Unemployment is probably the most widely feared phenomenon of our times ... (It) represents an enormous waste of human resources, reflects an important degree of inefficiency in economic systems, and causes a disturbing degree of social distress. (Organisation for Economic Co-operation and Development [OECD], 1994, pp. 4, 6) I. INTRODUCTION Unemployment is undoubtedly one of the most researched areas of economics. Determinants of the number of unemployed and duration of unemployment, the role of unemployment on the depreciation of human capital and on general life satisfaction, the interrelationship of unemployment and inflation and unemployment and output (Okun's law), and policies to reduce or even eliminate unemployment are just a few of the areas of research on unemployment. The underlying assumption in many studies on unemployment is that there is a very close correspondence between unemployment and labor market inefficiency as witnessed in the quote by the OECD just given. Although rarely if ever stated, there is also the assumption that a given unemployment rate indicates similar levels of inefficiency either across time and/or regions or countries. Unfortunately and somewhat surprisingly, this assumption has not been tested thus far. (1) Given the importance of comparisons of unemployment on the direction of macroeconomic policy of governments, it is very important to determine whether unemployment is indeed a good proxy for labor market inefficiency. The goal of the article, therefore, is to examine this issue. Thus we pose the following key questions: (1) To what degree do unemployment rates reflect labor market efficiencies? (2) If the goal is to measure relative labor market inefficiency, is it meaningful to compare unemployment rates over time? and (3) To what degree do differences in unemployment rates across countries reflect differences in labor market inefficiencies? Basic economic theory shows that inefficiency in any market depends not only on the wedge between supply and demand but also on the elasticities of demand and supply. The calculation of so-called Harberger triangles (Harberger 1954) of deadweight loss (DWL), which has a long history in the industrial organization and public finance literatures, indicates that for a given wedge between supply and demand, the DWL can be very different depending on the elasticities of supply and demand. Interestingly, however, studies examining the costs of unemployment do not focus on this basic view of welfare loss. Studies such as Gordon (1973), Ahsenfelter (1980), Hurd (1980), Clark and Oswald (1994), Murphy and Topel (1997), and Leslie and Blackaby (1999), among others, examine the costs of unemployment from an individual perspective but do not measure inefficiency from a market perspective, which is the focus of the Harberger triangles. Except for a publication by Fortin and Bernier (1988), (2) we know of no other application of calculating Harberger triangles to determine the level of inefficiency in the labor market. Therefore, this article uses labor market data from 1931 to 1996 in the United States and United Kingdom to calculate the DWL of unemployment and then compares the estimates of DWL to the unemployment rate to determine how closely related inefficiency and unemployment are. (3) The following section investigates the theoretical relationships between unemployment, labor supply and demand elasticities, and DWL. From this discussion we propose two alternative measures of inefficiency. The first is the Unemployment Loss Index, DWL divided by gross domestic product (GDP), which indicates how large labor market inefficiency is relative to total output in the economy. The second measure is DWL per unemployed person, a measure of the potential average gains to employing an unemployed individual. Section III outlines the data and methodology used to estimate the elasticities of labor supply and labor demand over time in the United States and United Kingdom, allowing for changes in the elasticities over time. …