Economics explains discrimination in the labour market
Economics explains discrimination in the labour market

Start this free course now. Just create an account and sign in. Enrol and complete the course for a free statement of participation or digital badge if available.

Free course

Economics explains discrimination in the labour market

6.4 Empirical analysis

Three key hypotheses have been the focus of empirical evaluation of the segmented labour market theory. First, that the labour market can be represented as comprising at least two well-defined and self-contained segments. Second, that the labour market behaviour of workers and firms in each segment requires a different set of behavioural hypotheses. Finally, that there is limited mobility between the segments reflecting institutional and social barriers in the labour market rather than a lack of productive ability among lower segment workers.

The proposition that there is a clear and evident separation between a primary and secondary segment of the labour market represents a principal hypothesis of the dual approach. It is one, however, that receives only modest support in the empirical literature (see McNabb and Ryan, 1990). In general, there is little evidence of clusters of firms into core and periphery groups on the basis of the various characteristics that supposedly define the two segments. For the UK, however, has found that while firms do not cluster into distinct groups or sectors, there is clear evidence that some characteristics, in particular characteristics such as monopoly power, size of establishment and union density (variables that are associated with the core sector) are strongly correlated, and that the correlations identify features of the underlying industrial structure that are consistent with a segmented labour market approach (McNabb and Whitfield, 1996). Similarly, the employment of women and casual workers, and the use of part-time labour are also correlated, highlighting the importance of gender in identifying differences between firms. Moreover, the pattern of correlations that emerges from this work is also found to affect labour market outcomes such as the incidence of low pay, employment stability, and so on. Thus, while there is little evidence to support the notion of duality or even of identifiable segments, the industrial structure, the product market and its link with the labour market are consistent with a segmentation approach.

Much of the empirical work on the segmentation approach has, however, dealt with the unequal treatment of comparable workers between segments. Initial attempts to test this hypothesis focused upon whether incremental changes in labour quality are more highly rewarded in the primary segment than in the secondary segment. The basic empirical approach adopted to test this hypothesis involves two steps. First, a sample of workers is divided into two groups to represent the primary and secondary segments. Second, multiple regression techniques are used to estimate wage equations (similar to the ones considered earlier) so that the basic hypothesis can be tested. In general, the proposition that education is rewarded more for primary workers than for employees in the secondary segment is supported by the results of a number of studies (McNabb and Ryan, 1990). Typically, these find negligible gains in annual earnings among (variously classified) secondary segment workers from increases in years of schooling and work experience. This contrasts with the marked benefits recorded for both in the primary segment. Some contradictory evidence may, however, also be found in the literature. Several studies report that the returns to schooling and age, while lower in the secondary than in primary segment, are nevertheless economically strong and statistically significant.

The example detailed below, based on work by McNabb (1987), tests for segmentation using earnings functions estimated for core and periphery industry groups. The basis for this division is two variables typically associated with disadvantaged employment: the proportion of women employed and the proportion of employees not covered by a collective agreement. On the basis of these two variables, industries are defined as either being core or periphery.

The data in the example came from the 1975 General Household Survey in which male employees of ages 16 to 64 were interviewed. The key variables were:

Annual earningsthe dependent variable
Schoolingyears of schooling
Experienceyears of work experience
Experience2the experience variable squared
Weeksweeks worked per year

The schooling and experience variables provide an indicator of the human capital of each worker. The more experience and education an individual has, the higher the quality of his labour. It should be noted that if the employee has worked for only a few weeks this will adversely affect his annual earnings, thereby giving a false picture of the relationship between these earnings and human capital. The weeks variable has been included to control for this.

Two of the wage equations reported in McNabb (1987) are shown in Table 6; the first is for the periphery segment, the second is for the core segment. The dependent variable is the logarithm of annual earnings. Note also that the weeks variable is expressed in logarithms (log Weeks).

Table 6: Industry earnings functions

Core segmentPeriphery segment
Intercept1.392.12
Schooling0.67*0.63*
Experience0.65*0.55*
Experience2−0.0009*−0.0008*
log Weeks1.21*1.10*
R20.47070.4517
Sample size16413373
Source: McNabb, 1987, p. 262, Table 1
* denotes that a coefficient is significant at the 1% significance level or better (we can have 99% confidence in its statistical significance)

For both segments of the labour market, the human capital characteristics of workers are positively correlated with earnings. In the periphery segment, for example, a coefficient of 0.67 shows the positive relationship between years of schooling and earnings. Note that there is a negative sign on the Experience2 term. This is a standard result in labour economics showing that beyond middle age, as workers get older, the effect of their experience has a diminishing impact on their earnings. Since a square of 50 years experience is much higher than the square of, say, 5 years experience, the squared term puts a greater weight on many years experience, thereby picking up the declining productivity of older workers.

It should be noted that all of the variables, apart from intercepts, are reported to be significant at the 1% significance level, which means that we can have 99% confidence in their statistical significance. In addition, the R2's for both equations are just under 0.5, which means that nearly 50% of the variation in wages is summarised by each regression equation.

We can now turn to a comparison of the coefficients for each equation. The dual labour market approach predicts that the return on human capital characteristics should be less in the periphery segment than in the core segment. In fact, Table 6 shows that the return on human capital is slightly higher in the periphery segment. The return on schooling is 0.67 in the periphery segment compared to 0.63 in the core segment. Similarly, the return on experience is 0.65 in the periphery segment and 0.55 in the core segment. In this example data, however, the coefficients are, in fact, very small, and it cannot be concluded that there is any significant difference between the return on human capital in the two segments. This particular evidence shows that a dual labour market does not exist for segments defined according to the proportion of women employed and the proportion of employees not covered by a collective agreement.

Evidence of segmentation can, however, be found if we look at the earnings of particular occupations. Table 7, again based on the analysis of the 1975 General Household Survey data by McNabb (1987), reports wage equations for professional and semi-skilled manual workers.

Table 7: Occupational wage equations

Semi-skilled manualProfessional
Intercept−0.0542.89
Schooling0.047*0.007
Experience0.077*0.034*
Experience2−0.011*−0.0006*
log Weeks1.79*1.10*
R^20.5560.500
Sample size254796
Source: McNabb, 1987, p. 264, Table 3
* denotes that a coefficient is significant to the 1% significance level or better (we can have 99% confidence in its statistical significance)

Activity 4

Comparing the two wage equations in Table 7, examine the evidence that professional workers enjoy a higher return to their human capital than semiskilled manual workers.

Discussion

You can see that Table 7 provides some evidence for the existence of occupational segments. Workers in semi-skilled manual occupations are treated differently from professional workers. As they gain more human capital, in the form of schooling and experience, semi-skilled workers are not reimbursed, for each unit of human capital, to the same extent as professional workers. This confirms the insight of segmentation theorists that workers are rewarded differently according to the jobs they do. Even if workers have the same human cognition characteristics, the amount they are paid depends on the job they do. This conflicts with neoclassical theory which predicts that workers will be rewarded proportionately according to their level of human capital.

The final hypothesis that has received attention in the literature concerns the alleged lack of mobility between primary and secondary segments. This issue has received some attention in recent years as longitudinal data has become available. Concerning the rate of movement between segments, studies which impose clear frontiers between primary and secondary segments have generally found rates of upward movement in excess of the low levels suggested by the descriptive dual labour literature. Moreover, a considerable proportion of this upward mobility can be associated with the possession of increased labour quality.

D319_1

Take your learning further

Making the decision to study can be a big step, which is why you'll want a trusted University. The Open University has 50 years’ experience delivering flexible learning and 170,000 students are studying with us right now. Take a look at all Open University courses.

If you are new to university level study, find out more about the types of qualifications we offer, including our entry level Access courses and Certificates.

Not ready for University study then browse over 900 free courses on OpenLearn and sign up to our newsletter to hear about new free courses as they are released.

Every year, thousands of students decide to study with The Open University. With over 120 qualifications, we’ve got the right course for you.

Request an Open University prospectus