Policy Coordination: How Is It Affected by Uncertainty?

For the latest thinking about the international financial system, monetary policy, economic development, poverty reduction, and other critical issues, subscribe to Finance & Development (F&D). This lively quarterly magazine brings you in-depth analyses of these and other subjects by the IMF’s own staff as well as by prominent international experts. Articles are written for lay readers who want to enrich their understanding of the workings of the global economy and the policies and activities of the IMF.


For the latest thinking about the international financial system, monetary policy, economic development, poverty reduction, and other critical issues, subscribe to Finance & Development (F&D). This lively quarterly magazine brings you in-depth analyses of these and other subjects by the IMF’s own staff as well as by prominent international experts. Articles are written for lay readers who want to enrich their understanding of the workings of the global economy and the policies and activities of the IMF.

IN AN increasingly interconnected world, the question of whether countries can improve their chances of successfully coping with the effects of economic uncertainty by coordinating policies is an important one. A recent IMF study suggests that, in the face of uncertainty, coordination may be more effective than independently conceived national policies.

Why should governments coordinate their macroeconomic policies? At the most general level, the case for coordination rests on the fact that the policies of individual countries do affect neighboring economies. In fact, the effects of spillovers are dramatic enough that governments acting independently may succeed only in neutralizing one another’s policies. Policy coordination, whether it takes the form of explicit, institutionalized rules for the international monetary system or ad hoc agreements—for instance, to prevent exchange rates or budget deficits from moving beyond target ranges—allows governments to better achieve crucial macroeconomic goals, both individual and collective.

The competitive devaluations of the 1930s provide an instructive example of what can happen in the absence of policy coordination. These devaluations caused soaring unemployment and a spectacular contraction of world trade as countries attempted to use their exchange rates to gain an unfair advantage in the face of a global depression. Another less serious failure to coordinate policies occurred in the aftermath of the 1979–80 oil price shock, when several countries used exchange rate appreciation to help lower their inflation rates, exacerbating inflation in other countries and leading to a severe contraction of output.

However, even well-thought-out, coordinated policies are subject to macroeconomic shocks, suggesting that it is not possible to know precisely what effects the policies will have. The 1978 G-7 (Group of Seven) Summit in Bonn, Germany represented one of the most important attempts at coordination among the major industrial countries, but the policy measures that were agreed to there are generally judged to have been overtaken by the second oil shock. The exchange rate parity changes decided at the Smithsonian in December 1971—the Smithsonian Agreement—did not save the Bretton Woods system of adjustable pegs, because the effect of these changes on trade flows had been greatly overestimated.

It can be seen from the above that uncertainty has both positive and negative effects on coordination. We therefore did a study that looked at the way such macroeconomic uncertainty affects both gains from policy coordination among industrial countries and efforts to reach and monitor multilateral agreements in order to form an overall view of the gains from coordination. We used econometric models to estimate the effects of different policy decisions on economies experiencing a variety of exogenous shocks to discover whether, in the face of uncertainty, coordination is better than independent policymaking. Our findings illustrate the benefits—and hazards—of policy coordination for the world economy.

Does uncertainty affect gains?

Economists disagree about the precise definition of coordination. In the purest sense, a coordinated regime is one in which each government chooses its macroeconomic instruments in order to achieve a common set of goals, but it is unlikely that macroeconomic policies have ever been set in a manner that precisely satisfies this definition. A less demanding criterion for judging the extent of policy coordination might be that coordination occurs when policy choices take into account effects on other countries’ welfare. Some economists have argued that what purport to be coordinated policies among the G-7 countries are really no different from the policies the national authorities of each country would pursue independently.

Yet substantive international economic policy coordination does take place; negotiations among governments generally do lead to out- comes that governments acting alone would not have reached. Political scientists have, for example, concluded that the agreement reached at the 1978 G-7 summit would have been impossible without coordination. Such agreements are negotiated with an eye to meeting mutually acceptable objectives, in contrast with policies that have clearly been chosen independently, without regard for the goals of other governments.

Government policies are, of course, planned and implemented in an environment of uncertainty about the prospects for national economies and, in particular, about the effects of policy changes. One view holds that this uncertainty makes formulating agreements much more difficult and may in fact eliminate any gains from coordination. This view has been expressed by former Chairman of the US Council of Economic Advisers Martin Feldstein, who observed in 1983:

“Economists armed with econometric models of the major countries of the world can, under certain circumstances, identify coordinated policies that, quite apart from balance-of-payments constraints, are better than uncoordinated country choices. But in practice, the overwhelming uncertainty about the quantitative behavior of individual economies and their interaction, the great difficulty of articulating policy rules in a changing environment… make such international fine tuning unworkable,”

(The Economist, June 11,1983)

Yet the view that uncertainty necessarily reduces the gains from policy coordination is much too simplistic. Though it is true that any policy, coordinated or not, may prove in retrospect to have been misguided because its effects were not correctly anticipated, under certain circumstances greater uncertainty may provide an additional incentive to coordinate policies. For example, even if each government has only one target (inflation) and is using only one instrument (monetary policy), uncertainty renders it unlikely that any country will hit its target exactly, since it is impossible to gauge the exact effects of its own or others’ policies. Coordination can improve the effectiveness of monetary policy by reducing some of the uncertainty, in particular that part related to the actions of other governments.

The transmission effects of policies onto other countries are not precisely known, and joint decision making can help ensure that the effects of foreign policies on domestic output and inflation are not ignored. In contrast, governments acting independently would not be able to minimize these effects. Therefore, gains from policy coordination may actually increase with growth in uncertainty related to transmission multipliers (the effects of each country’s actions on other countries).

The stock market crash of October 1987 provides a clear example of the positive effects of uncertainty on incentives to coordinate policies. Shifts in portfolio preferences leading to sharp declines in stock markets worldwide provoked concern about the stability of the financial systems in a number of countries. In general, during such a shift—in this case, out of equities or bonds into safer assets such as US Treasury bills—central banks may want to increase liquidity to avoid the bankruptcies among securities dealers that could threaten the financial system’s stability. But the monetary authorities, fearing exchange rate depreciations, may be unwilling to lower interest rates. Following the 1987 crash, the central banks of the major industrialized countries circumvented the problem with a series of coordinated interest rate cuts, averting both a financial crisis and a downturn in activity.

This action is an interesting example of coordination that is partly ad hoc and partly institutionalized. Not only was it a response to a unique and unexpected event, but it was undertaken at a time when the G-7 coordination process had been strengthened by the Plaza Agreement and the Louvre Accord, so that an institutional framework was in place.

What the study looked at

While the case for policy coordination seems strong in theory, even when the precise effects of the policies to be implemented are unknown, in practice the question of how large the policy gains will be remains unanswered. We formulated plausible, alternative econometric models to gauge the extent to which being “wrong” about the economy can affect gains from policy coordination among industrial countries. The models describe the functioning of the US economy and the remaining industrial countries as a group. Four alternative structures—representing prominent views within the economics profession—were used to specify the models, whose parameters were estimated using data for the 1970s and 1980s. It should be stressed that the actual parameters are unknown and that estimates of their values have standard errors associated with them.

We used the estimated models to simulate situations involving values of shocks to the world economy consistent with historical experience, including a sudden increase in oil prices, a speculative attack on a currency, a shift in consumer preferences, and a decline in money demand. The predictions concerning the effects of these shocks differ with each model, as do the policy implications. The simulations assume that governments, not knowing which is the correct model, attach probability values to each. The models allowed us to use the different kinds of policy responses governments might resort to in coping with the various shocks to measure the gains or losses from coordination. We were then able to gauge the extent to which being “wrong” affects gains from policy coordination.

Differences in the models relate to the causes of inflation, the form of money demand, and the effect of shocks on potential output. For instance, in all the models except one, there is a vertical long-run Phillips curve, meaning that an attempt to raise output above its potential produces accelerating inflation. In one of the models, prices are purely flexible, so that monetary policy does not systematically affect output. In another of the models, shocks can cause output to diverge from an earlier trend. Welfare is assumed, for each country, to depend on output compared with potential and inflation compared with price stability. When considering joint welfare, the United States and other industrial countries are treated equally.

Is coordination worthwhile?

Our findings showed that, in general, policy coordination is desirable, even in the presence of uncertainty, provided that the policymakers’ view of how the economy works is not too wrong or is revised to take experience into account. However, coordination in the form of simple, nonactivist policies that are more cautious and do not respond as actively to macroeconomic developments may be better than attempts to agree on fully optimal policies.

Coordination is the best policy. The simulation experiments indicate that coordinated policies are a significant improvement over policies conceived independently, despite the presence of model uncertainty. A good example is the stock market crash discussed above. However, policymakers wanting to gain the most from coordination must know which model is correct—or at least attribute a large probability weight to it.

Being wrong can result in large welfare losses. The simulations show that, in some cases, being very wrong (selecting the incorrect model) results in huge welfare losses, because a policy chosen on the basis of the wrong model destabilizes the world economy. This finding can be illustrated with a simple example. One of the models postulates a long- term trade-off between output and inflation, so that policymakers can choose higher output if they are willing to incur higher inflation. In the other three models, since faster monetary growth produces higher inflation but does not stimulate long-term output, this trade-off does not exist. If policymakers believe in the first model, then both optimal coordinated and uncoordinated policies will try to exploit the trade-off, which will produce accelerating inflation if one of the other models is in fact the correct one. However, since this result holds true for uncoordinated as well as for coordinated policies, it is not an argument against coordination. Instead, it suggests that simple, nonactivist rules should be used.

“…in general, policy coordination is desirable…provided that the policymakers’ view of how the economy works is not too wrong or is revised to take experience into account”

Simple, nonactivist rules may be best when uncertainty is great. Several of these rules were also evaluated: fixing the growth rate of the money supply and allowing the exchange rate to float freely, targeting nominal income, fixing the nominal exchange rate, fixing the real exchange rate, and a synthetic rule in which the money supply is adjusted as a decreasing function of output and producer prices and as an increasing function of interest rates. The rules performed well when policymakers were very wrong.

To return to the example of the previous paragraph, the fixed rules do not try to exploit the output-inflation trade-off and so are not destabilizing. But even these simple rules would probably need to be coordinated to some degree, because otherwise countries would be tempted to stray outside the rules to gain competitive advantages over their trading partners. In addition, targeting a joint variable (like the exchange rate) requires mutual agreement. Such simple rules may provide a useful framework for coordination.

Policymakers can learn from observation. Policymakers may be able to learn which model is correct by updating the probabilities applied to each. In other words, governments are assumed as before to set policies on the basis of their assessments of the likelihood that each model is correct, but they also update those assessments each period. Observing macroeconomic outcomes, and assumed to know the distribution of shocks affecting the world economy, they recalculate the probabilities that the four models could have generated those outcomes. Updating the probabilities in this fashion reduces uncertainty and makes it less likely that policy coordination will be destabilizing.

These simulations do not assume that policymakers eventually learn the true model in the sense of knowing precise parameter values; rather, policymakers endogenously learn the basic structure of the model, and parameter values have associated variances. Learning occurs at a rapid enough pace for gains to be realized by coordination relative to either uncoordinated optimal policies or simple non- activist rules. In the perverse cases in which activist policies are destabilizing and welfare losses unbounded, it is hard to believe that policymakers would not modify their views about the functioning of the economy in light of manifestly incorrect predictions of the model they thought was correct. If they did so, then coordination would still dominate the other policies.

A further interesting experiment consists in assuming that none of the four models is in fact correct. Instead, a completely atheoretical, time-series model is fitted to the data and used in the simulations to generate the data. As before, policymakers attempt to learn from observed outcomes which of the four models is correct, even though none of them really is. Once again, fixed, nonactivist rules generally come out quite well.

Information sharing

It is sometimes argued that information sharing, not coordination, is the major source of gains from cooperation among the major industrial countries. Using a simple model, it is possible to show that information exchange alone may actually lower welfare. For instance, if the incentives for beggar-my- neighbor policies exist and have not been removed by an attempt at coordination, knowledge about other governments’ actions may lead to increased efforts to use domestic policy in a way that can have further negative effects abroad. This possibility is not in itself an argument against information exchange; rather, it is an argument for recognizing the inherent limitations of forms of international cooperation that do not entail actual coordination of macroeconomic policies.

In practice, there may be good reasons why information exchange is accompanied by coordination. Information exchange typically results from the active consultation and negotiation that are part of the process of coordination among governments. One political scientist, Robert Keohane, has argued that only by entering into agreements to carry out specific policies can governments actually reduce uncertainty through the provision of information: “Uncertainty pervades world politics. International regimes reduce this uncertainty by providing information, but they can only do this insofar as governments commit themselves to known rules and procedures and maintain these commitments even under pressure to renege.” Sharing information without any policy commitment may be valueless, because the temptation to mislead exists, making information about future policies less than credible.

Obstacles to coordination

If the above gives a rosy view of the value of coordination, it is also important to take into account the obstacles uncertainty presents to achieving coordination.

Verification. One potential obstacle to coordination is verifying that agreements are actually kept. A well-known example of such a problem is the “prisoner’s dilemma,” which posits two prisoners, each with the incentive to accuse the other. The best policy for both the prisoners is to remain silent, but, even if they agree to this tactic, it is impossible to enforce compliance. In this example, and also in real-world international economic policy coordination, the parties to agreements have incentives to renege. But while the temptation to cheat may be overwhelming if the agreement is considered in isolation, such behavior would of course lead other governments to refuse to cooperate in the future, so that cheating should be a rare event: governments, like individuals, care about their reputations.

Unfortunately, macroeconomic uncertainty makes it more difficult to observe cheating because policymakers can always claim that outcomes are the result of events beyond their control. This concern apparently resulted in a switch at the 1977 London summit to framing macroeconomic policy agreements in terms of policies, not outcomes. Such a procedure does not, however, really solve the problem because the choice of instrument settings will generally depend on forecasts of outcomes for macroeconomic variables. Germany, for example, may be urged to stimulate its economy (as it was in the late 1970s and in the mid-1980s) but may counter that the economy is about to pick up, making stimulus inappropriate.

Misrepresentation. Such problems in verifying that governments are not misrepresenting their positions or the state of their economies may provide a rather strong argument for international organizations that provide independent inputs into the process of coordination. Because governments have asymmetric information, in the sense that they know more about their own economies than about foreign economies, there may be a suspicion that they are in some circumstances exploiting that information. A former G-7 deputy, Wendy Dobson, has argued forcefully that the G-7 Finance Ministers should draw on the services of a secretariat that would provide unbiased analysis and forecasts.

Just as there may be incentives to misrepresent private information, there may be strategic reasons for governments to misrepresent either their own preferences or their views about the effects of their policies in the face of model uncertainty. For instance, the United States may be urged to reduce its budget deficit, as it was in 1983 and 1984, but may respond that because its budget deficit is not the cause of high interest rates and the soaring dollar, reducing the deficit would not have the desired effects. US trading partners, however, are likely to consider these arguments a disingenuous attempt to shift the gains from any coordinated agreement in the United States’ favor. The problems of bargaining over the exact outcome of a coordinated agreement when there is asymmetric information about beliefs or preferences are in fact so severe that a coordinated regime may be unattainable. Again, it may be important in achieving coordinated policy agreements for a respected, unbiased organization to confront various theories and evaluate the empirical evidence, narrowing the extent of differences in views.

Final thoughts…

Uncertainty about the effects of policies, far from making coordination undesirable, is likely to provide an incentive to coordinate. It emerges from our simulations that the lack of precise knowledge about the world economy should not discourage attempts at coordination, as gains are likely to exceed losses. Although efforts at fashioning fully optimal policies (i.e., those that are best by some criterion) may lead to problems, coordination around simple policy rules can still be beneficial, even if policymakers are wrong about some aspects of the way the economy functions (e.g., the precise links between monetary policies and economic activity). As for the obstacles to achieving coordination, international organizations such as the World Bank and the IMF can play a role in making them less severe.

This article is based on a book by the author and Atish R. Ghosh, Economic Cooperation in an Uncertain World (Oxford: Basil Blackwell, 1993). The extensive literature on policy coordination is surveyed in “Scope and Limits of International Economic Cooperation and Policy Coordination,” by Jocelyn Home and Paul Masson (IMF Staff Papers, June 1988).