Two routes to full employment.

Britton, Andrew

The unemployment figures now show nearly three million people out of work. No one doubts that this represents a waste of resources on a vast scale and a social problem of prime importance. Nevertheless, despite very real public concern, it cannot be said that full employment is now an objective of government policy, neither was it given a high priority by any of the political parties at the general election last year. A common assumption, implicit or explicit, is that mass unemployment is here to stay--and the humane response is to make the condition easier to bear.

For the last ten years at least the rise in unemployment has been accompanied by subtle changes in public attitudes to work. On the one hand those who have jobs seem powerfully motivated to keep them: they seem to work much harder than they would have done twenty years ago and with the single-minded purpose of individual economic gain. On the other hand there are those who question the work-ethic: there are those who learn to play the welfare system, those who make a virtue of 'dropping out', and those who accept early retirement. Both those two extremes can be seen as reactions to the experience of mass unemployment, contrasting with the old view that work is both a right and a duty. Customers enjoying improved standards of courtesy, managers appreciating a more co-operative workforce may have concluded (privately at least) that unemployment, or the fear of unemployment, was not without advantages. During the recent recession however attitudes seem to have changed again. As the reality of unemployment has been brought home to more of the population, pressure has again built up for 'something to be done about it'.

Chart 1 shows the percentage unemployed in this country over the past hundred years (although not quite on the same definition, unfortunately). Historical experience suggests that full employment need not be an unrealistic goal. Prior to the first world war unemployment was highly cyclical with an average rate around 4 per cent and no clear trend. Between the wars unemployment was very much higher, with a peak rate of about 15 per cent in 1932. But after world war two full employment was maintained continuously until about 1970. It is only in the last twenty years that the trend has been rising.

Table 1 shows the unemployment figures for a selection of advanced industrial economies in 1971, 1981 and 1991. The large increase in the UK figures is broadly matched by those for several other countries, including France, Belgium, Canada and Australia. But the pattern is by no means universal. The figures for the United States show a much more modest increase. Moreover, in a number of countries, unemployment at the beginning of the 1990s was far lower than it is in Britain, for example in Japan, West Germany, Austria, Portugal, Sweden and Switzerland. This suggests that mass unemployment is not the inevitable result of modern technology or of a free market system.

Table 1. Standardised unemployment rates

 per cent

 1971 1981 1991

UK 3.6 9.8 8.7
US 5.8 7.5 6.6
Japan 1.2 2.2 2.1
Germany 0.9 4.2 4.3
France 2.7 7.4 9.5
Italy 5.3 7.8 9.9
Canada 6.1 7.5 10.2
Sweden 2.5 2.5 2.7
Australia 1.9 5.7 9.5

Source: OECD.

This note will consider what would need to be done to restore full employment in Britain. If this were the only consideration one could use conscription or abolish benefits. The real problem is to identify routes to full employment which are consistent with the kind of economic system and welfare provision we take for granted. Full employment will not be achieved without any changes to the structure or institutions of the economy; the question is the scale of reform which society would accept as the price of full employment.

It is useful to distinguish two approaches to the problem. The first, or narrow, approach concentrates on the unemployed themselves, their skills and motivation, the conditions on which benefits are available to them, the extent of their effective participation in the labour market. The second, or broad, approach looks at the state of the economy as a whole, the international competitiveness of British industry, the setting of wages and prices, the education and training of the labour force as a whole. The narrow approach focusses on labour economics and puts the Department of Employment in the lead; the broad approach is macroeconomic and suggests that the Treasury and the DTI must take the main responsibility.

The rest of the note is organised as follows: first the concepts of full employment and the natural rate, or NAIRU, are compared; then some calculations are presented of the growth rate of output and the time scale that would be required to achieve full employment. Then an attempt is made to assess the merits and practicalities first of the narrow approach and then the broad. The two approaches are not mutually inconsistent, and a final section considers how they might best be combined.

Full employment and the natural rate

William Beveridge defined full employment as a situation where 'those who lose their jobs must be able to find new jobs at fair wages within their capacity, without delay' ('Full Employment in a Free Society', page 20). 'Idleness' he added 'is not the same as Want, but a separate evil, which men do not escape by having an income. They must also have the chance of rendering useful service and of feeling that they are doing so'. This is a social objective which will not necessarily be achieved at the point of equilibrium between the effective supply and demand for labour. Indeed all the evidence of the last twenty years suggests that at the point of equilibrium the rate of unemployment has been much too high to fit Beveridge's description of full employment.

It is easy to speak loosely of an equilibrium point, or a natural rate of unemployment, but it is much more difficult to define it precisely or to provide numerical estimates. One approach is to define the natural rate (or NAIRU) as the level of unemployment which would now be consistent with a constant rate of inflation. If that definition were taken literally then we must now be close to the NAIRU since the rate of inflation has been roughly constant for the last few months. In practice something more subtle is required. An alternative definition would be the sustainable level of unemployment, that is to say the level which would be compatible on average over the next few years with the rate of inflation staying broadly constant (or perhaps staying within a target range).

The original Friedman definition of the natural rate of unemployment did not refer just to constant inflation, but to a rational expectations general equilibrium for the economy as a whole. But to make that concept operational one needs to know what time-scale of adjustment is involved. Is this a short-run or a long-run equilibrium? For example is it the level of unemployment possible with the existing capital stock or that which would be made possible by full adjustment after some years of additional investment? For practical purposes Friedman's reference to 'the Walrasian equations' does not take us very far forward.

As the economy emerges from a long and deep recession, the actual unemployment rate is almost certainly above the rate which could be sustained in the medium term, but perhaps not very far above. Unemployment last reached three million in 1986, and then fell sharply, during a period of exceptionally rapid growth to just 1 1/2 million in 1990. It might seem that all we need is to repeat that experience, and perhaps continue the boom a little longer this time to reach full employment. Unfortunately, the history of the last few years strongly suggests that this approach would not work. Despite the 'supply side' reforms of the 1980s the evidence since then suggests that the sustainable rate of unemployment has not been reduced. The most obvious signal that the economy was over-stretched at the end of the 1980s was the deterioration in the balance of payments. An increasing proportion of firms were constrained by shortages of plant or skilled labour. The rate of inflation speeded up. The recession which followed may be seen as a necessary correction (an over-correction in fact) to the over-heating of the economy. Following such a violent cycle it is difficult to know what level of unemployment would now be sustainable, but recent history suggests that the answer lies somewhere between 1 1/2 and 3 million; it may well be in the upper half of that range.

The labour market obviously does not in fact reduce the level of real wages to the point where the unemployed can all find a job 'without delay'. In the Institute's macroeconomic model this impasse is written into the equations. One interpretation is that the labour force, or the unions that represent them, will press for unrealistic wage increases unless the fear of unemployment acts as a discipline. Another interpretation is that many of the unemployed do not really count in the balance of the labour market, because employers do not think that they are suitable to fill the vacancies which exist.

The output increase implied by full employment

We define full employment as recorded unemployment of about one million or less. In the post-war period unemployment was typically less than a half of a million, but that might not be an appropriate figure to aim at now. Some frictional unemployment is unavoidable; as the economy develops old jobs are destroyed and new ones created, and some job changes will inevitably involve a short spell of unemployment. The economy may now require more flexibility of this kind than it did in the 1950s or 1960s. It is easier now both to hire and to fire. Firms are less willing (or less able) to support employment levels in excess of their immediate requirements. The same is increasingly true of the public sector. There will also be some cyclical unemployment on average over a run of years unless the economy becomes much more stable than it has been since the 1970s. An average of one million unemployed over the cycle would imply just a few hundred thousand at the peak. A level of one million unemployment would amount to about 4 per cent of the labour force. It would put this country towards the low end of the ranking of OECD countries, but still above Japan, Austria, Iceland, Luxembourg and Switzerland (on 1991 figures).

Another way of characterising full employment is in terms of the duration of unemployment. Beveridge spoke of new jobs being found 'without delay'. Currently, as it happens there are about one million people who have been unemployed for less than three months. No doubt the duration profile would change if the labour market picked up, but one consequence of reducing total unemployment ought certainly be to reduce the average duration substantially. (It is often argued that the long-term unemployed, since they are relatively unlikely to find employment, do not exert any 'useful' downward pressure on real wages. This appears to be the case, although the fear of becoming unemployed for a long spell must be a more powerful deterrent to rash pay claims than the fear of becoming unemployed for a few weeks.)

To reduce unemployment from nearly three million to about one million would, of course, imply a substantial rise in output. The scale of that rise, and its consequences for other macroeconomic variables, can be explored by constructing alternative scenarios for the future, using the Institute's model. The rise in employment would be more than the fall in unemployment as the higher demand for labour attracted more workers into the labour force. The percentage increase in output would be less than the percentage increase in employment if the productivity of the unemployed who found work was less than the average. On the other hand some high productivity jobs would have to be created as well, in particular to make the exports needed to match the extra demand for imports.

Table 2a below shows the base projection. This is not a forecast so much as a smooth extrapolation of trends which the model has identified in the data for the past. The economy expands at about 2 per cent a year, with unemployment at about 2.7 million and inflation about 4 1/2 per cent a year.

Table 2b shows the very different outcome made possible by a much lower sustainable unemployment rate. The model has been changed so that the sustainable rate falls steadily to reach 1.2 million or 4 per cent by the year 2001. This is done simply by 'doctoring' the wage equation. There is no attempt to specify the policies which would be required to bring this about. Interest rates and exchange rates are assumed to remain the same as in the base case.

The model indicates the way the economy would generate the demand necessary to bring actual unemployment down. The main route is through a reduction in the growth of wages which, with a fixed exchange rate, improves competitiveness and hence increases exports. Real wages remain below their 1992 level until 2003, by which year they are 13 1/2 per cent below the level in the base run. Initially both consumption and investment are cut back, because real incomes are lower and real interest rates are higher. Once the adjustment is complete the economy follows a new growth path with output about 9 per cent above base, and investment now substantially TABULAR DATA OMITTED higher so as to build up the required stock of fixed capital. Employment has been increased by about 2 1/2 million and unemployment reduced by about 1 1/2 million. This large increase in activity has taken place spontaneously according to the model without the need to assume more expansionary fiscal or monetary policies. There is no balance of payments problem, because TABULAR DATA OMITTED growth is export led, following the fall in nominal earnings.

The simulation illustrates the narrow approach to achieving full employment, for example by enabling (or requiring) the unemployed to compete more effectively for jobs in the labour market. The broad approach would produce a different scenario, because it would involve reform of the economy as a whole making the whole labour force more productive.

Table 2c illustrates the consequences of achieving full employment by the broad route. The employment equations on the model are changed so as to reduce the number of workers required to produce a given level of output by some 5 per cent. That on its own would not reduce the sustainable level of unemployment in the model at all--in fact it would increase it. The exports equations are also changed so that a higher volume is sold at any given level of relative prices. This could represent some 'non-price competitiveness' gain. The wages equation is also changed, as it was for Table 2b, so as to make the same level of unemployment sustainable.

The prospect now is much brighter, perhaps implausibly so. Output growth for the first five years averages 4 1/2 per cent, an extraordinary recovery by any UK standards. This means that real wages can rise substantially over that period instead of falling as they do in Table 2b. (The time path for real wages is different from that in the base run, but their level from about 2003 onwards is virtually the same.) Inflation is higher than in Table 2b, but still lower than in the base case. The problem of the balance of payments has been solved by the assumption of a much better export performance--in the absence of such an improvement the real wage would have had to fall so that relative price competitiveness could improve, and that would have made the fall in unemployment more difficult to achieve.

In showing what success would look like, these scenarios TABULAR DATA OMITTED assume that the aim of full employment is attainable. They suggest however that the process will have to be a slow one. The narrow approach, concentrating on the unemployed themselves, would require actual cuts in average nominal wages, if the time scale for the return to full employment was too short. The broad approach would require impossibly rapid growth of output, and especially of exports, if the timescale was too compressed. The simulations allow nine years for the reform to take place, in both the narrow and broad approach, with a further seven years or so before the target is approached. It would be useful to bear this minimum timescale in mind when considering the two approaches in more detail.

The narrow approach

We now have a great deal of experience of special employment measures in this country. The list of schemes gets longer and longer, many of them being variations on the same themes. They involve some combination of retraining, assistance with searching for work, selective employment subsidies, temporary work in the public sector and a tightening up of the conditions for receipt of benefit. At the most, the schemes may have reduced unemployment by a few hundred thousand each year since the mid-1970s.

Proposals to introduce similar schemes on a much larger scale run into problems of administrative feasibility. When unemployment is low it is possible to consider each case individually, to direct each claimant to the appropriate job vacancies or training schemes. If relatively small sums are being spent on job creation or selective subsidy in the private sector it does not matter too much if some of the money is not very well spent. When the numbers in receipt of training allowances are relatively low, the allowances can be quite generous. But if literally millions of cases are involved then the task becomes unmanageable, and the expense prohibitive. We experienced this situation in this country in the 1970s. A contemporary, and more sudden, change of the same kind can be observed in Finland, where a successful scheme of employment creation has been swamped by the numbers who have lost work as a result of the collapse of export markets in the old Soviet Union.

This does not mean that the narrow approach to full employment is impossible. Some combination of subsidy with compulsion could probably be devised which would result in the creation of even two million low-paid jobs, although there would be no question of giving individual attention to the needs or the abilities of two million 'clients'. The most perplexing issues concern the social implications of such a large-scale scheme. It involves creating a large low-productivity sector in the economy, perhaps a regimented sector, rather clearly differentiated from the rest. Those who worked in that sector would tend to be uneducated and unskilled; they would be poor relative to the rest of society and many of them would still be dependent on public support (even if it was indirectly through a subsidy to their wages as in the new 'Workstart' scheme). The segregation of the labour market is most obvious with a 'Workfare' system, but applies to some extent to any introduction of permanent special employment measures on a large scale. The likelihood must be that many individuals will be permanently (or at least frequently) employed on such schemes. This is not at all what Beveridge had in mind in defining full employment. One kind of social marginalisation and alienation could simply be replaced by another.

The broad approach

Those who follow the broad approach argue that the whole economy must be strengthened and reformed before full employment can be restored. The situation could be likened to that of Eastern Europe, although happily the adjustment needed is not on the same scale. The point of similarity which might be argued is that unemployment in Russia or the eastern lander of Germany cannot be tackled effectively except in the context of 'root and branch' reform. The British economy needs to be made more efficient and competitive so as to be able to participate in the open markets of Europe or the world as a whole.

One merit of the broad approach is that the necessary wage restraint (the doctoring of the wage equations in the model simulations) seems more likely to be forthcoming in the context of rising wages and rising living standards. The weakness is the scale of the improvement in competitiveness which this approach requires. The immediate and direct consequence of higher output per head is to reduce employment not to increase it. That effect must be offset by higher output (mainly of exports) before any reduction in unemployment can occur.

But is the broad approach feasible in practice? And how could productivity be raised? All we can do is offer some suggestions as to what might or might not be important as ingredients in a strategy with the aim of reaching full employment by this route. A series of studies by the Institute has suggested that one of the main reasons for the differences in output per head and in quality of production between Britain and Continental Europe lies in the proportion of the labour force with vocational qualifications and training. This gap could probably be closed, since there is no reason to suppose that there is any difference in innate aptitudes; but of course it would take a long time.

The first aim would be to improve standards of secondary education and vocational training for everyone, not just those most obviously at risk of unemployment. Here Continental practice could act as a guide, although this might involve the sacrifice of other educational objectives. The employment of school-leavers in unskilled jobs could if necessary be stopped, either by regulation or by incentives. Even so there would be a need for new investment, research and marketing if firms were to take the opportunity to move up-market created by a better-trained workforce.

An interventionalist government could use the reform of training as an opportunity to influence the structure and efficiency of British industry. The government would provide or subsidise training in the skills it thinks will be needed in the future, and then channel the skilled labour into the firms which were prepared to take the opportunities for innovation which had been created. This would require a break with the policy of disengagement which has characterised industrial policy since the 1980s. It would open the process of industrial development to political pressure in a way which might prove a serious handicap. The alternative would be to leave the direction of advance to market forces. Competition between firms should ensure, eventually, that the skills are used effectively. This process might be helped by the inflow of foreign capital and by the continued transfer of Continental or Japanese production methods to this country.

At one time it was widely believed that full employment could be achieved if the influence of trade unions was reduced sufficiently and if bargaining was conducted more at a local level. This was one motive for the employment legislation of the 1980s. It may be too soon to be confident of the effects of that legislation, but it appears at this stage to have made the labour market more flexible, so that unemployment can rise and fall more rapidly--but it does not seem to have made a lower average level of unemployment compatible with price stability.

In the past prescriptions for full employment (from a different school) have often included an element of wage restraint, incomes policy or national bargaining. The experience of the 1970s is mixed and may not be a good guide to policy now. The reform of employment law and negotiating practice during the 1980s has changed the British labour market fundamentally. The move to workplace bargaining and the fall in union membership has largely destroyed the institutional basis on which incomes policies were once built. It is still essential that the implications of pay for jobs are appreciated and taken into account in bargaining. This requires the cooperation of workers and their representatives, at least at local level, which is more likely to be forthcoming if pleas for wage moderation are made in the context of plans for growth and improvements in training (the 'broad' approach rather than the 'narrow'). But international experience suggests that a formal incomes policy is neither necessary nor sufficient for full employment.

An assessment of the two routes

In the 1930s Keynes presented his solution to the problem of unemployment as a triumph of the creative intellect over prejudice and ingrained habits of thought. This note makes no such claim. The two routes to full employment described here involve no innovations of economic theory, just the application of familiar ideas to familiar problems, with a shift of emphasis giving much greater priority to the goal of full employment than it has received for many years. If the shift of emphasis is sufficient the problem can perhaps be solved, but progress will be neither easy nor quick.

The narrow approach on its own would risk creating a permanent underclass; the broad approach on its own would take several decades and might initially raise unemployment rather than lower it. The best strategy would involve both approaches at once. Initially a much expanded system of special employment measures would be introduced, reducing unemployment and also the growth of real wages. Meanwhile better education and training for the population as a whole would improve productivity and competitiveness. New jobs would be created which would be filled by the newly trained workforce. This would leave vacancies which the unemployed would fill. Thus the unemployed would find 'real' jobs indirectly as a result of a general economic expansion. Eventually the special employment schemes could be run down.

The timetable required in total is probably of the order of ten to twenty years. This suggests that the strategy would need bipartisan political support if it was to be successfully brought to a conclusion. It is not clear that the objective has sufficient public support for it to be given the priority it would need. But the experience of the last few years does at least re-open the question. The possibility of full employment can still capture the imagination if only as a distant goal.