New Research on Gentrification

(cross-posted from

This month, the Federal Reserve Bank of Philadelphia came out with a study on “Gentrification and Residential Mobility in Philadelphia.” This paper included some findings about displacement (or the lack thereof) in gentrifying neighborhoods.

The study described 56 Philadelphia census tracts as “gentrifying” based on a variety of criteria. Some of the questions the study sought to answer were:

  • Are poorer residents of these tracts more likely to leave these neighborhoods than other city neighborhoods? Persons with the lowest credit scores (the study’s surrogate for poverty) were about equally likely to leave gentrifying and non-gentrifying tracts. So if moving equals gentrification-related displacement, it does not seem especially frequent in gentrifying areas. Similarly, longtime residents (those living in the tract for over five years) were actually less likely to leave gentrifying tracts than other city neighborhoods.
  • Were low-credit-score movers from gentrifying tracts forced into suburbia? A map of movers showed that high-credit-score movers were scattered throughout the region, while almost all low-credit score movers stayed within the city of Philadelphia. So it appears that despite media coverage of suburban poverty, few poor people moved from gentrifying areas to suburbs. (Of course, this pattern may be less true in high-cost central cities like New York and Washington.)
  • Did low-scoring movers from gentrifying tracts shift to areas poorer than those gentrifying tracts? The study’s answer to that question: No. The average low-score mover moved from a census tract with a median household income of $33,000 to one with a median income of $44,000. When intercity movers are excluded, this gap is narrower; even so, the movers wound up in areas with about a 10 percent higher median income than that of their former home (from $32,000 to $35,600). On the other hand, the “receiving” neighborhoods had lower home values than the gentrifying tracts, and fairly similar crime rates, and movers from some neighborhoods were less likely to move to richer areas. (Postscript: I also should have noted that intracity low-credit-score movers from non-gentrifying tracts improved their neighborhood status to a greater extent than those from gentrifying tracts; the average mover in this category left an area with a median income of just over $26,000 and moved to an area with a median income of almost $32,000- a 20 percent increase but still a lower median income, I note, than the movers from gentrifying tracts.  This point seems to have created some confusion among readers: because movers from gentrifying tracts improved neighborhood quality to a lesser degree than movers from nongentrifying tracts, readers may have believed that movers from gentrifying tracts were objectively worse off than they had been had they not moved.)

In sum, it appears that at least in Philadelphia, many gentrification-related stereotypes are not consistently true. Low-credit score residents of gentrifying tracts were not especially mobile, nor did they always move to poorer areas when they did move.

Supply, Demand and Housing Prices, Part 2

(cross-posted from

Throughout the United States, rents (and in some cases housing prices) have been skyrocketing over the past several years. Under the economic law of supply and demand, one* obvious answer to this crisis is: expand supply (by which I mean, eliminate zoning regulations that prevent landowners from building more apartments.) The latter argument, of course, has met a lot of resistance from a variety of quarters. Some months ago, I critiqued what I called “supply and demand denialism.

But since then, I’ve heard a variety of interesting new arguments on this point, which I would like to address in this blog post.

Argument 1: “Where land costs are high, housing costs will always be expensive.” I am not persuaded by this argument, for two reasons. First, the cost of land is not the same as the cost of housing. For example, suppose that the cost of land costs $5 million per acre. Does that mean that no housing unit can cost less than $5 million? Yes, if the government only allows one housing unit to be built. But if government allows ten housing units to be built, the landowner can sell each for $500,000 and break even. And if the government allows 100 housing units to be built, the landowner can sell each for $50,000 and break even. So if people were allowed to build more housing units, the price of housing would come down—high land costs or no high land costs.

Second, as long as some land is zoned for uses other than housing, the supply of land is not static, and, thus, the cost of land is not static. Government could zone additional land for mixed uses (or just for housing), effectively creating more land.

Argument 2: “Greedy developers build only for rich people.”** But this argument begs the question: why is it that developers build only for rich people in New York but build for middle-class people in Pittsburgh or Indianapolis? Are the developers any less greedy there? Of course not! High housing costs mean that developers can get more money for their housing in New York. So this argument creates a vicious circle: government regulation causes a housing shortage, causing high housing costs, causing developers to charge more for new housing, causing citizens to use this very fact as a reason not to allow new housing, causing the housing shortage to get even worse.

More importantly, most housing units aren’t new units. When I lived in Manhattan, I lived in a gentrifying zip code (10018) with high rents and newer housing than many other neighborhoods. But even there, my last search found that only 9 of 47 available rentals were built before 2000. So rather than asking “why are new units more expensive?”, perhaps we should be asking “why are older units so expensive?”

It seems to me that when the number of new units is restricted, only the most affluent renters can afford them. This in turn means that the upper middle class and the “ordinary rich” are then forced into the older unit market. They bid up the price of older units, forcing up the price of older units. By contrast, if more new units were permitted, the “ordinary rich” and the upper middle class could afford them, making older units more affordable for everyone else. To put it another way, if there were more newer units on the market, the aging walk-ups that now cost $2000 a month might cost, say, $1500 or $1000 per month or even less.

A variation of this argument is that “greedy developers only build condos.” I’m not sure I find this persuasive because condominia and apartments aren’t completely separate markets. Condos can sometimes be rented, and condos and rental units compete with each other because if condos are abundant enough and cheap enough, people who would otherwise rent buy condos (thus causing less price pressure on rentals, causing rents to become cheaper). So I’m not sure whether this argument changes the underlying economic rules.

A third argument that I mentioned, but did not address as fully as I could have, is the claim that new housing raises housing costs by making a neighborhood more desirable. This argument assumes that gentrification follows new housing, rather than vice versa.  Although more research on this issue would be useful, it seems to me that the reverse is often the case: historic neighborhoods throughout the country have been gentrifying over the last decade or two,*** and new housing follows the gentrification when zoning permits it. For example, in Greenpoint (one of Brooklyn’s most rapidly gentrifying areas) about half of all housing was built before 1950, while newer areas on Brooklyn’s outskirts are not doing so well.

In sum, it seems to me that, in a world with less zoning, fewer density regulations, and more building, rents would eventually stop exploding—and that neither land costs nor the existence of rich people necessarily prevents that.

*But not the only one. Another alternative is to build huge amounts of public housing- perhaps not a bad idea. But given the dismal reputation of public housing, the unusually high tax levels of some of the most expensive cities, and the high level of taxophobia in the United States, I cannot imagine this happening anytime soon.

**Not dissimilar to “greedy developers only build for rich foreigners,” critiquedhere.

***See generally William Lucy’s book “Tomorrow’s CIties, Tomorrow’s Suburbs” (pre-1940 neighborhoods experiencing greater income gains than postwar suburbs).

The Economics of Rent Control

Economists oppose rent control almost as unanimously as climate scientists oppose attempts to deny the reality of client change. 93 percent of economists (including liberals like Paul Krugman) agree that ceilings on rents reduce the quantity and quality of housing. And economists’ views seem borne out by experience: most major cities with rent control (such as New York and San Francisco) tend to have very high rents, indicating that rent control is perhaps not functioning as effectively as its creators had wished. And yet urban planners and citizen activists are much more closely divided. What do they not understand?

To understand the risks of rent control, imagine the strictest possible policy. Government freezes rents forever, prohibiting any landlord from increasing rents. As the costs of labor and utilities increase, landlords realize they are losing money by continuing to be landlords. So eventually landlords start to abandon property; if they cannot legally do so by converting apartments to condominiums, they do so illegally by letting property decay until a suspicious fire wipes out the landlords’ problem. No new apartments replace these housing units, because no one will invest in housing if they can’t make a profit by doing so. Eventually, renter-dominated areas turn into wastelands, like the South Bronx in the 1970s.

Of course, no city has ever enacted such a strict policy. A more plausible form of rent control treats landlords as public utilities: government could limit rent increases rather than freezing rents, by providing that a landlord could increase rents by only a certain amount this year (say, the costs of increased landlord expenses). In this scenario, landlords will not abandon property as rapidly as in the first scenario. Nevertheless, this policy will still cap housing supply; even if existing landlords are less likely to jump ship, new landlords will be unlikely to enter the market. Here’s why: rental housing competes with other, non-regulated investments. That is to say: if I have money that I could invest in (a) a business where my profits were capped by government (e.g. rental housing in this scenario) or (b) a business where government wants people to make huge profits (e.g., the stock market, single-family housing), I would, other things being equal, crazy to invest in (a). In this scenario, rent control creates housing shortages, much as Soviet centralized planning created shortages of consumer goods.

Having said that, in reality rent control is not quite as toxic as in my hypothetical scenarios. In the United States, most cities with rent control do not regulate quite as intensively as in my hypotheticals; for example, in New York only older unitsare covered. Such “moderate rent control” does not reduce housing supply as much as would universal rent control, because it only caps rents in a limited slice of the market. On the other hand, this means that “moderate rent control” also doesn’t do much very much good—rents in New York are in fact sky-high for most people other than longtime residents. In other words, moderate rent control doesn’t control rents very much. So what’s the point?

Defenders of rent control point out that the correlation between rent control and housing supply is incomplete; for example, places that have removed rent control (such as Boston) have not always experienced construction booms, while highly regulated New York built lots of new housing in the 1950s. But this is to be expected from the above discussion of moderate rent control: if rent control is too weak to actually reduce rents, it is also going to be too weak to restrain supply as much as a policy that did reduce rents.

And where rent control is relatively weak, it is just one factor affecting the supply of new housing: for example, zoning laws generally treat apartments as an undesirable use, and thus may restrain housing supply far more aggressively than moderate rent control. Conversely, if government allows lots of new housing construction, new housing will be built as long as rent control is not so strict as to prevent it. Similarly, if government subsidizes new housing or builds new housing itself, the positive results of such construction may be more important than the effects of moderate rent control.

In sum, the difference between strict rent control and moderate rent control is like the difference between smoking two packs of cigarettes a day and smoking two cigarettes a day. Smoking two packs of cigarettes a day, like strict rent control, is likely to create toxic results. Smoking two cigarettes a day, like moderate rent control, is far less harmful though still not ideal.

What HUD’s Been Up To

by Michael Lewyn

There has been some controversy about the federal government’s new “Affirmatively Furthering Fair Housing” (AFFH) rule. Supporters hope, and opponents fear, that the rule will integrate lily-white suburbs and eliminate exclusionary zoning. However, there is reason to believe that the rule’s impacts will be fairly minor.

The Department of Housing and Urban Development (HUD) which enacted the rule, did so in order to implement sec. 808(d) of the Fair Housing Act, which requires federal agencies “to administer their programs…relating to housing and urban development…in a manner affirmatively to further” the policies of the Act—in other words, to affirmatively further fair housing.

In the past, HUD has sought to implement this statute by requiring grant recipients (such as local governments and public housing agencies) to draft an analysis of impediments (AI) to fair housing. An AI typically described impediments to racial integration, such as exclusionary zoning and racial disparities in mortgage lending. HUD decided that the AIs were not tremendously successful, because they did not contain enough data and were not adequately linked to other planning documents. (80 Fed. Reg. 42348).

The new rule requires grantees to create a new document called the “Assessment of Fair Housing” (AFH) every five years. The AFH will address a community’s barriers to integrated housing, such as “integration and segregation; racially or ethnically concentrated areas of poverty; disparities in access to opportunity, and disproportionate housing needs based on race, color [and other factors]” (80 Fed. Reg. 42355). The AFH will summarize any current litigation, analyze relevant data, and identify major factors limiting housing opportunity. The grant recipient must also set goals for overcoming the effects of these factors.To receive funding from HUD, a grantee must certify that it will affirmatively further fair housing, which means that it must promise to take meaningful actions to further these goals.  (80 Fed. Reg. 42316).  In other words, the grant recipient has to create paperwork stating: “This is why our city/county/area is more segregated than we would like, and this is what we would like to do about it.” Continue reading

Cities, The Middle Class, and Children

by Michael Lewyn

In a recent article, Joel Kotkin critiques the work of Jane Jacobs; he points out that Jacobs idealized middle-class city neighborhoods, and suggests that because cities have become dominated by childless rich people, middle-class urbanity “has passed into myth, and… it is never going to come back.” He suggests that Americans are “moving out to the suburbs as they enter their 30s and start families” because central cities are only appropriate for “the talented, the young, and childless affluent adults.” This claim rests on a couple of assumptions: 1) that cities have little appeal to families and 2) that the only Americans whose preferences are typical are those middle-class families.

The first claim has an element of truth: families do tend to prefer more suburban living environments. But what Kotkin overlooks is that the tide is turning (at least a little). Although American suburbs clearly have more children than cities, the most desirable city neighborhoods are more appealing to parents than was the case a decade ago.

For example, Kotkin writes that Greenwich Village (where Jacobs lived) “today now largely consists of students, wealthy people and pensioners.” But according to the Furman Center’s neighborhoood-by-neighborhood surveys of New York housing, the percentage of households with children actually increased in New York’s more desirable urban neighborhoods. For example, in Jacobs’s own Greenwich Village, 15.1 percent of all 2013 households had children under 18—lower than in most places to be sure, but higher than in 2000, when only 11.4 percent had children. Similarly, the “households with children” percentage increased from 11.4 percent to 15.1 percent in New York’s financial district, from 14.6 percent to 17.8 percent in the Upper West Side, and from 13.3 percent to 16.6 percent in the Upper East Side. Continue reading

Walkability No Guarantee of Healthy Housing Market

This graph shows how the midsized cities (excluding Arlington) with Top 10 walkability rankings score in WalletHub’s latest ranking of cities with the healthiest real estate markets. Sad to say: High walkability seems to be correlated with moribund real estate economies. The cities are (from left to right): Jersey City, Newark, Hialeah, Buffalo, Rochester, St. Paul, Cincinnati, Richmond and Madison. (Click for more legible image.)

This graph shows how the midsized cities (excluding Arlington) with Top 10 walkability rankings score in WalletHub’s latest ranking of cities with the healthiest real estate markets. Sad to say: High walkability seems to be correlated with moribund real estate economies. The cities are (from left to right): Jersey City, Newark, Hialeah, Buffalo, Rochester, St. Paul, Cincinnati, Richmond and Madison. (Click for more legible image.)

There is an interesting juxtaposition of news items today. Redfin, the real estate brokerage website, has published a list of the Top 10 most walkable midsized cities in the country. Arlington County (a highly urbanized county) scored third and Richmond scored ninth, based on their Walk Score rankings.

Arlington won kudos for its Ballston-Virginia square neighborhood, where residents can walk to an average of 13 restaurant, bars or coffee shops within five minutes. While the Washington metropolitan area is notorious for its traffic, many Arlington residents live car-free, opting to get around on foot, bike and public transportation.

Richmond earned recognition for the revitalization of neighborhoods surrounding downtown, including Jackson Ward, Shockoe Bottom, Monroe Ward, the riverfront and Manchester. The Fan and Carytown neighborhoods to the west of downtown also stood out for their walkability.

To many urban theorists, walkability is a critical determinant of a community’s livability, ranking close behind the cost of real estate, the quality of schools and the level of taxes in what people take into account when deciding where to live. But it’s no guarantee of prosperity or rising real estate values…. which brings us to the other news item. Continue reading

Kotkin and the “Assault on Suburbia”

by Michael Lewyn

A recent article by Joel Kotkin tries to stir up a stew of resentment about alleged “attacks on suburbia”.  Kotkin’s article is in black; my comments to the article are in gray.


BY JOEL KOTKIN – July 10, 2015

The next culture war will not be about issues like gay marriage or abortion, but about something more fundamental: how Americans choose to live. In the crosshairs now will not be just recalcitrant Christians or crazed billionaire racists, but the vast majority of Americans who either live in suburban-style housing or aspire to do so in the future. Roughly FOUR IN FIVE HOME BUYERS prefer a single-family home, but much of the political class increasingly wants them to live differently.

Theoretically, the suburbs should be the dominant politically force in America. Some 44 million Americans live in the core cities of America’s 51 major metropolitan areas, while nearly 122 million Americans live in the suburbs. In other words, NEARLY THREE-QUARTERS of metropolitan Americans live in suburbs.


Yet it has been decided, mostly by self-described progressives, that suburban living is too unecological, not mention too uncool, and even too white for their future America. DENSITY is their new holy grail, for both the world and the U.S. Across the country efforts are now being mounted—through HUD, the EPA, and scores of local agencies—to impede suburban home-building, or to raise its cost. Notably in coastal California, but other places, too, suburban housing is increasingly relegated to the affluent.