British Journal of Industrial Relations 28: zyxwvutsr 1 zyxwvutsr March 1990 0007-1080 $3.00 Employment Segregation: Can We Measure It zyx and What Does the Measure Mean? zyxw Zafiris Tzannatos * 1. INTRODUCTION Measuring qualitative aspects of the labour force has always been a difficult task. One solution is to resort to various ad hoc indices whose construction is at times arbitrary and interpretation ambiguous. This short note examines the properties and possible (mis)interpretations of two commonly cited indices of employment segregation and offers new insight to understanding them. I draw my examples from employment segregation between women and men in Britain. 2. THE HAKIM INDEX A measure proposed by Hakirn (1981) is based on the observation that some sectors (occupations, industries, regions and so on) employ proportionately more women (or men) than the overall ratio of the two sexes in the total labour force. Thus, the ‘expected’ sex ratios in each and every sector, which should be equal to the overall ratio if men and women were identical in all respects, are usually different from the actual ratios in these sectors. If the actual ratio of women to men in a given sector is higher than the expected ratio, then the sector is said to be over-represented; if the contrary, the sector is said to be under-represented (always with respect to female employment). From this observation, Hakim argues that the difference between the ratios of women actually employed and women expected to be in the over- and under-represented sectors can be utilized to show the extent of sex segregation in the labour market; i.e., where zyxwvut H stands for the index, zyxwv F is the number of women workers, superscripts a and e indicate actual and expected values, and subscripts i and j are, respectively, the female over-represented and under-represented sectors. This index takes the value of 0 if the actual sex ratio in each and every (1) H= (ly/E) - (lyq *University of Buckingham