Transaction costs are minimized by using a common currency in the same way that communication costs are minimized by using a common language. The permanent transaction costs savings are likely to dominate the one-off changeover costs in the eyes of all but the most myopic government.

Printing non-interest bearing currency can be a cheap source of finance for the state but, for example, the monetary revenue extracted by the UK government is negligible – just 0.15% of GDP in 1993–4. If it joined economic and monetary union (EMU), the UK would be entitled to its share of the monetary revenues of the European Central Bank, which could well exceed the amount raised independently today.

Moreover, the unanticipated inflation tax – the capacity to impose unexpected capital levies on holders of fixed interest sterling government debt – has no place in the arsenal of a reasonably well-meaning government managing a tolerably well-functioning economy.

The macroeconomic management argument against monetary union is that national monetary policy is a valuable stabilization instrument. Those who are unimpressed by this argument make two points. First, national monetary autonomy is only a potentially useful stabilization instrument – in the hands of malevolent authority, it can destabilize as well as stabilize.

Second, even its maximum potential usefulness has been overstated by opponents of monetary union. EMU irrevocably fixes the nominal exchange rates of the member currencies vis-à-vis each other. An exchange rate is the relative price of two currencies. The ability to manipulate the nominal exchange rate or to allow it to respond to market forces is of no significance whatsoever unless it has an impact on the real exchange rate, the relative price of domestic and foreign goods or the relative price of traded and non-traded goods.

The ability to set the nominal exchange rate is without value unless there are nominal rigidities, that is, rigidities in money costs or prices that prevent domestic money costs or prices from responding to shocks to the economy. Theoretical considerations and empirical evidence suggest that if such nominal rigidities exist, they are strictly transitory or temporary.

There is no long-run trade-off between inflation and unemployment. Within the range of inflation rates historically experienced by the UK, there is no relationship between inflation rates and real growth rates. Real competitiveness cannot be improved in the long run by devaluing or depreciating the nominal exchange rate.

But how long is the long run? A stylized fact for the US, the national economy with probably the highest degree of money wage rigidity, is that after two years a monetary shock will have worked its way through to the general price level and other nominal variables. The UK has less nominal cost and price rigidity than the US, and continental West European countries less again. For real rigidities, which are not affected by monetary and exchange rate policy, the ranking of the US, the UK and continental Western Europe is reversed.

With long-run real economic performance essentially independent of the exchange rate regime, the benefits and costs of exchange rate flexibility are restricted to a transitory impact on the adjustment to real and monetary shocks. Here nominal exchange rate flexibility can be either a limited blessing or a limited curse.

The good news about exchange rate flexibility, when there are short-run rigidities in money prices or costs, is that it permits adjustments in international relative prices and costs that are warranted by fundamental real developments and shocks to be achieved more quickly and with smaller transitional costs.

No doubt it is easier and quicker to achieve a necessary improvement in real competitiveness though a nominal devaluation, with unchanged paths for domestic and foreign money costs, than by keeping the rate of domestic cost inflation below that of one’s competitors with a fixed exchange rate. In both cases, however, the same long-run real adjustment has to be made. The example of the Netherlands in recent years shows that it is quite possible to be tied to Europe’s strongest currency and to improve competitiveness while reducing unemployment at the same time.

The bad news about exchange rate flexibility when there are short-run rigidities is that it will cause purely monetary disturbances (shocks to money demand or supply), resulting in temporary changes in international relative prices and costs that are unnecessary and unwarranted by the underlying real fundamentals.

The common assertion that asymmetric or country-specific shocks make exchange rate flexibility desirable is therefore only half true. If the shocks are real shocks to consumption, investment, public spending or production, exchange rate flexibility helps minimize transitional pain. If the shocks are monetary shocks, a fixed exchange rate provides, under a high degree of international capital mobility, the ideal mechanism for absorbing and neutralizing these shocks.

Since the exchange rate regime only makes a transitional difference to real economic performance, it is not true that a large-scale fiscal redistribution mechanism among EU members is required to make up for the loss of national monetary and exchange rate independence. All that is required would be a scheme for effecting transitional, self-liquidating transfer payments – nothing like the EU’s regional or cohesion funds, which engage in quasi-permanent redistribution. Likewise, the only kind of international labour mobility required to make up for the loss of exchange rate flexibility would be a strictly temporary, reversible mobility.

My next point may come as a shock both to Mr Eddie George and to Mr Tony Blair. The Governor of the Bank of England and the Leader of the UK Labour Party both misunderstand the relevance of convergence in real economic performance for the desirability of monetary union. The truth is simple: real convergence or divergence is irrelevant for monetary union. As an illustration, over the past 15 years there has been significant real divergence between the economic performance of the rich and the poor in both the UK and the US: income and wealth inequality have increased markedly. It would have been no different had the poor and the rich in the two countries each been given their own currency in 1980.

Does anyone really believe that the problems of Italy’s Mezzogiorno would have been alleviated if southern Italy had been given its own currency after World War II and had decided to float the southern lira? There is no reason whatsoever why regions characterized by persistent differences in productivity growth or even by persistent differences in real earnings growth that are unrelated to productivity growth differentials could not or should not be locked together in a common currency area.

The same holds for regions with different burdens of non-wage labour cost, different demographic structures and different saving rates, etc. The real economic performance of a region where real earnings growth systematically outstrips productivity growth would be dismal with a common currency or with a floating exchange rate.

The key point to remember when considering the choice of exchange rate regime is that it ‘only’ concerns monetary policy. Monetary policy can do transitional good or transitional harm: it is not the stuff of which the wealth of nations is made. Central bankers, city financial analysts and opponents of EMU like Sir Alan Walters or Bernard Connolly tend to forget that.

It is up to those of us who are not professionally blinkered to point out that the monetary emperor wears no clothes - in the long run, at any rate. If there is a Labour government in the UK after the next election, it could save itself anxious months or even years of being tested by the international financial markets by closing the door on national monetary autonomy and throwing away the key. Two cheers for EMU.