Teaching tools: why workers should want mandated benefits to lower their wages.

Lee, Dwight R.

I. INTRODUCTION

I began this paper as Congress was debating whether health reform legislation should require employers to pay 80 percent of their employees' health insurance, or only 50 percent. While this debate no doubt had political relevance, economists know that the percentage of a mandated benefit that employers are required to pay directly is irrelevant to how the cost is ultimately shared between employers and employees. Every student of economic principles should be able to show that the cost of a mandated benefit is partially paid by employees through lower money wages and partially paid by employers though higher compensation expenses, with the division of the cost being determined by the relative slopes of the supply and demand curve for labor.(1) The legislative allocation of the cost of a mandated benefit is completely irrelevant to the actual allocation of that cost.

That politicians spend time debating the cost allocation of mandated benefits suggests that they do not understand, or feel they can ignore the possibility that the public understands, the vacuousness of such legislative determinations. Political discourse on mandated benefits does, however, occasionally acknowledge that mandated benefits may result in lower wages. Invariably this possibility is emphasized by those who oppose the mandated benefit and downplayed (or denied) by those who favor it. Social Security provides a good example of the disagreement on the effect of mandated benefits on wages. Those who would radically reform, or abolish, Social Security argue that most, if not all, of the Social Security tax is paid by employees. For example, Ferrara [1982, 13] states, "Although the payroll tax technically is assessed half against the employer and half against the employee, the employee really bears the entire burden of the tax." On the other hand, Robert J. Myers, who served as Chief Actuary of the Social Security Administration from 1947 to 1970, discounts the argument that employees pay some of the employers' Social Security taxes by questioning "on grounds of general reasoning, whether this can actually be proved."(2)

The implicit assumption in these debates is that the more employees end up paying for a mandated benefit through lower wages, the less they gain from having the benefit mandated. It seems reasonable that the less workers sacrifice in wages for a mandated benefit, the better off they are.(3) In fact, exactly the opposite is true. I shall show with simple supply and demand analysis that the more wages decline because of a mandated benefit the better off workers are. The strongest case proponents of a mandated benefit can make is that workers would pay more for the benefit, with lower wages, than it would cost to provide. And conversely, the strongest case opponents can make is that the mandated benefit would increase wages. The key to this result is in recognizing that a critical influence on the wage effect of a mandated benefit is the value workers place on that benefit. Everything else equal, the more a mandated benefit is worth to workers, the more wages will decline when it is provided.

The value of mandated benefits to employees has been recognized as important in evaluating their desirability. For example, Summers [1989] points out that the more employees value a benefit the less the deadweight loss associated with mandating it when compared to having government provide the benefit and financing it though general taxation. But even economists seem to have ignored an important implication of a mandated benefit's value by failing to point out that the more employees pay for the benefit, relative to its cost, the stronger the case for the benefit from the perspective of employees as well as employers.

II. THE ANALYSIS

I begin by considering an unregulated equilibrium in a labor market; that is, one prior to the enactment of any mandated benefits. This equilibrium is shown in Figure 1, where the number of workers is measured along the horizontal axis and the money wage is measured along the vertical axis. The demand curve for labor is given by D, the supply curve is given by S, and the market clearing wage is W. The number of workers hired is L and the workers' surplus is given by area abW. Against the benchmark of this unregulated equilibrium, we can see the effect of imposing a mandated benefit under different assumptions about the net value of the benefit.

Consider first a mandate that employers provide a benefit that each employee values by an amount exactly equal to the per-employee cost of providing it. The effect of this mandate is to shift both the demand curve for, and the supply curve of, labor down by the per-employee cost of the benefit.(4) The shifted demand and supply curves are shown in Figure 1 as D[prime] and S[prime] respectively. As shown in the figure, the wage declines from W to W[prime], exactly the amount of the value and cost of the benefit per employee, and employment remains at L, leaving general economic welfare unchanged. From the perspective of workers, although the composition of their compensation has changed, its value has remained the same, as has their employee surplus.(5) Workers are paying the full cost of the benefit, but this does not constitute a case either for or against mandating it.(6)

Making Work a Pleasure

When the cost workers pay with lower wages for a mandated benefit is greater than the cost of providing the benefit, the case for the benefit is the strongest, and the workers receive the most benefit from its provision.(7) To consider an extreme case of such a possibility, assume that employers could arrange for their workers to benefit from an extremely enjoyable form of on-the-job entertainment with no productivity loss or cost of any other type. Because the cost of providing this employment-related benefit is zero, the demand curve for labor continues to be represented by D in Figure 1. The supply curve, however, shifts down significantly to S[double prime], as workers are willing to work for far lower monetary compensation than before. Indeed, as shown, [L.sub.S] workers are willing to pay (accept a negative wage) to go work, and therefore represent the amount of labor supplied at a zero wage. The new equilibrium finds L[double prime] workers employed at a wage of W[double prime]. Although the wage has declined, workers are better off, with employment expanded and worker surplus increasing from area abW in Figure 1 to area a[double prime]dW[double prime]. Furthermore, even workers originally employed at the higher wage are better off, as their workers' surplus increases from abW to a[double prime]b[double prime]cW[double prime]. If someone proposed to mandate this entertainment benefit, it would be economic nonsense to argue against it on the grounds that far more than the cost of providing it would be passed on to workers in wage reductions. And it would be just as nonsensical to argue for the proposal by attempting to deny that a large wage decline would result.

Of course, when it is possible for employers to provide job-related benefits at a cost that is less than the value employees receive, a mandate is not likely necessary. Employers are better positioned to discover value enhancing fringe benefits than are remote political authorities and need no prodding to provide such benefits when they are discovered.

The Most Plausible Case

More likely a proposed mandated benefit is worth less than it costs. In this case providing the benefit harms workers, even though they pay less for it through lower wages than it costs. This situation is illustrated in Figure 2 where we begin with demand and supply curves D and S respectively, with the initial equilibrium calling for a wage of W and employment of L. The mandate causes the demand curve to shift down to D[prime], with the vertical distance between D and D[prime] equal to the cost of providing the benefit to each worker. In response to the mandate the supply curve shifts down to S[prime], with the vertical distance between S and S[prime] equal to the value of the benefit to each worker, which by assumption is less than the vertical distance between D and D[prime]. The new equilibrium results in a wage of W[prime] and employment of L[prime]. It is easily seen that the decline in the wage, given by distance db[prime], is less than the cost per worker of the benefit, given by distance be. But the total workers' surplus has also declined, going from area abW in Figure 2 to area a[prime]b[prime]W[prime], in part because of a reduction in employment from L to L[prime]. But even workers who maintain their jobs suffer a loss of workers' surplus when the benefit is mandated. The L[prime] workers employed with the benefit realized a workers' surplus of acdW before the benefit was provided, which is clearly larger than the surplus of a[prime]b[prime]W[prime] realized after it is provided.

The less workers pay for the benefit in terms of a lower wage, the worse off they are, even the ones who keep their jobs. If, for example, workers see the benefit as worth less than they originally thought, the amount they pay for it will decline as the labor supply curve shifts back up to the left and the wage increases. For example, assume that workers reconsider the value of the benefit just discussed, and decide it is worthless. This shifts the supply curve back to S in Figure 2 and increases the wage from W[prime] to W[double prime]. But not only does this reduce total workers' surplus as employment declines from L[prime] to L[double prime], it reduces the surplus of the L[double prime] workers who keep their jobs. Given the previous evaluation of the benefit, the surplus to these L[double prime] workers is given by area a[prime]fgW[prime], but when the benefit is seen to be worthless the surplus to the L[double prime] workers is given by the smaller area ahW[double prime].

Higher Wages Mean Negative Benefits

To reinforce the idea that the less workers pay for a mandated benefit the less they gain (or the more they lose) from it, I consider next a situation in which the mandate actually increases the wage. Going to the opposite extreme from great entertainment on the job, consider a mandated "benefit" that takes the form of periodic and painful electric shocks on the job. The cost of providing this "benefit" would exert a downward influence on wages by shifting down the demand curve for labor. But it is easy to construct a situation where this downward influence is more than offset as the negative value workers place on the "benefit" is reflected in an upward shift in the supply of labor. And it is also easy to show that employment would decline in this situation and that those who maintained their jobs would be worse off than they were before the "benefit" was provided and their wage increased.

Admittedly, it would take a political process even more perverse than the one with which we are familiar to require applying painful electric shocks to workers and call them mandated benefits. Regardless of the plausibility of the example, however, the point is clear. A mandated benefit that avoids the "problem" of lower wages for workers (which advocates of mandated benefits often claim is the case) would be a mandated benefit that is worth less than it costs and would harm workers.

III. PAYING LESS FOR SOCIAL SECURITY

I now return briefly to the issue of Social Security. In accordance with the above analysis, the amount that workers pay for Social Security has probably decreased over the years, with the situation falling more in line with that often claimed to be the case by Social Security supporters; i.e., that workers do not pay their employers' Social Security taxes. For those who entered into the program in its early years, Social Security contributions came with the credible promise of a significant rate of return. Because the Social Security contribution of employers began as a valuable benefit, workers were willing to pay for it in terms of lower wages, and no doubt they did. The value of Social Security contributions has declined over the years; few entry-level workers now have confidence that they will receive a positive rate of return from these contributions. Since they attach little value to their employers' contributions, workers are no longer willing to accept lower wages in return for them.

So advocates of Social Security are increasingly correct when they argue, or simply assert, that employers pay their legislatively mandated share of the Social Security tax. But they are increasingly correct for reasons that few of them are likely to use in making their case; i.e., Social Security is not worth much anymore.(8)

IV. CONCLUSION

Opponents to a mandated benefit commonly argue that much, if not all, of its cost will be passed on to workers in lower wages. Supporters commonly deny, or downplay, this argument with claims that most of the cost is paid by employers. It probably makes sense politically for the two sides to take the positions they do. But simple economic analysis shows that the opponents are making the stronger case for the mandated benefit, with the advocates arguing that the benefit is worth less than it costs and that mandating it would make workers worse off.

Everything else equal, the less workers end up paying for a mandated benefit though a lower wage, the less it is worth to workers and the less advantage (more harm) would result from mandating it. And if mandating a benefit leaves the wage unchanged (or increased), then workers must be placing a negative value on the benefit - they would be willing to take a lower wage to rid themselves of it.(9) Only if workers pay more for the mandated benefit with lower wages than it costs to provide will they benefit from having the benefit mandated, in which case it probably does not need to be mandated.

Those who see a government mandate as necessary to get employers to provide a benefit to their workers are more correct than most of them realize if they argue that the cost of the benefit will not be passed on to workers. The more worthless the benefit, the more correct their argument is.

1. It is typically stated that the division in cost is determined by the relative elasticities of the supply and demand for labor. Graves, Lee, and Sexton [1994] have shown, however, that comparing the slopes of the supply and demand curves is sufficient to determine the incidence of taxes or mandated benefits.

2. See Myers [1981, 361]. More commonly, supporters of Social Security simply assume that the statutory distribution of a tax burden is the actual distribution. For example, in a laudatory account of the Social Security system Schottland [1963, 187] states, "In workmen's compensation in virtually all states, the employer bears all of the cost; in unemployment insurance, the employer bears the cost; in old-age, survivors, and disability insurance, the cost is shared equally by employer and employee."

3. Of course, the less the decline in wages the larger the number of workers who will lose their jobs. I ignore the welfare implications of this unemployment effect here to give the proposition (the lower the wage decline, the better off workers are) the benefit of the doubt. When mandated benefits cause some workers to lose their jobs (which, as we shall see, is not always the case), I consider the well-being of only those who remain employed.

4. I assume that all adjustments in employee compensation take place in money wages. This ignores possible adjustments in nonmandated fringe benefits, but allows the pass-through effect from a mandate to be easily represented. Also, I will maintain throughout the assumption that all workers place the same value on the mandated benefit and that the cost of providing the benefit to another worker is constant. This assumption simplifies the discussion by allowing the demand and supply effects of a mandated benefit to be represented with uniform shifts in the respective curves.

5. Employee surplus is given in Figure 1 by area abW before the benefit is mandated and by area a[prime]b[prime]W[prime] afterward. These two areas are obviously equal in size. We ignore here the well-discussed implications of the fact that fringe benefits are typically untaxed while money income is taxed.

6. I ignore here the cost of enacting and enforcing any mandated benefit, even one that otherwise leaves no one better or worse off.

7. Indeed, it is obvious from a minor alteration in the previous analysis that only when workers pay more in lower wages for the benefit that it costs to provide will providing the benefit be economically efficient and improve the welfare of workers. The only way to reduce the wage in Figure 1 by more than the per-employee cost of the benefit is for the supply curve to shift down by more than the demand curve, which implies that the benefit is worth more than it costs to provide.

8. Using data from as long ago as the late 1960s and early 1970s, Hamermesh [1979] estimated that only about one-third of the employer Social Security tax was being passed through to workers.

9. This assumes, as opposed to the on-the-job entertainment example, that the cost of providing the benefit is positive. With a positive cost of provision it is also possible that mandating it would cause a decrease in the wage even though workers place a negative value on it. And, of course, mandating a benefit can make workers worse off even if the benefit has positive value.

REFERENCES

Ferrara, Peter. Social Security Reform: The Family Plan. Washington, D.C.: The Heritage Foundation, 1982.

Graves, Philip, Dwight Lee, and Robert Sexton. "Slope Vs. Elasticity and the Burden of Taxation." Forthcoming, Journal of Economic Education.

Hamermesh, Daniel. "New Estimates of the Incidence of the Payroll Tax." Southern Economic Journal, 45(4), 1979, 1208-19.

Myers, Robert. Social Security, 2nd. ed. Homewood, Illinois: Richard D. Irwin, 1981.

Schottland, Charles. The Social Security Program in the United States. New York: Appleton-Century-Crofts, 1963.

Summers, Lawrence. "Some Simple Economics of Mandated Benefits." American Economic Review, 79(2), 1989, 177-83.

Bernard B. and Eugenia A. Ramsey Professor of Economics and Free Enterprise at the University of Georgia, Athens. I gratefully acknowledge comments from Richard McKenzie and Robert Moore while I was preparing and revising this paper. I would also like to express my thanks for the useful comments from students in an economics class at Brigham Young University where I presented this paper in November. Any errors or absurdities are the entire responsibility of the author.