Next Saturday world leaders meet in Washington to discuss new rules for the global financial system (though little will be achieved with President-elect Obama absent). So far, thinking about this matter has scarcely got beyond calls for better banking regulation: a microeconomic issue that is doubtless important but misses the main economic plot. The Bretton Woods system of 1944 was set up to "promote a stable system of exchange rates". This system has gone. But any new agreement, will need to be equally ambitious in addressing the problem of exchange rates, because the prevailing "non-system" has played a major role in the wild credit boom that has led to the financial crisis.
The old system broke down because creditor countries such as Germany and France found it more convenient to accumulate dollar reserves than revalue their currencies. This enabled the US to run balance of payments deficits financed by printing treasury bills. But since these were ultimately exchangeable for gold, the system was unstable and came crashing down in 1971, leaving a world of floating exchange rates.
It is now east Asian countries, especially China, that have accumulated US treasury bills, paid out to finance current account deficits - recently running at $700bn a year. Asia's savings glut has its counterpart in the consumption glut of western states such as the US and UK.
This destructive conjunction can be traced to two interrelated projects of Asian governments. One was to accumulate US dollar reserves to insure against a repeat of the capital flight of 1997/8 and to avoid the humiliating conditions that the IMF imposed for rescue packages. The second was to keep exports growing rapidly to boost employment and growth. Exchange rate undervaluation, prohibited in 1944, was the policy weapon used. Asian governments intervened massively to buy dollars and resist market pressure for currency appreciation. Moreover, they "sterilised" their dollar purchases, preventing domestic price increases that would have eroded their export competitiveness. So balance of payments adjustment was blocked.
The excess Asian savings have been shovelled into housing bubbles in both the US and Britain - not directly, but by enabling our governments to pursue expansionary monetary and fiscal policies that stoked up credit expansions. In this way the global imbalances have contributed directly to the meltdown. For the past 10 years the US has, in effect, had no budget constraint. And what was true of the country as a whole was true of all those companies and individuals who piled up debt on the back of inflating asset prices. The IMF was marginalised because its central purpose, which was to prevent these huge imbalances from occurring, had been negated.
Two lessons follow from this story. First, a way must be found of meeting countries' reserve needs for crisis insurance, while avoiding the use of national currencies as international reserves. The simplest way would be to activate and build upon the IMF's ability to create special drawing rights as an international reserve asset. The IMF might also become an international asset manager, pooling countries' reserves and making them available to deficit countries.
Second, the global exchange rate system must help rather than hinder the correction of excessive imbalances. Ever since the IMF articles were amended in 1978, each country has been allowed unilaterally to choose any exchange rate regime that suits its goals and circumstances. We have ended up with a free-for-all that is radically flawed from a systemic viewpoint. The US and Europe are floating but the Japanese and Chinese currencies, as well as those of a number of other Asian countries, are undervalued, but closely tied to the US dollar.
Undervaluation may make sense for a small nation but it is dangerous collective nonsense if practised by a key country or by several significant countries. In the next quarter-century, as emerging countries catch up with the west, large exchange rate changes will be in response to fast productivity growth. Smooth realignments are unlikely if the free-for-all continues. It is crucially important, therefore, that the major countries agree on a common exchange rate system that promotes balance of payments adjustment. (Small countries tend to follow one or other major country.) A fixed rate between major currencies is one possibility, provided sterilisation is disallowed. But it would involve a loss of monetary independence - probably politically unacceptable as well as economically painful and inefficient. So the major currencies - the dollar, the yen, the euro, the yuan - will have to float. But unmanaged floating can lead to prolonged and manifestly insane exchange rate movements (for example the US dollar bubble in 1984/85) that can themselves cause macroeconomic instability.
That leaves only two realistic options. Exchange rates between major currencies could float in unmanaged fashion most of the time, but with occasional policy cooperation and coordinated intervention to prevent gross misalignments. Or more ambitiously, the major countries could decide to practise managed floating of a structured kind. They could periodically agree on exchange rates that are appropriate for global adjustment, intervention being permitted only to influence market exchange rates in the direction of the agreed rates. It is no good relying on the IMF. The organisation has to be led by a group of key economies that have significant weight in the world economy (currently the US, Europe, Japan and China).
The most we can expect of Saturday's conference will be an agreement to increase the liquid resources at the disposal of the IMF. Exchange rate reform will have to wait on the end of the crisis. But it is important to start thinking about it now.
Robert Skidelsky and Vijay Joshi, a fellow of St John's College, Oxford and a professor emeritus of political economy at Warwick University respectively.