On the measurement of mismatch

Abstract

We use a definition of an equilibrium rate of unemployment as a mismatch indicator. An application of the indicator to nine Organization for Economic Cooperation and Development (OECD) countries leads to diverging results. Most of the considered countries have experienced increasing mismatch in the 1970s and decreasing mismatch in the 1990s. The latter result is somewhat surprising, since mismatch was expected to be increasing in the 1990s. However, the estimates for Germany are against this international trend. Mismatch is not in general lower in countries with a more flexible labour market, but is decreasing in some of the countries which have chosen more flexibility on the labour market.

Notes

¹ See, for example, Pissarides (2000) for a discussion of that point in detail.

² See Thisse and Zenou (2000) for a similar discussion.

³ For Germany, the Institute for Employment Research publishes the estimated share of official reported vacancies. For the UK this ratio is taken from Nickell (1998).

  ⁴ Hall (1994) proposed that k should have a preferably high value in the initial situation. It should be reduced subsequently, until the difference with the highest lag is significant.

⁵ The significant second difference results are not reported here. Entorf (1998) also found I(2) for unemployment rates of other European countries.