Inequality at the Margins - Russell Sage Foundation



Inequality at the Margins

The Effects of Welfare, the Minimum Wage, and Tax Credits on Low-Wage Labor Markets[1]

by Michael Hout

RUSSELL SAGE FOUNDATION

UNIVERSITY OF CALIFORNIA, BERKELEY

March 1997

One of my favorite bits of Berkeley graffiti is a collaborative effort. A member of the Revolutionary Communist Party painted the slogan "ABOLISH WAGE SLAVERY" in

black block letters on the white-washed side wall of a moving and storage company.

An unknown party sprayed "Please don't" above and "or I won't get paid" below.

With the Personal Responsibility and Work Opportunity Act of 1996, the United

States promises not to abolish wage slavery any time soon. Former welfare mothers

will work for wages or they won't get paid. As a consequence, millions of women who

currently have jobs will find themselves in competition with millions of others who will

be mustered off the welfare rolls. Elaine McCrate's paper shows how the new labor

force entrants will bid down the wages of women already at work. In the past social

scientists have expressed doubts that welfare cuts could spark wage competition.

McCrate's data erase those doubts. Declining welfare support not only further

impoverishes those who depend on welfare but also cuts into the wages of people with

jobs.

[pic]

The welfare bill that trumpets work will actually lower the wages of the working poor. State-to-state variation in the combination of AFDC, food stamp, and Medicaid benefits for families with children indicates that the wages of young women who have not been to college drop 3 percent for each $100 cut in this benefit package. To illustrate McCrate's important result, I have plotted each state's wages against its combined AFDC and food stamp package for a family of three. Figure 1 shows wages for all women with no more than a high school education.[2] The gray line shows the 3 percent relationship revealed by McCrate's analysis.[3] Thus, in the cross-section, the anticipated relationship holds: women in states with low benefits work for lower wages (even after statistically adjusting for the tendency of low-wage states to supplement benefits less than high-wage states do).

LOW WAGES AND INEQUALITY

The link between benefits and wages offers an important clue to the mystery of how U.S. income inequality has increased since 1974 (surging between 1988 and 1993). State-to-state variation mimics national trends over time. As states frantically try to woo capital and shoo labor, they engage one another in a rush to the bottom in wages and benefits. McCrate's findings about states aid us in interpreting recent trends for the whole country.

Leading explanations for rising inequality include institutional settings, technology, and

immigration (see Fischer et al. 1996, Ch. 1 and 6 for a discussion). "Institutional

settings" is an umbrella term that spans a number of arguments about how the

organization of markets affects market outcomes. This point of view stresses how

conventions that used to support the low end of the wage distribution have been

removed or weakened. Proponents cite the decline in the real value of the minimum

wage (Card and Krueger 1995, chapter 9) and unionization (Freeman 1992) or both

(DiNardo et al. 1996) as key causes of wage inequality. The fall of the real minimum

wage in particular exacerbated inequality trends by lowering the lowest rung of the

U.S. wage ladder.

The main alternative to the institutional perspective on labor markets is the idea that

technological change has divided the labor market between a high-skill, high-wage

stratum and a low-skill, low-wage substratum. This "skill-biased technical change" view

holds sway among many influential economists these days (e.g., Danziger and

Gottschalk 1995, pp. 148-150). Immigration, too, may increase inequality if low-skill

immigrants will work for lower wages than low-skill natives will accept. However,

neither the earnings nor the unemployment rates of native-born workers differ much

between the "gateway areas" where many immigrants settle and other parts of the

country (Borjas 1994). However, immigration may be significant in reducing the pay of

high school dropouts (Borjas, Freeman, and Katz 1996).

McCrate's finding that welfare levels affect the wages of nonrecipients adds new data

to this debate. Low benefit packages push people who would not otherwise compete

in local labor markets into the fray. Their poverty makes them inexpensive to

employers.[4] States with low welfare benefits have lower wages for workers because

the increased supply of low-wage workers is not offset by any increase in the demand

for their labor. They all end up competing for all too few low-wage jobs.[5]

While the institutions-technology-immigration puzzle may not have a definitive solution,

McCrate's analysis is an important addition that favors the institutionalist view.

McCrate not only adds a new institutional factor - declining real welfare benefits - to

the list of vanishing social arrangements that used to support the lowest rung, but her

analysis also includes fresh evidence of how minimum wage protection and

unemployment rates matter. According to Table 1A of her paper, a 10 cent drop in the

real value of the minimum wage lowers less-educated women's wages by 2 percent; a

one percent increase in the unemployment rate lowers their wages by between 2.1 and

2.5 percent.

While the unemployment rate has cycled up and down three times since the early

1970s, the value of welfare benefits and the minimum wage have both dropped off

sharply. In tandem they have put pressure on wages and made it harder for people to

escape poverty simply by finding a job.

[pic]

The purchasing power of Aid to Families with Dependent Children (AFDC) eroded after 1974. Some of the erosion came about because many states reduced their AFDC support in the 1980s. Increased food stamp allotments helped offset much of the impact of these cuts (Wiseman 1996). More pernicious, though, was the less visible loss of real value to inflation. Monthly benefits were never indexed to the cost of living, so inflation made support packages worth less over time even though they were nominally unchanged. These are not trivial changes. From 1970 to 1993 the purchasing power of the average AFDC benefit (per family) fell from $665 to $370.[6] The welfare mother of 1993 had to get by on benefits that bought little more than half of what her counterpart had to live on in 1970. The AFDC benefit lost value in the 1970s because of inflation; it lost value in the 1990s because of decisions to cut benefits. Figure 2 shows the precipitous trend.

The real value of the minimum wage also fell. And for the same reasons: the minimum wage is not indexed to the cost of living and workers lack the political clout necessary to get Congress to raise it every time prices rise. Congress phased in a rise from $2.65 to $3.35 in 1978 (the last increase took effect in 1981) and did not act again until 1989. That is no action on the federal minimum wage during the Reagan years. By then inflation had eroded one-third of its purchasing power. The $295 decrease in AFDC and the $2.25 erosion of the real minimum wage combined to drag down the wages of less-educated women by 14 percent.[7]

[pic]

The minimum wage establishes the floor for the wage structure of all workers. When the real value of the minimum wage drops, low-wage earners fall behind the rest of the workforce. Figure 3 shows how the depleted minimum wage contributed to increased inequality from 1983 to 1993. The vertical lines are proportional to the fraction of prime-age (25-64 years old) working men and women that makes a specified wage (in 1993 currency).[8] The minimum wage in a given year is black in contrast to the other (gray) lines; a label expresses its value in current dollars. As the real value of the minimum wage drops, the wage distribution spreads to its left side indicating that the poorest workers lost out and inequality grew. The distance between the minimum wage and the average wage grew over time. If the minimum wage had kept pace with inflation, then the low end of the wage distribution would have remained a fixed distance from the average and

inequality would have grown less (if at all). Women were more exposed to this dynamic than men were.

Most minimum wage workers are younger or older than prime-age workers. They are also, almost by definition, more likely to be paid on an hourly basis instead of on salary. Figure 4 splits all persons with wage and salary earnings in 1993 into four age groups (combining men and women) and shows how their wages are distributed. The minimum wage holds up the low end of the distribution among both the youngest and oldest workers. If the minimum wage were lower, presumably the wage distributions of workers under 25 and over 65 years old would look as symmetrical as the ones of prime age workers - and there would be some people earning very low wages. On the other hand, if the minimum wage were higher (as it will be in September 1997), these workers would be earning an amount that is closer to the median. Then there would be less inequality.

[pic]

The minimum wage has almost no effect on salaried workers, whether they are full-time or part-time. Figure 5 shows the wage distribution of persons who are paid hourly or some other way. The salaried workers are split according to whether they work full-time or part-time. The minimum wage affects inequality only through its effect on hourly workers.

[pic]

In all of these charts, round numbers stand out. Wages of $5.00 and $10.00 appear to be especially popular. Unfortunately, we cannot tell if that is because employers pay round wages, e.g., exactly $5.00 per hour, or respondents round off their wages when they report them to the CPS. As the minimum wage moves above $5 later in 1997 between 15 and 20 percent of hourly workers can expect a raise. Moving that many workers closer to the median will reduce inequality if it happens. All attempts to manipulate the minimum wage are predicated on the belief that the added labor cost will not motivate employers to eliminate workers. The experimental evidence so crucial to the recent debates involved a minimum wage rise that affected fewer workers (Card and Krueger 1995).

THE WORKING POOR

The crisis of current policy is that it relies on work to redress poverty in a labor market

that is not up to the task. Too many jobs pay poverty wages. One-sixth of the 25-34

year-old fathers who work full-time do not make enough to raise their families out of

poverty (Fischer et al. 1996, p. 112). Nor do the low-quality jobs lead to future

opportunities in most cases; women who work their way off welfare in jobs that do not

offer education or skills training experience no earnings growth despite years of

continuous work effort (Harris 1997). Ending welfare and putting people to work in

this kind of labor market will make child poverty worse, not better.

The Earned Income Tax Credit (EITC) is a popular and effective antidote to the

problem of jobs that don't pay enough. By giving workers who file federal tax returns a

refund that exceeds the taxes withheld from their paychecks, the EITC program links

income support to work effort. The present system works by paying a credit on every

dollar a worker earns up to a specified maximum credit. The maximum differs for

parents and childless poor people. Workers are entitled to that maximum credit unless

their earnings exceed a threshold set for their family type. Above the threshold amount,

the EITC is phased out at half the rate at which credits accumulate for the poorest

families. The parameters of the system for policy-making purposes are the rate at

which benefits accumulate, the spread of incomes eligible for the maximum benefit, the

value of the maximum benefit, and the differential between parents and childless

workers. It is a special case of a negative income tax (NIT) as explained by Block and

Manza.

Figure 6 shows how the current EITC program supplements the earnings of parents

who are the sole earner in a family with two or more children. The implications of EITC

change as wages rise; the figure shows this key feature of EITC by repeating the

calculations for a single parent earning minimum wage and one earning $9 an hour.[9]

Regardless of the wage rate, a parent's credit is 40 percent of her earnings up to a

maximum credit of $3,556 ($2,152 if she has just one child). At the new minimum

wage of $5.15, a mother with two children but no EITC could not lift her family out of

poverty unless she can find enough work to average 46 hours per week over the

course of a year. With the EITC, her family reaches the poverty line if she averages 33

hours of work per week. If she averages 40 hours of work per week, her earnings will

add up to $10,712, her EITC will be $3,556, so her total income will be $14,268 (14

percent above the poverty line). Significantly, she can always make more money by

working more hours or by accepting a higher paying job. EITC overrides the

problematic disincentives of the old welfare system.

[pic]

The EITC becomes less significant for higher-wage jobs. The parent who makes $9

per hour reaches the poverty line and the maximum benefit at 19 hours of work per

week. The EITC begins to phase out at 23 hours, is down to $2,069 at 40 hours and

$1,086 at 50 hours. The parent who averages 40 hours per week at $9 per hour will

earn $18,720 and get an EITC of $2,069 for a total income of $20,789.

Parents of two or more children whose earnings exceed $28,495 get no EITC (the

limit is $25,078 if they have one child). Persons without children can get an EITC of up

to $323 if both their earnings and their adjusted gross income fall below the limit of

$9,500.

The EITC gets money to needy families without offering incentives to curtail work

effort. Combining wages and a parent-of-two's EITC turns a $5.15 wage rate into a

$7.21 per hour rate for the first 33 hours; the effective wage rate falls slowly after that

toward the nominal rate of $5.15 (which is only reached at 106 hours per week). The

worker whose wage is $9 per hour makes what amounts to $12.60 an hour for the first

19 hours each week then converges to the $9 wage rate (but does not actually reach it

until 62 hours of work per week).

CONCLUSION

America prescribes work as the cure for poverty. But work at low wages is ineffective.

Already 9.5 million Americans work at jobs that pay them too little to raise their

families out of poverty (U.S. Bureau of the Census 1996a). Elaine McCrate's paper

shows that crowding labor markets with welfare recipients pushes wages down. On

that basis it is reasonable to expect more crowding and lower wages in the late 1990s.

In this post-welfare world the Earned Income Tax Credit (EITC) and rising minimum

wage offer some hope for impoverished families. By playing within the mandatory work

system, EITC transfers money to the working poor by refunding more than they paid in

taxes (when and if they file). It can be effective in raising living standards for families

that rely on wages at or near the minimum wage level because it does not require them

to work more than 40 hours per week to climb out of poverty. Without EITC families

that rely on the earnings of a minimum-wage worker will be poor unless that worker

puts in more than 46 hours per week. With EITC they will not be poor if the worker

can average 33 hours per week. If the minimum-wage breadwinner secures a 40 hour

per week job, EITC will increase her income by one-third.

To maintain optimism, though, we have to pretend that the supply of jobs will be

sufficient to get every breadwinner on a payroll. Right now 7.2 million Americans are

looking for work; another 1.5 million are available for work but have given up looking

in the past 12 months (U.S. Bureau of Labor Statistics 1997). And 5 million

households rely on benefits as their sole source of income (U.S. Bureau of the Census

1996b, Table 11). The minimum wage and the EITC cannot reach them because they

have no wages to boost. And these are good times with rising wages and low

unemployment. Dare we speculate about what will happen to a job-based strategy

during the next recession? In the post-AFDC world, all poor families will be working-poor. Without wage assistance - EITC and minimum wage boosts - they won't get paid enough to escape poverty. Without a sufficient number of jobs, they won't get paid at all.

REFERENCES

Borjas, George J. 1994. "The Economics of Immigration." Journal of Economic

Literature 32: 1667-1717.

Borjas, George J., Richard B. Freeman, and Lawrence F. Katz. 1996. "Searching for

the Effect of Immigration on Labor Markets." American Economic Review

86:246-251.

Card, David, and Alan B. Krueger. 1995. Myth and Measurement: The New

Economics of the Minimum Wage. Princeton: Princeton University Press.

Danziger, Sheldon, and Peter Gottschalk. 1995. America Unequal. Cambridge:

Harvard University Press and Russell Sage Foundation.

DiNardo, John, Nicole M. Fortin, and Thomas Lemieux. 1996. "Labor Market

Institutions and the Distribution of Wages, 1973-1992." Econometrica 64:

1001-1044.

Fischer, Claude S., Michael Hout, Martín Sanchez-Jankowski, Samuel R. Lucas, Ann

Swidler, and Kim Voss. 1996. Inequality by Design: Cracking the Bell Curve

Myth. Princeton: Princeton University Press.

Freeman, Richard. 1994. "How much has deunionization contributed to the rise in male

earnings inequality?" Pp. 100-121 in Uneven Tides: Rising Inequality in America,

edited by Sheldon Danziger and Peter Gottschalk. New York: Russell Sage

Foundation.

Harris, Kathleen Mullan. 1997. Teen Mothers and the Revolving Door. Philadelphia:

Temple University Press.

Heckman, James J. 1974. "Shadow prices, market wages, and labor supply."

Econometrica 42: 679-694.

U.S. Bureau of the Census. 1996a. Poverty in the United States: 1995. Current

Population Reports, P60-194. Washington, D.C.: U.S. Department of Commerce.

U.S. Bureau of the Census. 1996b. Money Income in the United States: 1995.

Current Population Reports, P60-193. Washington, D.C.: U.S. Department of

Commerce.

U.S. Bureau of Labor Statistics. 1997. Employment Situation: February 1997.

Washington, D.C.: U.S. Department of Labor.

Wiseman, Michael. 1996. "Knowing welfare as we end it." Paper presented at the

Visiting Scholars Workshop, Russell Sage Foundation, New York.

NOTES

1. This paper expands on comments delivered at the Politics and Society conference

entitled "After AFDC: Reshaping the Anti-Poverty Agenda," New School for Social

Research, New York, 16 November 1996. I benefitted from comments by conference

participants and by Paula England and Robert K. Merton. The Russell Sage

Foundation provided financial support.

2. McCrate uses the same educational restriction but analyzes black women's and

white women's wages separately.

3. The scatter of points appears to incline even more than the gray line suggests. That is

because other factors, statistically controlled by McCrate, produce state-to-state

variation that, though related to variation in welfare benefits, cannot be ascribed to

them. Most notable among the other factors is state-to-state variation in the minimum

wage.

4. There is a school of thought that suggests that they are implicitly more expensive

than otherwise comparable individuals. The argument stresses that they have chosen

not to work at the wages that prevail for individuals like them; therefore their implicit

wage must be higher than what employers offer them (if it were not they would take the

job offer). This is a misunderstanding of "reservation wages" (Heckman 1974). The

idea of a reservation wage applies when the people choose between working for pay

and leisure. It does not apply when they have the choice between working for pay and

welfare. For people on welfare, the benefits exceed their reservation wage. A cut in

benefits gets some of them to leave welfare and take a job that holds them at the

standard of living that welfare provided before the cut.

5. Of course there might be a reverse effect, too. In the long run, higher wages mean

higher prices for important commodities like housing and food. State legislators often

consult these prices in deciding how much to allocate for aid. McCrate designed her

statistical analysis to control for this reverse effect.

6. The 1970 amount was adjusted to 1993 currency using the Consumer Price Index

for urban consumers (CPI-U).

7. McCrate estimates the net effect of $100 differences in AFDC and food stamps to

be 3 percent. A drop in AFDC of $295 that was not offset by increased food stamps

would reduce wages by 9 percent. She estimates that the minimum wage is worth 2

percent for every dime difference. A $2.25 drop would cut wages by 5 percent.

Together these add up to 14 percent. Estimates of this type are not always additive,

but they are in this case because her model contains both factors.

8. The data source is the Current Population Survey's "outgoing rotation groups" file.

The CPS is the monthly survey that provides important government statistics like the

unemployment rate and the consumer price index. Each month one-fourth of the

respondents are asked about the wages of working members of their household. They

report wages per hour for members of their household who have jobs that pay them on

an hourly basis and weekly wages or salary for other kinds of jobs. They also report

how many hours per week each worker puts in at that job. To make Figure 3 and the

other figures based on this data file, I calculated hourly wage as equal to the reported

wage for those paid on an hourly basis, as the ratio of weekly wages or salary to actual

hours worked for part-time salaried workers and so-called "full-time" workers who put

in less than 40 hours, and as reported weekly earnings divided by 40 for those who put

in 40 hours or more. For reasons known only to the designers of the CPS, the

self-employed are not included in the wage data.

These data include some wages below the minimum. Mostly they are the wages of

people who are not paid hourly. Details are scarce, but a few points are probably true.

At low hours of work, subminimum wages are probably "piecework" payments

converted to pay per hour. There is also the problem at the other end of people

working 70 or more hours; their hourly rate looks quite low. And, of course, some

establishments are exempt from minimum wage laws; their employees are an unknown

proportion of the subminimum wages in Figures 3 and 4.

9. Families with two or more earners qualify for EITC. They must file a joint return and combine their earnings as any other couple filing a joint return would

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download