Exchange Rate Choices for ‘Mature’ Economies
Robert Skidelsky
International Monetary Convention | Sunday, May 11, 2003

 
The purpose of this paper is to question the near-unanimous view of economists that floating is the best policy for mature economies.
 
It is true that European Single Currency was established on the belief that fixing is best.
 
But the European fixers share one important premise with the Anglo-American floaters. That is that fixing has to be irrevocable if it is to be credible. Since irrevocable fixing on a world-scale is unfeasable, the conclusion is that the major currencies (including the European Single Currency) have to float against each other.
 
I have long felt there was something paradoxical in this sequence of reasoning. Most floaters admit stable exchange rates between the major currencies would be to the world advantage, even as they advocate floating for the sake of national or regional advantage.
 
They are driven to this position by the thought that anything short of irrevocable fixing will not be credible –that currencies provisionally fixed will be subject to damaging speculative attack in face of asymmetric shocks.
 
But if this is the danger facing currencies which are only provisionally fixed –and I admit the danger exists, though I think it is a much greater danger for immature than for mature economies –it should surely stimulate efforts to set up a system which provides for mutual support to maintain agreed values and agreed escape clauses in face of asymmetrical shocks. This is what the Bretton Woods system tried to do.
 
In short, I accuse economists of giving up too soon. They talk about the impossibility of making provisionally fixed systems credible when capital is free to flow between members, when they should be talking only about the difficulty of achieving this. They sacrifice the enormous advantages of stable exchange-rates on the rock of a logical ‘trilemma’ which has no exact real world counterpart.
 
I am sometimes asked how, as Keynes’s biographer, and someone sympathetic to his views, I can support an approach that downplays the importance of national autonomy in monetary and fiscal policy. The answer is that Keynes never believed that policy could, or should, be read off from a logical theorem. He believed that the possibilities of effective thought and action were much wider than those suggested by a particular economic theory, and that the task of economic statesmanship was to effect a reconciliation between the claims of theory and those of commonsense.
 
I believe that he would have regarded the triilemma doctrine as an example of a logical dilemma, which policymakers should take into account, but not be governed by.
 
It is claimed that Keynes himself never believed that the system of fixed but adjustable exchange rates which he supported at Bretton Woods could work without capital controls: and indeed I quote him to this effect in my paper. However,we must be clear what he meant by capital controls.In the second and third drafts of his Clearing Union plan, dated 18 November and 15 December 1941, Keynes advocated that all central banks should retain the wartime machinery of capital controls, but argued that this need not preclude ‘open general licences…for capital transactions’ of ‘indefinite duration’ by agreement between the central banks concerned. (Keynes, 1980) In short, capital movements could be as restricted or free as circumstances allowed. He was particularly concerned with the need for Britain to exercise such controls immediately after the war, and more generally with the need to distinguish between movements of ‘hot money’ and ‘genuine new investment’ and between equilibrating and ‘speculative’ movements. When historians talk about growth of capital markets in the 1960s as a breach in the Bretton Woods system, they should remember that no permanent restriction of capital movements was envisaged by Keynes, or indeed by the American co-architects of the system.
 
As I explain in my paper, the ‘trilemma’ theorem is intended as a stylised summary of 20th century monetary experience. It purports to show that fixed-exchange rate systems are impossible to maintain consistently with government accountability to electorates for economic performance and freedom of capital movements. I have questioned this summary on two grounds, historical and theoretical. As regards the first, I do not think it is right to explain the collapse of the gold standard by reference to the growth of democracy; or the collapse of the Bretton Woods system by the revival of international capital markets. Both systems were brought down by exogenous political shocks. No system of fixed exchange rates could have survived the first world war and the bungled peacemaking which followed it, which contributed to the Great Depression. The Bretton Woods system was doomed when America put foreign policy considerations ahead of the defence of the $35 gold price.
 
These historical considerations reinforce a point I have long regarded as fundamental to the analysis of monetary systems: namely, that the explanation of their rise and fall cannot be abstracted from the state of international politics. But this is exactly the dimension which the ‘trilemma’ excludes.
 
Secondly, I have questioned Michael Bordo’s contention that ‘inflation targeting’ makes ‘exchange-rate targeting’ irrelevant as a policy goal for ‘mature’ economies. Bordo argues that if the main countries (or regions) have the same inflation rates, their nominal exchange-rates will tend to be stable, and change only in response to ‘real’ shocks, thus achieving the outcomes desired by the architects of Bretton Woods without incuring the costs of trying to defend a fixed-exchange rate system.
 
In rebuttal, I concentrate on three main points:
 
a.I question whether the actual inflation rates we have enjoyed in the last twenty years have been importantly the result of monetary policy.They seem to be much more outcomes of the changing ‘structure of the economy’ associated with globalization. As I put it: ‘it is not low inflation targets which have maintained low inflation; it is low inflation which has made possible low inflation targets’. This raises the question of the value of monetary independence. In Japan continued promises to print money to get inflation going have had little effect. ‘The Bank of Japan suffers from a chronic lack of anti-deflation credibility’ remarks one analyst, but he still assumes that the Fed. can pump up US inflation if it needs to. (FT 10/11 May,,2003)
 
b. I question whether inflation targeting is more credible than exchange-rate targeting. Inflation targeting is a purely national policy, involving no international commitments. (That is why so many prefer it.) A fixed-exchange rate system is more like an international treaty, in which the participating countries undertake obligations to each other in return for promises of support in difficult times and, ideally, on the basis of agreed escape clauses. There is a strong argument for saying that the second type of commitment is likely to be regarded as more credible than the first.
 
c. Bordo’s argument that with a common inflation rate, nominal exchange rates among the main countries should tend to remain stable, with changes in nominal rates tracking changes in real exchange rates only, is not borne out by experience. Volatility has been much greater than can be explained by inflation or productivity differentials. Against the view that nominal exchange rate changes are needed to absorb real shocks, it may be argued that many real shocks are produced by speculator-led nominal exchange rate changes.
 
For all these reasons, I agree with Richard Cooper that a ‘cost-benefit calculation for flexible versus fixed exchange rates will gradually alter the balance against flexibility, even for large economies’. I expect that some years hence, probably after another nasty experience of volatility, agreement will be reached to keep the exchange-rates of the three main currencies stable. Bretton Woods required the agreement of two countries. A new system requiring the agreement of three should not be much more difficult to set up, once the need is felt.