Time to repair the wage floor

Raising the minimum wage to $6.65 will prevent further erosion of its value

by Jared Bernstein and Jeff Chapman

The federal minimum wage has lost 10% of its buying power since it was raised to $5.15 in September 1997. Unless Congress acts soon to increase the minimum wage, the 1997 increase will be fully eroded by next year. And as a result of general wage growth, the minimum wage has also fallen in relative terms. In 1997, it stood at about half the median wage, but in 2001 it had fallen to 40% of the median.

Congress periodically repairs the damage done by inflation to the national wage floor by legislating a new increase. At this time, support is building for a new $1.50 increase, raising the minimum in three increments from $5.15 to $6.65 by 2004.1 Such legislation could help lift the earnings of millions of low-income workers who depend on the minimum wage. Specifically:

• Raising the minimum wage to $6.65 will lift the wages of 5.6 million low-wage workers.
• An increase in the minimum wage will be especially helpful in boosting the earnings of women and minorities.
• Most of the increase will go to families with low incomes and low weekly earnings.
• An increase in the minimum wage will work in tandem with the Earned Income Tax Credit (EITC) to close the gap between earnings and needs.
• Due to relatively fast recent wage growth for many low-wage workers and the fact that 12 states have their own minimums set above the federal level, the current proposal would affect less than 5% of the workforce-a smaller share than the last increase. Even with unemployment still high in the wake of the last recession, an increase of this magnitude is unlikely to result in job losses among affected workers.
• As Congress works to reauthorize the welfare program, the minimum wage can help address the low earnings faced by former welfare recipients.

Fixing the hole in the wage floor
Unlike many government policies, the minimum wage is not automatically adjusted each year to account for inflation. The result is that the purchasing power of the minimum wage falls each year that Congress does not act to raise it.

From 1938, when the first minimum wage was passed, until 1981, Congress passed regular increases, keeping it on a track of slow and steady growth (Figure A illustrates this trend from 1950 forward). After the 1981 increase to $3.35, Congress neglected to raise the minimum wage again until 1990. During this time, prices rose an average of 3.9% per year, and the real value of the minimum wage fell by 26%. Despite minimum wage increases in the 1990s, its real value is still 19% lower today than it was in 1981.

Congress last increased the minimum wage from $4.25 to $5.15 in 1996 and 1997. The additional purchasing power afforded to low-wage workers by this increase will have disappeared by 2003, when the real value of the minimum wage will have sunk back to its 1995 level. Even with this proposed increase to $6.65 by 2004, the minimum wage will still have less purchasing power than it did in 1981. This increase would, however, represent a significant step toward putting the value of the minimum back to where it was before the long decline over the 1980s.

Characteristics of workers affected by the increase
If today’s minimum wage was $6.65 instead of $5.15, then 8.9 million workers would now have higher wages. But because of the planned phase-in-$0.60 after passage, $0.50 in January 2003, and $0.40 in January 2004-fewer workers will ultimately be affected by the increase. One reason for the lesser coverage under the phase-in is that wages are expected to grow somewhat over the next few years, so a given wage increase will cover fewer workers as time passes. In addition, seven states raised their minimum wages this year, and three states have scheduled increases for 2003 and 2004.2

Given these trends as well as other assumptions about the growth rate of low wages, inflation, and the labor force (see the methodology section for details), the proposed increase will most likely reach 5.6 million workers, or about 4.5% of the workforce. Table 1 shows the characteristics of the workers who would directly benefit from the proposed increase.

• Most of the workers (68%) who would directly benefit from the proposal are adults rather than teenagers (16-19 years old).
• Only 23.4% of the affected workers work less than 20 hours per week, and 42% of the affected workers work full time (35 or more hours per week).
• Sixty-one percent of affected workers are women, while less than half of the total workforce is female.
• About one-third of the affected workers are African American or Hispanic, while these groups make up less than one-fourth of the total workforce.
• Workers affected by the increase are over-represented in low-wage industries and occupations: 45.9% work in the retail trade industry (which employs only 17.1% of the total workforce) and 56.0% work in sales, service, or food preparation (occupations which employ 22.9% of the total workforce).
• Only 4.0% of the affected workers are covered by union contracts, compared to 15% of the total workforce.

In addition to the 5.6 million workers directly affected, an additional 8.4 million workers will be earning less than $1 above the new minimum wage. Various studies of “spillover effects” find that this group may also benefit from the increase.³ As shown in Table 1, these workers are even more likely than minimum wage workers to be adults working full time.

Compared to past increases in the minimum wage, this $1.50 increase affects a relatively small share of the workforce (4.5%). The last increase, for example, affected twice that share (8.9%, Bernstein and Schmitt 1998), and did so without leading to any measurable disemployment among those affected. As shown in Figure B, higher increases would raise the share of workers affected. An increase of $2.25, from $5.15 to $7.40, would reach 8.2%, still a smaller share than the 1996-97 increase, and also unlikely to generate negative employment effects. Thus, in historical terms, the current proposal is modest.

Most of the benefits reach low-income workers
Because payment of the minimum wage is not tied to family income, some opponents of the policy raise concerns that it is poorly targeted, that is, some who benefit from the increase don’t actually need the increase. The data show, however, that most of the benefits of the increase will flow to low-income working families.

But from a policy perspective, the minimum wage should not be judged solely on the basis of its efficiency at “hitting a target” (i.e., the extent to which it reaches low-wage workers in low-income families). Since its receipt is not based on income, and wages and income are imperfectly correlated, it is not designed to score highly on this criterion. Instead, it is designed to impose a wage floor on the labor market, below which covered workers cannot legally be paid. Its purpose, as conceived by its framers in the Roosevelt Administration and as stated in the Fair Labor Standards Act, is to prevent market forces from driving the wage of the least skilled down below a level deemed to be minimal by Congress. The government says, in effect, that it doesn’t want the people in this country to be paid less than a certain amount. Given the weak bargaining power of
low-wage workers, this aspect of the policy remains as relevant as ever.

Nevertheless, an increase in the minimum wage is a fairly well-targeted way to improve the lot of lower-earning households (Table 2). Overall, 59% of the gains from the proposed increase would go to the bottom 40% of prime-age households, and over three-fourths of the gains would go to the bottom 60% of prime-age earner-headed households. This number includes only households headed by workers between the ages of 25 and 54, and excludes households with no earnings and households headed by older or younger persons (who tend to be less connected to the workforce). If all working households are considered, the proposal is even more targeted at low-earning households, with 61% of the gains from the proposal going to the bottom 40% of households.

While the proposed increase will affect less than 5% of the total workforce, the earnings of affected workers are significant to their families. In 2001, minimum wage workers contributed 49% of a family’s total weekly earnings, on average (see Table 3). Not surprisingly, low-income families relied even more heavily on minimum wage workers; families with incomes of less than $25,000 relied on minimum wage workers for 76% of their total weekly earnings, on average.

The minimum wage, EITC, and making work pay
The EITC is an important piece of the ongoing strategy to make work pay, but its effectiveness in raising the incomes of the working poor above the poverty line depends in part on regular increases in the minimum wage.

The EITC is one of the most substantial and popular anti-poverty programs. One reason for its popularity is that it is based on family income and is therefore easily targeted to poor families. In addition, it encourages work because the wage subsidy increases with earnings until it reaches the maximum ($4,008 for a family with two children in 2001). The amount of the maximum credit is adjusted for inflation each year.

Consider a single mother of two children working 40 hours per week year-round at the minimum wage. In 1997, this worker would have earned $9,893 after Social Security and Medicare taxes-only 77% of the poverty line. But since she would have been eligible for the maximum EITC of $3,656, her family income would surpass the poverty threshold (which is adjusted annually for inflation). (The first bar in Figure C represents this case.)

A single mother with two kids working the same number of hours in 2001 at the minimum wage would earn the same wages in nominal dollars, but since the poverty threshold is adjusted for inflation, her earnings would only equal 69% of the poverty line. She would still have been eligible for the maximum EITC-$4,008 in 2001-but the EITC would no longer push her income above the poverty line (see the second bar in Figure C).⁴

Because the income level at which a family reaches the maximum EITC is adjusted for inflation while the minimum wage is not, this mother would begin to fall even further behind beginning in 2004. That same year she would no longer be eligible for the maximum credit and her EITC would begin to shrink in real value (third bar of Figure C).

The proposal to raise the minimum wage to $6.65 by 2004 fixes this problem (fourth bar). Her higher earnings would lift her back up to the EITC “plateau”-the earnings range wherein she would again be eligible for the maximum benefit-and these two policies would thus work in tandem to raise the family’s income to 11% above the poverty line. Note that, since the proposal does not index the minimum wage to inflation, this would not be a permanent fix. Future regular increases would still be required.

What about the weak recovery?
Although the recession that began in March 2001 has not been declared officially over, most economic analysts believe a recovery is underway. However, many indicators reveal that recovery to be a weak one. The unemployment rate, which rose to 6.0% in April 2002 (up from a low of 3.9% in October 2000), provides the most salient evidence that labor demand remains relatively slack. Given the current situation, is it safe to raise the minimum wage?

The evidence regarding the impact of past minimum wage increases, including an increase that occurred over the last recession (1990-91), suggest that Congress can implement the proposed increase without fear of disrupting the fledgling recovery.

First, note that the current proposal phases in slowly, and is not fully in place until 2004, when the recovery will likely be stronger than it is at present. Also, the relatively small share of affected workers, compared to the number affected by past minimum wage increases, means that, as a share of the nation’s wage bill, this increase will be small. This year, about 2 million workers will be affected. Assuming the first increase is made effective October 1, the estimated cost to low-wage employers this year will be about $0.2 billion, or about one-hundredth of one percent of the nation’s wage bill in the fourth quarter of last year. The cost of the increase in 2003 is estimated to be about $2.2 billion, or about four-one hundredths of one percent of the nation’s wage bill last year. In 2004, when the full increase is implemented, the total cost will be about $4.1 billion, still only about seven-one hundredths of a percent of the nation’s wage bill. The cost of the increase will decrease each subsequent year. An increase of this magnitude will be swamped by other macroeconomic phenomenon that determine the strength of the recovery, such as Federal Reserve interest rate policy and trends in aggregate consumption, investment, government spending, and the balance of trade.

Second, there is little evidence for any negative effect on employment from past increases in the minimum wage, regardless of the business cycle. By analyzing changes in the employment status of affected workers before and after minimum wage increases, economists have rigorously searched for, but generally failed to find, these negative employment effects.⁵ The estimates from the empirical literature show that the impact on employment is either statistically insignificant or slightly negative, a finding widely accepted by economists. And even in cases where there is evidence of job losses, the number of workers negatively affected are tiny compared to the number who get an hourly pay raise.

Take, for example, the last minimum wage increase, from $4.25 to $5.15, enacted in two steps between 1996 and 1997. Opponents argued that the increase would lead to diminished job opportunities for low-wage workers, but following the increase, conditions in the low-wage labor market did not deteriorate. In fact, this segment of the labor market improved more than it had in decades, in part due to the wage increase itself, but largely due to the low overall unemployment rate that prevailed over this period.

Recent evidence also sheds lights on the impact of increasing the minimum wage during a recession. In 1990 and 1991, while the economy was in a recession, the minimum wage was raised from $3.35 to $4.25. A highly regarded analysis of the increase’s impact (Card 1992), which controlled for overall economic conditions, showed that the increase had no negative effects on employment. Another study found that “although the 1990 and 1991 minimum wage increases led to significant earnings gains for teenagers and retail-trade workers in many states, these wage increases were not associated with any measurable employment losses” (Card and Krueger 1995, 114-5).

Note that these claims do not imply that low-wage workers are immune to the problems caused by a slow-growth recovery,
but rather that the difficulties faced by these workers will be due to weak macroeconomic conditions, the same underlying conditions that are leading to the weakness of the current recovery.

Methodology
Figure A: Value of the minimum wage
From 1978 to 2001, we use the CPI-U-RS as a measure of inflation in order to convert dollars into current values. The CPI-U-RS is both historically consistent and is considered to be a more accurate measure of inflation than the CPI-U (inflation grows somewhat more slowly in the RS than in the CPI-U.) Since the RS exists only back to 1978, we “backdate” it using the ratio of the RS and the CPI-U-X1 (an earlier experimental index) in 1978 (when both are available) times the X1 for earlier years. For 2002-04, we use Congressional Budget Office (CBO) projections from The Budget and Economic Outlook: Fiscal Years 2003-2012. These are 1.8% for 2002 and 2.5% for 2003-2004.

Table 1 and Figure B: Workers affected by the proposed increase
The data in this section come from the 2001 Current Population Survey Outgoing Rotation Group. We restricted our sample to persons aged 16 or over who were employed in the week prior to the survey. We exclude self-employed persons. For persons who reported weekly wages, but not hourly wages, we derived their hourly wage using the number of hours worked. To account for growth in the workforce, we used the growth projection from the Bureau of Labor Statistics of 1.14% per year. Assumptions concerning nominal wage growth can have a dramatic effect on the number of workers predicted to be affected. To account for the fact that some workers will not experience wage growth, we held the wages of workers earning exactly the minimum wage constant while allowing the wages of other workers to grow nominally at 4% from 2001 to 2002, 3% for 2002-03, and 2% for 2003-04. These growth rates are derived from a model of low-wage growth in Bernstein and Baker (forthcoming).

For each worker, we found the applicable minimum wage for each year from 2001 to 2004 in the absence of a federal minimum wage change. The applicable minimum wage is the higher of the federal minimum wage and the state minimum wage. We included scheduled increases in state minimum wages and gave workers those scheduled increases. Also, the minimum wages of Alaska and D.C. automatically adjust in response to federal increases. In Alaska, the state minimum wage is set at $0.50 higher than the federal wage; in D.C., the minimum wage is $1.00 higher.

Some workers are not covered by the minimum wage and therefore may not be affected by the increase. We assumed that all workers who earned less than their applicable minimum wage in 2001 were not covered by the minimum wage and therefore would not be affected by the increase. Since coverage rules differ between state wage laws and the federal law, this may overstate the number of workers not covered. In 2001 there were about 900,000 workers that earned at least $5.15, but less than their state minimum wage. On the other hand, this may be offset by the possibility that there may be some workers earning 5.15 or above who wouldn’t be covered by the proposed increases.

In order to be affected by the proposed increase in a given year, workers had to meet three criteria:

1. They had to earn above their applicable minimum wage in 2001.
2. The proposed minimum wage had to be higher than their state minimum wage would be in absence of a federal increase.
3. Their wage in the given year (after nominal growth) had to be less than the proposed minimum wage.

Note that any worker affected in 2002 would also be affected in 2003 and any worker affected in 2003 would also be affected in 2004. Therefore, the total number of workers affected is equal to the number affected in 2004. Also, note that the proposed minimum wage would be higher for Alaska and D.C. than the other states.

In order to be in the group that may receive spillover effects, the workers had to meet three criteria:

1. They couldn’t be directly affected in any year of the proposed increase.
2. They had to earn above their applicable minimum wage in 2001.
3. Their wage in 2004 (after nominal growth) had to be greater than the proposed minimum wage, but less then the proposed minimum wage plus $1.00.

Table 2: Distribution of gains by total household earnings
For workers who report hourly wages, but not weekly earnings, we derive weekly earnings using the number of hours worked. The gain from the proposed increase is equal to the difference between the total household weekly earnings with the proposed increase and the total household weekly earnings in absence of the increase.

Table 3: Share of weekly earnings contributed by minimum wage workers, 2001
Minimum wage workers are defined here as those workers earning at least their applicable federal or state minimum wage, but less than the level of the proposed increase. The level of the proposed increase is $6.65, except in the District of Columbia and in Alaska. For those workers, the proposed increase would automatically raise the applicable minimum wage to $7.65 and $7.15, respectively. Only families with at least two people are included.

Figure C: Minimum wage workers and the poverty threshold
In order to estimate the poverty line and the income levels used to figure the EITC for 2002-04, we used the CBO inflation projections noted above.

Endnotes
1. The proposal calls for a $0.60 increase 30 days after passage, a $0.50 increase on January 1, 2003, and a $0.40 increase on January 1, 2004.
2. This includes Washington State, which has a minimum wage that is indexed for inflation.
3. See, for example, Spriggs and Klein (1994).
4. Figure C does not include the new refundable child tax credit that was implemented in 2001. For this mother, the child tax credit would equal only about $71 and thus would not have a significant impact on the analysis.
5. See, for example, Bernstein and Schmitt (2000).

References
Bernstein, Jared, and John Schmitt. 1998. Making Work Pay: The Impact of the 1996-97 Minimum Wage Increase. Washington, D.C.: Economic Policy Institute.

Bernstein, Jared, and John Schmitt. 2000. The Impact of the Minimum Wage:Policy Lifts Wages, Maintains Floor for Low-Wage Labor Market. Washington, D.C.: Economic Policy Institute.

Card, David. 1992. “Using Regional Variation in Wages to Measure the Effects of the Federal Minimum Wage.” Industrial and Labor Relations Review. Vol. 46, No. 1, pp. 22-37.

Card, David, and Alan Krueger. 1995. Myth and Measurement: The New Economics of the Minimum Wage. Princeton, N.J.: Princeton University Press.

Spriggs, William, and Bruce W. Klein. 1994. Raising the Floor: The Effects of the Minimum Wage on Low-Wage Workers. Washington, D.C.: Economic Policy Institute.