countries resulting in a pooled time series data for a handful of countries limiting
variations on the key independent variables. Data on gender breakdown of occupational
categories are hard to come by—as frequently lamented by experts of occupational sex
segregation. In addition, multiple year data on the institutional variables used in this
study (OECD employment protection legislation index, generosity of family leaves) are
only available for circa 1990 and a more recent year, making a pooled time series
analysis virtually impossible. In order to address the issue of change across time, I have
elsewhere conducted a comparative study by means of carefully matched case studies.
This paper, however, is primarily concerned with cross-national variations at a particulat
point in time.
The rest of this paper is organized in four sections. Section one reviews three
families of relevant theories and discusses their insights and limitations. Section two
presents an alternative theory of sex segregation—what I call the skill-based theory of
occupational segregation. Section three provides empirical evidence in support of the
skill-based theory. Section four, the concluding section, summarizes the findings of this
paper and discusses normative implications.
I.
THEORIES OF SEX SEGREGATION
:
THEIR INSIGHTS AND LIMITATIONS
Before we can review the existing theories of sex segregation, it is necessary to
clarify what sex segregation means. Despite the common use of dissimilarity index as a
way of capturing degrees of sex segregation, the reality of sex segregation is a
complicated multi-faceted phenomenon impossible to express in one single aggregated
index such as dissimilarity index.
Dissimilarity index simply measures the extent to
4
Convention