International finance and macroeconomics.

• Author: Frankel, Jeffrey A.
• Position: Program Report

This report on the activities of the Bureau's program in International Finance and Macroeconomics (IFM) is the first since the original program in International Studies was divided into IFM and International Trade and Investment two years ago. It discusses the program's research in six areas: the international integration of financial markets; the determination of exchange rates; open economy macroeconomics; European monetary integration; the world monetary system in historical perspective; and international aspects of economic reform and growth.

The International Integration of Financial Markets

The increasing integration of financial markets across national boundaries has been a central aspect of the Bureau's international research. We have approached the subject of integration from a number of directions.

One common approach is to examine the ability of international capital flows to eliminate differences in expected rates of return across countries. This process is known as arbitrage. In his work, IFM associate Richard C. Marston has continued to test interest rate parity, focusing on Japan-U.S. differentials in one study, and on the pre- and post-1973 periods in another.(1)

An alternative approach is to look at the extent to which "representative agents" residing in different countries are able to smooth the variability of consumption over time, by taking advantage of the ability to borrow and lend internationally. This approach was pursued recently by Maurice Obstfeld, and by Andrew K. Rose and Assaf Razin.(2)

Stock markets, too, are becoming increasingly important as a vehicle for international financial integration. Work by Bernard Dumas has considered the implications of various empirical regularities in international equity markets for the choice between two models that describe those markets: the idealized representative agent model, in which all investors are the same; and an international version of the Capital Asset Pricing Model (CAPM).(3)

One of those empirical regularities is the tendency for investors in every country to hold far more domestic than foreign equities, despite the degree of internationalization that has taken place in recent years. Linda L. Tesar and Ingrid M. Werner document this phenomenon, which is known as home country bias.(4) To some degree, home country bias is created by the risk of exchange rate fluctuations. But that risk can be hedged easily in the forward exchange market. Kenneth A. Froot recently has shed new light on this strategy for the case of equities, and Richard M. Levich and Lee R. Thomas have done the same for the case of bonds.(5)

Logically, investors should pay a price for hedging, which would show up as a discrepancy between the forward exchange rate and the expected future spot rate. Three recent NBER papers by Bennett T. McCallum, Martin D. D. Evans and Karen K Lewis, and Richard H. Clarida and Mark P. Taylor, have examined the well-known failure of the forward rate to forecast the future spot rate.(6) In related work, Dumas and Bruno Solnik show how to implement the international CAPM empirically with an allowance for exchange risk and hedging.(7) Tesar and Warner, however, argue that home country bias is too great to be explained by exchange risk.

Another possible explanation for the failure of investors to diversify fully is the existence of remaining restrictions on cross-border investment. Rene M. Stulz and Waiter Wasserfallen explore such restrictions, looking specifically at Switzerland.(8) The ultimate test of intermarket frictions is to study prices of shares in the same company, in cases where they are cross-listed, as Alan W. Kleidon and Werner have done.(9)

The Bureau recently held a conference on "The Internationalization of Equity Markets," which examined many of these issues in depth. A summary of the conference appears in the Fall 1993 issue of the NBER Reporter. Many of the papers, including an editor's introduction, will appear soon as NBER Working Papers, and the volume should be published next year by the University of Chicago Press.

The Determination of Exchange Rates

Some Bureau economists recently have studied the empirical aspects of exchange rate determination. Martin, S. Eichenbaum and Charles Evans focused on those changes in the money supply that can be identified specifically as the result of policy changes, and found a significant (lagged) effect on the exchange rate.(10) Vittorio U. Grilli and Nouriel Roubini extend this approach to include monetary policy in other major industrialized countries.(11) Michael Dooley, Peter Isard, and Mark Taylor found some predictive ability from gold prices, as well.(12)

A number of Bureau researchers also have examined intervention in the foreign exchange markets by central banks, since it became more common with the 1985 Plaza Accord. The conventional wisdom is that--precisely because international financial markets are well developed, highly integrated, and subject to diversification--central bank purchases or sales of foreign exchange are unlikely to be large enough to have much of an effect on the exchange rate, except to the extent that they change money supplies, in which case they are simply a variety of monetary policy. Some years ago, however, Michael L. Mussa--an IFM program member currently on leave at the International Monetary Fund as Director of the Research Department and Economic Counsellor--suggested the signaling effect of intervention. This channel requires that intervention is reported to market participants and that they then interpret it as conveying information on future monetary policy. Michael W. Klein offers evidence relevant to the first proposition, and Graciela L. Kaminsky and Lewis offer evidence relevant to the second.(13) Kathryn M. Dominguez and I have found evidence of intervention effects through both the signaling channels and the traditional portfolio channel.(14)

Still, the overall empirical track record for macroeconomic models of exchange rate determination remains poor, despite the isolated examples of success with regard to specific aspects. Recently, Robert P. Flood and Rose have shown that the failure of the macro models transcends particular problems of parameter estimation.(15) Charles M. Engel, too, has found yet another example of the surprising difficulty in forecasting exchange rates better than a "random walk."(16)

Such findings are persuading some IFM members to turn to an entirely different sort of approach: microstructure models of the foreign exchange market that include variables, such as trading volume, bid-ask spreads, and trader heterogeneity, that the macro models omit. Richard K Lyons, in particular, has made progress with this approach.(17) The Bureau is planning a conference on the new topic of "Foreign Exchange Microstructure" next June, cosponsored with the Bank of Italy and the Centre for Economic Policy Research.

Open Economy Macroeconomics

International financial markets have implications for the real economy. Razin recently reviewed a popular theory of the current account measure of the balance of payments, in which it is determined by...