Since Adam Smith, economists have questioned how the actions of individuals in a decentralized market economy could be effectively coordinated. Can a system of prices, responding to supply and demand alone, lead to a desirable market equilibrium ? In recent years they have discovered that whether prices alone can do the trick depends in part on the presence and nature of externalities and public goods - spillover effects from the actions of one agent on the welfare of another.

Since the advent of the flexible exchange rate system in 1973, the international economy has relied much more on prices: exchange rates and interest rates. Can these prices coordinate the international economy ? Have they done so in practice ? The unsatisfactory performance of the world economy in the past decade has motivated analysis of whether a system of flexible financial markets is sufficient to deliver satisfactory performance, without explicit coordination among governments taking into account the spillover effects of their policies on other economies. The Economic Summit held in London in June highlighted this question.

The Centre for Economic Policy Research and the National Bureau of Economic Research recently held an important conference on this subject - 'The International Coordination of Economic Policy'. The conference took place on 28-29 June in London, and was jointly organized by CEPR Programme Director for International Macroeconomics, Willem Buiter of LSE, and Richard Marston, of the University of Pennsylvania and the NBER. Major financial support for the conference was provided by the Ford Foundation, with additional assistance from the Bank of England and the UK clearing banks. Participants were drawn largely from the United Kingdom and the United States, but others came from France, Greece, Israel and Italy. Eight papers were presented, and the meeting concluded with a panel discussion on the prospects for international policy coordination.

Many of the issues discussed throughout the conference were raised in the first paper, 'On Transmission and Coordination Under Flexible Exchange Rates', by Max Corden of the Australian National University. He considered whether there was a need for macroeconomic policy coordination under flexible exchange rates and the nature of the equilibria that might be reached without such coordination. Corden's two-country model was an attempt to formalize the popular argument that coordinated expansion would be easier than expansion by any one country on its own, so lack of policy coordination leads to more deflation than would coordinated policies.

His model was short-term in nature, neglecting capital flows. It included a non-vertical, short-run Phillips curve, thus a trade- off between inflation and unemployment. The principal link between countries in the model was through the terms of trade - the ratio of import to export prices. An economic expansion by one country improved the terms of trade of the other country, shifting the latter's Phillips curve in a favourable direction. This in turn permitted the latter country to pursue a more expansionary policy. Expansion by one country thus induced expansion by the other, so there was positive transmission. Without policy coordination, the equilibrium had a contractionary bias, in the sense that both countries could be better off if they both expanded beyond that equilibrium. In the case where one country exercised leadership in setting policy, the equilibrium still had a contractionary bias and mutual expansion still would be beneficial. Corden noted that the transmission effect in his model was through the effect of the terms of trade on output - there was no direct transmission of inflation effects. But if inflation had longer-run real effects, these too might be transmitted.

There was some criticism of the microeconomic basis of Corden's model, in particular his specification of the Phillips curve. The short term nature of the model and its neglect of capital flows was also criticized. Some participants argued, however, that this absence of capital mobility, together with the symmetric treatment of the two countries, made the model a useful tool for the study of European policy coordination.

This paper also prompted a vigorous debate concerning the nature of equilibrium and 'solution concepts' in policy coordination models. That discussion continued throughout the conference and was at times highly technical, drawing on concepts from game theory. Much of it hinged on how 'players' are assumed to react to the decisions of other 'players'. Elementary economic theory avoids this question - it assumes that each consumer or producer takes market prices and the decisions of other economic agents as given and beyond its influence. This approach may be inappropriate for international economics, where there are countries which are large enough to influence world prices and exchange rates by their policy actions.

Analysing international policy coordination therefore requires some assumptions about how countries will react to the policies of other countries. Suppose each country assumes that whatever policy it announces, other countries will not react by changing their policies. This assumption will lead to a 'non-cooperative Nash equilibrium'; it may be appropriate where no country is significantly larger than any other. What assumption might be appropriate in a situation with one large country and many smaller countries - for example the United States and the rest of the world? Here the one large country may assume 'leadership', i.e. it will take into account the effects of its policies on the rest of the world, but the other countries assume that they are too small to have any influence. This will lead to what is called a 'non-cooperative Stackelberg equilibrium'. The choice of equilibrium or solution concept is further complicated by questions of 'credibility' and 'trust', which arise in situations where policies are formulated not just on one occasion, but repeatedly over time.

The combination of fiscal and monetary policies pursued recently by the United States is often thought to have especially powerful effects on the rest of the world. Jacob Frenkel, University of Chicago and NBER, and Assaf Razin, Tel Aviv University, gave a theoretical perspective on this issue in their paper 'Fiscal Policies, Interest Rates and International Economic Interdependence'. In their model, world capital markets were integrated, with a single world interest rate. It was through the financial markets and the common interest rate that one country's policies had international effects. Their analysis thus differed sharply from the more short-run model presented by Corden. Frenkel and Razin dealt with a two-country world, in which markets are assumed to clear and individuals behave rationally. The key feature of their analysis was its detailed consideration of the intertemporal budget constraints faced by both governments and individuals.

Does it matter whether the government decides to finance its spending through bond issues rather than taxation? An argument tracing back to Ricardo suggests that if individuals take full account of the future tax liabilities resulting from the bonds, it will make no difference to the real behaviour of the economy whether the spending is financed by taxes or by bonds. Budget deficits will not therefore affect the real economy. We should look only at total government spending, not the means by which it is financed, whose real effects are equivalent.

Frenkel and Razin show that this proposition does not hold if individuals and governments have different 'discount rates' - the rates at which they substitute consumption now for consumption in the future. The gap arises because individuals are mortal, while the government is not. They examine the international and the domestic effects of a change in taxes, holding the level of government spending constant. They find that this will raise world interest rates and domestic wealth, while lowering foreign wealth. Expected future deficits will be transmitted negatively to the rest of the world through the integration of capital markets and interest rates.

What happens if instead the level of government spending is changed? Frenkel and Razin considered an increase in government spending with a matching increase in taxation - a balanced budget. They show that a temporary increase in spending now would raise interest rates and lower domestic and foreign wealth, while a temporary increase expected in the future would lower interest rates, reduce domestic wealth and increase foreign wealth. Suppose there was a permanent shift in spending? Then the effect on the interest rate depends on whether the domestic economy is a net saver or a net dissaver in the world economy - i.e., whether it is in surplus or deficit on its current account. If in surplus, then a rise in government spending raises world interest rates; if in deficit, then the rise lowers interest rates.

The key to their results is the difference in individual and government discount rates, which in turn is based on the finite lifetime of individuals. Some participants doubted whether in practice this alone could account for the apparent impact of current government deficits on the world economy. Perhaps other factors, such as liquidity constraints on borrowing or the privileged position of sovereign borrowers would prove more important in explaining the effects of these deficits.

The impact of US budget deficits on the international economy was the subject of Patrick Minford's paper, 'The Effects of American Policies - A New Classical Interpretation'. Minford, from Liverpool University and CEPR, linked 9 small models, each representing an OECD country. The structure of each of the country models is 'new classical' in its inspiration. Expectations are formed rationally, and financial markets in particular are efficient. All markets clear, the labour market by virtue of the residual, non-union sector. There are information lags, as well as strong wealth effects. The model was not an estimated one - some coefficients were estimated, while others were assigned values thought plausible. The model is presented in CEPR Discussion Paper No. 11 and summarized more fully in CEPR Bulletin No. 2.

Minford used these linked country models in an ambitious first attempt to simulate the effects of US fiscal and monetary policies. For example, a bond-financed US deficit, lasting one year and equal to 1% of US output, caused a sharp rise in interest rates. US output fell; output in the rest of the world fell, then rose slightly. Minford also simulated an expansion in the US money supply, together with a budget deficit. He found that this had strong short-run effects on the level of both US and world output, as well as more lasting effects on the world inflation rate.

Some participants found these simulation results unexpected, occasionally implausible, and raised questions regarding the structure of Minford's models. Was it appropriate to use the same model structure for all the major OECD countries? Many of the results appeared to come from the very strong wealth effects in his expenditure and demand for money equations, whose empirical basis was discussed.

There has been increasing interest in the interwar period and the lessons it might suggest for current policy (see the Workshop reports in CEPR Bulletin No. 2 and below). Barry Eichengreen, Harvard University and NBER, writing on 'International Policy Coordination in Historical Perspective: A View from the Interwar Years', reconsidered the history of the international financial system to see how it illuminated international policy coordination. He first examined how contemporaries viewed the role for policy coordination at the start of the interwar period, taking as a case study the Genoa Economic and Financial Conference of 1922. Eichengreen argued that the advantages of policy coordination were in fact well understood in the 1920s but that political disagreements impeded moves towards a framework for cooperative action.
What effect did the resulting non-cooperative behaviour have, once the gold standard was again in operation? The gold exchange standard resembled a rules-based regime, an attempt to minimize the need for policy coordination by limiting policymakers' discretion. Eichengreen argued that the gold standard constrained but by no means eliminated the authorities' options. He developed a two-country model designed to highlight the strategic behaviour of governments under the gold standard. This described the behaviour of the banking sector, as well as aggregate supply and aggregate demand in each country. Capital was perfectly mobile and financial assets were perfect substitutes, resulting in a common interest rate in the two countries. The model was completed by equations describing each government's policy objectives: both domestic price stability and their share of the world gold stock.

What effect did the resulting strategic behaviour have on the international economy ? Eichengreen found that this imparted a deflationary bias to the system. Governments increased their discount rates in order to capture more of the world's gold, depressing the level of domestic activity. But not all governments could increase their share of gold reserves at the same time! Their attempts to gain more gold were offset by higher discount rates in other countries, resulting only in more deflation for everyone. Cooperative behaviour, through coordination of discount rate policies, yielded less deflation than non-cooperative behaviour. The model also helped in understanding why cooperative solutions proved so difficult to achieve, though it lacked dynamics and did not treat expectations explicitly.

The lessons of the experience with non-cooperative strategies were reflected in the next attempt to reconstruct the international monetary order: the Tripartite Monetary Agreement of 1936. This resembled closely the Genoa Resolutions of 1922 but was more successfully implemented, because political circumstances had changed and the scope for collaboration was more tightly circumscribed. Thus, Eichengreen argues, the history of international financial collaboration in the interwar period sheds light not only on the rationale for policy coordination but also on the political and economic circumstances which might promote it.

The discussion focussed on the nature of the spillover effects in such models. It was claimed that the spillovers in the interwar period were negative, but Eichengreen argued that they were positive during the gold standard and negative during the period of floating exchange rates which followed. Some participants stressed that not only governments, but also other sectors of the economy could engage in strategic behaviour. Models in which policy coordination between governments was beneficial might yield quite different conclusions if other sectors were allowed strategic behaviour as well.

Suppose that the need for policy coordination were accepted? How would one devise sensible coordinated policies for governments?
Marcus Miller, Warwick University and CEPR and Mark Salmon, Warwick attempted to clarify some of these issues in their paper, 'Dynamic Games and Time-Consistent Policy in Open Economies'. They used a two-country model in which capital is perfectly mobile and expectations are formed rationally. Their model thus exhibits substantial 'policy interdependence' through the effect of monetary policy on real interest rates and real exchange rates.

Here governments formulate policy not just on one occasion, but repeatedly over time. This is more realistic, but if there is forward-looking behaviour, then markets will try to take account not only of current but also future government policies. The government may announce its future policies - it may state that no matter what level unemployment reaches, it will continue with a restrictive monetary policy - 'there will be no U-turns'. But is this announcement credible, given that the government cannot actually pre-commit itself to carry out all its announced policies? Would the government carry out such a policy, even if no unforeseen events occurred ? Paradoxically, many policies which are optimal over time without pre-commitment are 'inconsistent', in the sense that even in unchanged circumstances, the government would find incentives to depart from its previously announced policies. It is argued that policies which are 'inconsistent' in this sense are therefore not credible and could not form the basis of sensible policy recommendations in a world where governments are not bound to carry out the policies they announce.

Miller and Salmon examine non-cooperative policy in their two- country model and compare it with the outcome of policy coordination. They rule out policies which are time-inconsistent by assuming that the path of the real exchange rate is taken as given when the policy is designed. They find that with substantial 'spillover' effects via international financial markets, changes in the strategic behaviour of governments produce considerable changes in the dynamic behaviour of the system. Models in which capital is highly mobile possess much more policy interdependence than models with floating exchange rates but not capital flows. They find that time-inconsistent, coordinated policies could be beneficial. But their assumption that the future path of the real exchange rate was taken as given, which ensured the time consistency of the policies, was strongly criticized. Miller and Salmon replied that they were exploring alternative assumptions which gave time-consistent solutions.

The design of coordinated policies was also considered by David Currie, Queen Mary College and CEPR, and Paul Levine, Polytechnic of the South Bank, in the paper 'Macroeconomic Policy Design in an Interdependent World'. Discussions of policy coordination usually involve models with only very simple dynamics for reasons of tractability. Currie and Levine argued that models with more dynamic structure would yield more useful insights into policy. The model they consider includes wage and price dynamics, asset accumulation, a floating exchange rate and aggregate demand and supply lags. They first analyse a single open economy and a variety of simple policy rules, in which interest rates are adjusted in response to changes in an indicator such as the money supply, nominal income, the exchange rate or the price level. The price rule performed best and even compared well to the performance of optimally designed monetary policy.

These results did not carry over to an interdependent world. Both the monetary and the nominal income rules for adjusting interest rates performed better than the price rule - the reverse of the ranking for a single economy. Moreover, when applied by all countries simultaneously, the price rule performed very poorly, in some cases completely destabilizing the system. Why should this happen ? The price rule relied heavily on manipulating the exchange rate, via monetary policy, to stabilize the domestic economy, at the expense of sharp exchange rate movements. In the aggregate, this caused interest rates to over-react to inflationary pressures, leading to instability.

Currie and Levine argue that these results highlight the externalities inherent in macroeconomic policy design in an interdependent world. They also note that countries may adopt 'free-rider' behaviour, reneging on previously agreed cooperative policies. This temptation is particularly strong for the price rule, so they argue that this and other rules which rely on the exchange rate to influence domestic prices are inimical to international cooperation.

It was suggested that the simple rules considered by Currie and Levine were arbitrary, and that they might instead have considered simplifications derived from the optimal policy rules. They were also criticized for assuming that individuals were very much more sophisticated and knowlegeable than were governments. The authors conceded that this was somewhat implausible, but replied that modelling learning by the private sector in such models was very difficult. One participant felt that increasingly sophisticated game theory and control theory were applied to models whose economic specification was somewhat arbitrary.

If policies could be coordinated, what magnitude of benefits might one realistically expect as a result? Gilles Oudiz, INSEE, and Jeff Sachs, Harvard and NBER, considered some of the methodological problems raised by such a question in their paper 'International Policy Coordination in Dynamic Macroeconomic Models'. Suppose that two countries are each attempting to adjust their policies along a short-run Phillips curve. It may be that each country will choose contractionary policies no matter what the other country does, though the policy of joint expansion would be preferable to the non-cooperative solution in which both countries contract. This situation arises naturally under flexible exchange rates, since by contracting while the other country is expanding, a country can appreciate its currency and export some of its inflation abroad. Cooperation, in the form of a binding internatonal commitment to expand, may move the countries to a more efficient equilibrium. Will these gains persist in a longer-run model? Oudiz and Sachs argue that in most models, policies which lead to a short-run real appreciation also lead to a long-run real depreciation, or at least a return to the initial exchange rate. In the longer run, therefore, appreciation may be less attractive since the inflation will eventually be reimported. To this extent, the payoffs to these policies look very different in one-period and multiperiod analyses, and thus so do the gains to coordination.

Oudiz and Sachs also maintain that policy coordination may actually reduce economic welfare if governments are myopic in their policymaking, as is sometimes claimed. Policymakers may have a short- run expansionary bias if expansion shows up as increased output today and increased inflation only in the future. To some extent, the fear of currency depreciation following a unilateral expansion keeps this bias in check: the flexible exchange rate disciplines short-sighted governments. With policy coordination, the fear of currency depreciation can be removed by a commitment of all countries to expand. In this way, policy coordination may give incumbent governments a free hand to undertake overly inflationary policies. Conversely, there may be circumstances in which policy coordination ties the hands of governments and prevents self-serving and myopic behaviour.

The authors also note that governments cannot bind the actions of later governments, and they investigate how this fundamental constraint might alter the gains from policy coordination. Some authors have given examples in which the inability to commit future governments imparts an inflationary bias. Oudiz and Sachs conclude that this may not be an important problem, since governments may feel obliged to honour the commitments of their predecessors in order to preserve their 'reputation'.

Have there been recent examples of successful macroeconomic policy coordination, from which some lessons might be drawn ? In the final paper, 'Rules versus Discretion: the EMS Experience' Tommasso Padoa-Schioppa, Bank of Italy, claimed that the European Monetary System was in many ways a successful case of institutional cooperation for the purpose of macroeconomic policy coordination. He analysed what the EMS has achieved in the light of the 'rules versus discretion' debate in international policy coordination.

Padoa-Schioppa discussed the conflict between national and international trends in the authorities' use of rules and discretion and maintained that more discretion was needed at the international level. He analysed 'institutional cooperation', as opposed to 'ad-hoc cooperation', and argued that institutions should play a greater role. To extend discretion in the management of international macroeconomic problems required strengthening institutions.

He also considered the experience of multicountry monetary cooperation within the EMS and presented statistical evidence on the System's ability to achieve its objectives. Whatever criteria were used, exchange rates had been more stable during the period of the EMS than they had been previously. Padoa-Schioppa also described key features of the EMS such as parity changes, adjustment decisions and the disciplinary effects exerted by the System on private as well as official agents.

What lessons can be drawn from the EMS experience for managing the international monetary system as a whole ? Padoa-Schioppa argued that increased discretion is necesary in managing international economic policy, but that this should be exercised at the level of international and not national institutions.

The conference concluded with a panel discussion, chaired by William Branson, Princeton University and NBER, on the prospects for international policy coordination. The panel included Richard Cooper, Harvard University; Michael Emerson, Commission of the European Communities; Louka Katseli, Centre of Economic Planning and Research; and Stephen Marris, Institute for International Economics.

Richard Cooper noted that governments, like other economic agents, responded to their environment, and that policy could be analysed using many of the same tools used by economists to study the behaviour of the private sector. Many of the questions asked were very similar. Did a 'policy' equilibrium exist ? Would coordination of policy be beneficial, or was decentralized policymaking desirable? Cooper noted that the justifications for coordination included the public good aspects of international policies and their externalities or spillover effects. Sucessful examples of international cooperation did exist. Some, such as the metric system or the international public health system, were clearly public goods which could only be provided cooperatively. These examples indicated that international coordination was only likely to be successful when the ratio of benefits to costs was high and clearly recognized to be so by all concerned. Cooperation could take a variety of forms, however, ranging from the adoption of common standards to continuous joint decision- making by governments.

He saw a number of problems which could obstruct international economic policy coordination. There were differences among governments in their objectives. Indeed it had been argued that the only sucessful economic summits had been those in which there was strong disagreement within governments, leading to the formation of coalitions across government lines by means of which agreements were concluded. There was little agreement on the way in which economies actually worked, and therefore little agreement on the benefits which might arise from policy coordination or even what the coordinated policies should be. Cooper noted that over 50 years elapsed from the recognition of the problem of communicable diseases to the introduction of the quarantine system, since there was no agreement on the way in which diseases were transmitted. He thought that international policy coordination would require at least as long to develop.

Michael Emerson was somewhat more optimistic. He believed that international policy coordination was both a very real and a very important activity. European and American agendas for such a programme were quite different, however, and he saw few prospects for policy coordination at present, since the United States did not see coordination as advantageous to it. Should difficulties arise in 1985 or 1986, this attitude might change. In general, he thought that any policy coordination would also have to involve Japan as well.

Commenting on the prospects for European policy coordination, Emerson argued that the first three years of the European Monetary System had been disappointing, but that recent realignments had been more serious, and had delivered more convergent policy changes. He thought that European policy was now better coordinated, albeit in a restrictive direction - but this may have made the policy more credible and therefore less costly in terms of lost output and employment.

Louka Katseli emphasized the asymmetries which characterize the current international system and international decision-making. There are asymmetries in the origin of the shocks affecting different countries, as well as in the impacts of the shocks on various markets. In particular, there is a great asymmetry in the monitoring of outcomes and performance of different countries. Katseli remarked that rules which apply to the less-developed countries seemed not to apply to the United States, even when both were running deficits. Moreover, there is no monitoring of the position of creditor countries as there is for debtor countries. She suggested that more attention be paid to the structures of decision-making in international organizations, and that this too might be modelled as a game. She argued that a theory of 'optimum coordination' areas, like 'optimum currency areas', might be usefully developed.

Stephen Marris noted economists' recent interest in policy coordination, but with mixed feelings, since he found little enlightenment in their writings. The official view of flexible exchange rates had shifted very suddenly in the 1970s. What had not been taken into account then was the effect of flexible exchange rates on the Phillips curve. The experiences of the weak currencies in the 1970s had reflected this and had shown that it was difficult for one country to expand alone. More recently coordination had served mainly to keep up the joint resolve to fight inflation, at the expense of a recession.

He argued that the major current problem was not the inflationary or deflationary bias of the system, but simply policy divergence between the United States and the rest of the world. World economic prospects if the US deficit were reduced were quite favourable, and this should be evident to all. But US policy- makers did not believe they were much dependent on the rest of the world. Only an outbreak of economic cholera, to use Cooper's example, would convince them otherwise - but this might be on the horizon.

The papers presented at the conference and described in this article will be available in a volume to be published by Cambridge University Press in early 1985.



STEPHEN YEO