The gender pay differential in America is a product of labor market behavior of workers and employers, and a number of market anomalies and failures.

Historically, the gender wage gap has been about 40% (women made salaries on average that were at the level of 60% of the salaries of men).

historic-gap

This longstanding gap seemed fixed, causing some to fasten onto the idea that it was God’s will, or prophesy:

prophesy

But, alas, it has finally changed, in the workforce of the 1980s and 90s. The gap between men and women’s salaries has shrunk from the historic norm of about 40% to about 20% today.

up-to-80

This gap has shrunk because of many things, but mainly:

1. Increase in labor force participation of women— largely enabled due to the availability of birth control after the mid 60s

2. This brought a flood of college educated women into the workforce, women who largely did not enter the workforce in earlier years (the earlier female workforce was less educated, earned less, and were largely women who had to work to support their families and didn’t have the luxury of being at home).

In 1970 about 22% of the female workforce had some college or degrees. By 2010,  nearly 2/3 of the much larger female workforce had college educations.

3. Many more women began preferring professional education and careers in law, medicine, business and other highly paid occupations (where previous preferences favored nursing, teaching, social work, etc).

4. Less overt discrimination in pay, largely due to deterrent effects from legal actions (eg equal pay laws) .

This change in the female labor force following the the introduction of birth control technology, and the flood of interest in labor force participation that followed was remarkable.

flood

Today, the wage gap is somewhere around 20%, with women making about 80% of men salaries across the workforce.

The following chart shows the wage differences by occupation. Obviously, there are large differences in the gender gap situation across occupations. In service worker and clerical jobs women make more than men. While there are considerable gaps favoring men in many of the ‘professions’.

occupation-gap

Before reviewing these possible sources of the gender gap, we will first review the way markets are thought to work for setting wages.

 

Wage Determination

Generally, labor markets work like other markets, with supply forces interacting with employer interests (demand) to determine the equilibrium wage and employment level. Demand by employers is called a derived demand, because, in part, it is determined by the demand for the employer’s product by customers. The demand for labor is determined by the physical productivity of workers, and the price that the product can be sold for by the employer. The demand for labor at any particular wage level increases when:

• The firms demand for its product shifts out (more demand) due to things like competitors raising price, income rising, an effective branding process, etc.
• The firm has more or better machinery, which increases the productivity of labor, shifting the demand for labor to the right
• The workers are trained, or become more experienced and more capable, shifting the demand for labor to the right
• The availability of new sources of cheaper labor—may shift the demand for labor inward (or replace it altogether). Trade agreements, for example, made it possible to “outsource” production to China and other places where labor and manufacturing in general are cheaper—which shifted the demand for many blue collar manufacturing jobs to the left (reducing wages, and lowering the level of employment for domestic workers).

Generally, the demand for labor slopes down because of three reasons;

• When more labor is hired and it produces more, this additional output can only be sold if prices for the final product are reduced somewhat (demand curves slopes down)
• Diminishing marginal returns of labor. Firms can afford to hire more labor only if wages are lower, because the marginal productivity is falling due to diminishing returns
• Labor has substitutes. If wages were going up, they will get by with less labor because of this, preferring to substitute other kinds of (cheaper) labor, or use more machinery.

Households supply labor. They do it to earn money by selling time. Supply curves generally slope up because more work will be supplied only if wages are higher—this stems from the fact that the opportunity costs (what you have to give up) becomes higher and higher as you have less and less time for other things in your life (fun, family). Supply curves shift for several reasons:

• When a job becomes significantly more or less attractive, for example. Say, that the newspaper reports that some hazardous chemical used to be present in the building the employer now owns. This will cause workers to offer less labor at all wage levels (a shift upward in the supply curve, because of the hazard). On the other hand, the job may become more prestigious, or fashionable than before, which will attract more workers wanting jobs, shifting the supply to the right (more workers at all wage levels).
• The job now requires a license, or some kind of exam, or special training paid for by the worker. These things reduce the supply of labor by making it more expensive or more limiting of potential workers. They will supply less labor at each wage level when these kinds of things are imposed on the labor market.
• New technologies are introduced that may shift the work-life balance preferences of workers. The possibility of working from home thru the internet, or the birth control pill, or construction of public transportation to the community where employers are located—are all examples that will change willingness to supply labor to the labor market (shifting supply out)

Workers are said to be paid what they are “worth” economically to the firm”. The main drivers of wages being: the level of demand for the product the workers are producing, the price paid for the product, and the extent to which workers can be substituted for by other resources. On the supply side, wage drivers are the attractiveness of the job aside from the economic benefits, the extent of training requirements to be acquired by the worker, that the attractiveness of alternative activities (to this job) for the worker.

Subject to these forces, markets work by trying to return to equilibrium, if something changes. So, if the demand for nurses is shifting out as the population ages and demands more and more health care, then the pressure will be on wages to rise. As demand shifts to the right, then shortages appear (at the existing wage level employers cant fill all the jobs they have because not enough nurses are willing to work at that wage— a shortage occurs). So, employers will begin to raise wages (or offer signing bonuses, more flexible hours, or other attractions) to attract the nurses they need. Wages, in general will rise and eventually equilibrium will return, with a higher wage and a higher employment level of nurses.

Human capital

Human capital is the embodied skills and knowledge in workers. Firms sometime “invest” in training some workers. And, workers often “invest” in themselves. Both make such decisions based upon the costs they incur, and the future stream of benefits they expect to earn. That is, decisions about whether or not to invest in such capital are presumed by economics to be made exactly the same way by decision makers who are faced with other decisions about lasting assets. The idea is that decision to buy/not are made on the basis of ROI, where the annual stream of net cash flows are discounted to present value, and compared to initial investment outlays to determine if net present value is positive. The critical elements in this analytic framework are:

1. the stream of annual productivities (benefits) associated with the trained worker
2. the stream of salaries which will need to be paid the worker (1-2 is the annual stream of exploitive returns to the firm)
3. the discount rate used to discount to the Present Value the net returns (1-2). typically the discount rate is the weighted average cost of capital
4. the up front training cost investment
5. the number of years the worker is expected to stay with the firm, allowing the firm to recover returns on the investment

Human capital theory was developed by Gary Becker, who won a Nobel prize for this and his theory of discriminatory behavior—both applications of pretty basic economics to what heretofore were intractable problems for economics and presumed to be problems of “preference”, not economic decisions.

So finance people have long used the theory of decision making that was adapted by Becker and applied to labor economics. Why do people demand education? How is the decision made? Why do people decide to pursue healthy lifestyles? Why do firms pay women less than men under some but not all circumstances. We could say that the answer to all these important questions is that “ they prefer to do it”. Unfortunately, this is unhelpful to understanding what drives decision making behavior, and economics of human capital offers much more systematic insight into what is going on in these cases.

The idea that was expressed in class by one of the students, and echoed by others, is that decisions about which MBA program to enter were driven, at least in part, by using the ROI model of decision making, where the idea was to plug in the wage before and after the MBA reported by various schools, discount those benefits for many years of employment history, and compare the Present Value to the outlays for tuition, books and foregone income for 18 months. Sounds like an investment decision making approach. Used by many of you to decide which school to choose. Other benefits and costs can enter the decision too, though not numerically. The model works well to describe firm’s investment decisions too. Typically, the weighted average cost of capital for the firm is used as the discount rate.

So, the human capital investment decisions by the firm (whether to train, pay for schooling, or pay for other investments that might be made in new managers) can also be described by this same model. Indeed, we contend that the observations we can make about labor market differences by gender can be explained by this presumed approach to decision making by business firms. What we observe is:

1. businesses often pay women less than men for the same position when the job requires training (this helps explain why women are hired at all when jobs require firm-provided training)
2. in jobs requiring training we often see women with higher levels of education or more experience than men (this compensates for the shorter tenure on the job)
3. Such differentials are not seen for jobs where the firm does not have to invest in training.

The observations are exactly what would be predicted if we believe that firms are making their decisions using the human capital model. The model suggests that if firms believe that women will not stay as long as men on the job following training, then firms will have fewer years to recover investment returns for the women. Other things the same, they’d prefer men. Indeed, the model predicts the circumstances under which the firms would hire women for jobs that require training would be when they (1) could hire the woman for less money, or when (2) the woman was more productive, cet par. Both are frequently observed in such labor markets.   When jobs do not require the firm to invest in human capital, such as for teachers, professors, nurses, and most service workers—we do not see any gender discrimination. The discrimination follows from the necessity of the firms investment, which requires some accommodation to equalize the gender Returns On Investments (ROI) under the situation where the firm expects females to have shorter tenure on the job.

This theory of human capital does not say what “should” be done by firms. It seeks only to explain what “is” evident in practice.

Discrimination

Firms are said to discriminate in hiring if they do not choose workers based on their comparative economic worth to the company. Gary Becker wrote his dissertation on this topic.

The argument can be made simply about what discrimination is, and how it can occur. Think about a firm hiring 5 new workers for a particular job to meet the needs of an expansion. Say they have 6 candidates for the job based on applications, credential reviews, interviews and other data. Now they need to decide who to hire. Rank ordering the candidates by level of expected productivity we might get the following:

Candidate           Expected daily net productivity ($)         Race

1                                                 100                                                W

2                                                   95                                               W

3                                                   90                                               B

4                                                    90                                               W

5                                                   85                                               W

6                                                   80                                               W

Based on the data, no discrimination would result in the first (top) five candidates being hired and yielding a total productivity of $460. If discrimination occurs, and the B worker is not hired, then the five W workers would produce a total of $450. This behavior reduces the profitability of the enterprise, and occurs under conditions of (1) some compensating benefit occurs to the firm owner to make it “worthwhile” to give up profit to satisfy a stakeholders preferences for discrimination (firm customers, other employees, others) and (2) sufficient excess profits are present for the firm, allowing them to give up some of those profits in order to discriminate. In highly competitive markets we wouldn’t expect discrimination to occur because profits are barely enough to able to keep the capital invested in the firm.

Fringe Benefit economics.  

Say a fringe benefit (like firms providing free daycare) is added to a labor market that had no such thing before. The cost to the firm to provide it is some amount per worker. The impact on market wages and employment will depend on how much the workers value the benefit, and what happens to the numbers of job applicants for this job.  Say, the increase in supply is small, and the workers do not value the benefit as much as it costs to provide. If this was a voluntary fringe benefit, firms would not offer it. Why not? They offer a benefit voluntarily only if the workers are willing to pay for it; eg. only if the workers will accept wages that are lower than before the fringe was offered by an amount lower than the employer’s added fringe costs.

But, if the fringe benefit is more valuable, and the number of applicants seeking work expands a lot, then this will drive the wage down, possibly enough to compensate for the added cost of the benefit. In such a case, firms would offer voluntarily, such benefits, because workers are willing to pay for them in lieu of wages.Indeed, what does “making profit on fringes” actually mean? Essentially it means that the firm has found a “second” product/service to provide to a set of customers, who happen to be employees!

Obviously employers are fast to identify such benefits that are valued more than they cost to provide, because it creates an HR department full of qualified job candidates from which to pick and choose! In the case where employees value the benefit = to the cost of providing it, the supply curve shifts to a new equilibrium, and the drop in wages is exactly equal to the cost of providing the fringe.

What happens if the government mandates the offering of a benefit that employees don’t value very much? The employer and the workers share the cost.As with any other fringe benefit, it increases labor costs, and will tend to reduce the number of jobs offered by the firm. This is the classic case of the mandate causing cost sharing between the employer and the workers. even though the workers do not see full value in the benefit.

A final example of fringe issues concerns the economics of “flexibility” in work arrangements, or offers of on site day care, or humane work rules for part time employees. These “conditions of the job” vary across employers because they are voluntary benefits. Some employers may see benefit in providing them, others not. For those that offer them, employees value them enough to take lower pay. So, other things the same, a firm that offers “job flexibility’ as a fringe benefit will be able to get workers though they pay lower salaries to them. If some firms do this, and others do not, then there is a mechanism for wag gap impacts.

If it happens that such voluntary benefits are differentially preferred by gender, what does it mean. What will the effect be? It means that in instances where the benefit is offered, the wages will be lower (or they wouldn’t voluntarily offer it). If women happen to care more often than men about such “fringes” at work, then they will be found to have lower average wages than men, other things the same. That occurs because the women will differentially seek out and apply for jobs that have such benefits. These jobs in these firms will end up having lower wages because supply of workers is more robust.Men, who may not care as much about such benefits, will prefer jobs where they are not offered, and where wages are higher. This will drive a wedge between gender salaries .

This is likely a source of a small portion of the gender wage gap (now about 20% across the workforce). If the government mandated the benefit be provided by all firms, then women would not differentially prefer some of these employers over the others, and cost sharing would occur in probably all firms.

Remember, any kind of employer provided work benefit (bigger offices, personal secretaries, health insurance, contribution to retirement plan, company car, etc) will create these same kinds of impacts: the impacts of offering the voluntary fringes will, other things the same, be fewer jobs offered but at higher total compensation levels. The cost of the fringe will be borne partly by the employer, and partly by workers who agree to take a lower salary (than otherwise) in order to get this benefit. For workers who don’t happen to value the benefit offered by the firm, then they will be reluctant to take a job that requires them to help finance, and will seek jobs with benefits more to their liking.

Tenure Differences in Employment

Women do not stay as long in jobs as men. This longstanding fact has diminished in magnitude over the last generation, again due mainly to the birth control pill. But, small differences remain between the sexes. The theory of human capital contends that when jobs require employer-financed training, these differences in tenure limit the ability of the employer to recoup their investments and will result in one or more of several responses: don’t hire women at all, if hired, pay them less, or find female employees who don’t need as much training as the men. All of these are to be found in the evidence. When jobs require no training, there should not be any basis for any of these impacts. Evidence from research suggests that this is responsible for only a small portion of the wage gap in affected jobs. The tenure-on-the-job differences between men and women is responsible for only about on tenth of the wage differential[3].

Today, the tenure differential has become rather small. The chart below shows that the tenure differential has shrunk a lot, almost disappearing. This is largely a consequence of the technology of birth spacing, and resultant changes in labor force participation by women, discussed more below.

tenure

Market Failure due to Asymmetry of Information

Asymmetry of information (employers know the wages and levels of productivity of all workers, and don’t know about the true skills of job candidates) is something that, if kept secret, can allow exploitation to occur. Equity adjustments between comparable workers are slow to be made if salaries are kept secret. So, there are two mechanisms by which asymmetry may affect the gender gap: (1) lack of transparency may create a gap in new hire salaries if women, in the absence of information about other salaries, do not negotiate as aggressively as male counterparts. And (2) as time passes on the job, the lack of transparency does not tend to dissipate between equally productive peers who are paid differently (for whatever reason).

The argument is that if salaries were transparent, then some unknown portion of the wage gap would be eliminated.

Resultant Gender Pay Gap of 20%— why does it exist?

The gender gap in salaries has shrunk considerably, but has plateaued again, at about 20%.

The gender wage gap that remains is probably due to

(1) discrimination (enabled by market failure of monopoly, wherein firms can afford to be a non meritocracy and leave potential profits on the table),

(2)  due to gender differences in preferences in employment conditions.Women preferring jobs that have some “flexibility” to accommodate the disparities women face in family child care responsibilities.

(3) Some small portion of the wage gap is also likely due to somewhat shorter tenure with firms than men (the average difference is only one quarter of a year).

(4) part of the gap may also be a product of market failure due to asymmetric information. The argument goes that gender pay differentials result, in part, from employees not having information about what employers are paying others for similar jobs. This lack of transparency allows differentials to persist, when otherwise they would disappear.

Worth reading Citations

G. Gaumer, “Sex Discrimination and Job Tenure.” Industrial Relations, February 1975.

Deborah Eisenberg, “Money, Sex and Sunshine” University of Maryland School of Law, Working Paper: http://ssrn.com/abstract=1801238