An Analysis of the Causes of Unemployment
Cloda Lane (Senior Sophister)
The current unemployment crisis in the advanced capitalist countries is primarily a European phenomenon. The rise in the EU unemployment rate from 3 percent in 1973 to more than 10 percent today has been in marked contrast to the experience in the rest of the OECD. Cloda Lane investigates the numerous causes of unemployment and analyses the regional disparities.
Throughout the OECD, there has been a dramatic rise in unemployment in the 1980s and 1990s. The problem has been more severe in some countries than others and has not been evident in a minority of economies (notably Japan and the US). There is thus an increasing awareness of the need to tackle this serious economic, social and political problem. Before successful policies can be formulated and introduced, however, there is a need to be fully aware of what the problem being treated is. In particular, an in-depth knowledge of the causes of the disease is needed so that the illness rather than merely its symptoms are treated. The task of determining the causes of unemployment is by no means a simple one but this essay will attempt to provide some guidance as to particular areas that policy-makers should be aware of and that economic theorists should be attempting to understand. Before analysing potential causes, however, it is important that the reader is aware of just how severe the unemployment problem has become.
Activities in the labour market are determined by the interaction of labour demand and labour supply. It is thus crucial that we examine how each of these variables have been changing over the past 15 years. The size of the labour supply is determined by the labour force and there has been positive labour force growth across the OECD in the 1980s. We can see from the table below that there has been positive labour force growth across the OECD in the 1980s. This has remained true for most countries up to 1994, Finland and Sweden being notable exceptions with average annual rates of labour force decline since 1991 of 0.7 percent and 1.3 percent respectively. Ireland had a relatively low growth rate of 0.1 per cent in the 1980s but this is expected to increase in the 1990s. The influence of labour supply on the market, while significant, does not explain the existence of unemployment. To do this we must also look at labour demand. The level of employment is the variable that best measures this component of the labour market. From table 1 we can see that there was positive employment growth in most OECD countries. Ireland stands out with a negative growth rate, however, indicating a fall in employment. The 1990s saw a swing in most countries towards a similar trend of employment growth resulting in a fall-off in labour demand. Japan is a notable exception as positive employment growth was maintained over the period. The disturbing picture emerging is that for many countries, particularly in the 1990s, labour supply growth has by-passed the growth in labour demand. This has automatically resulted in an increase in unemployment across the OECD, particularly in Europe.
Table 1 - Employment growth in the OECD area (annual percentage change)
Source : OECD (1993), Employment Outlook, Table 1.2
The full extent of the problem is clear from the table 2. Spain had the highest unemployment rate in 1993 (22.4 percent) with Ireland next in line with 15.8 percent. Most countries have witnessed a growth in unemployment between 1983 and 1993, the problem being most severe in countries such as Finland, Italy, France and Sweden. Even where unemployment rates have declined, the rates are still too high. In contrast, the United States has a declining unemployment rate due to a sharp decline in the rate of population growth, while, in the case of Japan, this decline was so great that unemployment remained low despite a marked slowdown in employment growth.
Table 2 - Standardised unemployment rates (per cent of total labour force)
Source : OECD (1994), Employment Outlook, Table K
Given the severity of the unemployment problem and, perhaps more importantly, given that it has persisted since the 1970s, it is important to determine the underlying causes of the problem. I will focus in this discussion on those causes which are commonly put forward as explanations of unemployment and attempt to see if they apply to the OECD.
Causes of Unemployment in the OECD To examine the causes of unemployment in the OECD I will analyse two distinct, but interdependent, phenomena. These are (a) structural rigidities that exist in the labour market and (b) the influence of a changing global economy on the labour market. Taken together, these two factors go some way to explaining why unemployment has risen and why it persists in the OECD.
In relation to the structural rigidities in the labour market I will examine the influence of the rigid real wage, inflexible labour supply and inflexible labour demand.
Rigid Real Wage
The most common theoretical explanation of unemployment is that the real wage is fixed in the labour market. This means that "price movements" will not adjust the excess supply in the labour market that the resulting unemployment will remain until the self-equilibrating process of the labour market is restored. The purpose of this analysis is to examine why the wage rate is sticky. There are two possible explanations for this. First, firms may not be able to adjust the wage downward because of pressure from trade unions and because of the way the wages are determined. The other explanation is that employers may maintain wages at a high, non-clearing level in an attempt to reap efficiency gains for the firm.
The prime reason for rigid real wages in the labour market is that insiders, represented by a trade union, bargain with the employers for higher wages. We would thus expect that the more unionised the economy the more powerful the trade unions will be and, hence, the more upward pressure will be put on wage rates and the higher unemployment will be. This argument seems to be supported by comparing the European and US experiences. In the former, unionisation has been rising or remaining constant since the 1950s but unemployment has been increasing. In contrast, unemployment in the US has fallen due to a decline in the number of unions. The relationship is not so evident, however, when we look at the two variables on a country by country basis in the 1980s. As the table shows there was an increase in union density in most countries in the 1970s. This was not the case, however, in the 1980s when there was a fall in unionisation in nearly all countries, with the notable exception of the Nordic countries. This occurred when unemployment continued to rise. The positive relationship between unemployment and the strength of trade unions does therefore appear to be absent in the OECD. This does not mean, however, that trade unions do not influence the unemployment rate.
A central problem is that the density of unionisation is not sufficient for examining the impact of trade unions on the labour market. Two other factors are more crucial. These are the extent of coverage of union agreements and the process by which wages are determined. High wages are more predominant in the economy than the density figures would suggest. This is because many non-union members are covered by union wage agreements. The extent of real wage rigidity is thus more worrying than would have been initially assumed. The influence of trade unions on unemployment comes via the wage bargaining process. The labour market determination of wages is, in most circumstances, a bilateral monopoly and exhibits the familiar feature of uncompetitive price determination. This results, like all non-competitive markets, in reduced demand for labour and high, non-clearing prices (i.e. wages). The extent of the problem depends on whether wages are determined at the firm, sector or national level. In practice, collective wage bargaining at the centralised level is generally deemed to be preferable. At the decentralised level the level of unemployment is taken as given and trade unions tend to ignore the effect of their demands on job opportunities for others in the labour force. In contrast, with centralised bargaining, one trade union represents the whole workforce and takes account of all job opportunities. This results in more moderate wage claims. The problem in the OECD is that very few wage bargains are truly centralised. Partial-centralisation and multi-tier bargaining are more evident. This is a worse case scenario. Either a high degree of corporatism (as exists in Sweden and Austria) or a low degree (as exists in US and Japan) would be preferable. Across Europe, sectoral-level bargaining coexists with plant-level bargaining. There are also cases where centralised bargaining sets a suggested wage and these decisions feed through into discretionary sectoral- and firm-level arrangements. In these circumstances, it is difficult to identify at which level the main wage changes are being made and, hence, who is imposing the rigidity on the labour market. There is thus much circumstantial evidence to support the claim that non-centralised, across-the-board wage agreements have imposed a rigidity on the labour market and, hence, may be causing the rise in unemployment (or at least preventing it from falling).
The other reason why wages remain fixed at a non-clearing level is the increased existence of efficiency wages. These wages exist because they provide a benefit, in the form of an overall reduction in total labour cost, for the firm. If productivity is related to real wages then employers can improve performance by paying a high real wage. This can reduce labour costs by reducing turnover costs and/or reducing employee shirking. Salop (1979) looks at efficiency wages as a means of reducing turnover costs. Wages will be set at a level that will reduce quits and hence lower training costs. Shapiro and Stiglitz (1984) provide an alternative explanation They focus on the use of wages as a disincentive to shirking. An employer will set a high wage (which takes account of the probability of being caught shirking, the probability of getting another job, the wage attainable in an alternative job and the unemployment benefit level) so as to discourage workers from being completely unproductive.
It is very hard to find evidence of efficiency wages in the OECD. This is not because they don't exist but because it is hard to pull away all the factors influencing wages and determine that it was a voluntary decision by firms to set them at a high rate. The only evidence that we have that they are becoming more predominant is the fact that the growing sectors of the economy are those that would generally be expected to have efficiency wages. That is, growing areas such as financial services, high-tech industries and client-based work require productive, well-trained, high quality and diligent staff. The norm, to induce these characteristics, is to provide wages above the supply price of labour so as to maintain an efficient output from the worker.
To conclude this section I will look at the trends in 'compensation per employee' in the business sector as a proxy measure of wage trends in the 1980s and 1990s. In the 1980s wage growth was higher in those countries that experienced high unemployment. Included in this are Italy, Ireland and Spain. At the same time, those countries with low unemployment rates (Japan, Norway, Sweden and US) were experiencing low wage growth. In the 1990s, however, there was a decline in wage growth throughout the OECD. This has not fed into reduced unemployment rates, however. These declining growth rates probably reflect a cross-causation effect from high unemployment rates which act as a dampener on the wage rate. While not growing, the wage levels have not declined and are still high and above market-clearing levels (as indicated by the widespread existence of unemployment). Thus, their influence, while reduced, is still present.
Rigid Labour Demand
The labour market can also be classified as inflexible because labour demand does not automatically react to changes in the real wage or in labour supply. This is because there are exogenous factors influencing the hiring decision. Two such factors are employment protection legislation and the influence of activities in the product market.
The existence of employment protection legislation is thought to dampen the rate of job creation in an economy. This is because much of the legislation involves the imposition of rigid regulations and high firing costs. The central problem is that, while employment protection legislation aims to protect the jobs of those who are employed, it discourages firms from hiring new staff. The effects of security legislation are ambiguous. If it didn't exist there could be random, haphazard patterns of hiring and firing which could result in a decrease in average employment. At the same time, however, its existence seems to promote the persistence of high rates of unemployment. The enforced high price of dismissal makes employers very cautious about hiring workers with uncertain skills or little experience. They would rather take people who are working elsewhere than take a chance with an unemployed individual.
The decision of a firm to hire workers is dependent on their expectation of the future. If they are sure that demand and production are going to be buoyant in the future they will hire the necessary extra workers knowing that the need for dismissal will be low. If, as is more likely, the future is uncertain employers may prefer to provide more overtime hours to those already employed when production increases than to hire new workers. Alternatively, they could use a different form of hiring, such as operating through a temporary employment agent, which allows them to avoid firing costs. The influence of employment protection legislation depends on how strictly it is enforced. Ireland and the UK have the lowest levels of security legislation. This implies that its influence on labour demand, and hence unemployment, is not very pronounced in these countries. In contrast, Spain and Italy have very strict legislation and it must be concluded that their unemployment rates are partially determined by this.
The other factor that causes labour demand to be inflexible is the behaviour of firms in the product market. As we know, from general equilibrium analysis, there is a direct interaction between the labour market and the product market. This is not surprising given that, on aggregate, it is the firms producing in the goods market who determine labour demand. Where there is no competition in the product market, firms may choose supernormal profits over increased employment. The crucial point here is that in an imperfectly competitive market prices are set high, production is constrained and, hence, employment levels are reduced.
The impact of the imperfectly competitive product market is most pronounced when employers 'rent-share' with their employees while retaining some profits for themselves. The rent-sharing could arise for two reasons. First, powerful trade unions may force the employer to share some of the excess profit with them. Second, a high wage may be voluntarily provided by the employers to ensure improved efficiency and, hence, higher profits. Imperfectly competitive product markets thus reinforce real wage rigidities in the labour market. They also constrain the labour demand in the effected sectors. It is very difficult to determine, however, the full extent of this factor in the OECD. It is worth noting that, although competition policies are becoming more widespread, the legislation does not cover all abuses of market power or restrictive business practices and many activities are exempted from its influence. In particular, many labour-intensive service industries (banking for example) are protected from the legislation. While this legislation remains a partial phenomenon imperfectly competitive markets will continue to influence the rate of unemployment.
Rigid Labour Supply
Labour supply responds to factors that are exogenous to the labour market. These factors reduce the flexibility of the labour supply and dulls its reactions to changes in the real wage and/or labour demand. One of the most frequently discussed and most significant factors is the level and duration of unemployment benefit.
Unemployment benefits directly influence an individual's decision to work. The influence is particularly strong for those whose after-tax average wage is very close to the unemployment benefit. Increasing marginal tax rates and improved benefit levels in the 970s and early 1980s increased the number of people who fell into this category. One of the reasons for the existence of unemployment benefits is to provide some security for the unemployed so that they can undertake effective job search. If the benefits are excessive, however, the search becomes overly 'choosy'. The only job search comes from the minority who are not eligible for the benefits but the off-setting effect of this is minimal.
Table 3 - Maximum duration of unemployment benefits and replacement rates
Source : OECD (1993), Employment Outlook, Table 3.11
The link between unemployment and the benefits system has become blurred in the 1980s. In the past there was a clear positive correlation between the two variables. Over the 1980s, however, while unemployment continued to rise the benefits system was being overhauled in many countries, with eligibility criteria becoming stricter and amount and duration being reduced. This can be seen where duration has not increased in most countries and replacement rates have generally declined. This suggests that the increase in unemployment is probably not explained, to any significant degree, by unemployment benefits.
The OECD suggests, however, that there is more to the analysis. They argue that benefits may have a lagged effect on unemployment levels. This means that while they may have little influence on the current unemployment rate, benefits do have an important impact on the persistence of unemployment. There is a lag of between 5 and 10 years between rises in entitlement and rises in unemployment for Canada, Ireland and Finland. The lag was 10 to 20 years for Norway, Sweden and Switzerland. Unemployment benefits do, therefore, impact on labour supply but their overall influence is ambiguous. In particular, rising benefits seem to result in rising unemployment but a fall in benefits has not manifested itself as a decline in the unemployment rate. This may be due to the persistence effect induced by the lagged influence of benefits on the unemployment rate.
The Changing Global Economy
Having looked at some of the factors that impose rigidities on the labour market I will examine how continuous changes in the global economy may effect the unemployment rate. In particular, I will look at technological change and the expansion of international trade. Both of these changes are not, in themselves, primary causes of unemployment. Rather it is the failure of economies to adjust to these changes which has altered the face of employment and has hence impacted on unemployment in various sectors.
The increased openness of economies and the massive growth of globalisation has meant that all OECD countries are trading more with each other, and with non-OECD countries. This increased trade has had a significant impact on behaviour within these countries. The increased competition arising from trade is forcing firms to be more efficient and economies are altering the structure of their employed labour force so that the nation can have a competitive advantage. Trade is changing the nature of jobs in the economy by displacing labour intensive jobs and supplementing them with jobs in the capital-intensive sector. In a fully flexible market this would have no effect on overall employment as suitable adjustments would be made. We have already seen, however, that the OECD labour markets are not flexible and hence unemployment can emerge as a by-product of the changing nature of jobs.
To examine the influence of international trade on employment we can examine the case study of OECD trade with Southeast Asia. This has been increasing rapidly in the 1980s and 1990s. In particular, there has been a growing trend towards importing labour-intensive goods from these nations. The cheap, low-skilled labour of these countries is depressing demand for the same factor in the OECD. Similarly, there is a reduction in employment of unskilled labour because OECD multinationals, who produce similar goods, are moving to South East Asia to take advantage of their lower wage levels.
Despite this, there is no certainty that an increase in international trade is directly responsible for increased unemployment in the OECD. In fact, there has been increases in employment arising from economic growth that are directly attributable to improved trade balances. The problem is, however, that the new jobs that are emerging are in 'new' expanding industries and are not attainable for the low-skilled workers who lost their jobs because of the growth in international competition. Hence, international trade has led to unemployment in some sectors and compounded the problem of mismatch.
Technological improvements allow for economic growth which is a necessity for increased standards of living in an economy. This is a long-run effect, however, and in the short-term the introduction of new technology can result in unemployment. The impact on employment in different sectors is widespread. There has been a marked change in the skills required for existing jobs and new jobs. There has been a shift in demand towards more skilled employment in manufacturing and in other industries. There is thus a widening of the gap between those who have suitable skills and those who don't. This results in unemployment arising from mismatch.
The influence of technological change on employment is ambiguous. If we compare the US with Europe it is difficult to see how technological change has directly caused rising unemployment. Unemployment in the US is at around 6.5 per cent (as it was in the early 1960s) but there has been a huge investment in technology there. In Europe, however, a lower degree of investment in technology is matched by a much higher level of unemployment. Looking at the economy from a general equilibrium perspective the unemployment generated by technological change may not emerge in the economy as a whole; that is, even though fewer people will be employed in certain (mainly manufacturing) jobs. The resulting increase in productivity and output will allow for improvements in other markets and will thus create extra jobs. Thus while unemployment rises in some sectors there is the possibility that it will fall in other areas. This indicates that the direct impact of technological change on the level of employment is going to be minor compared to the impact of other factors.
The impact of technological change on the unemployment rate is dependent not on the level and the extent of this change but, more fundamentally, on the ability of the economy to accommodate the change. The economy must be able to upgrade the skills of the labour force and redesign firm organisation if the technology improvements are to efficiently integrated into industries. The US and Japan, which are both characterised by low unemployment, are among the fastest growing economies to have adapted their economy's structure to one based on high-tech and knowledge-based activities. This is what many other OECD economies have failed to do and hence the job-loss element of technological progress has dominated.
My analysis indicates that there does not seem to be any single cause of the rise in OECD unemployment. Rather, there have been a number of adverse developments, in particular an increasing rigid labour market and a failure, especially in Western Europe to adapt to new technology and the competition of a more global market, which have all acted in the direction of raising unemployment. Much of the thrust of policy advice emanating from institutions such as the OECD has been to enhance labour market "flexibility" by limiting union power, improving retraining and so forth. The United Kingdom has probably gone the furthest in enacting such structural policies, although so far with little beneficial effect on unemployment. The moral is that, although such policies may improve the functioning of the labour market over the medium term, additional specific measures may be required to reduce the already high unemployment that has been allowed to develop during the 1980s. These could take the form of active labour market policies such as those pursued in the Nordic countries, where an exceptional sense of social solidarity and an appreciation in the benefits from the welfare state has allowed resources to be diverted to expanding state service. It is only through such creative policies to replace the "benefit principle" with an "employment principle" that the current crisis will be overcome.
Carlin W and Soskice D (1990), Macroeconomics and The Wage Bargain
Layard R, Nickell S and Jackman R (1991), Unemployment - Macroeconomic Performance and the Labour Market
OECD (1994), Jobs Study - Evidence and Explanation : Part I & II
OECD (1994), Employment Outlook
OECD (1993), Employment Outlook
Salop S. (1979), "A Model of The Natural Rate of Unemployment", American Economic Review, 69
Shapiro C and Stiglitz J (1984), "Equilibrium Unemployment As A Worker Discipline Device", American Economic Review, 74