The Minimum Wage Fiasco

There has been tension between St. Louis City Democrats and the Republican State Legislature for many years. This tension has most recently become apparent as a dispute over minimum wage laws in St. Louis.

In my last couple of posts, I’ve focused on crime in St. Louis. There are many, including myself, who would say that this crime is deeply related to the lack of legitimate sources of income for poor and low skilled people in St. Louis City. During the recent announcement by Gov. Greitens and Mayor Krewson, the Governor was drowned out by minimum wage (and anti-gun) protestors. Clearly, the minimum wage is a divisive issue into which many people are emotionally invested.

Just to catch you up on the issue. During the 2015-1016 session, the Board of Aldermen passed Ordinance 70078 establishing a minimum wage in St. Louis City. This minimum wage would only affect businesses which make more than $500k a year and/or hire 15 or more workers. As written, it would have gone into effect upon the Mayor’s signature on 10/15/15. The minimum wage would be increased to $8.25/hr in St. Louis City. It would increase each year by about $1 until reaching a final value of $11.00/hr on January 1st, 2018. It would continue to be increased annually to keep up with inflation. In contrast, the state’s minimum wage is currently $7.70/hr.

Ordinance 70078 was not allowed to go into effect. Several groups, including the MO Chamber of Congress, sued the city claiming that it did not have the authority to increase the minimum wage. These groups believed that increasing the minimum wage would not only end up hurting low wage workers more than helping, but that it would also create a bad business environment. Eventually, that injunction was overturned by the MO Supreme Court. The new wages finally went into effect May 5, 2017. However, the victory didn’t last long. MO House Bill 1194 passed through the state legislature making it illegal for municipalities within Missouri to create minimum wages above the state’s minimum. Governor Greitens, declaring a disdain for both liberals and career politicians, has chosen not to sign the bill, so it won’t take effect until August 28 due to a constitutional procedural rule.

And that’s where it stands at the moment.

This brings up an interesting question: Is the minimum wage good or bad for St. Louis City and MO?

Unsurprisingly, Conservatives and Liberals have taken opposing sides on this issue. Conservatives hold that increasing the minimum wage cuts low skilled workers out of the labor market by removing the only competitive edge that they have, low cost. They assume a competitive labor market and claim that minimum wage laws hurt the poor more than they help.

Liberals hold that increasing the minimum wage increases income for the poor, giving those people more money to spend in the market and, therefore, improving the economy as a whole. Liberals assume a monopsonistic labor market, where employers hold all the cards in the labor market and pay low wage workers less than they deserve.

Who’s right?

Early Minimum Wage Research

Federal minimum wage was first enacted with the Fair Labor Standards Act of 1938. It was immediately opposed by many economists, most notably George Stigler in his article The Economics of Minimum Wage Legislation. Stigler argued that minimum wage:

1) Decreases employment in covered industries which outweigh gains in income.

2) Doesn’t precisely target low income families. For example, many low wage workers are teenagers in non-poor families.

3) Forced uncovered sectors to absorb those unable to find work, and so would face steeper competition. If there were no uncovered sectors, those people would go jobless.

Despite the opposition, the FLSA was passed and a federal minimum wage went into effect. The effects of this federal minimum wage were studied prolifically over the following years. These studies were summed up best by Brown, Gilroy, and Kohen in their 1982 paper, Effect of Minimum Wage on Employment and Unemployment. According to BGK, the literature showed a weak correlation between low wages and membership in low income families. It also showed that, with a 10% increase in minimum wage, teenage employment decreased by 1%-3%. They qualified their findings by noting that most studies used data describing teenagers (16-19 years old). Effects on adult (20-24 year old) employment was studied much less often and made it difficult to draw conclusions for that cohort. Generally speaking, the BGK paper gave a slight negative employment rate with increasing minimum wage, but couldn’t base that finding on much more than teenage studies.

The BGK results didn’t go without opposition. Opponents found problems with the techniques used by most of the BGK studies. First, many of the studies used aggregate U.S. data to look for effects. This was viewed as a shortcoming. Other economists raised concerns about a lack of a well-defined control case to compare the data against. They asserted that, without a control case, you could not state with any accuracy whether minimum wage increases had any effect on employment.

That was about to change.

The Infamous Card and Krueger Study

In the 80s and 90s, individual states began to set minimum wages above the federal minimum. This introduced new data sources and allowed additional experimental designs that included control cases. One of these new studies was the famous case study by David Card and Alan Krueger. In 1992, Card and Krueger published one of the most widely cited studies for the pro-minimum wage argument. Case Study of the Fast-Food Industry in New Jersey and Pennsylvania. Card and Krueger surveyed 410 fast food restaurants in New Jersey and Eastern Pennsylvania before and after New Jersey increased its minimum wage from $4.25/hr to $5.05/hr. Since Pennsylvania did not increase minimum wage, it served as a proper control. When the study was complete, Card and Krueger concluded that, despite the increase in wages, full-time employment actually increased in New Jersey as opposed to Pennsylvania. This directly conflicted with the findings of BGK. Proponents of minimum wage were vindicated. This paper is still used by Democrats today to argue the merits of minimum wage. Unfortunately, the Card/Krueger study has been completely discredited. The New Jersey study collected data imprecisely by using phone surveys. The data was later reconstructed using more precise payroll records. These newly constructed results pointed to there being a significant(4.6%) employment decline rather than an increase. Despite being discredited, this study is still pointed to as a proof of the minimum wage’s efficacy, a fact which never ceases to irritate conservatives. Democrats would do well to never bring this study up again.

Minimum Wage Studies in the 90s and 00s

Throughout the 90s and 00s more and more case studies were performed using payroll data between states with differing minimum wages. This literature is reviewed by Neumark and Wascher. Neumark and Wascher concluded that, while most of the studies leaned towards negative employment, there was high variability among the studies. Some of them showed no employment effects and some showed positive employment effects. However, NW pointed out that data over longer periods of time found negative employment effects. The studies with positive employment effects, in contrast, were usually short studies or case studies in specific industries, which aren’t necessarily reliable sources due to the many omitted variables. NW, like BGK, pointed out that most of the studies were on teenage employment, which isn’t the most interesting group to study. NW also noted substantial evidence that the lowest skilled minimum wage workers are typically replaced by better skilled minimum wage workers. This means that the results did not adequately describe the negative effects on the lowest skilled individuals. Additionally, they noted a potential endogeneity on minimum wage effects. Endogeneity is a situation where the outcome of an experiment causes changes to the input variables. For example, does minimum wage typically get raised in situations where employment is already decreasing?

The conclusion of the Neumark and Wascher review was that, while we still needed more information to truly determine whether the low wage market is monopsonistic or competitive, most of the data pointed to there being a negative relationship between minimum wage and employment.

Things have since gotten much murkier in the minimum wage debate.

Keep in mind that the economy is an incredibly complicated combination of variables and outcomes. Employment can be affected by a variety of things: business cycles, local political environments, population idiosyncrasies, or a multitude of other unknowable variables.

How are we to determine whether or not minimum wage has any effect whatsoever on employment?

The Equation Wars

There are some studies coming out of Berkely which are refuting all of the prior case studies. Dube, Lester, and Reich in their paper Minimum Wage Effects Across State Borders argue that case studies on differences between states don’t properly control for all of these variables. Most of the studies reviewed by Neumark and Wascher used data from the “Current Population Survey” put out monthly by the Bureau of Labor Statistics. DLR argue that employment effects vary across time and location. The term they use is “spatial heterogeneity.” The aggregate state data doesn’t control for this spatial heterogeneity, and therefore isn’t specific enough to draw any conclusions. In contrast, DLR use the “Quarterly Census of Employment and Wages,” a dataset composed of ES-202 filings (used to calculate unemployment insurance). This data is reported on a county by county basis. This allows DLR to study the effects of minimum wage increases on individual counties separated by a state border. They reason that, by using counties connected geographically, they can control for many variables which aggregate state data cannot. They also mathematically include a correction factor which, they argue, controls for spatial heterogeneity.

DLR found that, when controlled for spatial heterogeneity, minimum wage increases had a positive effect on employment.

And this is where we get into the weeds. DLR’s results depend on a statistical correction. Of course, this is nothing new. Every modern economic study has a variety of statistical corrections factored into their findings. Not surprisingly, different economists using different correction variables come up with different results, some positive and some negative.

For example, the debate over Seattle’s minimum wage increase has people arguing both sides. Both sides are using different correction variables to arrive at their conclusions. How can any of us, assuming we are laymen, legitimately weigh in on this debate?

The answer is, we cannot. This is an issue that is still being hashed out by the experts.

An Unsatisfying Conclusion

This leads me to a final, and unsatisfying conclusion to this debate:

Minimum wage, historically, has been deemed to be slightly negative in respect to its effects on employment. However, maybe that’s not true? We don’t know yet.

So how do we help poor people? Maybe the answer is the Earned Income Tax Credit. That seems to be the conclusion of the Congressional Budget Office. Here’s a little more information on that front.

Maybe the answer is Universal Basic Income?

At the end of the day, I think the hearts of the St. Louis City Board of Aldermen was in the right place. However, they didn’t fully grasp the inconclusiveness of the literature on minimum wage. Democrats would probably do better focusing on another more direct way of helping the poor.