EFFECT OF EITC ON WORK EFFORT LIKELY TO BE SMALL, CRS REPORT SAYS.

                       CRS REPORT FOR CONGRESS

Recent criticisms of the Earned Income Tax Credit (EITC) include a concern about the possible negative effects of the credit on work effort. Although the EITC would be expected to discourage work effort for some individuals already in the work force due to the phase-out of the credit, the EITC should, at the same time, encourage more individuals to enter the labor force. The overall effects of the EITC on work effort, while uncertain, are likely to be small and could be either positive or negative. The EITC is different from more traditional anti-poverty programs, whose adverse work incentives are more clear.

INTRODUCTION

The EITC is targeted to the working poor: it pays a percentage of wages up to a ceiling, then becomes a flat amount (a plateau) across some wage levels, and finally is phased out as income rises. ¹ The ongoing expansion of the EITC enacted in 1993 was designed to ensure that a typical family could, even though working at a minimum wage job, earn sufficient amounts to be above the poverty line. Indeed, the EITC performs some of the same functions as would a higher minimum wage. A minimum wage can raise the wages of lower income workers but implicitly carries a cost (if it is binding) in that it can also reduce employment of workers whose productivity falls below the minimum. ²

Because it is paid only to low-income workers, not those who do not work, the EITC is often characterized as a work incentive. However, some commentators have criticized the EITC, and the expansions adopted in recent years, on the grounds that it decreases work effort for many individuals. ³ One critic, in fact, claimed that many families would actually be better off (have higher money income) without the EITC. ⁴ These criticisms have frequently concentrated on the potential reduction of hours for those already working and who would be working if there were no EITC. These criticisms, however, disregard those who entered the labor force because of the EITC.

The effects of the EITC may be contrasted with other programs for the poor. Traditional transfer programs, which normally consist of a flat grant, or a benefit like medicaid, that is phased out with income, unavoidably discourage work. While it is possible for the EITC to reduce overall labor supply, the EITC can increase labor supply and should increase labor force participation of single parents and mature single workers.

Moreover, any change in work effort in response to the tax is voluntary on the part of the individual worker, in contrast to the minimum wage, which can cause involuntary unemployment. An individual could not be made worse off financially by the EITC unless he voluntarily decides to work less and presumably would, in any case, be trading paid work for leisure or unpaid work (e.g. providing services in the household).

EFFECT OF THE EITC ON WORK

THEORETICAL CONSIDERATIONS

To understand the likely consequences of the EITC for work, it is easier to begin with an analysis of the effect of wage changes. (Tax changes normally produce similar effects.)

When wages rise, two opposing effects on labor supply can occur for those who are already working. First, the higher earnings permit the individual to maintain or increase earnings (and consumption) while working fewer hours (and thus enjoying more leisure). This "income effect" discourages work. Secondly, the higher wage makes the "price" of leisure (in terms of forgone consumption) higher and thus discourages leisure. This "substitution effect" encourages work. Thus, higher wages can either increase or decrease hours worked.

There is also a labor participation decision -- a wage change may cause individuals to enter or leave the work force. When there are no other family earnings, a wage increase encourages entry into the labor force. (Secondary workers in a family (usually married women) may have an income effect due to the earnings of the primary worker if the wage change affects the husband.)

The effect of the earned income credit is a bit more complicated. Up to a certain point in earnings, the EITC is a percentage of wages and acts just like a wage increase. Then there is a plateau, where the credit is a flat amount and acts just like a grant. Finally there is a phase-out range, where the credit is gradually reduced, which is the same, at the margin, as a tax.

In fact, there are four categories of workers who face different effects from the EITC.

(1) Individuals not already in the work force and who have no family members receiving the credit, whose participation will, in theory, unambiguously increase because of the substitution effect. (For secondary workers, participation may not increase if the primary earner receives the EITC).

(2) Individuals who would be working without any EITC, whose credit is a percentage of wages, where the effects on hours worked are ambiguous due to offsetting income and substitution effects.

(3) Individuals in the plateau who should reduce work effort because there is only an income effect (the credit is not affected by changes in hours).

(4) Individuals in the phaseout range who should reduce work effort because the EITC produces an income effect that discourages work, and it also produces a negative substitution effect (each additional hour worked reduces the credit). In this case, both the income and substitution effects discourage work.

The aggregate effects, depend, of course, on both the distribution of individuals in these categories and the magnitude of their responses.

Before turning to the empirical evidence, it is important to clarify two potential misconceptions about the EITC and work effort.

First, when we observe the current recipients of the EITC, we find that most of the recipients (about two-thirds) and most of the earnings (about 80 percent) are in the phase-out range. ⁵ But that does not mean that all of these individuals had a disincentive to work or that the EITC discourages work. If most individuals who enter the labor market obtain full time jobs and earn wages somewhat above the minimum wage, we would expect most EITC recipients to fall in that category, but we would not know how many of them decided to enter the labor market because of the EITC or how their hours of work were influenced. This observation tells us nothing about the overall effect of the EITC.

Secondly, even if families have less disposable income with the imposition of the EITC, these families do not have their economic conditions worsened by the EITC; rather, they have made a voluntary choice of having more leisure time which also has value to them. Moreover, the families where this effect is most likely to occur are those toward the top of the phaseout range -- that is, those who are likely to have incomes above the poverty line.

EMPIRICAL ESTIMATES

There is a large body of literature that tries to estimate the labor supply response to wage changes or even directly to negative income taxes. On the whole, this literature has suggested a small overall response for men and widely varying responses for women. A number of studies have used these estimates to try to determine the consequences of the EITC, or changes in the EITC, for work effort.

A recent study by Browning ⁶ has recently been cited by critics of the EITC. Browning suggests that the credit discourages work and that an increase in EITC benefits causes a significant proportion of the lower income population to have a large enough response that their disposable income is reduced. Browning, however, examines a small change in the credit while maintaining the phaseout range. This type of change is most likely to produce a negative effect on work. He indicates that about 40% of families in the phaseout range will eventually have less disposable income with a small increase in the rate, and that the distortions are very large relative to the budgetary cost of the rate change. Browning, however, does not explicitly simulate any actual policies and cites no labor supply responses that would be relevant to the 1993 expansion of the EITC.

There are some reservations about the Browning results, even for the limited experiment he is considering. First, he focuses on individuals in category (4) described above. Although he does include a section providing some quantitative estimates for individuals in categories (2) and (3), be does not address the participation decision (other than a qualitative discussion) by individuals in category (1). Secondly, his labor supply elasticities are somewhat large. In calculating his substitution elasticity of 0.3, he averages in a very large elasticity for women (0.75) with a small elasticity for men (0.1). This female wage elasticity is drawn from the empirical literature on response to wages which reflects both a negative participation response and a negative hours response to lower wages. Yet the EITC would provide a positive participation response, not a negative one, and Browning's analysis focuses on hours, not on participation. Moreover, at least some researchers now believe that the estimated female hours response is too large. ⁷

Other studies have suggested that the effect of the EITC on work is small and that the EITC may increase rather than decrease work.

Dickert, Houser and Scholz present direct estimates of the labor supply effects of the changes in 1993, incorporating both participation decisions and hours worked, with a range of elasticities. ⁸ Their results suggest modest effects ranging from an increase in labor supply of 0.37% to a reduction in supply of 2.05%. Their results, therefore, suggest negligible effects on labor supply (that could be either positive or negative).

A simulation study by Holtzblatt, McCubbin and Gillette ⁹ also suggests relatively small effects for families with children. (They do not consider the effects on workers without children.) The effect of the 1990 revisions was estimated to reduce earnings overall by 1.7 percent; the combined effects of the 1990 and 1993 changes were estimated to reduce earnings overall by 2.4% -- for an overall difference of 0.7% for the 1993 changes. These effects derive from a change in hours worked for those already working. This study did not include participation responses, which would be positive and would, therefore, lower the negative overall effects, or perhaps cause an overall positive effect.

A recent study by Eissa and Liebman ¹⁰ tried to directly examine the response of single women, by comparing the change in behavior of single women with and without children following the Tax Reform Act of 1986, which expanded the credit (not available to families without children until 1993). They find that the EITC increased participation in the labor force while having little effect on the number of hours worked. The authors also speculate that the lack of effect on hours worked may be because individuals were not aware of the marginal effects, since the EITC is usually received as a lump sum.

There have been a few other studies of the EITC and labor supply, although none specific to the 1993 change. These are discussed in the Browning and the Dickert, Houser and Scholz articles.

It should be noted that despite the extensive examination of labor supply response, these studies confront many difficulties. Some of these studies rely on estimates from "cross-section" analysis that compares the labor supply of many individuals at a single point in time. Measuring the wage, particularly the expected wage that confronts non-participants (primarily women) is not easy, and is made more difficult when taxes are considered. Separating the substitution and income effects are also difficult. The Eissa and Leibman study is different in that it examines response to a single event in time. This approach confronts a different type of problem -- their results would hold only if there is no other unidentified factor that could explain their results. In particular, these types of studies cannot control for trends.

Although there are difficulties associated with the empirical research, it suggests modest, and perhaps positive, effects of the EITC on labor supply.

EVALUATION

There are several mitigating factors to consider in the concern recently expressed over adverse work incentives of the EITC.

First, the EITC may increase overall work effort when entry into the labor force as well as hours of work are considered. In any case, reductions in hours of work, even if they were to lead to reductions in disposable income, are not an indication that the EITC is harming the individuals affected.

Second, the possibility of adverse work incentives is inevitable in any anti-poverty program, and is caused by the necessity of phasing out the benefits as income rises (in the case of traditional programs) or the fundamental nature of the approach (in the case of the minimum wage). The only way to eliminate adverse incentives is to eliminate anti-poverty programs. When the EITC is compared to traditional transfer programs, it is clearly much more favorable to work effort (although it may fare less well in other criteria used to evaluate transfer programs). ¹¹

Third, for policy purposes, increased participation may be more important than the contraction in hours worked for those already in the labor force.

² Most evidence suggests that the minimum wage has modest effects on unemployment; these results can occur because the minimum wage differs little from the market wage or if there are no close substitutes for unskilled labor. In the latter case, the cost of the minimum wage falls on other factors of production (capital and skilled labor).

³ Marvin Kosters, The Earned Income Tax Credit and the Working Poor, The American Enterprise Institute, (May-June, 1993), pp. 65-72.

⁴ Bruce Bartlett, Weighing the Worth of the EITC, Washington Times, July 24, 1995. This observation was based on calculations by Edgar Browning (Effects of the Earned Income Tax Credit on Income and Welfare, National Tax Journal, Vol. 48, March 1995, pp. 23-43) that indicated that some families would have smaller disposable incomes because of the EITC.

⁵ Edgar K. Browning, Effects of the Earned Income Tax Credit on Income and Welfare, National Tax Journal, Vol. 48, March 1995, pp. 23-43.

⁶ Ibid.

⁷ Thomas Mroz, The Sensitivity of an Empirical Model of Married Women's Hours of Work to Economic and Statistical Assumptions, Econometrica, vol 55, July 1987, pp. 765-800.

⁸ Stacy Dickert, Scott Houser, and John Karl Scholz, The Earned Income Credit, Presented at the 1994 National Bureau of Economic Research Conference on Tax Policy. (This paper will appear in the 8th of a series of conference volumes published by the MIT Press).

⁹ Janet Holtzblatt, Janet McCubbin, and Robert Gillette, Promoting Work through the EITC, National Tax Journal, vol. 47, September 1994, pp. 591-607.

¹⁰ Nada Eissa and Jeffrey B. Liebman, Labor Supply Response to the Earned Income Tax Credit, National Bureau of Economic Research Working Paper 5158, June 1995.

¹¹ As compared to a traditional poverty program, the EITC does not have a minimum floor, is less "target-efficient," is generally less precise in adjusting for family size, and may be more difficult for poor individuals to utilize. It is also typically received as an annual lump sum payment, which some view as a serious problem for the budgeting of household expenditures.