The Post that was Promised

In my last post I said that real wages for a given class of income have fallen in hyperdense cities since 2000 due to rising rents, and that professional class real incomes there are much lower than in other cities. If you haven’t read that one, read it first because here I’m just going to make good on that promise.

Real wages by location are difficult to measure because of selection bias. Suppose you wanted to look at the wages of college graduates in two places. You find that real wages are higher in, say, Berkeley than they are in Tulsa, and conclude that another wave of Okies will take flight to the San Joaquin, this time with diplomas in hand.

But that’s probably wrong. The college graduates in Berkeley aren’t making more than those in Tulsa because the opportunities in Berkeley are that much better. It’s mostly because the college graduates in Berkeley are smarter, more go-getting, and better trained than the graduates in Tulsa or Dayton.

The other issue is this: cost of living is much higher in Berkeley. Largely, this reflects the high cost of rent, but high rents raise prices for any goods and services that require a physical footprint (everything but porn and Netflix).

When studies look at consumption patterns, they tend to find that uneducated people have much better consumption opportunities (which I will henceforth call real wages) if they move to a cheaper place, even though their paychecks get smaller numbers on them. For college graduates, however, these studies tend to find similar real wages across different commuting zones.

But I disagree with these studies, because of a proverb my father described to me as a child.

“My son, if a paper tried to control for a confounder, they did not fully control for the confounder.”

Even if you really have measured cost of living, the problem with comparing real wages for college graduates by location is that college graduates differ vastly. Papers try to correct for this by identifying human capital, but in my opinion they are all doomed to fail.

In a good new paper on this topic (the one linked above), let’s check the variables they used to distinguish college graduates so they could form comparison classes:

1. Age
2. Gender
3. Race
4. Hispanic Origin
5. Marital Status
6. Number of Children

You will notice that none of those are the number you would really want, which is the USNews ranking of the college that the person attended, the grades they got, and the subjects they majored in. Obviously this paper had no way of measuring actual human capital, and had to proxy based on years of education, which only gets you some of the way there.

So the attempt to identify how much each American household could be making in different locations is probably doomed. You would need pretty granular information about each person’s human capital, and I’m just not sure that exists (there might be some good way to do this using military assignments).

So how do we predict what a person’s real wages would be if they moved?

Probably you should just look at a class of college graduates who are basically the same everywhere and see how their wages vary by place.

I choose pharmacists. No offense to pharmacists, but, like, pharmacists are all the same dude. They all receive similar training at similar schools and they all do a similar job. They have very little premium on working special hours or having a stellar career history. There are no hotshot pharmacists, and there are no sleazy ambulance chasing pharmacists. So I think an average Tulsa pharmacist is probably pretty similarly productive as a New York pharmacist, and by seeing the wages of pharmacists we can get a good idea of wages for comparable college graduates by location. Another useful thing about pharmacists is that you basically need the same number of pharmacists anywhere you go, because the need for pharmacists is demand for medication, and age/health distributions just don’t differ that much across metros.

Pharmacist wages are also set by a super competitive market. They work by the hour, usually aren’t unionized, and their skills easily transfer across firms. So their wages aren’t subject to many screwy distortions and probably basically reflect supply and demand.

(GPs and civil engineers might also work for my purposes so I could include them in the analysis, but lol screw you guys, this is just a blog).

I have two claims about pharmacists:

1. Real wages for pharmacists are lower in New York City and the Bay Area than elsewhere.
2. Real wages for hyper-urban pharmacists have fallen since 2000.

Can we test these claims?

Luckily the BLS collects data on wages by location and they have a category for pharmacists.

Here are pharmacist wages in New York, Boston, San Francisco metros and the nation compared to cost of living in 2022 (since that’s the year I can find). 

	New York	Boston	San Francisco	U.S.
Median Wage (Hourly)	$59.62	$60.15	$77.5	$61.88
Cost of Living	168.6	153.4	244	(100)
Real Wage	$35.36	$39.21	$31.76	$61.88

As you can see, once you adjust for cost of living, real wages for pharmacists are half as high in New York and San Francisco as elsewhere.

So, one of two things are true. Either all the pharmacists are trapped in New York by a malevolent power, or professionals will make enormous sacrifices to their material well-being for the chance to live in the city. I think the latter.

Now let’s look at what sacrifices they would have made in 2000:

	New York	Boston	San Francisco	U.S.
Median Wage (Hourly)	$31.83	$32.62	$39.48	$34.11
Median Wage (2020 Dollars)	$48.35	$49.55	$59.97	$51.81
Cost of Living	118.5	120.8	145.6	(100)
Real Wage	$40.80	$41.02	$41.19	$51.81

These are much smaller. Real wages for pharmacists in the country have grown a lot since 2000, but in New York, Boston, and San Francisco they’ve actually fallen.

These numbers on the cost of living are kind of shocking, but I honestly believe them. The change for New York is hard to stomach at first but remember that the biggest thing happening here is a close of the gap between the outer boroughs (where it used to be pretty cheap to live and there were lots of poor people) and Manhattan (where it was always expensive and there weren’t any poor people south of maybe 110th). The change for SF does not surprise me at all. 

Also, there is a limitation on my study. Pharmacists can tell us how much professionals will pay to live in a location because they are similar people across locations. But we can’t just compare their real incomes over time. We want to look at the change in real incomes for a typical professional by location, but pharmacists have moved up the income distribution as health care has eaten the entire economy.  Pharmacist real wages grew 19.4% over that time, but average wages in the 4th quintile (remember, we’re talking about yuppies here) increased only 13% in that time. Supposing pharmacist real wages had only grown 13%, pharmacist wages would be 5.4% lower than they are now, and the change in real wages would have looked like this (all in 2020 dollars):

	New York	Boston	San Francisco	U.S.
Median Wage in 2000	$48.35	$49.55	$59.97	$51.81
Cost of Living (2002)	118.5	120.8	145.6	(100)
Real Wage (2000)	$40.80	$41.02	$41.19	$51.81
Median Wage if 5.4% lower (2020)	$56.40	$56.90	$73.32	$58.54
Cost of Living (2020)	168.6	153.4	244	(100)
Real wage if 5.4% lower (2020)	$33.45	$37.09	$30.05	$58.54

So, while actual real  wages for pharmacists in hyper-dense cities fell by only about 5-25%, they would have fallen by 10-30% if pharmacist incomes hadn’t outpaced the rest of the upper middle class. At the same time, real incomes for pharmacists nationwide increased 20%, and would have increased about 13% if pharmacist wages had just kept up with the rest of the upper-middle class.

You might doubt my particular choices of cost of living indices, but the general conclusion is robust. CPI-adjusted incomes for professionals in big cities have increased maybe 15% since 2000. Have prices in big cities increased that little in that time? Absolutely not; more like 25-30%. Beer at bars used to be 50% more expensive in New York than elsewhere, now it’s usually twice as expensive. 

It is unquestionable: pharmacists make large material sacrifices to live in New York, and they will now make much larger sacrifices to live in New York than they would in 2000. 

Of course, it is possible that pharmacists are effete social butterflies or have gotten way cooler since 2000, and fail to represent more practical, sensible groups of professionals who wouldn’t make these massive sacrifices. But, like, come on. It’s pharmacists.

The fact is that everyone, even boring pharmacists, have spent the last 20 years in a desperate dash to the coasts.

So why is everyone fleeing Milwaukee?

I doubt this really fits anyone’s political narratives that well. Liberal narratives fail. The non super-dense part of the country has gotten a lot more socially liberal and ethnically diverse since 2000, so people aren’t chasing liberalism. And I think conservative narratives probably fail too. The new liberalism in middle America is probably not why people hate living there so much, because the most diverse part is where they want to go.

Nice urban amenities are also not the answer. Coffee and restaurants have gotten a lot better in middle America since 2000, but coffee in New York was always good. I would argue the museums have gotten worse, but maybe I’m just crotchety and insta-ready exhibits are actually profound.

In sum, even though they’ve gotten more expensive, New York and SF haven’t improved in any material way. Outside of literally just Jersey City, regulation and zoning rules have kept the built environments of America’s hot locations in cryogenic stasis, and anyone who’s lived in New York for a long time can tell you that other than the slow removal of families and the poor and the rise and fall of crime, nothing has really changed.

I also want to reiterate that “taste have changed, now people like being in cities more” is not an answer to this question. Are young professionals now happier to be resource-poor in cities while they weren’t before? Probably not. Why would they be? What would be making them happier? And yet they take the deal, which means they wouldn’t be happy elsewhere, either. Even if “taste” is why they hate the Midwest now, they still hate the Midwest now and are forced by markets to live there.

I have to think this is a story about the collapse of social life. As Americans spend more time and home and become increasingly uncomfortable socializing with strangers, the critical mass of people required for educated to have a satisfying circle of friends has increased to 5-10 million people all connected to the same train system.

Bad, bad, bad, bad, bad.