The best case for raising the minimum wage and where it breaks

Twenty-five cents. That was the federal minimum wage when it was born, in the Fair Labor Standards Act of 1938, and a man who got it in his pay envelope in that first year was supposed to feel the government had done him a great kindness. In a way it had. The law also set a maximum work week and required overtime pay, and those were real changes for people who had been working sixty hours for a pittance.

I start there for a reason. The minimum wage is the most popular economic idea in American politics and the one conservatives have the hardest time arguing against. Say you oppose raising it, and the headline writes itself: "Republican thinks a single mother should live on $X an hour." I know the routine. I am going to try something different. I am going to make the best case for raising the wage floor, the case a smart and decent person would make, and then I am going to show you where it breaks. If the conservative position cannot survive the strongest version of the other side, it does not deserve to win.

Let me say one more thing up front. Nobody on our side thinks low-wage work should pay badly. Everybody who works should be able to make ends meet. The argument is about how to get there, and which tool does the least damage to the people it is meant to help.

A short history of the floor

Congress passed the Fair Labor Standards Act in 1938, and President Franklin Roosevelt signed it that June. It set the minimum at 25 cents an hour, with a rising schedule over the next several years. The Supreme Court upheld it in United States v. Darby (1941). Earlier, in Adkins v. Children's Hospital (1923), the Court had struck down a minimum wage for women in the District of Columbia, and in West Coast Hotel Co. v. Parrish (1937) it upheld a state minimum wage law, overturning Adkins. Since then Congress has raised the floor by statute many times, and many states have set a higher floor of their own. The current federal number is a quick lookup at the Department of Labor, and I would rather you check it there than trust a figure that may have moved since I wrote this. What matters for the argument is that the federal floor has been a fixed dollar amount that Congress changes only by vote, which means its real purchasing power erodes between increases.

The best case for raising it

Here is the argument at full strength.

Start with the nature of the labor market. The textbook story says that if the price of labor is pushed above the market level, employers hire fewer workers. But that story assumes many employers competing for workers, so that no one of them can set the wage. In real life, a low-wage worker often has few choices: one big employer in a small town, a restaurant row where everyone pays the same, a commute that limits where you can go. Economists call a buyer with the power to set the price a monopsonist; the term comes from Joan Robinson's 1933 book. In a market like that, an employer pays less than the worker's contribution to output, and a higher floor can raise wages without reducing jobs, because the employer was holding the wage down in the first place.

And there is evidence. In 1992, New Jersey raised its minimum wage from $4.25 to $5.05 an hour, while neighboring Pennsylvania did not. The economists David Card and Alan Krueger surveyed about 400 fast-food restaurants in New Jersey and eastern Pennsylvania before and after, and published their results in 1994. They found no sign that employment in New Jersey fell relative to Pennsylvania. If anything, it looked slightly higher. The study was a landmark, and it upended a generation of textbook certainty. Add the moral case, which is that a person who works full time should not be poor, and a floor that loses value every year to inflation, and you have a coherent argument for raising it.

That is a good argument, and it is honestly put. It has a respected intellectual pedigree, and David Card won the Nobel Prize in economics in 2021 in part for his work on natural experiments like this one. A conservative who dismisses it with a sneer has not done his homework. So let me engage it.

Answer one: the evidence is contested, not settled

The Card-Krueger finding was challenged almost immediately. In 2000, David Neumark and William Wascher, two economists, re-examined the same question using payroll records from the restaurants rather than telephone surveys, and found that employment in New Jersey had declined relative to Pennsylvania. Card and Krueger replied, using other payroll data, and defended their results. The argument has gone on for a quarter century. Neumark and Wascher later surveyed the research literature and concluded that the bulk of the studies find negative effects of minimum wages on employment, especially among the youngest and least skilled workers.

Read what I just wrote carefully. I am not telling you that Card and Krueger were wrong, and I am not telling you that Neumark and Wascher are right beyond question. I am telling you that the economics profession has not settled this. A person who says "the research proves higher minimums cost no jobs" is overstating the evidence, and a person who says "the research proves they always cost jobs" is doing the same. The honest reading is that small increases, in a labor market where employers have some wage-setting power, may have little effect on employment, and that large increases relative to local wages carry greater risk.

The Congressional Budget Office has framed the issue in just that way. In its analyses in 2019 and again in 2021, it described the effect of raising the federal minimum as a trade-off: higher pay for most low-wage workers who keep their jobs, and a loss of jobs for some, with the size of each depending on how high the new floor is set. The agency also stressed the uncertainty of its estimates. That is the right frame. It means the question is not whether a higher floor helps or hurts, but how many people win, how many lose, and whether the losers are the ones we most want to protect.

Answer two: the floor prices out the least skilled first

Here is the mechanism, and it is simple enough for a kitchen table. An employer hires a worker only if the worker's output is worth at least what the worker costs. A minimum wage does not make a worker more productive. It makes the cheapest legal hire more expensive. Whoever is worth less than the floor to the employer is the person who does not get the job.

Work a hypothetical example, adding the employer's share of payroll taxes of 7.65 percent. Suppose a worker with little experience would add $12.50 an hour in value to a small business.

• At a wage of $11, the employer's cost is $11 times 1.0765, which is $11.84. The worker adds $12.50 and costs $11.84. The employer hires.
• At a wage of $12, the employer's cost is $12 times 1.0765, which is $12.92. The worker adds $12.50 and costs $12.92. The employer does not hire.

That $0.42 gap (the $12.92 cost minus the $12.50 in value) is the difference between a job and no job. The worker who gets hired at $11 gets a paycheck, a reference and a start. The worker who is just below the line at $12 gets nothing, and what he loses is the thing no wage law can give him back: the first rung on the ladder. A first job teaches you to show up, to follow instructions, to deal with a customer. Wages rise from there. Take away the first rung and the ladder is shorter at the bottom.

This is why economists across the spectrum agree that the people most at risk from a large increase are the young, the inexperienced, and those with the fewest skills. Those happen to be the workers a minimum wage is supposed to protect.

Answer three: one national number ignores very different local costs

America is not one economy. Rent, wages and prices in a rural county differ greatly from those in a big city. A single national floor is either too low to matter in the expensive place or too high to be safe in the cheap one.

A hypothetical pair of towns makes the point. In Town A, the typical hourly wage is $22. In Town B it is $14. Now impose a national floor of $15.

• In Town A, $15 divided by $22 is about 0.68, so the floor equals 68 percent of the typical wage. Most employers already pay near or above it, and the effect is modest.
• In Town B, $15 divided by $14 is about 1.07, so the floor is 107 percent of the typical wage. It sits above what the average worker there earns, so it forces the biggest change on the place that can least afford one.

The federal floor affects the two towns completely differently, and no one in Washington can see the difference. That is an argument for letting states and localities, closer to the facts, set their own. Many already do, and many index their floors to inflation. For the general case for decentralization, see Three Towns, Three Rule Books and the Case for Letting States Differ. Federalism is not a way to avoid helping workers. It is a way to match the help to the place.

Answer four: someone pays the bill

Every wage increase has to come from somewhere. Take a hypothetical small employer with 10 workers who each earn $10 an hour and work full time, 2,080 hours a year (40 hours times 52 weeks). The floor rises to $12.

• The extra wage per worker is $2 times 2,080, which is $4,160.
• For 10 workers, that is $41,600.
• The employer also pays payroll tax on the increase: 7.65 percent of $41,600 is $3,182.40.
• The total added annual cost is $41,600 plus $3,182.40, or $44,782.40.

Where does a business with narrow margins find $44,782? It can raise prices, which falls on customers, often including low-income ones. It can cut hours, trim benefits or reduce training. It can invest in a kiosk or an automated system that replaces a position. It can hire fewer people, or accept a lower profit, or in the worst case close. A large chain with scale may absorb the change. A mom-and-pop shop might not. The point is that the cost does not vanish, and the workers who keep their jobs gain at the expense of others who now do not have one or who see fewer hours. For a worked look at the gap between what employers pay and what workers take home, see why your raise feels smaller than the number your boss pays out.

Answer five: the minimum wage is a blunt tool for fighting poverty

Here is a point the other side rarely wants to discuss. The link between a low hourly wage and a poor household is looser than it looks. Many workers at or near the minimum are young people in households that are not poor, and many poor households have no worker earning the minimum, whether because they have no full-time worker, or because the earner makes slightly more than the floor. A minimum wage is a tax on the employer for each hour of work, and it lands on every minimum-wage job whether or not the household behind it needs help. If the goal is to help low-income working families, a measure aimed at the family hits the target more precisely than one aimed at the job.

The conservative alternative: the Earned Income Tax Credit

That tool exists, and it is one conservatives have supported for a long time. The Earned Income Tax Credit, or EITC, was enacted in 1975 and has been expanded several times, including in the tax reform of 1986 under President Reagan, and again in 1990 and 1993. It is a refundable credit for low- and moderate-income workers, larger for those with children, and it phases in as earnings rise. In other words, it pays only if you work.

Why do we like it? Several reasons.

1. It rewards work. The credit is tied to earnings, so a person gets more by working than by not working.
2. It is targeted at families. The amount depends on household income and the number of children, not on the hourly rate of a single job.
3. It does not put the cost on one employer. The money comes from the whole tax base, so a small business is not made to bear the full price of a social goal. And it does not price anyone out of a job.
4. It has an intellectual pedigree on the right. Milton Friedman proposed a negative income tax in the 1960s, the idea that the government should supplement the income of low earners through the tax system rather than by regulating wages.

I owe you the other side of this one too. The EITC costs the taxpayers real money. It has an error problem, because the rules are complicated and some of the payments go to people who are not eligible. It has phase-out ranges where an extra dollar of earnings reduces the credit, which can blunt the reward for working more. And some economists argue that a wage subsidy can lead employers to offer lower pay, because more people are willing to work for less. These are real issues, and a serious conservative position includes fixing them. But as a way to raise the income of working families without pricing anyone out of a job, it has a better record than the floor.

How to make this argument and still be heard

This is where many of us blow it. Here is a better way.

1. Begin by agreeing on the goal. Say that work should pay, and that a full-time worker should be able to support a family. Then ask which tool gets there with the least collateral damage.
2. Grant the research. Admit that small increases in some places have shown little effect on jobs, and then point to the uncertainty and the larger risk at higher levels.
3. Tell the story of the person who never gets hired. You cannot see him in the data, but he exists.
4. Offer your plan. Local floors set by those who know local costs, plus a strong EITC. Do not just say no.
5. Lean on the price system. The classic essay on how nobody plans your breakfast, I, Pencil and the price system, how nobody plans your breakfast, explains why a price set by decree rather than by buyers and sellers sends the wrong signal, and the long history of attempts is in From Diocletian to Nixon, Every Price Control Ends the Same Way. A wage is a price, and the lesson applies.

For the numbers behind the argument, learn how the Bureau of Labor Statistics measures employment and wages, which how the jobs numbers work, answered in eleven questions explains. And if you want to see where each dollar of a paycheck goes, the worksheet in a paycheck worksheet that shows where every dollar of your pay goes does that.

The bottom line

• The case for a higher floor is serious: employers in some markets can set wages, and the best-known study of New Jersey found no job loss. Concede it, and then answer it.
• The evidence is contested. Neumark and Wascher found job losses using payroll data, and the Congressional Budget Office describes a trade-off, not a free lunch.
• The floor prices out the least skilled first. In our hypothetical, a worker worth $12.50 an hour gets hired at $11 and not at $12, once payroll tax is counted.
• One national number fits poorly. A $15 floor is 68 percent of the typical wage in one hypothetical town and 107 percent in another.
• Back the Earned Income Tax Credit and local floors, and say so out loud. Work should pay, and the best tools raise family income without closing the first rung of the ladder.