Sunday, May 1, 2016

Housing: Part 143 - Closed Access and Inequality

After posting part 142, I realized that I could improve on the graphs.

Here is the distribution of incomes, before and after rent expense, both for Closed Access cities and for zip codes everywhere else.  This includes 1670 Closed Access zip codes,  (Closed Access cities here are NYC, LA, Boston, San Francisco, San Diego, and San Jose.) and 14,154 other zip codes.

Non-Closed Access cities are a veritable worker's paradise.  In the Closed Access cities, we can see both the very fat distribution of incomes far above the median and the fat distribution of incomes after rent far below the median.  Although it would be difficult or impossible to measure, the incomes far above the median reflect a combination of (1) incomes due to selection bias where highly skilled workers are drawn to the lucrative labor markets of the Closed Access cities and (2) excess income that those highly skilled workers collect because Closed Access housing policies limit competition in their fields.

Here, we can see how far up the income scale the housing constraint problem reaches.  For zip codes with roughly double the median income, discretionary shifts in housing consumption allow them to retain, proportionately, their incomes after housing expenses.  We can see here how, given the ability, households revert to a stable level of housing expenses.

 Closed Access = Red
Next is the scatterplot comparing incomes before and after rent (these are both in log scales).  Here, compared to the previous post, I have color coded the plots by Closed Access and Other.  Isn't it amazing how sharp this pattern is?  I have also added marks delineating income quintiles.  (These are zip code quintiles, not household quintiles.)  We can really see clearly here how the migration into and out of Closed Access cities creates the statistical artifact of a hollowed out middle class.

Highly skilled workers at the top of the income distribution move into Closed Access cities, increasing both their incomes both before and after rent.  Because housing is constrained, households at lower income levels must move.  This happens passively, as housing expenses ratchet up, eventually causing enough distress to force a household into the Open Access part of the country.  You can see here how sharp the difference between Closed Access and Open Access markets is.  When households make this geographic shift, they lower their gross incomes but they raise their discretionary incomes after rent.  This compositional shift makes it look like the middle incomes aren't growing as much as upper incomes.  This doesn't affect the lowest quintile, though, because there can be no downward shift out of that quintile.

Here is a graph of the relative change, in real dollars, of incomes, by quintile.  (The top quintile is divided into deciles.)  This is from the Survey of Consumer Finances.

The next two graphs show the change in zip code incomes, arranged by beginning income levels.  The first is for St. Louis - a typical non-Closed city.  The other is for LA, a typical Closed Access city.

From 1998 to 2006, a 1 point increase in mean zip code income (an increase of 172%) in the open access cities correlated to an additional total nominal income growth of about 3% over the entire period.  This is roughly the difference between the top limit of the lowest quintile zip code and the bottom limit of the highest quintile zip code.  Compare this to the difference in real household incomes in the previous graph, which shows the lowest and highest groups rising by about 20% more over the same period, compared to the middle quintiles.  This hollowing out doesn't show up anywhere in an individual city.

Let's compare a typical low income zip code in St. Louis with one in Los Angeles.  Here I will use log incomes of 10.5 and 11.5, which correspond to \$36,316 and \$98,716, as starting incomes for a typical zip code.  This table shows how those zip codes fared.  Both neighborhoods in LA fared better than in St. Louis before rent expense.  But, the high income zip code fared much better.  And, remember, in the high income neighborhoods, housing expenses are much less of a drag.

This is where some statistical analysis of the housing boom gets into trouble.  Migration is the story here.  The composition of households and their movement between cities is the story.  So, if we're trying to be good, objective statisticians here, we would normalize all of the city data, to get rid of local effects.  Then, if we ran a regression of local incomes and local home prices, we would find that the zip codes with low and declining incomes would correlate with zip codes that had the highest increases in home prices.  This would appear to confirm that the housing boom was created by an explosion of credit to low income households.

But this gets it entirely wrong, because the story was edited out of the regression by those standard statistical procedures.  All of those zip codes that supposedly had declining incomes and rising home prices are from Closed Access cities.  They only look like they had declining incomes because their rising incomes were adjusted out of the data.  In-migration was causing incomes to rise throughout the Closed Access cities.  The low income neighborhoods that had rising home prices actually had high income growth, but it was growth mostly coming from a migration pattern where low income households were moving away and high income households were moving in and bidding up houses.

The pattern of low income neighborhoods with falling incomes and excessively rising home prices doesn't show up outside the Closed Access cities.  In all the Closed Access cities, there is a sharp pattern where income growth was weighted to high income zip codes and home price growth was weighted to low income zip codes.  I have included the St. Louis and Los Angeles graphs here to give an example.  This is how all the Closed Access cities look.  And, the other cities typically look like St. Louis.  Even Las Vegas and Phoenix look like St. Louis, except of course home prices across the board rose at a much higher pace than they did in St. Louis.  Here, I'll throw in the Phoenix graph, just because you, understandably, probably don't believe me.

As in St. Louis, there is little difference between home price appreciation in low and high income zip codes during the boom - any difference amounts to about 5% or less.  Maybe credit expansion increased home prices in low income zip codes outside the Closed Access cities by 5% relative to income growth.  What we do see, however, in most cities, whether open or closed, is a collapse in home prices that is weighted toward low income zip codes.  We did that to them.  There is no getting around it.  The persistent drop in home equity has been targeted on low income zip codes, because the housing bust was imposed through constraints on mortgage credit.  We're socking it to them coming and going.  And, it has nothing to do with wages or negotiating power.  First we refused to build houses, then we refused to fund them.  We've been patting each other on the back for 10 years about how those low income rubes were being led like lemmings into homes above their means and how we put a stop to it and how the bankers did this to us and how they need to pay.  It never happened.  First policy makers in Closed Access cities created a systematic high cost refugee crisis, then, because we blamed creditors for the housing problem instead of supply constraints, we pulled the rug out from under households that depended on credit access to fund reasonable housing consumption  (households that were going so far as to pick up and move cities, by the millions, in order to maintain reasonable housing consumption levels), and we continue to clamp down on the mortgage industry so that a decade later those households can't fund reasonable housing transactions and they are sitting on properties that are significantly undervalued with much less equity than they rightly should have.  And those that don't own homes have sharply rising rents because the housing shortage has been exacerbated.  We did it to them.  It's on us.  Do we have the honesty and integrity to fix it?  Can a country fix a decade long error that it has been this emotionally committed to?

1. Really superb blogging.

No, the USA will not "get it."

The housing bust story has become a narrative or set pieces on both the left and right wings.

The real story, of elitist property zoning in and around closed access cities, and a stupid contraction of housing finance, is not one Americans want to hear.

Americans embrace property zoning and tight money.

1. I meant to mention in the post, but forgot, that as much as I tend to focus on San Francsisco, the scale of the problem, just due to the size of the cities, is heavily weighted to LA and NYC. That scatterplot wouldn't change much if I took out San Francisco, but you would notice if I took out LA or NYC. It's amazing how tight the patterns are and what a bifurcation there is. Very little in-between.

2. "We did it to them."

Um, who's this "We" you speak of? The post would be much stronger if you used the third person to specify exactly who did this. Perhaps you personally were involved in this. I don't know. Probably some of your readers were too. But equally likely lots of your readers had nothing to do with this.

1. You make a good point. I am usually wary of that sort of rhetoric, too. And, it probably does turn off some readers, unnecessarily. On the other hand, in my defense, I would argue that few points of view have gathered such universal and vociferous support as the policies that led to the Great Recession and the death of lower middle class homeownership. The idea that there was a bubble, created by excessive lending, and that a collapse was unavoidable, exists above the realm of empirics. It is only mentioned as that thing which we all know, which does not require evidence - whether in a newspaper article, academic research, or public policy statements. The sentiment is so strong that, even after basically destroying the mortgage market, the Fed mostly has to defend itself against complaints that it has been too friendly to the banks. Is there a single person anywhere who has had the temerity to suggest in public that mortgage lenders were the main potential source of support for the economy since 2005? Can you imagine any room anywhere in this country with more than 5 people in it where that idea could be raised without being scoffed at? If there is any idea that can be attributed to "we", I think this might be it.

3. I am out of touch with is New York City. But I don't understand why they are restricting development, since they have a subway.

In Los Angeles everyone hates development as they connect it to impossible traffic, life-sucking traffic. Also, most people in Southern California do not believe that single-family detached neighborhoods should be bulldozed.

But Manhattan? Shouldn't they want super high-rises?

1. Yeah. It's crazy. They'll build commercial as dense as you want, so it's even worse than that. Building residential on the island would actually reduce the need for subways. Manhattan is full of buildings to work in but the workers all have to sleep in Queens and New Jersey. It's such a universal Western problem - from London to Sydney - though, that it must have its roots deeply in the modern sense of local political power and Western culture, in general - or at least Anglosphere culture. The American exceptionalism here is the fortunate fact that we still have Dallas and Atlanta and Phoenix, although Canada may have a similar dynamic.

4. So question.

America has progressive taxes. And while we can argue about the degree of progressiveness, the tax bill on an income of \$150K is way, way higher than the tax bill on an income of \$75K. Even as a percentage of income.

And your argument here is basically that rich people are moving into places where they can make \$150K instead of \$75K, getting dinky condos, and then going on cool vacations with the money from not buying a million-dollar house. While pushing out the people making \$35K and sending them to places where they only make \$25K, but the rent isn't completely stupid.

What happens when you add on taxes? Both higher CA rates, and also just generally that you're spending a lot of time hanging out in the federal 28% (or even 33%) brackets that you normally wouldn't be?

Is it still as much of a net gain?

1. Great point. And, I think that is part of what makes this issue so difficult to notice and difficult to address. I think we can uncontroversially say that the basic trend is true, because high income households are willfully moving into these cities, and other households are moving out, willfully or not. I'm getting to where I can't remember what I've posted on the blog and what I've put in the book, but something I may not have written about here is that in 2005, net in-migration, just from California, was more than 1% of the populations of the surrounding states.
But, there are two parts of that exchange - the top end includes a jump in very high incomes. I have written about the question of what is due to true productivity and what is due to limited competition. You mention here another factor - how much is kept, after taxes. This portion of the exchange, the top end, probably greatly overstates the added income inequality, to the point that it might not amount to that much, just as you say. Though, again, the migration pattern suggests that the net gains are significant.
At the bottom end, for those who remain in the closed access cities, income inequality is much worse than it looks, because they have much lower incomes after rent is accounted for. On the other hand, many of them can remain because they are taking large subsidies through rent control policies, so it probably isn't as large a factor as these market rates make it out to be.
For those who migrate, inequality is overstated by the gross numbers because their discretionary income after rent probably tends to rise, even as their gross incomes fall.
The whole thing comes down to Closed Access. It is very difficult to analyze an economy that is based on limited access to certain assets. Our models are all based on the assumption that patterns of consumption and production are sensitive to price signals, and these policies undermine those assumptions. The compositional effects overwhelm the normal factors that would normally give us clues about economic outcomes. I think there is very little we can say other than that the migration pattern is a strong indicator that this benefits the top end.
The only way to stop this pattern, which is what we did in 2006-2008, is to slow down economic growth itself, which reduces the opportunities for the high income in-migrants. Raising taxes on them would probably reduce the migration pressure through the same mechanism, by keeping productive workers away from matching with productive creative labor markets in those cities.

2. Don't forget the home mortgage interest tax deduction. Worth more in higher tax brackets.

3. Yes. My God our governance is such a tangle of targeted subsidies and taxes, it's impossible to conclude just about anything, isn't it?

4. FWIW, \$72K/year in the Bay 3 years (and \$1300/month in rent increases) ago was about the same as \$30K in Detroit. I paid \$22K in taxes, he paid \$4K in taxes, I had a new car and roommates, my father had a used car and his own 3BR condo, and I had to beg my mother to fly me home for Christmas.

Every dollar after that gets taxed at about 40-45% average between FICA, CA taxes, and just being in the 25/28/possibly 33% bracket for the Feds.

That's probably a decent assumption to start off on. What happens when you cut out 40% of the marginal income gains?

5. Interesting. Things like higher taxes, I think, will tend to slow down the in-migration a little, cause local wages at the top end of the distribution to rise faster, and hasten the inevitable arrival of the date where the economic rents become unsustainable and San Francisco follows Detroit. That's my guess. It seems impossible that San Francisco could be Detroit, but there was a time when it seemed impossible for Detroit to become what it has.

5. But I don't understand how a city like Manhattan Beach, a closed access city, can have a rise of 35 thousand dollars in income in 13 years but the actual house prices have risen by 400,000 dollars. That seems like there is speculation at work, not just income. From city data: Estimated median household income in 2013: \$135,822 (it was \$100,750 in 2000)
Manhattan Beach:

\$135,822
CA:

\$60,190

Estimated per capita income in 2013: \$80,327 (it was \$61,136 in 2000)

Manhattan Beach city income, earnings, and wages data

Estimated median house or condo value in 2013: over \$1,000,000 (it was \$669,800 in 2000)
Manhattan Beach:

over \$1,000,000
CA:

\$373,100

6. I don't understand how in Manhattan Beach, a closed access city, incomes could rise by 35 thousand dollars from 2000 to 2013 yet house prices have gone up from 600k to over a million dollars. I think that must also be because investors are competing with residents, right? This is from city data

Estimated median household income in 2013: \$135,822 (it was \$100,750 in 2000)
Manhattan Beach:

\$135,822
CA:

\$60,190

Estimated per capita income in 2013: \$80,327 (it was \$61,136 in 2000)

Manhattan Beach city income, earnings, and wages data

Estimated median house or condo value in 2013: over \$1,000,000 (it was \$669,800 in 2000)
Manhattan Beach:

over \$1,000,000
CA:

\$373,100

1. Price/Income went from 6.7 to around 7.5 or 8? That's not that much of a difference. A little expected persistent rent inflation can explain that.

7. Kevin and Kevin:

I often tell people I think living standards were higher in SoCal in the 1960s than today. Oh sure, the medical care and technology is amazing.

But the cost of housing and taxes (state, local, federal and FICA) means people have less after-tax, after-housing income.

The two-income family is the norm, not the Dad works family. Young people live in clumps in apartments, unable to start families.

This decline in American living standards seems the usual in all the popular, closed access cities Kevin Erdmann has identified.

Even with good government, some aspects of the SoCal lifestyle would be doomed, such as low-density development near the water. In free markets, the density would be highest by the water, bulldozing GOP-Newport Beach and Donk-Santa Monica.

The days of being a beach bum in Venice or Hermosa Beach or down in OC would be over in any even good scenario, and middle class families who wanted to live near the beach would have to buy a condo---but at an affordable price.

The world discovered SoCal and the perfect weather incredible environs---beaches, mountains, winter deserts.

The only way to "help" the middle class and poor would be to keep building housing. And cut taxes.

1. I've been looking a little more finely at zip code income growth, by city, and I think you're right. Real income growth at the bottom quintile, from 1998 to 2006, is close to zero in all cities. It's over 15% for the top quintile in the closed access cities and about 10% for the top quintile in the other 17 cities I am using. That is using city-specific CPI deflators. But, as the incomes after rent graph above shows, rent inflation in the closed access cities is much harsher on low income households (unless they have rent control). That, plus the fact that low income migration out of the closed access cities probably causes a compositional change that causes zip code average incomes to rise faster than incomes for individuals who don't migrate, probably means that the low quintiles in the closed access cities have had very negative real income growth since the mid-1990s - maybe double digit. Unless, of course, they have a rent controlled unit that has protected them for the meantime.

8. Kevin--

Are these median household income stats?

One of my point is that the two income household has supplanted the one-income household. So sure, some income gains--but egads, if two bedraggled adults coming home at 7:30 pm after an hour in traffic to the kids is a "higher living standard"....

1. These are reported adjusted gross income to IRS. Married filing jointly should show up as a single unit, I think. I don't have data on the number of income earners, by MSA, but your intuition seems right.

9. Kevin, In your first graph, it looks as if the fat part of the distribution of after-rent AGI for closed-access cities is above, not below, the median. Am I looking at it wrong?

1. Sorry David. My verbiage may have been a little sloppy. I meant, relative to the other cities.

2. Ah, now I get it. So it shifts it more. BTW, my daughter, living in SF on a modest income, is dealing with this crap. She made an astute observation about it the other day, which I may put on a blog post.