Time to review Maastricht blueprint on convergence

Even those in Britain who favour a single European currency have recently been arguing that the convergence criteria included in the Maastricht treaty are insufficient to ensure success. From both sides of the House, and from the Bank of England, there have been calls for "real" conditions to be added to the Maastricht list. But what are these extra criteria, and are they really needed?

Maastricht is very much a creature of its time. It assumes that monetary policy can determine the inflation rate independently of the level of output in the long run. If this were true, the European central bank could aim simply to achieve a low and stable inflation rate across the Union, and could leave both output and employment to reach their equilibrium levels. These would be determined by supplyconditions in the economy and by the structure of the labour market - factors unaffected by monetary union.

Several things follow from this. First, because countries cannot permanently solve an unemployment problem by adopting an expansionary monetary policy, they are not abandoning anything important by ceding monetary policy to a supra-national entity. Second, because monetary policy affects inflation and nothing else in the long run, countries might as well cede control of that policy to an entity that would pursue a low inflation objective with maximum credibility. For all members of the Union, other than Germany, there is nothing more difficult to reverse, and perhaps therefore more credible, than signing a treaty that puts monetary control outside their own direct jurisdiction.

Third, in the Maastricht world, there need be no convergence in unemployment rates between countries, since either inside or outside a monetary union, differences in unemployment would be determined entirely by structural factors in the labour market (for example, the level of unemployment benefits).

Fourth, if the Maastricht philosophy holds, then countries cannot make permanent gains by devaluing their currencies since any decline in the exchange rate will be quite rapidly washed away by a rise in domestic prices, leaving the real exchange rate unaltered.

Evidence in favour of this comes from the fact that real exchange rates within Europe have tended to remain broadly unchanged over long periods of time. As a result many economists have concluded that governments are not losing any genuine degree of freedom by joining a single currency. (In the past, I have fallen for this trap myself, but I now think it is wrong, for reasons set out below.)

The official convergence criteria included in the Maastricht treaty are based directly on these beliefs. The criteria require that countries should have similar inflation rates; that interest rate and foreign exchange markets demonstrate that inflation convergence is permanent and credible to the private sector; and that governments are debarred from putting strains on the monetary authorities by running excessive fiscal deficits.

The treaty does not require that unemployment rates, or the behaviour of output, need be convergent across the Union before the single currency can be introduced. In the Maastricht world unemployment and output are both assumed to adjust to a common monetary framework, so there is no need to mention them specifically. Other countries seem content with this approach, but Britain has become increasingly concerned about it. And Britain is right.

It would be quite easy to design "real" convergence criteria that would parallel the "nominal" criteria in the Maastricht treaty, and the accompanying table takes a crack at doing this. By and large these real criteria would exclude from the single currency countries like Italy, Spain, Sweden, Portugal and Greece - countries that are anyway having difficulty passing the official Maastricht tests. But criteria such as these are inevitably arbitrary, and anyway might miss the main point.

The key point is this. It may be true that economies in a given year are convergent in both real and nominal terms. It may also be true that, in the very long run, output and employment are both determined by non- monetary factors, and that the real exchange rate is fixed. Nevertheless, there can be an important role for changes in monetary policy to stabilise each of the individual economies in the Union in the face of "temporary" shocks. The word "temporary" in this context should not be taken to mean small, or even short-lived. A temporary shock might, for example, include German reunification, or the housing-induced boom-bust cycle in the UK in the late 1980s.

Outside a single currency monetary policy can be adjusted to respond to such shocks, and this can lead to short-run deviations in exchange rates from their "real" equilibria. The decline in sterling against the mark since 1992 is a case in point. Interest rate and exchange rate adjustments of this type can cushion the level of output and unemployment from the worst effects of the shock, so that the real economy is more stable than it would be in a monetary union. In the very long run the real exchange rate may move back onto its original path, but that does not negate the value of allowing it to adjust for short periods in order to stabilise the system.

Of course, these arguments only apply if the type of shock that is likely to hit the system will have different effects in different countries. If this is not the case, then a monetary policy set for the entire Union will work perfectly adequately, since the required monetary response is common to all countries. As Patrick Minford has correctly pointed out, this depends not on the degree of convergence between economies, but on the underlying degree of integration in their industrial structures.

For the core group of northern economies (Germany, France, Benelux and, excluding the oil sector, the UK) the industrial structures are so similar that most shocks may be expected to have similar effects (though as our examples demonstrate, this is by no means assured). But even this does not establish that the economies are sufficiently well integrated to justify a single currency. At least two other issues need study.

First, even assuming that the first round impact of a shock on output and unemployment will be identical, the second round response of wages and prices to the change in output may be very different, and this may justify a different response from monetary policy. So not only do real economies need to be integrated, but the flexibility of their inflation mechanisms need to be integrated as well. Second, even if the output shock and the inflation response is identical across countries, the responsiveness of each economy to the same change in interest rates might be very different. For example, in the UK, the housing market is hugely affected by changes in short term interest rates, while in Germany and France, there is almost no impact. This means that the UK economy may be affected much more than other European economies by the same change in interest rates. Again, different interest rate responses would be appropriate.

These considerations have been swept under the Maastricht carpet, and our partners in the Union seem content to keep it that way. The UK has a responsibility to get them out into the open.