A parallel analysis applies in determining the number of additional workers a firm will hire when a credit is introduced without there first being a decline in labor demand. For example, in figure 3.2, if current employment is L2, the firm would hire L, -L2 additional workers if a VBEC is introduced with base L, and credit rate of 8 percent of the wage.

Thus far the analysis has been confined to the response of an individual competitive firm which perceives the supply of labor as being infinitely elastic. With over 7 percent of the work force currently unemployed, it is tempting to assume that the general level of wages will not have to rise to prompt an increase in the quantity of labor supplied.

The short-run aggregate supply of labor, however, is unlikely to be perfectly elastic. This follows because labor markets are heterogeneous, job information is imperfect, and the opportunity cost of entering the work force is the loss of transfer payment. Empirical evidence suggests that short-run elasticities of labor supply, while appreciably higher than long-run elasticities, are finite. Estimates made by Lucas and Rapping, for example, indicate that a 1 percent increase in the real wage will increase the quantity supplied of labor from 1.7 to 4 percent. Therefore, if the VBEC increases aggregate demand for labor, the wage rate must rise if employment is to increase. The increase in the tax credit rate required to achieve a specified increase in employment will be larger than for the infinitely elastic case and will depend, as it did in the macroeconomic model, on elasticities of supply and demand for labor. In general, the more inelastic labor demand and supply, the greater must be the credit to achieve a given increment in employment.

3.3. THE RELATIONSHIP BETWEEN A VBEC AND THE LENGTH OF THE WORKWEEK

The labor market analysis presented in previous sections presumes a direct, constant relationship between employment (number of workers) and man-hours of labor services. Man-hours, however, are the productive input which firms hire to produce goods and services and are the conceptually correct measure of labor services. A definition of man-hours is: man-hours equals utilization multiplied by employment, where utilization is measured as hours per day (week) and employment is measured as number of workers. Given constant utilization (for example, an 8-hour workday), an increase in the number of workers directly translates into higher man-hours and greater real output. Whenever it is possible for firms to substitute hours worked for employment, this simple, direct link no longer holds. Specifically, man-hours will decrease if utilization falls and employment stays constant; thus, there are different combinations of employment and utilization which equal the same number of man-hours. For example, 80 employees working 7 hours per day yield the same man-hours as 70 employees working an 8-hour day. While man-hours in the above cases are identical, 560, employment is 10 workers greater in the first situation.

2 Technically, manhours is a flow of labor services and employment is a stock. Utilization transforms the stock variable into a flow.

In terms of capacity of the economy, the conceptually preferred definition of unemployment is also a flow of man-hours of labor services equal to the difference between man-hours offered (supplied) and man-hours actually worked (demanded). Reported unemployment statistics, however, present a stock of workers rather than a flow of man-hours. Unemployment, as reported by the Bureau of Labor Statistics, is the difference between those who are offering labor services and those who are actually selling some positive amount of labor services. Specifically, a person who is willing to work more hours at a given wage than firms are willing to purchase at that wage is unemployed by the above definition; yet this unemployment is not reported in unemployment statistics.

During periods of declining business activity, many firms initially reduce the number of hours worked per day per worker. Only after this adjustment in utilization has been made are firms likely to begin to lay off workers. Consequently, unemployment as a flow of manhours, will rise more quickly during contractions and fall more quickly during expansions of business activity than reported unemployment. There is substantial empirical support for this proposition. For example, Rosen and Nadiri (p. 268) report that "*** patterns of adjustment indicate that utilization rates are truly variable inputs and that there is a hierarchy of adjustment speeds among stocks, ranging from fairly rapid adjustment of production employment to rather slow adjustment of capital stock."

Even after short-run utilization adjustments have been made during a recession, firms may retain currently unproductive workers. This behavior can be explained by the existence of certain nonwage related labor costs which are associated with changing the level of employment. The willingness of firms to hoard labor, that is, to maintain the level of employment even though these workers are not producing real output, depends upon the hourly wage, and nonwage costs of hiring, training, and firing (Clark). In general, the higher the ratio of nonwage costs to wage costs, the longer it is profitable for firms to hold nonproductive labor.3

Given this brief discussion of the relationship between man-hours, utilization, and employment, what are the implications for a VBEC? The answers depend, in part, on the policy objectives of the program. If the purpose of the credit is to reduce the reported rate of unemployment, it is possible to do this without increasing, and possibly decreasing, man-hours worked. If the purpose is to maintain or expand man-hours, this conceivably could be accomplished without large, initial effects on the number of unemployed workers.

If a credit is offered as some amount per eligible worker, firms may hire additional workers and reduce the length of the workday. That is, firms might replace full-time with part-time workers, and thereby receive tax credit even when man-hours and real output are falling. Here, an employment tax credit will simply encourage substitution of the eligible component of man-hours for the ineligible. Also, part

Extensive literature is developing which treats employment as a quasi-fixed input and attempts to explain the apparent "labor hoarding" that occurs during contractions of business activity; see Oi, Tinsley, and Clark. Under a rather restrictive set of production conditions, Clark derives the maximum length of time that it is profitable for a firm to keep a worker on the payroll without having him work as t* equals h plus f divided by W, where h is hiring cost, fis layoff cost, and W is the wage rate. In this formulation, a reduction of wages will increase the length of time a firm will hoard labor.

of the program's cost will fall on employed workers who are forced to take a reduction in hours worked. While it is unlikely that unions will tolerate significant reductions in the workweek which penalize established members and reward newer workers, this phenomenon might develop in less organized sectors where the credit is likely to have the largest impact.

If the credit is offered as some percentage of the hourly wage, as we propose, it is possible that a firm will increase man-hours by raising utilization for its current work force. Here, even if employment does not change, there is an actual increase in labor services. Furthermore, as employed workers begin to work the normal hours per week, the number of employed will be gradually adjusted upward.* Specifically, firms begin to hire only when there is upward pressure on utilization. As utilization begins to approach the normal rate, employment will typically increase since further efforts to expand utilization will entail increases in the hourly wage.

Legislative proposals sometimes suggest offering a credit on the hourly wage of eligible workers for a normal length workday. While this approach may avoid some abuses, for example, substitution of full with part-time employees, it also reduces a firm's ability to shorten the workday while maintaining employment, or at least to substitute utilization decreases for decreases in employment. A VBEC that credits the hourly wage permits firms to decide for themselves the appropriate mix between utilization and employment. Also, a percentage reduction of the wage rate will not distort the pattern of expenditure between unskilled and skilled workers, in contrast to a credit paid on each eligible worker.

3.4. ADMINISTRATION OF A VBEC

The most efficient way to administer an employment credit program is through the existing payroll tax system. As previously discussed, an increase in a VBEC, which offers a percentage reduction on the hourly wage, is identical to a short-run reduction in payroll taxes. Thus, the most comprehensive way to administer a VBEC is to provide employers a rebate against payroll tax liability. In effect, payroll taxes are thereby adjusted to stabilize employment and output, that is, they take on a countercyclical fiscal role. For political and psychological reasons, the rebate can be called an "employment tax credit,' and the bookkeeping can be separately kept from social security contributions.

Payroll tax rebates are preferable to providing a credit against corporate profit tax liability since many firms have zero or insignificant taxable income during periods of declining business activity, and those are the periods when the VBEC is most needed. While a credit against corporate profits can be deferred, it is easier and more immediate to reduce payroll taxes which must be paid regardless of business conditions.

Another advantage of attaching the credit to the payroll tax structure is the possibility of extending the coverage to encompass employment

For a complete, conceptual discussion of possible relationships between utilization and employment, see Grossman.

in nonprofit organizations and possibly State and local governments. These institutions employ a growing percentage of the labor force and presumably would respond to a wage bill subsidy in the same manner as profit-motivated businesses. For State and local governments, the subsidy against their wage bill can be considered a cheaper and more efficient alternative to public service employment.

19

4. SUMMARY AND CONCLUSION

The purpose of this paper has been to examine policy implications of a universally applied employment tax credit. The program is intended to be a short-term supplement to existing fiscal and monetary policy tools. The basic rationale of the program is to reduce the cost of labor to business firms and thereby initiate an increase in the utilization of labor. Presumably, the credit will be instituted during periods of declining business activity, and during periods when actual employment is below potential employment. The major conclusions in this investigation are:

(1) In the context of a standard multiequation macroeconomic framework, our analysis and numerical calculations indicate that a variable base employment tax credit will increase employment and real output without necessarily increasing the aggregate level of prices.

(2) In contrast to reductions in personal and corporate income taxes, which act to expand aggregate demand and thereby aggregate prices, a universal employment credit will increase both aggregate demand and aggregate supply. Thus, under a variety of financing strategies, the employment tax credit will have a dual impact on employment and output, expanding both supply and demand. This dual impact will typically help to mitigate price rises usually associated with an expansionary fiscal program.

(3) Much of the criticism of employment tax credit policy is predicated on the assertion that short-run labor demand relationships are not responsive to the price of labor services. We offer the following response to this argument. There is not a body of empirical evidence which points to zero price elasticities of demand for labor in the short run. Even in the more capital intensive sectors of the economy such as manufacturing, empirical evidence indicates that while price elasticities of demand for labor are low, they are not zero (Coen and Hickman, Rosen and Nadiri, Tinsley and Berndt, Kesselman and Williamson). Also, given the general pattern of rising wages and employment experienced in the United States during the postwar period, it is unlikely that price elasticities of demand can even be identified statistically. Further, there are no reasons to believe, either conceptually or empirically, that demand for labor is not responsive to changes in price in the more labor intensive sectors of the economy where over half the U.S. work force is employed.

(4) Another criticism of employment tax credit policy is that it will provide windfall profits to business firms, particularly during the expansion phase of business activity when firms are intending to increase their utilization and hiring rates anyway. There are three responses to this criticism. First, a variable base program calls for adjustment of the credit base to accommodate changing business conditions. The base can be decreased during declining periods of