The minimum wage, the great reset, and the research of Isaac Sorkin

From Free Exchange at The Economist:

In the first [paper] Isaac Sorkin of the University of Michigan argues that firms may well substitute machines for people in response to minimum wages, but slowly. Mr Sorkin offers the example of sock-makers in the 1930s, which took years to switch to less labour-intensive machines after the federal minimum wage was brought in. He also explains how this finding squares with other research. Most studies look at past minimum wage increases that were not inflation-proofed. Firms may decide not to go through the hassle of investing in labour-saving machines if the minimum wage will affect them less over time. But they could respond differently to a more permanent increase.

Mr Sorkin crunches the numbers, using a model of the American restaurant industry in which companies choose between employees and machines. He investigates the effect of a permanent (ie, inflation-linked) increase in the minimum wage and shows that the tiny short-run effects on employment normally seen are fully consistent with a long-run response over 100 times larger. The lack of evidence for a big impact on employment in the short term does not rule out a much larger long-term effect.

In a second paper, written with Daniel Aaronson of the Federal Reserve Bank of Chicago and Eric French of University College London, Mr Sorkin goes further, offering empirical evidence that higher minimum wages nudge firms away from people and towards machines. The authors look at the type of restaurants that close down and start up after a minimum-wage rise. An increase in the minimum wage seems to push some restaurants out of business. The eateries that replace them are more likely to be chains, which are more reliant on machines (and therefore offer fewer jobs) than the independent outlets they replace. This effect has not been picked up before because the restaurants which continue to operate do not change their employment levels, so the jobs total does not shift much in the short run.

The piece offers further points of interest.


I would like to see the results of the same papers but with "foreign workers" replacing the word "machines" in the first sentence of each of the 1st and 3rd paragraphs above. E.g.

"...firms may well substitute foreign workers for people in response to minimum wages, but slowly."
"...higher minimum wages nudge firms away from people and towards machines."

* towards foreign workers

(...and I supposed you'd have to replace "people" with "domestic workers," too)

Sooo you basically just wanna re-write the papers with new words, thus rendering their analysis meaningless?

Ooor he's just trying to make a point.

The BLS reports that restaurant employment has grown faster than total employment in every year since they started posting the data.

Actually, the BLS reports that restaurant employment has grown faster than real restaurant sales in every year since 1992, the year the data on real restaurant sales starts.

This data implies that the long run trend growth of restaurant productivity is negative.

The minimum wage has declined relative to average wages since 1992.

"The minimum wage has declined relative to average wages since 1992."

Are you sure?

From 1992 to 1999 they did, but since then they look pretty comparable.

Yes, compared to average hourly wages or compensation. Those have gone up faster than household income and inflation, so even though there was a large bump up in 2007-2009, it didn't reach as high as earlier levels. MW was about 39% of AHETPI in 1992, fell to about 36% in 1996, rose back to 39% in 1998, then dropped to 31% by 2006, hit about 38% in 2010 and has fallen back down to about 34% now.

Here's a post where I have a chart that includes the ratio of MW to hourly compensation. It's a little hard to read, but it got really low in 2006 - probably below the natural minimum wage and definitely much lower than it has been since good data has been available.

Alright, I looked. And it just looks like you chose a high point as a bench mark. I could just as well say that minimum wage has risen relative to average wages since 1989.

From the linked paper: "Fig. 1. Federal minimum wage relative to average hourly earnings in the private sector."

Jwatts, Spencer picked the starting year to reference, not me. I'm not sure what the problem is. Figure 1 is exactly the same data I'm talking about, except that it ends in 2009, so it doesn't show the decline since then. My point was just that we should expect growing low wage restaurant employment during a time when mw has generally been declining relative to average wages.

And, look again at Fig. 1, JWatts. Do you really look at that graph and fail to see the downtrend?

Interestingly, the relative performance of the S&P index of restaurant stocks has a strong positive correlation with real average hourly earnings.

Um yeah, workers earn more when the restaurant business is doing well....

Clarification: I mean restaurant workers earn more when the industry is doing well. Conversely, when national wages are up the restauraunt business probably does better (because people have more spending money). But restauraunt workers alone earning more does not cause the restauraunt industry to do better.

Restaurant stocks? What proportion of restaurants does this represent? Fast food and some chains, but there are vast numbers of owner operated restaurants out there.

People eat out more when they earn more.

The perfect machine replaces not only the employees, but also the customers and the investors.

I envision an AI-driven system which manages all physical resources in the solar system optimally, with the mandate to create positive experiences (e.g. eating at a restaurant with friends) as simulated brain states as efficiently as possible.

Are the friends going to be simulated too?

They could be. But they don't have to be. If the experience is positive, whatever yields the greatest efficiency per resource input should be used.

If you think about something less amicable, where the exerience is unpleasant for at least one party (as in many Star Trek holodeck examples), their brain states should obviously not be simulated.

The simulated people also don't have to have continuous existence or identity. At the stage where human input is no longer required for system function, the human obsession with selfhood and ego will have become obsolete.

I'm just thinking that the sort of nerds who would be interested in a simulated dining experience probably don't have any friends. So the machine would have to supply simulated ones.

On the other hand, I suspect that such a machine would have uses beyond a simulated dining experience.

At that point, I think any consciousness will be simulated. Perhaps there will be some biological humans in a lab or something. But not many.

Horses were replaced by automobiles, wild animal species become fewer and fewer, now we're working to make low-skilled humans obsolete.

Your status quibs won't save you, your way of life is not compatible with probable futures.

If you're unlucky, it will be obliterated completely and there won't even be simulations of the complexity and human experiences you find high status. (Not that I care.)

Friends are a waste of time - look at all the idle time when friendly coworkers talk about game results or their vacation or weekend plans or politics.

Dining is inefficient and can be eliminated by continuous electric power or by fast chargers.

Entertainment is not only unproductive but its costs money to produce nothing.

Sleep is unproductive and can be eliminated, so all housing for homes can be eliminated.

Eliminating manufacturing of food, cooking tools, construction materials, carpet, beds, mattresses will make the economy more productive. Along with food markets, shopping malls,...

Sex is totally inefficient when a 3rd printer can reproduce much faster at full size and eliminate all the waste on clothes and shoes that become too smalls

Legs and feet are inefficient and can be replaced with wheels and treads.

Mulp thinks that he is criticizing capitalism, but actually, he's criticizing his own preferred command economics.

More and more I think that we already are in a simulation.

I like the concept that higher minimum wages benefit McDonalds at the expense of the local burger joint.

Maybe politicians know this - New York minimum wage only affects chains.

Of course. The chain restaurants have standardized processes that bring down labor costs per establishment. For example, most chain restaurants only prepare the final stage of cooking on the premises. They don't do much if any preparation of (for instance) sauces or other ingrediends. Whereas a small independent establishment night have a bunch of things that have to be prepared locally each night, and some minimum wage worker has to do the work for all that. The franchise model harnesses economies of scale to make chains more efficient. Those things are also produced in some central location and shipped out to the franchises, so the workers involved wouldn't be affects by a restauraunt-only minimum wage law, nevermind the fact that they are probably produced with machines.

I think it's important to remember that capital used to replace labor because of the MW is capital that has been reallocated, not capital that has been created out of thin air.

Also, I wonder if capital allocations that come about from MW hikes tend to be allocations where capital is more of a substitute than a complement to labor.

Because there is a finite amount of capital, Kevin?

Is this a trick question?

Maybe I'm wrong about this, but I have the feeling that more people are changing the time scale involved in issues in order to get a result more in line with their beliefs. A similar problem occurs as regards wars or conflicts where the parties concerned date the beginning of the problems from a starting point more beneficial to their beliefs, such that the parties concerned can't even agree on when the conflict began. But maybe that feeling is a result of my beliefs.

* Mood Affiliation

Why are we only discussing restaurant wages here?

Is there an assumption that the MW will be selectively enacted?

and its much easier to switch production now than in the 30's.

If we are only talking restaurants, of course there will be little investment in machinery: you don't have to compete with China as a restaurant. But the local CNC milling shop?

"If we are only talking restaurants, of course there will be little investment in machinery"

That's not completely true. Various chains are already considering automation to counteract stagnant sales growth. A rapidly implemented $15 per hour minimum wage, would almost certainly result in McDonalds rolling out kiosks across it's American restaurants.

Assuming no backlash, I would expect it to end up being implemented across all major chains. It's already happened with grocery stores.

From the Self-checkout section in Wiki: "As of the end of 2008, there were 92,600 self-checkout units worldwide. The number was estimated to reach 430,000 units by 2014."

So nearly a five fold increase in 6 years.

I read that Walmart moved away from self checkout due to theft. It's easy to code expensive organic vegetables as bananas or potatoes. It's also easy to throw an item into the bag without paying for it and then throw your arms up for help from the human - the human who doesn't care enough to check that everything in your bag was already scanned.

In fairness, organic vegetables are already a form of theft.

In France, the minimum wage is almost $11/hr. What has been happening to restaurants in France?

1. They've been automating ordering and payment:

2. They've been automating food preparation (by serving more and more industrially produced ingredients and foods):

" The reason for restaurateurs’ consternation is a law that came into effect on July 15 intended to protect “real cuisine” from the onslaught of industrial food that’s served in up to 70 percent of restaurants in France, according to some estimates."

3. They've been closing up shop:

"In 1960, France had 200,000 cafés, said Bernard Quartier, president of the National Federation of Cafés, Brasseries and Discotheques. Now it has fewer than 41,500, with an average of two closing every day."

France is where the fully automated convenience stores originated. Investors are rolling th em out in out of the way places here. Think of it as a huge walk-in vending machine. Everything from soap to light fare, and you can pay with your phone.

In 2014 some 21.2% of restaurant workers were paid at or below the minimum wage.

The next highest industry is personal care where some 6.2% of employees are paid at or below the minimum wage.
Most industries seem to only have 1% to 3% of their workers earning at or below the minimum wage.

Why is it that we have one industry that seems to depend so much more on low wage workers than any other industries?

It clearly is not foreign competition.

What's the number for adjunct professors?

Many restaurant workers earn tips, which can be several multiples of their wage.

Actually the 21.2% of restaurant employees earning at or less than the minimum wage only 5.4% earn the cw while 15.8% earn less than the minimum wage.
That probably is almost all tipped workers.

Very little of it is probably teens earning the sub-minimum training wage.

Of course this opens another question, why do so few firms take advantage of the sub-minimum teen training wage?

Since the typical teen works in a fast food place for less than six months a fast food manager could probably get away with paying almost all of his teen workers the sub-minimum wage.

Why is it that we have one industry that seems to depend so much more on low wage workers than any other industries?

I have thought about this a bit and come up with the following guesses:

1. Demand curve is steep because people have the option to eat at home.
2. There is always excess labor for the restaurants to sop up. That is labor that cannot find better higher demand work that pays more.
3. Many people feel that they are qualified to run a restaurant. Lots of competition.
4. Some people like to cook, they like the art of it. Consider people go to culinary school getting debt for a low paid job.
5. Many people learn how to cook and clean (do dishes) at home and so can do the jobs.
6. It is not a licensed profession.
7. Restaurant are a relatively nice environment to work in compared with other jobs restaurant workers might get

> 7. Restaurant are a relatively nice environment to work in compared with other jobs restaurant workers might get.

Restaurants are very appealing to young, attractive, outgoing people (since they tend to get the best tips). If you're not really that concerned about earning a high salary or developing human capital, restaurant work is a great way to spend your 20s. You don't have to wake up early, you meet lots of other young and attractive people, work is generally fun, nobody really minds if you come in high and you often get to eat good food and/or drink on the job. A lot of people even when they have high-paying, professional jobs still choose to do some shifts bartending. You basically are getting paid to hang out in a social environment that others are paying to be in.

Air conditioning, no heavy lifting, no particular skills needed. I used to do volunteer work with young people. One of my kids, a high school dropout with no parents, was perhaps the laziest human being I've ever known. He preferred working in restaurants over everything else. The grocery job required him to go outdoors in the sun and collect carts, and he flatly refused a very good opportunity to work for a landscaping company. He always managed to get sacked from even his restaurant jobs because he took too long breaks. But other restaurants would hire him. He didn't much care about money, so long as he had food, shelter, and weed. That's just one anecdote, of course, though I have others (but not enough to be data). Still, seems plausible.

I'm not familiar, but have heard from many that working in kitchens can be really exhausting hard work in a hot environment. I guess it depends on the type of restaurant?

"Isaac Sorkin of the University of Michigan argues that firms may well substitute machines for people in response to minimum wages, but slowly."

IOWs, the long run demand elasticity is more elastic than the short run demand elasticity for minimum wage workers. I'm not the least bit surprised.

Wouldn't that be the other way around? IE the short run demand for MW workers is more elastic than the long run demand.

The trade-off of transferring income to some low wage workers at the partial expense of other low wage workers depends of course on long run not short run elasticities. A higher elasticity makes the alternative of using an EITC to transfer income much more attractive. How odd that people who oppose increasing the minimum wage do not advocate a higher EITC more strongly. Could it be that it is the redistribution that they oppose, not the loss of some low wage jobs? not advocate a higher EITC more strongly....

You make this claim repeatedly, even as you see people advocate a higher EITC all the time. And you get called on this all the time.

Is there any research on the minimum wage and the effects on displacement of low-skill workers by higher skill workers? Intuitively, if the minimum wage is now above the marginal productivity of your low skill workers, you try to higher more productive (and thus higher-skilled) workers if you can.

I was thinking about the same thing.

Isn't the high skill option path most likely to be replacing many low skilled workers with fewer high skilled workers and more machines, aka, automation?

Something tells me there aren't that many high sckilled workers out there desperate to work in fast food restaurants.

Also what's stopping you from hiring the most productive workers you can now?

No, they would be the high skilled people who aren't presently working.

Because there's so many of those, right?

What's stopping a business is higher wages and business model. Compare for example, Wal-mart and Costco. The former has very low-skill, low wage workers. The latter uses higher-skill workers at higher wages. Presumably there is less turnover, absenteeism, etc. at Costco and they are probably more productive in general.

Because Costco contracts with other companies to provide the for low paid workers in their stores.

The chains may be winning out but another group that may benefit might be the Mexican and Chinese restaurants in strip malls. There is no way that those people dressed in rags working day and night in the Chinese restaurants are here are making minim wage. And you know what else, no way that the worker's kids I see running around in those restaurants are going to be drop outs stuck in low paying jobs when they grow up due to their parents earnings.

Once in a while one of the restaurant of which I speak gets busted for paying so low.

Usually there's a nail salon wedged in between the Chinese and the Mexican restaurant.

Here's the link to Professor Mankiw's review in the NYT of Arthur Brooks' new book. The point Brooks makes in his book is that conservatives should put on a happy face about policies promoted by conservatives, including conservatives' arguments against increases in the minimum wage. Mankiw agrees with Brooks. According to Mankiw, Brooks "offers a recipe for how conservative politicians can revise their rhetoric to undermine the left’s monopoly on compassion and empathy." Some might call it propaganda. Brooks and Mankiw would argue that most people are just stupid and that propaganda is necessary to lead them in along the right path.

@rayward I assume that mean the word propaganda to have a negative connotation, so If the people demand that the minimum wage be changed to $25/hour by September 1st 2015, would it be bad (propaganda) to tell them reasons why that might be a bad idea?

He means that if "conservatives" were worried about the negative employment effects of a higher minimum wage, they would point out that a higher EITC could raise the income of low income workers without the disincentive of a higher cost of employment.

I think the "propaganda" point is that Burns is suggesting no changes in policy, just how it is presented/advertised/popagated.

"I think the 'propaganda' point is that Burns is suggesting no changes in policy, just how it is presented/advertised/popagated."

Considering that the party who is advocating an increase in the EITC as an alternative to a higher minimum wage are also trying to cut the EITC, this seems like a reasonable conclusion.

So he's sort of like the George Lakoff of the right, then?

Sounds more like Jonathan Gruber.

My speculation is that minimum wages might exact their greatest toll not in a time of sunny prosperity but in a recession when unemployed workers are trying to find new jobs and employers find them too expensive to be worth the gamble. A stranger is much more of a gamble than a retained employee.

"I gotta get me a Polish".

Those were the words of an Irish B&B owner who asked a friend to arrange for her to get a servant for her B&B to do the dirty work she was tired of doing.

In this case, minimum wages dictate the substitutability between owners own labor and the labor of third persons.

So, what happened when the Irish economy tanked.

Bye-bye Polish. The owner resumed washing her own floors and cleaning the bathroom. And, the B&B owner, strangely enough, became an anti-immigrant person who argued that the Polish were taking the Irish jobs.

"Those were the words of an Irish B&B owner": really? She spoke in American?

Yes, as told to me by the tour guide who got her the Polish.

Diearie, By the way, if you think "gotta" is just an Americanism, here is a link to an Irish blog:

I've lived too long near a Costco. My first thought was, "yeah, mmm. A polish. With ketchup."

The thought became less delicious with the second sentence. Sigh.

"In this case, minimum wages dictate the substitutability between owners own labor and the labor of third persons.
So, what happened when the Irish economy tanked.
Bye-bye Polish."

So with a lower minimum wage, the Polish laborer would have probably kept his job. But as it was, when economic conditions drove the value of the Polish laborer's service below the minimum wage, the laborer was fired.

Yes, that sounds about right.

I feel like this another case of economists thinking too much like statisticians and less like historians. The 1930s was a period where the population was growing rapidly, firm focus was on expanding production to meet demand and was less focused on cost savings. In today's lower growth environment, the focus of firms have been driving efficiencies rather than expanding capacity, so instead of hiring more sales staff and product developers, they are more likely to invest in IT systems that can save labor.

So a high minimum wage has little adverse effect in the short term on unemployment, but leads to large efficiency gains in the long term, so everyone's a winner. It's a magnificent policy.

I was thinking the same thing. "I support the minimum wage because I think we need to give our businessmen incentives to innovate," they could say.

But the problem then becomes, what happens to the workers? One could imagine them working fewer hours but that's almost unthinkable in the current ideological environment, as are things like a basic income.

With time, the workforce should adjust, and on average you'd expect jobs to pay better. I guess the adjustment consists partly of older people retiring and younger people getting the newer jobs working with the technology, and partly the same workers changing jobs.

I remember it happening when garbage collection trucks got the automated bin pick up arms on them. In places with high minimum wages, these things spread quickly in the 1990s. In other countries it took some time, and you would still see a truck with two people running behind loading bins into the 2000s.

Now the garbage trucks have one guy, not three, working them. When you're worth $10 an hour and MW is $12, the only long range result is that you are on the dole.

No, local governments are spending their money on manufacture and maintenance of the machinery and other things entirely unrelated to collecting garbage. End result is people spending their time working more productively.

Some activities are worth $10 per hour. People get paid whatever the available activities are worth. There is no such thing as a person who is worth any particular amount of money. In the middle ages I could not have earnt what I do now.

Dean Baker appreciates when people get the point of productivity growth.

It's only an efficiency gain if the capital being substituted would have been cost effective _before_ the minimum wage was raised: If an automated ordering kiosk is not cost effective at a $7 minimum wage, but it is cost effective at a $15 minimum wage, raising the minimum wage is still going to be a net loss.

Second, it's only an improvement for the customer is the labor to capital substitution improves their experience somehow. I don't think I have a better experience at McDonalds now that all the food is prepared offsite and just re-heated than I did when burgers were cooked on site. If that lowered the price I pay, that would be an improvement, but if that just limits the price increase that I would be facing in the face of a minimum wage increase, it's no benefit.

Finally, artificially raising the cost of labor might simply make some entire types of business unprofitable. Amazon's model of selling the same books for cheaper drove a lot of bookstores out of business, but that's not a bad thing for me as a consumer because I can still get the same books and at a lower price now. However, a minimum wage increase might make it impossible to do a sub $10/plate seated meal, driving the diners I like to eat at out of business. There will still be fast food, there will still be $20/plate restaurants, but those are completely different experiences than an $8-10/plate greasy spoon.

Why doesn't a low minimum wage require a lower maximum price?

And to get a higher maximum price requires higher minimum wages.

Or can wages be low while the cost of living is high, based on free lunch economic theory: paid $5 per hour for 30 hours per week is more than enough to pay $300 a week for rent, $100 a week for food, $50 for transportation, $25 a week for clean clothes, $25 a week for telecom to track hours required to be at work

What is missing from the entire discussion is the fact that the higher minimum wages increases the income of the workers who are employed. Most studies to-date suggest that for modest minimum wage hikes, the income increase to low wage workers outweighs the income lost from employment reductions

You can't make an omelette without causing a few working poor to become completely destitute.

"Most studies to-date suggest that for modest minimum wage hikes, the income increase to low wage workers outweighs the income lost from employment reductions."

That only addresses the immediate effect. What are the net results when you look at the future life time earnings of all the effected workers? The life time earnings of the unemployed may be drastically worse, because they may stay unemployed, whereas many of those boosted by the minimum wage would have gone on to better paying jobs after a few years at minimum wage levels.

Cape Cod makes an interesting case study in the minimum wage.

When I was young, college student flocked to the Cape for summer jobs. It was a great experience, I worked one summer at a resort hotel in New Hampshire. In addition to the minimum wage they provided room and board. But more importantly they hired many coed as waitresses. Every night was Saturday night and every morning was Monday morning.

But the cost of living got so high on the Cape that the college students had to spend everything on room and board and could not save anything for school. So they just quit coming to the Cape for summer employment. It got so bad that the Cape Cod businesses started going to Eastern Europe and the Caribbean to recruit summer workers. And this was despite Mass having a minimum wage higher than the national.

A highly relevant factor not considered. A extra high minimum wage or similar restriction even in a small area of the country will create a highly profitable niche market for innovation, which, once invented/developed, will move down the learning and cost curve and ultimately automate jobs at lower minimum wages.

For example, the $15/hr rate for hamburger flippers in Seattle and now LA, just more than doubled the potential value of a robotic system. That extra margin will allow a small innovative startup to get across the "valley of death" problem with a higher valued niche product, then expand down the learning/cost curve to the mass market in the rest of the nation, independent of the minimum wage. Innovation often needs high value stepping stones to become a reality in the mass economy.

Comments for this post are closed