell (1961). Mundell isolated the key economic factors that make two regions or countries part of a common currency area. These factors include trade interdependence and the degree of commonality in economic shocks.1 As in previous work (e.g. Bayoumi and Eichengreen, 1998, Hausmann et al., 2001 and Larrain and Tavares, 2000), we use these as explanatory variables in modeling bilateral exchange rate variability across countries. But in addition to the standard set of OCA variables, we add a further set of determinants measuring financial linkages between countries. Recent theoretical literature suggests that these variables may be of key importance in understanding exchange rate variability, especially for developing economies. Our central hypothesis is that for developing countries, high levels of financial linkages with a creditor country C (in the form of portfolio debt or bank loans) will, ceteris paribus, be associated with a lower level of bilateral exchange rate variability vis-à-vis country C. This hypothesis is derived from a substantial recent body of work that points to the importance of financial factors in understanding exchange rates in emerging market economies. Many writers have questioned the neglect of financial market structure in standard macroeconomic models. Bernanke et al. (1999) stress the importance of balance sheet effects in understanding the properties of business cycles. Among others, Krugman (1999), Cook (2000), Aghion et al. (2001), Cespedes et al., 2000 and Cespedes et al., 2001, Devereux and Lane (2001), Gertler et al. (2001) and Eichengreen (2002), have extended these ideas to the open economy. All these papers highlight a fundamental failure of the ‘Modigliani–Miller’ theorem: balance sheet effects matter for macroeconomic outcomes and especially for the exchange rate. One conclusion of this literature is that, combined with these balance sheet effects, the presence of external debt (denominated in foreign currency) may have an important effect on the way in which movements in the exchange rate impact on an economy. Fluctuations in exchange rates, in the presence of large stocks of un-hedged foreign-currency denominated debt may be important through its effects on the financial sector and corporate balance sheets. This introduces a cost of exchange rate variability quite separate from the traditional theory. Eichengreen and Hausmann (1999) suggest that many emerging market economies may have little ability to tolerate a high degree of exchange rate volatility against their major creditors. The observation that many countries, especially emerging market economies, display a ‘fear of floating’ (Calvo and Reinhart, 2002) offers supporting evidence for the hypothesis that OCA factors alone cannot provide a full account of the degree of bilateral exchange rate volatility that emerging market economies experience.2 Accordingly, the central empirical hypothesis of the paper is that in addition to the standard OCA factors, bilateral exchange rate volatility is related to the stock of bilateral financial claims across countries. For the ‘rich’ countries that are not constrained in international capital markets and can freely borrow by issuing assets denominated in their own currencies, it is unlikely that international balance sheet considerations have a large impact on their choice of exchange rate regime. But for developing economies that are subject to various borrowing constraints and must issue debt in foreign currency, exchange rate volatility may have an extra cost, beyond that suggested by the standard OCA criteria. Accordingly, we test the hypothesis that bilateral exchange rate volatility is especially negatively related to bilateral financial claims for developing economies. The paper begins by developing a simple model of exchange rate choice for a small open economy vulnerable to external terms of trade disturbances. Exchange rate policy matters due to nominal price stickiness. Our model allows for the exploration of two separate cases. For an economy free of credit constraints, we show that exchange rate adjustment is desirable, and the exchange rate should respond to external shocks according to OCA theory. We then show that the presence of credit constraints, in combination with external debt, leads to a significant decline in the optimal response of exchange rates to shocks. If the credit constraints are significant enough, it may be optimal to have essentially no adjustment of the exchange rate in response to external shocks. In our empirical estimates we examine the determinants of bilateral exchange rate volatility in a broad cross section of countries, using a number of standard OCA variables that have been employed in the literature, such as trade interdependence, differences in economic shocks, and country size. We then add a series of financial variables. One represents internal finance, capturing the degree of financial depth within countries. A second set of variables measures external financial factors. One of these comes from banking data (obtained from the BIS), and represents exclusively creditor-currency denominated loans, so it captures the importance of foreign currency liabilities. The second measure comes from the IMF’s International Portfolio Survey, and represents bilateral portfolio debt liabilities between countries. Our empirical results find that financial variables do play a significant role in explaining exchange rate volatility, in addition to the standard OCA set of variables. For the most part, the results indicate that the effect of OCA variables on exchange rate is consistent with standard theory. Greater bilateral trade reduces bilateral exchange rate volatility, and economic size increases volatility. This holds both for developed and developing countries. For the full sample, bilateral exchange rate volatility is reduced by both internal finance and by external financial linkages. But the results are sharply different for developed economies and the developing country sample. For developed economies, bilateral exchange rate volatility is either positively affected by external financial linkages, or affected insignificantly. By contrast, for the developing country sample, bilateral exchange rate volatility is significantly reduced by external financial linkages. Thus, the dichotomy between developed and developing economies suggested by the model is supported by our empirical results. In fact, we see the same dichotomy with respect to internal finance; this variable tends to increase exchange rate volatility for developed economies, but reduces it for the developing economies. The paper is organized as follows. The next section develops a simple model of the optimal exchange rate policy for a small economy, following in the tradition of the recent ‘new open economy macroeconomics’ literature. Section 3 outlines the data and the empirical strategy used in the paper. Section 4 discusses the empirical results. Some conclusions then follow.

Rather than repeating our results, here we emphasize some of the outstanding questions raised by our analysis. We have stressed some of the limitations of the empirical results arising from the lack of data. For instance, as emphasized in 8 and 10, we must attribute at least part of the observed bilateral exchange rate volatility to policy decisions. In addition, we must remain concerned about endogeneity in explanatory variables. In future work, it would be useful to further investigate the link between international financial linkages and exchange rate behavior. In this regard, examining the role of finance in determining exchange rate regime decisions (de jure and de facto) and considering alternative measures of volatility would be interesting. Data permitting, it would also be desirable to examine volatility over longer time spans, and to find better instruments for bilateral financial trade in order to minimize endogeneity concerns. Related to this point, the forthcoming publication of the second IMF Portfolio Survey will provide new financial data that may permit a panel estimation approach. Although our empirical results are preliminary, the findings so far suggest that economists may have to extend the list of variables important for understanding bilateral exchange rate volatility beyond those suggested by traditional optimal currency area theory.