Let the Facts Decide on Minimum Wage

Basic economic theory has very clear predictions on what should happen to a labor market with a minimum wage. If we assume an upward sloping labor supply curve and a downward sloping labor demand curve, a minimum wage will cause demand for workers to fall while supply increases. The result: an excess supply of workers, also known as unemployment.

The real world is obviously much more complicated than the standard Econ 101 textbook story. Perhaps the most important difference is that there is not just one equilibrium wage. No two jobs are exactly the same. Workers with different skill levels will face different labor markets. Location matters. The list of reasons why the real world doesn’t conform to the simplifying assumptions of economic theory is potentially endless.

However, those caveats do not necessarily invalidate the intuition of basic economics. I don’t think it’s a controversial statement that firms will attempt to substitute away from an input when its cost increases. In this case, that input is low skilled workers. As minimum wage increases, so does the necessary productivity of a worker who wants to be hired. A worker who only produces $10/hr of value for an employer will never be paid $15/hr regardless of the level of the minimum wage. Their choice is not between 10 and 15, but between 10 and 0 (unemployment).

But can we be sure that workers are actually paid based on their productivity? Couldn’t it be that they are simply being exploited, with firms pocketing the additional profits they generate? Under this scenario, an increase in the minimum wage could increase wages without hurting employment.

Here we see the limits of theory. Under some assumptions a minimum wage is good and under others it is bad. The clear next step is to look at the facts. Do minimum wage laws hurt or help low wage workers in the real world? Luckily, due to recent experiments with a $15/hr minimum wage in some cities, we have plenty of data to work with.

Supporters of minimum wage laws will be happy to find out that cutting edge research shows we have nothing to worry about (here’s the link to the full study). The increase in Seattle minimum wage to $13 (15 is being phased in over time) hasn’t had severe disemployment effects. There was a minimal decrease in employment, but overall, “results show that wages in food services did increase — indicating the policy achieved its goal.” So take that Econ 101. Minimum wage is great. Case closed.

Well, not quite. Because this morning, just 6 days after the study above came out, we have a new study looking at the exact same natural experiment (although with a different data set). The results are not so nice. They find that “the minimum wage ordinance lowered low-wage employees’ earnings by an average of $125 per month in 2016.”

Now what? Theory gives us conflicting results and so does data. How can an unbiased observer make a decision about the truth? Well, to be sure of the results of the studies above, they would need to sift through around 100 pages of dense statistical analysis. But of course, anybody untrained in statistics would first have to take a few classes to have any idea what they were talking about. And even with that training, they’d need to take a close look at the datasets used, weigh the pros and cons of each methodology, decide whether the results can generalize to other places, etc. Maybe after about two years of hard work they’d be able to have a qualified opinion on the two studies (never mind the dozens of other studies that have been done on the topic).

What is actually more likely to happen? Everyone who supported the minimum wage will cite the first study and look for flaws in the second (I would bet anything that Arin Dube – one of the biggest minimum wage scholars to support an increase – is scouring it right now looking for something to criticize). Everyone against minimum wage will do the opposite. Both will pretend they are letting the facts decide.

9 thoughts on “Let the Facts Decide on Minimum Wage”

  1. The question that comes to mind for me when I engage in this debate is why stop at $15 ?   Why not $50/hr ?  Most people would reject that out of hand…..but why ?  Intuitively most people understand that there are economic consequences (unemployment, inflation,etc) to creating artificially high wages.   So somebody in Seattle has figured out the exact right wage ? Certainly these are very complex systems but the market is much smarter than Al Franken in helping drive the optimal result.

    Attempting to create a ‘living wage’ or ‘affordable education’ by artificially increasing the supply has shown to have the opposite effect.

    I think you need to do a study of babysitting wages – something that is governed more by the free market than any other system.

    1. I think everyone agrees that there is a minimum wage that is too high, but that doesn’t guarantee that a nonzero minimum wage is optimal. There is a theoretical argument that if firms have monopsony power (like monopoly but instead of one seller it’s one buyer and many sellers) then they can set a below market wage and increase profits that will not be bid away by competition. If this is true then increasing the wage would actually move the market closer to the point it would be at in a perfectly competitive market. So then the question becomes empirical. What is the optimal wage? I agree with you that the government seems unlikely to find it, but others might be more optimistic and it is theoretically possible for a minimum wage to be beneficial.

      As for the babysitting market, it might not be as simple as you think! 


      1. Theory makes sense but I would think Seattle would be the farthest thing from a monopolistic situtation.  Are there good historiocal example of these situations that were not created via government protection/regulation ?

        Unfortunately the babysitting article is hidden behind a paywall – so unable to receive insight on why somebody would pay my daughter $20/hr for watching TV.   Although I have heard that the Economics tutoring wage scales are even more skewed. 

        1. I agree monopsony is probably not relevant in Seattle, but in some places there could be firms with some monopsony power.

          The babysitting article isn’t about minimum wage in particular but it shows how even private interactions can sometimes lead to coordination problems. In this case it was a babysitting coop that tried to implement a coupon system where people who babysat for others got coupons to use when they needed babysitting. The result was that nobody ever went out because they wanted to hoard coupons. Of course that still doesn’t imply government could have helped the situation. 

  2. Hey Chris, great timing with your blog post.

    In my opinion, there are two overlooked things about minimum wages:

    The first is that in labor markets where employers provide non-wage amenities, the relationship between wages and minimum wage laws is superficial. Rosen’s theory predicts that employers provide amenities when the perceived benefits exceed the cost. In response to minimum wage hikes, employers can simply eliminate amenities and pay more in wages. If the foregone amenities are roughly valued at cost, then supply and demand (hence employment) remain unchanged, earnings increase, but workers are no better off than before. Hence, our neglect to study non-wage amenities seems to create a tendency to find welfare benefits in cases where minimum wages have trivial effects. Perhaps amenities are uncommon, but I know that in retail and food services there appears to be a lot of variation in the amount of discounted goods offered to employees, time off, flexibility,

    The second issue is that minimum wages are not just, if at all, a mechanism to curb exploitation or force employers to create “better” jobs. Minimum wages are a mechanism for achieving a more equitable distribution of wealth. In theory, although minimum wages create unemployment, they almost undoubtedly raise the aggregate earnings of low-wage workers, in a manner similar to the rise in profits and loss in output of firms colluding on prices. I think the key question should not be whether minimum wages create inefficiency, but whether they are worse than alternative transfer mechanisms.

    1. If goal is redistribution, why not just redistribute directly? If they do generate inefficiency, there is a pareto superior outcome generated by allowing flexible wages and redistributing.

      Also, the new Seattle study does show an aggregate decrease in low wage workers so it might not even be working in the right direction.

        1. Probably not, but minimum wage seems like an especially crude method for redistribution. It pretty much ensures at least some redistribution from the lowest skilled workers (who lose their jobs) to slightly less low skilled workers (who are lucky enough to keep them). I don’t see how minimum wage effectively redistributes from rich to poor, which I assume is the goal, but I might be missing something. At least if we redistribute through the tax code it is more clear who we are taking from and who gets the benefits.

  3. The outcome that maximizes rents to a group selling a product generally involves a higher price and smaller quantity than the competitive outcome. Price collusion and minimum wages are similar, in this respect.

Leave a Reply

Your email address will not be published. Required fields are marked *