Compensating Wage Differentials
The theory of compensating wage differentials provides one explanation of wage differences across individuals and across occupations. This theory suggests that wage differentials exist, in part, to compensate workers for nonpecuniary characteristics of alternative types of employment.
Compensating wage differentials
Let's consider an example to illustrate this concept. Suppose that two occupations (X and Y) are initially perceived as being equivalent in all attributes (e.g., educational requirements, job stress, working conditions, and other characteristics). In this case, it would be expected that labor supply adjustments would equate wages between these two occupations (as illustrated below).
Suppose, though, that it is discovered that workers in occupation Y face a greater risk of suffering a fatal on-the-job injury than workers in occupation X (a perfectly safe occupation). This will induce some workers to migrate from occupation Y to occupation X. Migration continues until the wage difference between the two jobs is large enough to induce workers to stay in their current occupations. The diagram below illustrates this possibility.
The wage differential w"-w' is the amount that a worker must be compensated to accept the additional risk associated with employment in the risky occupation. This compensating wage differential can be thought of as the risk premium associated with employment in occupation Y. The left-side diagram below illustrates the magnitude of this compensating wage differential.
Ceteris paribus, it would be expected that a similar compensating wage differential would exist for differences in working conditions, job stress, educational requirements, and other characteristics of jobs that make them either more or less desirable. It is expected that more pleasant jobs will offer lower wages than less pleasant jobs, holding all other job characteristics constant.
Compensating wage differentials will reflect the market value of non-wage job characteristics if:
- workers attempt to select an occupation that maximizes their utility levels, not their income,
- workers have perfect information about all job characteristics, and
- sufficient labor mobility exists.
The hedonic pricing model
We will use a hedonic pricing model to explain the existence and magnitude of compensating wage differentials. Under a hedonic pricing model, a commodity is sold that possesses a set of characteristics that vary across the products that are offered in a particular market. The bundle of characteristics that describe a particular commodity is observed by both buyers and sellers of the commodity, as is the price of each particular object. The market price of each individual characteristic, however, is not directly observed, but may be imputed using econometric techniques.
In the labor market, each job can be described as consisting of a set of characteristics (e.g., the level of education required, the amount of risk associated with the job, the level of job stress, and so on) and an associated wage offer. As noted earlier, it is assumed that wage differentials across jobs (under the conditions listed above) compensate for differences in non-wage job characteristics.
Let's examine how firms and workers may jointly establish a market value for differences in the risk of injury faced on alternative jobs.
Human Capital Investments/ Economics of Education
The positive relationship between the level of education and the level of earnings is one of the most robust relationships observed by labor economists. Typically, this relationship is explained using the human capital model. Human capital, in this model, can be thought of as a measure of an individual's productive capacity. Under the human capital model, it is assumed that the level of an individual's earnings is determined by the individual's stock of human capital. An individual can increase his or her human capital by investments in:
- education,
- training, or
- health care.
For now, we'll focus on the first of these types of investment. Most economic models of educational attainment are based on the assumption that individuals select the level of educational attainment that results in the highest expected present value of lifetime earnings (net of educational costs). Simply stated, this means that a person will attend college only if the present value of the expected benefits exceeds the present value of the expected costs associated with this choice.
Costs of education
There are three types of costs associated with a college education:
- direct costs such as tuition, books, and supplies,
- forgone earnings (the opportunity cost of time), and
- psychic costs.
Notice that the direct costs include only those direct expenditures that a student would make only if he or she attends college. The costs of meals, dorm fees, etc., would not generally be a cost of education since these individuals would face costs of meals and lodging if they had been engaged in some alternative use of time (such as working). Room and board fees would partially enter as a cost only if these costs are higher than they would have been under the next-best alternative use of time.
As noted earlier, the forgone earnings associated with being a full-time student is usually the largest cost associated with acquiring a college or advanced degree.
The psychic costs associated with attending college include the stress, anxiety, and sometimes boredom associated with classes, exams, assignments, papers, etc.
Benefits of education
The benefits associated with acquiring a college degree include:
- higher expected earnings,
- more pleasant jobs,
- lower expected unemployment rates, and
- psychic benefits.
In general, college graduates receive not only higher pay, they also receive jobs that are more secure and involve less tedious work, less physical work, more pleasant work environments, better working conditions, higher social status, and so forth. The psychic benefits associated with education include the enjoyment that may be received by being in the college environment.
Optimal investment in education
An individual will acquire additional education as long as the present value of the marginal benefits from this additional education outweighs the present value of the marginal costs. Those individuals who have higher benefits and/or lower costs will acquire more education. The diagram below illustrates the effect of changes in MC and MB on the optimal level of human capital investment.
The costs and benefits associated with deciding to acquire a bachelors degree are represented in the diagram below. This diagram illustrates two possible earnings streams facing an 18-year old high school graduate. The costs associated with college attendance includes both forgone earnings (the upper portion of the area shaded in red in the diagram) and the direct costs of college (the lower rectangle that is shaded in red in the diagram below. Note that this diagram suggests that a 22-year old college graduate earns less than they would have at this age if they had gone to work directly after high school. On average, it takes approximately 6-7 years for the earnings of a college graduate to catch up to the earnings of a high school graduate with identical observable characteristics. The area shaded in light blue represents the increase in earnings that a college graduate would be expected to receive over the rest of his or her worklife. (These earnings streams, of course would differ across individuals due to differences in individual ability and costs.)
It is expected that an individual would attend college if the present value of the costs (the red area above) is less than the present value of the benefits (the light blue area above).
Factors affecting human capital investment
The human capital model suggests that the level of human capital investment is affected by:
- interest rates,
- the age of the individual,
- the costs of education, and
- the wage differential between high school and college graduates.
Since most of the benefits associated with a college degree occur relatively later in the lifecycle while the costs are borne more immediately, an increase in the interest rate facing an individual will be expected to lower the net benefit of education. (This occurs because an increase in the interest rate lowers the present value of more distant benefits and costs by more than it lowers the benefits of short-term benefits and costs.) Government subsidized student loan programs are designed to reduce interest rate differentials across households. (In the absence of these subsidized interest rates, low-income households would face substantially higher interest rates, resulting in a lower probability that children from such households will attend college.)
It is expected that individuals will tend to invest more in education at an earlier stage of their lifecycle because this results in a larger period over which the increased earnings may be realized. (Exceptions to this often occur when individuals change careers.)
The theory discussed above, of course, directly predicts that more people will attend college when the costs are lower and/or the benefits are higher.
Age-earnings profiles
As your text notes, age-earnings profiles, for a given level of educational attainment, are generally concave. This means that earnings increase at a progressively slower rate as the individual ages (holding constant the level of educational attainment). The simplest explanation for this is that individuals invest in a larger quantity of on-the-job training at earlier stages of their worklife. Evidence also suggests that those who invest in more education also invest in a larger amount of on-the-job training (for similar reasons). This results in a widening in the gap in earnings across educational levels as individuals age.
Gender, education, and age-earnings profiles
One of the reasons for the historically lower level of educational attainment for females is that females tended to have significantly shorter expected worklives than males. In recent years, however, increases in female labor force participation have narrowed the gap in expected worklife between males and females. This increase in expected worklife is one of the reasons for the rather dramatic increase in female college enrollment rates in recent decades.
Does college attendance pay off?
Estimates of the rate of return to education are determined by comparing the expected lifetime earnings streams that an individual could receive under alternative levels of educational attainment. Roughly speaking, estimates of this rate of return to a bachelors degree are derived by comparing the earnings streams of college graduates with the earnings streams of high school graduates who have equivalent observable characteristics. Estimates derived in this manner suggest a rate of return to investment in education in the range of 5-12%. Numerous studies suggest that this rate of return has increased in recent years.
There are, however, a few potential sources of bias in these estimates.
If college graduates differ from high school graduates in terms of unobservable differences in ability or motivation, those who attend college might have received higher earnings even if they had not attended college. The same argument suggests that high school graduates would not earn as much as college graduates do if they had instead chosen to attend college. In this case, the return to education would overstate the increase in earnings that would be received by individuals. This type of bias, called "ability bias," suggests that the observed difference in earnings between high school and college graduates overstates the return to education that would be received by a given individual.
The return to education, however, may be understated by a comparison of the earnings of high school and college graduates as a result of the non-pecuniary returns received from education. Some of the benefits from education involve increased productivity in non-market activities. College graduates also receive more pleasant jobs. These nonpecuniary benefits from education result in a smaller wage difference between high school and college graduates than would have been received if their jobs were equivalent in all dimensions except for the wage.
Still another possibility involves the existence of selectivity bias. Willis and Rosen (1979) found that those who attend college perform relatively well in the types of jobs that college graduates receive while those who do not attend college perform relatively well in the types of jobs that high school graduates receive. This suggests, for example, that a good lawyer may be a poor carpenter while a good carpenter may be a poor lawyer. In this case, the return to college is relatively large for those who attend college for two reasons: they do well in the types of jobs that college graduates receive while they would have received relatively low wages if they had not gone to college. Their results also suggest that the return to education is relatively low for those who choose to not attend college. Similar results in a later study were found by Kane (1986).
Is the level of educational investment socially optimal?
While there is substantial evidence that higher education provides a good investment for individuals, it is less clear that the level of educational spending is socially optimal. Education in the most countries is heavily subsidized by the government (at least through the secondary level). In the U.S., public education is free through high school and higher education is subsidized in both public and private institutions. Subsidies of this sort will result in an optimal level of investment only if the marginal subsidy equals the value of the marginal external benefits associated with education. Evidence on this is somewhat mixed. There is, however, substantial evidence that economies that invest more in education tend to experience higher rates of economic growth.
The signaling model, however, raises substantial questions about social returns to education. Under the signaling model, education does not raise any worker's productivity. Instead, it allows firms to sort workers according to their productivity. The signaling model suggests that firms cannot directly measure the productivity of individual workers (at least not when they are initially hired). Over time, though, firms observe that college graduates are more productive than high school graduates. This results in higher pay for college graduates and lower pay for high school graduates (as compared to a situation in which this educational "signal" did not exist). Under this model, the benefits to a college degree are the same for all workers (since all workers with a college degree receive the higher pay). Low ability individuals, however, are assumed to face higher costs of education (it requires more time and effort for a low ability worker to make it through a bachelors degree). Thus, only high ability find it profitable to attend college in an equilibrium in this model.
Since education does not raise any individual's productivity under the pure signaling model, society as a whole does not benefit from higher education under this model. If this model is correct, total output would increase if institutions of higher education were shut down.
It should be noted that, among economists, the human capital model is much more widely accepted than the signaling model.