Why has wage inequality increased more in the USA than in Europe? An empirical investigation of the demand and supply of skill

Abstract

During the past two decades the wage gap between high and low skill labour has increased more in the USA than in many European countries. In this paper, the correspondence between occupation and education is used to construct aggregates of skill supply, skill demand and unemployment by skill group that are comparable across countries. Using individual data for years 1983–1994, it is found that the relative demand for skilled labour has increased to a similar extent in the USA and in Europe and that wage inequality remained low in Europe partly because the European relative supply of skill increased much faster than in the USA, and partly because European relative wages were rigid, which caused an increase in unemployment among the low-skilled.

Notes

¹For two recent studies, see Haskel and Slaughter (2002) and Acemoglu (2002).

²Theories linking international trade to wage inequality have suffered from bigger criticism. Haskel and Slaughter (1999) finds that international trade, and not technical change, is the main driving force behind UK's increase in inequality during the 1980. Haskel and Slaughter (2003), however, find little evidence that international trade affects relative wages in the USA. Welsch (2004), finds that export growth in Germany during the second half of the 1980s is inversely linked to the high-skill intensity of labour. Green et al. (2003), argue that the relationship between trade and skills is weak and that computer usage is very strongly associated with the process of upskilling in the UK.

³Although this paper focuses on the changes in these four countries, other studies found similar differences among other OECD economies. See Freeman and Katz (1995), and Tachibanaki (1998) for various individual and cross country analyses.

⁴Acemoglu (2003a) is one of the very few exceptions.

⁵This last group of studies, sometimes using the same set of countries, has found either opposite or non-conclusive evidence when looking at the relative unemployment rate trends of the different education and experience groups. See for example Blau and Kahn (1996) and Krueger and Pischke (1997) for evidence in favour and against the role of institutions in Germany. See Edin and Topel (1997), Card et al. (1995) for other comparative studies.

⁶I thank those who made it possible for me to access these data. I thank Kevin M. Murphy for providing the CPS data, Antonio Filippin and Andrea C. Ichino for allowing me to access the BI data and Syracuse University together with the German Institute for Economic Research for providing the GSOEP data.

⁷‘Material from the GHS is Crown Copyright; has been made available by the Office for National Statistics through the Data Archive and has been used by permission. Neither the ONS nor the Data Archive bear any responsibility for the analysis or interpretation of the data reported here’.

⁸This is consistent with the fact that the supply and the employment of more skilled individuals increased during this period of time.

⁹In a regression that controls for education, occupation and experience, it is found that residual wage inequality did increase in the UK during this period of time. This seems to explain the observed increase in wage inequality across different deciles of the wage distribution. (Fig. 3).

¹⁰A different model would be needed to incorporate the effect of international trade, but the nature of the decomposition analysis presented later would be the same.

¹¹This flexible specification embeds more specific production functions easily. For example, Murphy et al. (1998) specify the following production function to study the different growth in relative wages between the USA and Canada:

where σ governs the elasticity of substitution between high and low skill labour. Assuming a linear time trend for the relative demand change, i.e., assuming , can write the following relative wage equation can be written

Since β₁ + β₂ = α₁ + α₂ and β₁ − α₂ = −(1/σ), changes in the relative demand schedule come solely from skill-biased technical change.

¹²The returns to occupation and education are more similar for Germany and the UK. This is reasonable considering that wage inequality barely hanged in these two countries from 1983 to 1994.

¹³In the case of constant relative wage and increasing employment share, a positive demand shift exists as long as the elasticity of the demand for that occupation is less than ∞ (in absolute value).

¹⁴The only exception being managerial occupation in the UK Given the nature of this occupation and the small decrease of its relative wage, it has been included it in the high skill group.

¹⁵In this case, the high skill group is formed by those individuals with more than high school and the low skill group by those individuals with high school or less.