The Unemployment/Vacancy Curve

Rising unemployment has become one of the most challenging problems for economic policy in many developed economies over the last fifteen years. In the second half of the 1970s and during the first half of the 1980s the labour market situation worsened dramatically. For the OECD area as a whole, unemployment as a percentage of the civilian labour force went up from 3.3 percent in 1974 to 8.1 percent in 1985. The increase in unemployment rates was even more pronounced for OECD-Europe, where it climbed from 3.3 percent to 10.5 percent in this period.

The basic idea of the U/V curve, which was initially defined as a negatively sloped relation between unemployment and vacancies¹⁾, can be found in the paper of Dow/Dicks-Mireaux (1958) “The Excess Demand for Labour”. In their study, which is primarily concerned with the question whether unemployment and unfilled vacancies are good indicators for trends in the demand for labour, they state (p. 3f):

“We must now consider more closely how changes in the pressure of demand can be expected to affect the statistics of unemployment and of vacancies. It is convenient to abstract from two other things which we have already seen may affect the figures, namely changes in the degree of over- or under-statement of vacancies and changes in labour maladjustment. An increase in the pressure of demand will then always increase the level of unfilled vacancies reported, and reduce unemployment. But even at very high levels of demand there would remain some unemployment, i.e. above a certain point unemployment is decreasingly sensitive to demand. The reverse is true of the statistics of unfilled vacancies. Even when demand was clearly deficient, some unfilled vacancies would remain, so that below a certain point the level of vacancies must be reckoned as decreasingly sensitive to demand.”

Essentially the U/V curve can be seen as a reduced form of the employment function, explaining employment by both demand and supply of labour.

Another approach to the foundation of the U/V curve is the one via matching models comprising elements of search theory. The superiority of this approach to the foundation via the employment function stems from three sources: firstly, from a dynamic view of the labour market transactions, secondly, from the capability of explaining the coexistence of unemployment and vacancies in any single submarket and, thirdly, from a microeconomic foundation of the macro U/V curve.

So far, we assumed in our analysis of the matching process that

Hysteresis phenomena have received renewed attention within the context of an explanation of the persistence of European unemployment; there is a large volume of theoretical and empirical literature on this problem (for an overview, see Emprical Economics Vol 14. and Möller, 1990). Hysteresis occurs when a fall in demand is accompanied — or followed — by adverse changes in mismatch, search effectiveness and other structural effects which persist for a longer time than the initial demand shock. In the context of unemployment the hysteresis debate has been especially stimulated by a series of papers by Blanchard and Summers (1986 and 1987). The analytical problems are evidenced by the fact that empirical estimates of the non accelerating inflation rate of unemployment (NAIRU) seem to follow the actual path of unemployment rather closely (see e.g. Coe 1985, Solow 1986, Franz, 1987, Blanchard/Summers 1987, Pichelmann, 1990). This finding was also obtained for estimates of the natural rate by U/V analysis (Christi, 1987). Thus the hysteresis phenomenon is another explanation of outward shifts in the Beveridge curve.

Our considerations so far have been based on the assumption that the matching technology in the labour market is linear-homogenous. This assumption has implicitly been made in practically all empirical U/V studies. But as we will see in this chapter further down, only if this assumption holds, shifts of the steady-state Beveridge curve can be regarded as an increase or decrease of structural and/or frictional unemployment. If, however, the matching technology exhibits increasing or decreasing returns to scale, movements of the curve, holding variations in the acceptance probability constant, may be due to an increase or decrease in the matching activity level.

In Austrian political discussion, as well as those of many other countries, arguments about increased structural and frictional unemployment occur frequently. Such arguments point out that employers are reporting a lack of skilled labour and some people maintain that generous unemployment benefits and limited regional and occupational mobility have reduced the search intensity of the unemployed. In this chapter we will try to use empirical U/V analysis to shed some light on these kinds of questions.

The number of studies on structural unemployment using the tool of the Beveridge curve has increased rapidly since the early eighties. The whole bulk of empirical work in this field is already so extensive that it cannot be discussed in all details here. Therefore, the following chapter will give a short description of some of the most prominent studies for various countries surveying the empirical results and the methods used.

Unemployment has remained a serious problem throughout the 1980s (although its role as a major target for economic policy has suffered during that time). Both in academia and in public, “structural factors” have increasingly been made responsible for the problem of the permanent high level of unemployment in Europe. In this context, the unemployment/vacancy curve has been widely used in order to identify extent and causes of structural and/or frictional unemployment in the labour markets.