Shrinking American cities

This post will be of little interest to Americans, but might surprise a few foreigners.  Suppose you live in a European country, or Canada, Australia, Japan, Mexico or Brazil.  Or anywhere but America.  Find a list of your 20 biggest cities in 1950, and then another list of the 20 biggest cities today.  I’d bet there’s lots of overlap on the two lists.

In 1950 the 4 biggest cities in America were New York, Chicago, LA and Philadelphia.  Today, these are still 4 of the 5 biggest cities in America.  No surprise.  But the rest of the top ten list is nowhere to be seen–they’ve all dropped out of the top ten.  Actually, I wasn’t all that surprised by that fact.  But what did surprise me a bit is that all but Detroit dropped out of the top 20!  (Detroit fell from #5 to #18, as its population declined from 1,850,000 to 717,000.)  And the second ten didn’t do much better, eight of the next ten also dropped out of the top twenty.  Only Houston (which jumped from #14 to #4), and San Francisco (which fell from #11 to #13) remained.  So there was a pretty complete wipeout of American cities just below the top tier.

It’s actually not as surprising as it seems, as two factors were at work.  First, the cities in the 5-20 range in 1950 mostly did do rather poorly.  Many weren’t quite big enough to re-invent themselves as post-manufacturing centers, which explain why they lost more people than the top 4 cities.  In addition, many were in the north, and lost populations to sunbelt cities.  But the population within city limits exaggerates this trend, as many newer cities were able to annex large suburban areas, while older northern cities were hemmed in by suburbs.  Thus some old cities that did become white collar success stories (Boston, Washington, etc) actually have fewer people than cities like El Paso, which annexed suburban areas.  Obviously, however, their metro areas are much bigger.

This will not happen again!  For complex reason, in 2070 the top 20 cities will look similar to today.

In addition to Detroit, here are some other cities where residents can enjoy more open space, uncrowded streets, nice museums and symphonies, and beautiful old homes at rock bottom prices:

St. Louis:  857,000 to 319,000,  Cleveland:  915,000 to 397,000, Buffalo: 580,000 to 261,000

I was born in 1955, so to me those three will always be big cities.  But they have fewer people than Mesa.

These cities are not in the top 20:

Baltimore, Boston, Seattle, Washington DC, Nashville, Denver, Milwaukee, Portland, Las Vegas, Kansas City, Atlanta, Miami, Cleveland, Oakland, Minneapolis

And these aren’t even in the top 50:

New Orleans, St. Louis, Pittsburgh, Cincinnati, Buffalo

Prediction:  Austin will be the fastest growing big city in America over the next 50 years.



20 Responses to “Shrinking American cities”

  1. Gravatar of MikeDC MikeDC
    11. July 2011 at 05:04

    I think you should look by MSA and CMSA. Those are the standard ways to measure population centers (especially from an economic context) rather than city limits. That gives a much more stable look at things.

    DC, for instance, is the 4th largest CSA in the country.

  2. Gravatar of Indy Indy
    11. July 2011 at 06:30

    MikeDC is totally right – the numbers are highly distorted by arbitrary municipal boundaries. This is a much better list.

    Houston is over 600 square miles (the MSA is 1300 – nearly as big as Rhode Island!) and Phoenix is 520 – whereas Detroit-proper is only 143 but surrounded by nearly four million people in the suburbs who live in different cities but are in truth part of the same Metropolitan area. Fully 84% of that MSA is not in the city limits. You can tell by the huge variance in density in the list.

    Kansas City is an even more extreme circumstance, because a state-line divides the population nearly in half and neighbors across the state-line road from each other “don’t live in the same city” – which is the absurd triumph of legal technicality over reality.

    What happened is that almost all metros grew, and people all around the country moved to suburbs, but there were older cities in the East and Midwest which has small boundaries, and newer cities in the West which, for Historical reasons, started out gigantic or annexed all their surrounding territory. When Houstonites moved to the suburbs, they were still all in Houston. When Detroiters did the same – they were almost all not technically in Detroit anymore.

    So the phenomenon at play here is very different from a “economic rise and fall and mass migration / labor-mobility” narrative. There is that, of course, but the bigger story is the abandonment, hollowing out, social decay, and de-facto racial resegregation of the urban cores of most of our major large cities since WWII. It’s a tragedy.

  3. Gravatar of Scott Sumner Scott Sumner
    11. July 2011 at 07:11

    MikeDc and Indy, Yes I did mention that the metro changes were less dramatic, and that city limits bias the picture.

    Still, in most foreign countries that’s also true, but you wouldn’t see such a dramatic shakeup of the top 20.

  4. Gravatar of Derrill Watson Derrill Watson
    11. July 2011 at 07:12

    I play around with the idea sometimes of trying to reintroduce city states in the US (e.g. New York CityState and Empire State) and was dismayed a couple years ago when I realized this. How can you make stable city-states if the largest cities are going to change that much?

  5. Gravatar of SHocking SHocking
    11. July 2011 at 08:33

    Prof Sumner,

    One of the thing I think that’s very interesting about praise of Sunbelt construction build-ups like that doled out by Ed Glaeser is that when you find yourself in a Detroit-esque situation, that excess supply of housing becomes a major liability. Vegas has seen a similar phenomena rock its housing markets.

    So during boom times, if the supply of housing is very flexible, then regions will naturally acquire a huge stock of housing to accommodate the incoming population. This will further boost the economy and in a feedback cycle, frothy housing markets and the dynamism supplied by new migrants will buoy these economies (at least in part). But when a secular shock (or decline in some other area of the economy) hits this same region, the extra houses they built becomes a major drag on the economy. As there’s already a huge supply, prices respond severely to fluctuations in demand. On top of that, because housing was driving the initial euphoria, once that engine of growth cools, the region becomes considerably less attractive.

    Finally, unoccupied housing and blighted neighborhoods impose major negative externalities on the welfare of residents, further discouraging new entrants and compelling exits. This potentially dangerous feedback cycle seems to suggest that some impediments to residential construction might be a good thing in the long view.

  6. Gravatar of anon anon
    11. July 2011 at 09:15

    SHocking, my observation is that “blighted” neighborhoods tend to re-gentrify over time if the fundamentals are strong. Detroit and Las Vegas are facing deeper issues which have nothing to do with housing affordability.

    Still, it makes sense to moderate speculative swings in housing prices. One straightforward way of doing that is taxing assessed ground rent, since much of the variability in housing prices is due to changes in expected land rents. Taxing rent would also internalize the benefits of local government services.

  7. Gravatar of happyjuggler0 happyjuggler0
    11. July 2011 at 17:14

    NYC had about 8 million people in 1950, and has about 8 million people now, and has had about 8 million people in every period in between.

    Now which clever econo-head can tell us why?

    The answer is easy if you know at least a little about NYC and at least a little about economics.

  8. Gravatar of Scott Sumner Scott Sumner
    11. July 2011 at 17:59

    Derrill, I’ve had exactly the same (sad) thought.

    Shocking, First lets take away all the massive subsidies, like tax deductions, the GSEs, FHA, etc. If we can’t get rid of deposit insurance, require 20% down from any institution whose deposits are insured, or who is considered too big to fail. That’s a good start.

    anon, Agreed, my hunch is that Vegas will have an easier time bouncing back than Detroit. Vegas is much likelier to be healthy in 10 years (although it’s also very possible it won’t.)

    happyjuggler0, Rent control? High prices? (i.e if prices start high, they can fall until new people replace those who moved out. Not so in Detroit. NYC has actually had a slight increase in population, which is surprising considering the big fall in family size. The typical fully built up American city might have lost 25% of its population merely from families going from 4 people in 1960 to 3 people today.

  9. Gravatar of happyjuggler0 happyjuggler0
    11. July 2011 at 18:58

    My answer was rent control. They’ve added a ton of office space and other business space in NYC during that time period, but really few new apartments or condos (let alone houses).

    At current market prices there is still a lot of demand to live in NYC itself, but precious little new supply is added to accommodate that demand except at the luxury level.

  10. Gravatar of happyjuggler0 happyjuggler0
    11. July 2011 at 19:02

    I also meant to fill in one more step to my last reply. Namely that relative to office buildings for the same spit of land, it makes no sense in most locations to add new housing units in NYC precisely because of rent control.

  11. Gravatar of Sean Brown Sean Brown
    12. July 2011 at 15:34

    Scott, I am from Austin and I doubt that we will be the fastest-growing U.S. city over the next 40 years. This is especially true if you are only counting within the city limits.

    As local development economics/land-use blogger Chris Bradford has written, anti-density development policies at the city council and city government have actually resulted in a drain of population from Austin’s city center.

    Most of the growth has been on the fringes of the city, the bulk of this coming in suburbs outside the city limit.

    We have a 7-member city council, and last month a strident anti-density candidate defeated a pro-density incumbent 56-44. One of the main campaign issues is a new water-treatment plant under construction which would probably be necessary if Austin is to be a fast grower over the next 40 years. The anti-density candidate implied she will explore ways to stop or modify its construction. While I think that making water a lot more expensive (Austinites water their lawns way too much due to excessively cheap water), there is a lot of NIMBYism here in general which will inhibit dense growth (high-rises, duplexes, apartment complexes, etc.), even in the city center.

    In terms of transit, traffic is bad, bus service is awful, and a promising light-rail plan was voted down in 2000 by a 51-49 margin, so we now have a $100m commuter rail which is essentially useless (1000 boardings a day, serving mostly suburban areas).

    Anyway, those are some major “choke point” reasons why Austin’s growth is likely to slow, particularly within the city limits. Of course, despite my inveterate pessimism things may work out more positively – developers may eventually be able to tear down “suburban” (single-family) inner-city neighborhoods to build apartments and mixed-use developments may make living within the city limits more attractive. However we are a long way away and Austin literally can’t keep growing at “prior trend” (1930-2000) of 40% – let alone the last decade’s 20% – without annexing most of the surrounding suburbs (politically conservative cities like Kyle, Cedar Park, and Round Rock). I doubt this will happen for political and taxation reasons.,_Texas

    The other way to grow is infill/densification, which has major political barriers due to the preference of many voters to keep the city of Austin like a huge suburb, not allow outsiders to park on or use streets near their houses, not allow new developments to possibly erode the scarcity value of their houses, etc.

    Anyway, Scott, what is the basis for your projection re Austin??

  12. Gravatar of ssumner ssumner
    12. July 2011 at 18:37

    Happyjuggler), I thought so.

    Sean, You know much more than me. Here was my thought process:

    Texas is best positioned for growth.

    Dallas and Houston will suffer from congestion, because of large population. Also their climate is slightly worse that Austin/San Antonio,

    Austin is hipper for young people than San Antonio, has the university, state capital, high tech, etc. Small enough that it can still grow fast in percentage terms.

    I meant metro area, I didn’t know about those problems within the city limits.

    It’s growing fast now, and leads some surveys on best places to move to.

    But the very liberalism I point to may slow growth, as you say it leads to NIMBYism. So maybe you are right.

  13. Gravatar of kiwi dave kiwi dave
    13. July 2011 at 07:18

    Not to pile on, but Indy and MikeDC are right: if you look at metro areas rather than the areas within city limits, the changes are much less dramatic. Comparing the list of the top 20 cities from 1950 to the table of the biggest MSAs from 2010, what is striking is actually how consistent it is.

    Of the top 20 cities from 1950, all are in the top 50 MSAs in 2010. 13 of them are in the top 20 MSAs now, four are in 20-30 place, one is 30-40, and two are 40-50.

    So although the shift from rust-belt to sun-belt is real, the much bigger story revealed by this data is a shift from cities to suburbs.

  14. Gravatar of Scott Sumner Scott Sumner
    13. July 2011 at 14:51

    Kiwi, Not to repeat myself, but I made exactly the same point in my post, as you Indy and Mike made. Still I was comparing to the within city limits population of other countries, so the comparison is valid. Other countries also have cities hemmed in by suburbs.

  15. Gravatar of Jaap Jaap
    13. July 2011 at 15:14

    hi Scott,

    great that you’ve started again, but now I have a lot of reading to catch up on!

    anyway, my thoughts here on your prediction:
    you seem to be falling for an extrapolation fallacy. what happened in the past, and is happening now, therefore must keep on happening in the future.
    this week, I read that in 2100 Nigeria will have a larger population than China (according to the UN-predictions). doesn’t that just sound absurd?

    what makes you think that an event prevents these new big cities from falling apart? it will not repeat exactly, but may rhyme. water-shortages may make some fast-growing cities undesirable for instance.

  16. Gravatar of kiwi dave kiwi dave
    13. July 2011 at 17:51

    Kiwi, Not to repeat myself, but I made exactly the same point in my post, as you Indy and Mike made. Still I was comparing to the within city limits population of other countries, so the comparison is valid. Other countries also have cities hemmed in by suburbs.

    Yes and no. I think this is one of those cases where people’s use of terminology differs between countries. In at least some other countries, (particular Australia and NZ, with which I am most familiar), when people talk about the population of a city they’re almost always referring to the metropolitan area, and not the core municipality. For instance, ask any Sydneysider, or any Australian, and they will say that Sydney is Australia’s most populous city, and (if they are moderately well-informed) they will say that Sydney has 4.5 million people. The actual City of Sydney has only c. 180,000 people and is far from the biggest municipality in New South Wales, let alone Australia, but everyone in Australia thinks that the boundaries of local government areas are arbitrary and mostly irrelevant. Although I am not as sure, I think many (most?) Europeans think similarly.

    Likewise, if you mention, say, LA, Boston, Washington or San Franscisco to many non-Americans, they will think of the entire metro areas, not the tiny core municipalities that only have about half a million people each.

    Which is a long way of saying, that I think your foreign readers may be signiicantly less surprised about the change in population rankings of American cities when the context is explained to them.

  17. Gravatar of ssumner ssumner
    14. July 2011 at 08:57

    Jaap, Actually, Nigeria is forecast to still trail China by 200 million in 2100, but thats pretty close.

    Water shortages don’t really inhibit growth, there’s plenty of water in the rural areas that can be diverted to cities. They been saying water shortages would stop growth in Phoenix for decades, and they are always wrong.

    Kiwi, I didn’t know Sydney had that few people. I stand corrected. I was also thinking of the fact that, at least in Europe, people don’t move around as much.

  18. Gravatar of Bob Bob
    15. July 2011 at 12:42

    The advantages of living in the urban center in the cities that are not in the top 5 are pretty much nil. Most of those large, but not large enough cities were designed with the car in mind, and have ludicrously bad public transportation systems. The advantages of being surrounded by other people just weren’t there, and the more time passes, the less advantages they provide.

    If all the advantages that one gets from living downtown disappear, of course people have to move to the suburbs. Compare this to NY, Chicago, or any large city in Europe: Living downtown is actually desirable for many people.

    In St Louis, the good jobs, restaurants and shopping centers aren’t downtown: They are in the inner belt of the country, where most people actually want to live.

  19. Gravatar of ssumner ssumner
    16. July 2011 at 10:10

    Bob, It depends how rich those medium cities are. San Francisco, Seattle, Portland, San Diego, Boston, Washington DC, Miami and many others have decent places to live in the city. Even my home town (Madison, pop. 200,000) does.

    St Louis is clearly near the bottom.

  20. Gravatar of TheMoneyIllusion » A few random thoughts on Detroit TheMoneyIllusion » A few random thoughts on Detroit
    27. July 2013 at 06:15

    [...] say about the Detroit bankruptcy, but I’ll throw out one possible factor.  A few years ago I did a post discussing the striking population loss of large American cities after 1950.  The top 4 continued [...]

Leave a Reply