The turnover point in the Productivity bucket matches what the math looks like from the floor. Recruitment and retraining are real line-item costs, and they're easy to under-count because they hide inside "normal" churn. When pay moves people out of constant-quit mode, a chunk of that "cost" was never as fixed as the model assumes. Useful to see it isolated in the data.
If I understand this correctly, the increased minimum wage had no effect on the number of available jobs relative to states that did not increase minimum wage. This suggests that any reductions in jobs would have happened regardless of changes to minimum wage. i.e., companies would make similar decisions about changing headcount whether or not the minimum wage changed. Maybe total expenses for minimum wage employees were relatively small compared to a firm's overall expenses. Or maybe the ways companies reduced headcount still made sense regardless of the minimum wage, e.g., if increased automation makes sense when paying $15/hr, it could still make sense when paying $7.25/hr.
Also fair to question whether the observed patterns over the past few years are likely to persist going forward. I believe consumer behavior was largely consistent the past few years despite high inflation, i.e., companies increased prices were largely paid for by consumers (price inelasticity). At some point consumers are likely to reach a breaking point on inflation and make significant changes to purchasing behaviors. At that point we may see states with higher wage minimums suffer more job losses than states at the federal minimum wage level.
Interesting article. I accept that the US data suggests minimum wage increases have not led to a clear fall in overall restaurant employment, at least within the range studied. However, I think the UK experience shows why the issue is more nuanced.
On anecdotal evidence alone, increased employment costs do appear to change business behaviour. In the UK, employer National Insurance has increased and the minimum wage has risen materially over the last five years. At the same time, many supermarkets now operate with only a small number of staffed tills, with the rest replaced by self-service checkouts. Restaurants and other labour-intensive service businesses are also under significant pressure.
Of course, employment costs are not the only factor. Rent, energy, food costs, business rates, finance costs and weaker consumer spending all matter. However, it seems difficult to argue that rising labour costs have no effect at all. The effect may not always show up as immediate job losses. It may instead appear through reduced hours, less hiring, more automation, higher prices, lower margins, or weaker business viability.
So I agree with the broad point that minimum wage rises do not automatically destroy jobs, but there still needs to be a balance. The question is not whether wages should rise, but how far and how quickly businesses can absorb the additional cost.
I think minimum wages are always destructive. I suspect Dube's study doesn't show the damage for various reasons I'll summarize below, but the UK is a good case study of how badly this policy can go wrong.
When the British policy was introduced in 1999, critics pointed out that (a) price controls always create shortages and (b) once the Rubicon of regulated wages is crossed, the minimum wage is likely to be constantly pushed upwards as a form of vote buying by politicians. Both things have happened. It is to the Republican's credit that they resisted pushing up the minimum wages, but that didn't happen on the other side of the Atlantic.
Economic studies that claim there is no impact of minimum wages don't convince me because there have been so many businesses in the UK that directly told me or family members that they're understaffed or shutting down entirely because of the minimum wage. For example, my parents movement is now seriously constrained because they have a dog, and yet there are no nearby dog kennels because the only one they had access to shut itself down after a minimum wage hike (the owner specifically said that was the reason). And my grandparents were in care homes that had continuous staffing problems, again, directly blamed on the minimum wage.
So why doesn't this show up in the US data? Here's my theory.
The damage caused by the min wage comes in several forms. Outright reduction in jobs is the last and most visible form of damage, indicative of a truly crippled sector as employers like restaurants have limited ability to automate/reduce employment. If you're seeing jobs disappear entirely it's the result of outright business closures which are the last resort.
But what happens before this - and UK employers complain about this problem constantly - is that wage gaps get compressed. If you have two jobs, one of which is more skilled or a worse job than the other, then there must be a wage gap between them to incentivize people to take the harder/worse job. When min wages go up, to keep the gap existing businesses have to increase _everyone's_ pay, but that is very expensive so often they can't do that. The result is that employees become demotivated and fewer people train for the harder job because it no longer pays any better than the min wage. This causes recruiting failures and generalized dysfunction as important jobs are either not done or not done properly. Yet the number of jobs doesn't seem to change.
Birmingham is an example of how badly this can go wrong. A court judgement decided that paying bin men more than catering workers and librarians was sex discrimination, as the work was of "equal value" yet garbage trucks are staffed almost exclusively by men and librarians are nearly all women. So the council was forced to equalize their pay. This is similar to what happens under min wage increases, and it resulted in immediate strike action by the bin men with the result that Birmingham no longer has working trash collection in much of the city, it piles up on the streets and huge rats are now a regular sight. Technically the city still employs the same number of bin men as before, but they're so demotivated they refuse to do the work.
This kind of dysfunction simply isn't captured by simple analysis of spreadsheets.
So to be clear, you believe minimum wages are *always* destructive, regardless of the data, based on some personal anecdotes supposedly relayed to you by interested parties. Your prime example of the destructiveness is not falling employment due to rising prices. It's garbagemen pitching a fit because they got paid the same as the wimmin. Have I missed anything?
There is one additional channel for costs to be absorbed that you didn't mention, non-wage compensation. For example, suppose there is a benefit that workers value for more than it costs the employer to provide. It would be a win-win to for the employer to provide that benefit and lower wages, if lowering wages is legal.
I first started thinking about this when I was working for minimum wage in a nursing home. Our schedules were non-negotiable and given to us 5 days in advance, making it impossible to make any weekend plans (you never knew if you'd have to work the next weekend). I couldn't understand why, because I thought giving schedules farther in advance would be cheap and the employees would like it much more, but I realized we were already paid above the market clearing price for the work so every non-wage benefit could be sacrificed.
Another way of putting this is that if the market clearing price is $5/hour and minimum wage is $9/hour, the employer can subject employees to $4/hour of inconvenience and annoyance. And if they can make any amount of money by doing so, they will
The US population has increased by nearly 25 million people since 2010. How does this reflect in the findings? In other words, has raising the floor affected the rate of restaurant openings/closings during the period under discussion? If wages were lower, would we see more, about the same, or fewer new restaurants than we do? In other words, are higher wages a significant restraint on capital for new business formation?
Curious: Even if some low wage restaurant jobs were displaced, that might also mean some number of workers went into higher value-add industries and gotten higher pay jobs, right? Like a shrinking restaurant share might not even mean lost jobs, and a static share might mean total job gain?
Nice one, Arin! Love your work.
Hours worked is a key part of the complete picture though, as well as cuts in employee benefits, which is another way to absorb higher labor costs.
The turnover point in the Productivity bucket matches what the math looks like from the floor. Recruitment and retraining are real line-item costs, and they're easy to under-count because they hide inside "normal" churn. When pay moves people out of constant-quit mode, a chunk of that "cost" was never as fixed as the model assumes. Useful to see it isolated in the data.
I am curious about how the minimum wage increases impacted
- total hours worked,
- automation adopted,
- business consolidation,
- businesses started,
- businesses closed,
- teen employment, and
- homelessness.
Nice analysis, thank you.
If I understand this correctly, the increased minimum wage had no effect on the number of available jobs relative to states that did not increase minimum wage. This suggests that any reductions in jobs would have happened regardless of changes to minimum wage. i.e., companies would make similar decisions about changing headcount whether or not the minimum wage changed. Maybe total expenses for minimum wage employees were relatively small compared to a firm's overall expenses. Or maybe the ways companies reduced headcount still made sense regardless of the minimum wage, e.g., if increased automation makes sense when paying $15/hr, it could still make sense when paying $7.25/hr.
Also fair to question whether the observed patterns over the past few years are likely to persist going forward. I believe consumer behavior was largely consistent the past few years despite high inflation, i.e., companies increased prices were largely paid for by consumers (price inelasticity). At some point consumers are likely to reach a breaking point on inflation and make significant changes to purchasing behaviors. At that point we may see states with higher wage minimums suffer more job losses than states at the federal minimum wage level.
Any chance of doing cross-border comparison for areas just on the boundaries of state with minimum wage increase and state without?
Interesting article. I accept that the US data suggests minimum wage increases have not led to a clear fall in overall restaurant employment, at least within the range studied. However, I think the UK experience shows why the issue is more nuanced.
On anecdotal evidence alone, increased employment costs do appear to change business behaviour. In the UK, employer National Insurance has increased and the minimum wage has risen materially over the last five years. At the same time, many supermarkets now operate with only a small number of staffed tills, with the rest replaced by self-service checkouts. Restaurants and other labour-intensive service businesses are also under significant pressure.
Of course, employment costs are not the only factor. Rent, energy, food costs, business rates, finance costs and weaker consumer spending all matter. However, it seems difficult to argue that rising labour costs have no effect at all. The effect may not always show up as immediate job losses. It may instead appear through reduced hours, less hiring, more automation, higher prices, lower margins, or weaker business viability.
So I agree with the broad point that minimum wage rises do not automatically destroy jobs, but there still needs to be a balance. The question is not whether wages should rise, but how far and how quickly businesses can absorb the additional cost.
I think minimum wages are always destructive. I suspect Dube's study doesn't show the damage for various reasons I'll summarize below, but the UK is a good case study of how badly this policy can go wrong.
When the British policy was introduced in 1999, critics pointed out that (a) price controls always create shortages and (b) once the Rubicon of regulated wages is crossed, the minimum wage is likely to be constantly pushed upwards as a form of vote buying by politicians. Both things have happened. It is to the Republican's credit that they resisted pushing up the minimum wages, but that didn't happen on the other side of the Atlantic.
Economic studies that claim there is no impact of minimum wages don't convince me because there have been so many businesses in the UK that directly told me or family members that they're understaffed or shutting down entirely because of the minimum wage. For example, my parents movement is now seriously constrained because they have a dog, and yet there are no nearby dog kennels because the only one they had access to shut itself down after a minimum wage hike (the owner specifically said that was the reason). And my grandparents were in care homes that had continuous staffing problems, again, directly blamed on the minimum wage.
So why doesn't this show up in the US data? Here's my theory.
The damage caused by the min wage comes in several forms. Outright reduction in jobs is the last and most visible form of damage, indicative of a truly crippled sector as employers like restaurants have limited ability to automate/reduce employment. If you're seeing jobs disappear entirely it's the result of outright business closures which are the last resort.
But what happens before this - and UK employers complain about this problem constantly - is that wage gaps get compressed. If you have two jobs, one of which is more skilled or a worse job than the other, then there must be a wage gap between them to incentivize people to take the harder/worse job. When min wages go up, to keep the gap existing businesses have to increase _everyone's_ pay, but that is very expensive so often they can't do that. The result is that employees become demotivated and fewer people train for the harder job because it no longer pays any better than the min wage. This causes recruiting failures and generalized dysfunction as important jobs are either not done or not done properly. Yet the number of jobs doesn't seem to change.
Birmingham is an example of how badly this can go wrong. A court judgement decided that paying bin men more than catering workers and librarians was sex discrimination, as the work was of "equal value" yet garbage trucks are staffed almost exclusively by men and librarians are nearly all women. So the council was forced to equalize their pay. This is similar to what happens under min wage increases, and it resulted in immediate strike action by the bin men with the result that Birmingham no longer has working trash collection in much of the city, it piles up on the streets and huge rats are now a regular sight. Technically the city still employs the same number of bin men as before, but they're so demotivated they refuse to do the work.
This kind of dysfunction simply isn't captured by simple analysis of spreadsheets.
So to be clear, you believe minimum wages are *always* destructive, regardless of the data, based on some personal anecdotes supposedly relayed to you by interested parties. Your prime example of the destructiveness is not falling employment due to rising prices. It's garbagemen pitching a fit because they got paid the same as the wimmin. Have I missed anything?
There is one additional channel for costs to be absorbed that you didn't mention, non-wage compensation. For example, suppose there is a benefit that workers value for more than it costs the employer to provide. It would be a win-win to for the employer to provide that benefit and lower wages, if lowering wages is legal.
I first started thinking about this when I was working for minimum wage in a nursing home. Our schedules were non-negotiable and given to us 5 days in advance, making it impossible to make any weekend plans (you never knew if you'd have to work the next weekend). I couldn't understand why, because I thought giving schedules farther in advance would be cheap and the employees would like it much more, but I realized we were already paid above the market clearing price for the work so every non-wage benefit could be sacrificed.
Another way of putting this is that if the market clearing price is $5/hour and minimum wage is $9/hour, the employer can subject employees to $4/hour of inconvenience and annoyance. And if they can make any amount of money by doing so, they will
The US population has increased by nearly 25 million people since 2010. How does this reflect in the findings? In other words, has raising the floor affected the rate of restaurant openings/closings during the period under discussion? If wages were lower, would we see more, about the same, or fewer new restaurants than we do? In other words, are higher wages a significant restraint on capital for new business formation?
Curious: Even if some low wage restaurant jobs were displaced, that might also mean some number of workers went into higher value-add industries and gotten higher pay jobs, right? Like a shrinking restaurant share might not even mean lost jobs, and a static share might mean total job gain?