Revised Chapter 12 Notes (May 30, 2007)

Chapter 12: Discrimination and Male-Female Earnings Differentials

economic theory can indicate the labour market effects of discrimination, which may be based on sociological or psychological preferences, and can shed light on the expected impact of alternative policies designed to combat discrimination in the labour market
the focus of this chapter is on discrimination that occurs when people who have the same productivity, or productivity-related characteristics (such as education), are treated differently in the labour market purely because of their personal characteristics such as sex, race, age, national origin, or sexual orientation
discriminatory labour market treatment could take the form of lower wages and reduced employment opportunities, including less training and fewer promotions
although the focus of our analysis is on gender discrimination in the labour market, discrimination can occur in other markets (such as the housing market) and can be based on a variety of factors
in order to better understand the economics of gender discrimination, the reasons and sources of sex discrimination are briefly noted
discrimination against females may result because:
males have a preference for working with, and being supervised by, males
customers have a preference for being served by males
employers have erroneous information on the labour market worth of females; for example, they may underestimate the productivity of females
information on individual workers is costly to obtain and employers may judge individual females on the basis of the average performance of all females
for reasons of job security, males may try to protect their high-wage jobs from low-wage female competition

Theories of Labour Market Discrimination

Demand and supply theories of discrimination
the demand for female labour may be reduced relative to the demand for equally productive male labour because of erroneous information, prejudiced employers, prejudiced male co-workers, or prejudiced customers who prefer the services of males (for example, a male car mechanic)
a reduction in the demand for female labour relative to male labour will result in a reduction in the employment of females and a reduction in the relative wages of females compared to males (assuming that supply of female labour is not perfectly elastic)
the crowding hypothesis argues that females tend to be segregated into 'female-type' jobs and the resulting abundance of supply lowers their marginal productivity and the relative wage for 'female-type' jobs
because of conditioning in a male-dominated labour market, females may also erroneously underestimate their own labour market worth and have a lower reservation wage, which would shift the labour supply curve vertically downward
in theory male-female wage differentials for equally productive workers are inconsistent with competitive equilibrium
as long as females could be paid a wage lower than that of an equally productive male, firms that do not have an aversion to hiring females could increase their profits by hiring females and the resultant increase in the demand for females would bid up females wages and eventually eliminate the male-female wage gap
firms that do not have an aversion to hiring females would maximize profits by hiring many females and firms that have an aversion to hiring females would be foregoing profits by employing only males (eventually such firms will go out of business as they forego profits to discriminate)
in a competitive labour market, discrimination leads to a segregation of males and females (into prejudiced and non-prejudiced firms), but not wage discrimination
Noncompetitive theories of wage discrimination
several noncompetitive theories of discrimination have been advanced to explain the persistence of male-female wage differentials (discussed below)
firms that pay an efficiency wage premium will have a queue of available workers seeking these premium wage jobs and such firms may ration these scarce, premium wage jobs by discriminating on the basis of sex (or race)
monopsony is another noncompetitive factor that can affect female wages relative to male wages; females may be relatively immobile (because they are tied to their husbands' place of employment) and may be paid less than a competitive wage by a monopsonist employer
systemic discrimination has been advanced to explain discrimination that may be the unintended byproduct of historically determined practices, such as informal word-of-mouth recruiting networks (the 'old-boy' system) and job requirements, such as height or strength, which may no longer serve a legitimate purpose

Productivity Differences: Choice or Discrimination?

female wages may differ from male wages because of productivity differences that arise from differences in acquired human capital (education, training, and labour market experience) and differences in absenteeism and turnover of males and females
it may be economically rational for females (or their employers) to be reluctant to invest in human capital that is labour-market-oriented
given their dual role in the household and in the labour market, women traditionally have a shorter expected length of stay in the labour market and thus a reduced benefit period from which to recoup the costs of human capital formation
in addition, a woman's time in the labour market can be intermittent and uncertain, which can lead to a depreciation of her labour market human capital and may inhibit her from acquiring continuous labour market experience
whatever the reason, females tend to acquire less labour-market-oriented human capital than males, and consequently their wages and employment opportunities are reduced in the labour market
productivity and hence wage differences may also occur because of higher rates of absenteeism and turnover of females
compared to males, absenteeism is somewhat higher for women than for men; Akyeampong (2005) documents that annual absenteeism excluding maternity leave for full-time employees is 10.9 days for women compared to 8 days for women
absenteeism for personal or family reasons is small for both sexes, but slightly higher for women (1.9 days compared to 1.5 days for men), and increases substantially for persons with preschool children (again, the gender differences are now found to be small, although earlier data from 1997revealed marked gender differences)
in addition to causing lower wages, higher absenteeism and turnover may be the result of low wages: cause and effect work in both directions, reinforcing each other
differences in household responsibilities and in the division of labour within the household can be quite pronounced
Gunderson and Muszynski (1990) find that when both husband and wife work full-time, the wife averages more than twice the number of hours per week spent on household tasks; this can lead to a situation where women develop a comparative advantage in household tasks and men develop a comparative advantage in labour market tasks

Evidence on Male-Female Earnings Differentials

empirical studies have employed a variety of techniques to estimate the male-female earnings differential and to see how much of the differential reflects discrimination
for recent annual data on the female-male earnings ratio, clickon:
for full-year, full-time (F.Y.F.T.) workers, the female-male earnings ratio in recent years is around 70%
the female-male earnings ratio is:
higher (about 80%) after controlling for the number of hours worked (see Exhibit 12.5 on page 366)
much higher for younger workers under age 25 (80.8% compared to an overall rate of 72.5% for full-year, full-time workers in 1997) and for single workers (91.8%) (see Drolet, 2001, “The Persistent Gap: New Evidence on the Canadian Gender Wage Gap”; available at
econometric studies have estimated the separate impact of the many determinants of wages with the intent of isolating a pure male-female wage differential that remains after controlling for the effect of productivity-related factors
human capital earnings equations are estimated separately for males and females and the overall average earnings gap is decomposed into an 'explained' component attributable to differences in the endowments of productive characteristics and an 'unexplained' component attributable to discrimination
as shown in Textbook Figure 12.1, a hypothetical woman is paid at point A and a man is paid at point D; the vertical distance between point D and A is decomposed into that part which is attributable to the man having more productive characteristics (the vertical distance DC) and that part which is attributable to discrimination resulting from a different pay schedule for women and men (the vertical distance AB)

empirical studies suggest that most of the wage gap in Canada between men and women cannot be explained by different endowments of productive characteristics and is thus attributed to discrimination
the productivity-adjusted wage gap tends to be smaller in the union sector and in the public sector; in the private sector, the productivity-adjusted wage gap tends to be smallest when product markets are competitive
there is some new evidence, however, suggesting that the male-female wage gap tends to be much smaller after controlling for factors that are typically unobserved in most data sets (see Exhibit 12.6 on page 374)

Policies to Combat Sex Discrimination

because of the variety of sources and forms of discrimination, policies to combat discrimination have taken a variety of forms

Conventional Equal Pay Legislation

all Canadian governments have passed laws requiring equal pay for equal work within the same job within the same establishment
the scope of conventional equal pay legislation is limited by the fact that it only deals with wage discrimination within the same job within an establishment and econometric studies have found that conventional equal pay policies have not had any impact on closing the male-female earnings gap, which is largely attributable to male-female wage differences across occupations and industries

Pay Equity and Equal Value Legislation

while pay equity legislation exists in various forms in all jurisdictions except Alberta (see Table 12.2), in most jurisdictions it is restricted to the public sector

only Ontario and Quebec have proactive systems and apply pay equity to the private sector (the most extensive pay equity system in the world)

pay equity requires equal pay for work of equal value, where value is determined by administrators and not by the market place
unlike the economic value of labour, which is determined by the marginal product of labour (the market value of output produced by an additional unit of labour), the pay equity value of labour is administratively determined by valuing inputs, such as skill and effort
pay equity explicitly rejects the notion that market forces should determine wages, arguing that the remuneration for a job should not be low simply because there is a large supply of labour willing to do that job or because there is little demand for that type of labour
the first step in a pay equity plan is to divide jobs into three categories: predominantly male, predominantly female, and other
a predominantly male (female) job is defined as a job where X% of jobholders is male (female), with X specified by the administrator (say X = 70)
the second step is to determine a point score value for each job; points are usually given for four 'input' factors: skills/knowledge, effort, responsibilities, and working conditions
the administrator must decide the weight for each factor (for example, skill may be three times more important than working conditions) and how many points to be awarded for components of each factor (for example, how many point to give for a high school diploma, a BA, an MA, and a Ph.D.)
the third step is to conduct a statistical/econometric analysis of the relationship between the wage paid and the point score for all persons (both sexes) in predominantly male jobs and for all persons in predominantly female jobs (see Textbook Figure 12.2)

in situations where equal pay legislation has been applied, statistical analysis typically reveals that a person with a given point score in a predominantly female job receives only 80% to 90% of the wage of a person with the same point score in a predominantly male job
if some of the female-dominated jobs with low points scores do not have direct male comparator jobs, the male dominated wage line may have to be projected backwards to provide hypothetical male comparator (phantom) jobs

the final step is to use the statistical estimates to adjust the wage of a person with a given point score in a predominantly female job so that it is equal to the wage paid a person in a predominantly male job with the same point score
as discussed on pages 385-386, there are different adjustment options available such as line-to-line, point-to-line, and point-to-point
many assumptions have to be made to determine the 'pay equity' wage adjustment for people with predominantly female jobs, and different administrators, statisticians, and econometricians can make different assumptions (for example, what value for X, how many points is a B.A. worth, which estimator and which adjustment procedure to use), which can produce quite different estimates for the size of a 'pay equity' adjustment
given the many statistical assumptions imposed, formulas to generate pay equity adjustments tend to be incomprehensible to anyone not in the 'pay equity' business
in the public sector average wage adjustments in pay equity settlements have typically been between $3,000 and $4,000 per recipient (some of whom are males in female dominated jobs) and have raised the female/male earnings ratio from around .78 to about .84 (pay equity wage adjustments only close about 1/3 of the female/male earnings gap)
see Exhibit 12.13 on page 392 for a discussion of the largest and most significant pay equity award in history, the 1999 federal pay equity settlement which took 15 years to resolve and provided $3.6 billion in back pay for 230,000 current and former federal employees (an average wage adjustment of $15,000 per employee)
in the private sector pay equity wage adjustments have been much smaller; pay equity legislation has had only a small (or no) effect on wages, employment, and the female/male earnings gap in the Ontario private sector (see Exhibit 12.15 on page 394)
the scope of pay equity policies to reduce the male-female earnings gap is limited by the fact that wage comparisons can be made only within the same establishment and can not reduce the portion of the male-female earnings gap that reflects the segregation of females into low-wage firms or low-wage industries
to the extent that pay equity laws raise the wage for female-dominated jobs, there will be fewer such jobs (because the labour demand curve slopes down) and an increase in the number of displaced or unemployed females workers

Equal Employment Opportunity Legislation

equal employment opportunity legislation, which is part of the Human Rights code of each province, is designed to prevent discrimination in recruiting, hiring, promotion and dismissals
unlike equal pay laws which increase female wages at the expense of their employment, equal employment opportunity legislation increases the demand for female labour, which increases both the wages and employment of females
equal employment opportunity legislation is likely to be most beneficial to new recruits or women seeking to change jobs, and may do little to help incumbent females in their existing jobs

Affirmative Action or Employment Equity

while equal employment opportunity legislation is designed to remove discriminatory barriers so as to ensure equality of opportunity, affirmative action programs focus on results (not just opportunities) and typically require some preferential treatment for women to ensure an equality of starting points; it is a more intrusive form of government intervention to overcome systemic discrimination
in Canada legislated employment equity currently exists only in the federal jurisdiction, which covers about 10% of the Canadian work force
federal employment equity applies to four designated groups (women, visible minorities, disabled persons, and Aboriginal people) which are particularly susceptible to systemic discrimination
U.S. studies have shown that affirmative action legislation has provided benefits to the target group (mainly black males) but at the expense of losses to other minority groups (such as white females) which were not targeted

Policies to Facilitate Female Employment

a variety of policies, such as improved availability of daycares and childbirth leaves, can expand the range of choices open to women and facilitate their participation in labour market activities in a nondiscriminatory fashion; however, these facilitating policies, such as subsidized daycares, are not without their costs and there is a legitimate debate over the efficiency of such policies
new institutional features in the labour market, such as flexible hours, child-care leaves and daycares arrangements, are emerging as endogenous responses to the increased role of women in the labour market