There is no simple explanation or cure for high unemployment in Europe, argue the authors of a recently published CEPR Report, Unemployment: Choices for Europe. Popular explanations, such as new technology, foreign competition or 'Eurosclerosis' do not provide convincing reasons why unemployment should have risen so dramatically, though all of them may have had some part to play. No single factor can fully explain why so many Europeans fail to find jobs, and many of the more widely touted proposals for reform are either likely to be ineffective or politically difficult to introduce.

For example, according to the Report, the notion that the developing countries have in some sense `stolen our jobs' is quite simply wrong: access to a cheaper source of some goods should allow Europe to shift resources into other lines of production in which it has a comparative advantage. Another common claim is that Europe's high unemployment is the consequence of an unduly rapid rate of technological change, associated with advances in information technology. The Report rejects this argument: Europe's problem is, if anything, too little, not too much technological change.

Historical evidence clearly refutes the idea that jobs are permanently destroyed by new technology: productivity today is ten times higher than in 1900, but the unemployment rate is not that different. This is because technical progress also opens up possibilities for new jobs elsewhere in the economy. But a high level of such `creative destruction' could be associated with more unemployment as more of the labour force is reshuffled between sectors. The data suggest, however, that there is a significant inverse relationship between the amount of job reallocation and average unemployment in the 1980s; so a high rate of job reallocation need not imply a high unemployment rate.

The Report argues that it may not be an unduly rapid rate of technological change that is the problem, but rather the reverse. The rate of growth of productivity and output is in fact much lower today than it was during the pre-1973 `Golden Age'. Such a slowdown reduces the returns from creating new `job slots' and thus raises unemployment. The evidence suggests that countries which experienced a greater deceleration in productivity growth also experienced a greater rise in unemployment. Since the high European growth rates after 1945 were a only transitory phenomenon (associated with reconstruction), it was inevitable that the historically low unemployment rates of the 1950s and 1960s could not last.

The expected length of unemployment spells is much higher in Europe, as is the percentage of long-term unemployment. Is this the result of Europe's relatively generous welfare state? The Report notes that despite Europe's more generous welfare provisions, unemployment rates in Europe were half US rates in the `Golden Age'. The European welfare state did not suddenly become more generous in the 1970s, and so its generosity cannot be the prime mover behind the rise in unemployment. It does, however, help to explain the persistence of high unemployment long after transitory disturbances (such as the oil price shocks) have disappeared.

It is often argued that the source of Europe's unemployment problem is the high level of job security and the consequently `sclerotic' nature of its labour markets. Many proposals to cure European unemployment, including the recent OECD Jobs Study, call for extensive deregulation to create US-style `flexible' labour markets. The Report notes that the reality of European labour markets is actually rather different. One in every six or seven jobs are created or destroyed each year, a rate only slightly lower than for the US and Canada. Furthermore, cross-country comparisons of average unemployment in the 1980s against an index of the strength of job security (a measure of rigidity) reveal no systematic relationship between the level of unemployment and rigidity. Countries with more `flexible' labour markets do, however, experience more variability in unemployment.

Proposals such as those in the OECD Jobs Study would, if fully implemented, change the face of European labour markets. But could they ever happen? The Report identifies three reasons why labour market reform is politically difficult to implement in democratic societies. First, there may be more losers than winners. The majority of the electorate are employed, and policy is in practice determined by their interests. Many of the rigidities that are thought to contribute to Europe's high unemployment (minimum wages, generous benefits, high firing costs and strong unions) actually benefit the employed and are therefore difficult to remove. Second, labour market reform may be difficult to implement because it is likely to benefit a very heterogeneous group, including shareholders, skilled workers, small entrepreneurs and unemployed workers. These people may have a common interest in labour market flexibility, but divergent interests on almost all other issues.

Uncertainty about who will benefit and who will lose from reform will lead to bias towards the status quo. Skilled workers and the unemployed with particularly poor prospects of finding a job know for certain that they will gain from reform: the latter will find a job sooner, and the former will enjoy higher wages because of the increase in unskilled employment. The unskilled, however, are far less certain to gain because their wages will fall and/or they might be forced to relocate to other sectors. The same is true of the unemployed with good prospects of finding a job quickly, who would not like to see the wage in their next job fall. If uncertainty about the impact of reform makes the gains too diffuse for the unskilled and the better-placed unemployed, then they might end up opposing reform.

The Report provides a simple analytical model to illustrate who is likely to gain or lose from market reform, and whether there is likely to be sufficient support to make reform politically viable. The population is divided into five groups according to income (and skill level) and the model is calibrated so that it initially replicates the current French distributions of unemployment rates and income across these groups. The aggregate unemployment rate before the reform is assumed to be 11.7 per cent, falling to 6.4 per cent for all skill levels following reform; the reform is also assumed to lead to a US-type distribution of income. The model suggests that support for reform is closely balanced, since slightly less than half the labour force (44 per cent) benefit. It also shows that reform generates large gains and losses, thus provoking strong opposition from the losers, who are more likely to mobilize. The stakes are also rather small for the decisive voters, generating a bias towards maintaining the status quo.

This analysis suggests that `the European model' is a political equilibrium which is very difficult to change. It need not, however, remain an equilibrium indefinitely. Just as the Swedish corporatist experiment collapsed as circumstances changed, so might the European model. But until such a time, it is pointless to talk about sweeping changes. The best that can be hoped for are incremental changes that improve the functioning of the European economy, while still commanding enough support to be politically viable.

The Report advocates a set of reforms which appear to be both politically feasible and economically desirable:
* Where possible reduce the duration for which (but not the level at which) unemployment benefits are payable.
* Erode the relative value of the minimum wage, particularly for young workers, and use in-work benefits to tackle poverty instead.
* Where the minimum wage cannot be reduced, restructure the payroll tax so as to reduce the cost of employing unskilled labour.
* Foster competition in product as well as labour markets.
* Improve training, but do not expect too much from it.
* Offer the long-term unemployed the option of turning their benefits into employment vouchers.
* Ensure that macroeconomic policies support supply side measures.