The overshooting theory of exchange rates seems ideally designed to explain some important aspects of the movement of the dollar in recent years. Over the period 1981-1984, for example, when real interest rates in the United States rose above those among trading partners (presumably due to shifts in the monetary/fiscal policy mix), the dollar appreciated strongly. It was the higher rates of return that made U.S. assets more attractive to international investors and caused the dollar to appreciate. The overshooting theory would say that, as of 1984 for example, the value of the dollar was so far above its long-run equilibrium that expectations of future depreciation were sufficient to offset the higher nominal interest rate in the minds of international investors. (Figure 1 shows the correlation of the real interest differential with the real value of the dollar, since exchange rates began to float in 1973.)

When international policymakers do not agree on the correct macroeconomic model, they will still be able to agree on a cooperative policy package that each believes will improve welfare; but the package may turn out to move the target variables in the wrong direction. Using ten leading econometric models that could represent U.S. beliefs, non-U.S. beliefs, and the true model, we find that monetary coordination improves U.S. welfare in only 546 cases out of 1,000.

[Survey data provide a measure of exchange rate expectations superior to the forward rate in that no risk premium interferes. We estimate extrapolative, adaptive, and regressive models of expectations. Static or "random walk" expectations and bandwagon expectations are rejected: current appreciation generates the expectation of future depreciation because variables other than the contemporaneous spot rate receive weight. In comparing expectations to the process governing the spot rate, we find statistically significant bias.]

Until recently, there was an unusual degree of consensus among economists that intervention by central banks in the foreign-exchange market did not offer an effective or lasting instrument for affecting the exchange rate, at least not independently of monetary policy. This consensus was largely shared among policymakers and participants in the financial markets as well. The 1982 G-7 economic summit at Versailles commissioned a study of intervention, known as the Jurgenson report, which found that the effects were small and transitory at most.' We think that the time is ripe for new statistical testing of the question. Many policymakers and foreign-exchange traders believe that the intervention operations that have taken place since the Plaza Agreement of September 1985 have had an effect, especially when operations are coordinated. Moreover, the theoretical case against the effectiveness of intervention is not as clear as a reading of the economics literature might suggest. The academic literature is predicated on the distinction between intervention operations that are sterilized and those that are allowed to affect the money supply. We study the intervention operations that actually took place between 1982 and 1988, regardless of whether they were sterilized. However, we do begin in Section I with a review of the issues involved.2 There are two possible channels through which intervention (whether sterilized or not) can influence the foreign-exchange rate: the portfolio and the expectations channels. Intervention can, even if sterilized, influence exchange rates through the portfolio channel, provided foreign and domestic bonds are considered imperfect substitutes in investors' portfolios. Intervention operations that, for example, increase the current relative supply of mark to dollar assets which private investors are obliged to accept into their portfolios, will force a decrease in the relative price of mark assets.3 Intervention can also influence exchange rates, regardless of whether foreign and domestic bonds are imperfect substitutes, through the expectations channel. The public information that central banks are intervening in support of a currency (or are planning to intervene in the future) may, under certain conditions, cause speculators to expect an increase in the price of that currency in the future. Speculators react to this information by buying the currency today, bringing about the change in the exchange rate today. In Sections II and III we describe the econometric problems that arise in the standard portfolio-balance estimation equation. We derive an alternative portfolio-balance *Dominguez: Kennedy School of Government, Harvard University, 79 J. F. Kennedy Street, Cambridge, MA 02138; Frankel: Department of Economics, University of California, 787 Evans Hall, Berkeley, CA 94720. We thank three anonymous referees for valuable comments and suggestions; Julia Marsh and Julia Lowell for research assistance; and Franz Scholl at the Bundesbank, Jean-Pierre Roth at the Swiss National Bank, and officials at the U.S. Treasury and the Board of Governors of the Federal Reserve System for making the daily intervention data available. 1Many of the econometric results, finding little or no effect, were reported in Kenneth S. Rogoff (1984) and Dale W. Henderson and Stephanie Sampson (1983). 2For authoritative statements, see Henderson (1984) or Maurice Obstfeld (1990). 3The exchange-rate reaction to an increase in the relative supply of outside foreign assets may be reduced if there is an increase in their expected rate of return that induces a corresponding increase in demand.

In their paper in this Review, Jeffrey A. Frankel and Katherine E. Rockett (1988) show that when international macroeconomic policymakers do not agree on the correct macroeconomic model, they will still be able to agree on a cooperative policy package that each believes will improve his welfare. Yet the package may turn out to move target variables in the wrong direction. From extensive simulation experiments with ten empirical models, Frankel and Rockett conclude: ...the bargaining solution is as likely to reduce welfare as to improve it. But more definitions of cooperation should be investigated... (p. 338). In this paper we propose an alternative definition of a policy bargain to that investigated by Frankel and Rockett and show that results in a higher success rate and higher expected utility in cooperation exper-iments. It follows from our finding that, given uncertainty or ignorance about the true model, a measure of disagreement is beneficial because facilitates a simple robustness check for proposed policy bargains. As Frankel and Rockett (1988 p. 328) acknowledge: it is the countries' failure to perceive the true model, not their failure to agree with each other per se, that alters the standard conclusion regarding coordination (i.e., is uncertainty, not disagreement, that leads to failure). We go further: given the inevitable failure to perceive the true model, some disagreement is better than unanimity in error. Extreme disagreement about the nature of reality, however, makes robust bargains impossible.