Housing has several distinct problems with different causes and solutions.1. Homes Could Be Larger or Nicer.This problem is caused, chronically, by low productivity in construction, and recently more acutely by the 2008 mortgage crackdown. Fixing this would make us all marginally better off in ways that aren’t earth-shattering, but aren’t nothing, either.2. Homes Could Be in Better Locations with More Local Amenities.Every metropolitan area in the country has arbitrary limits on urban development that block the realization of locational amenities created by city-building. As with the first point, these would include marginal improvements in lifestyle and quality of life.Some of those improvements could be significant in terms of equitable human improvement, because cities are mostly valuable for poorer households. I think there is a bit of an odd disconnect here. Several cities around the country have been reforming urban land use rules. I was recently in St. Petersburg, which appears to have relatively recently created a residentially active downtown with a lot of condos in the million dollar range. And that is fantastic. There was value in the location waiting to be tapped. Why leave it untapped?The problem is that we interrupted the process of city-building for a century. There would have been a complex path of development, depreciation, distribution of families across neighborhoods according to needs and incomes, and, as a result of that, neighborhoods with all sorts of values, ages, and structural characteristics. There would be relatively affordable, dense neighborhoods in the mix, in some cases, because high-tier new neighborhoods from 60 years ago had depreciated. Since we don’t have that complex array of interconnected neighborhoods, there are million dollar condos that share the space with a homeless population, surrounded by single-family neighborhoods.This leads to complaints about gentrification, as if development has caused this condition rather than a century of obstructed development.I’m not going to pretend here that I have all the answers, or even a fraction of them, to manage the complicated task of city-building. But, I think all those complicated answers would need to be built on the realization that the most important cost of freezing the structural form of our 20th century cities in place wasn’t just the mathematical loss of potential locational value. It has been a deep interruption in equitable quality of life improvements because cities couldn’t serve their purpose to their most important potential constituents - households with below average incomes.3. More Homes Could Be Built in Some Cities With Regional Amenities.New York City traditionally provided value for the world’s aspirational poor. Even today, in spite of the impressions given by the “superstar city” literature, many of the denser parts of New York City are notable for housing households with relatively low incomes, who live there in spite of the expense because there are amenities in New York City for households with lower incomes that other, more recently developed, metropolitan areas have made sure not to develop.As Figure 1 shows, there are 2 characteristics that make New York different from Los Angeles: (1) A lot of very dense neighborhoods with low incomes, and (2) a lot of very not dense neighborhoods with high-ish incomes. In most cities that have grown in the post-zoning era, there is little or no relationship between density and income. Where we have actual cities because they pre-date zoning, density and income are negatively correlated because most urban amenities are inferior goods.4. Homes Could Be More Affordable.Even in cities with regional amenities and agglomeration value, most of the affordability problem is related to an acute condition of a shortage, in the technical sense of the word. And, since 2008, ironically triggered by the mortgage crackdown in point (1) that is directly related to the most benign aspect of the housing crisis, the shortage is nationwide. When, either through zoning or through lending obstructions, supply conditions become so poor that many families perennially have to make hard decisions about whether to trade down or not, families resist that. And, under those acute conditions, the families choose to pay more to avoid trading down or being displaced. Only under this acute condition does nominal spending on housing increase to uncomfortable levels.In cities like New York City and Los Angeles, this is a permanent condition created by local zoning rules. Until those rules change, large numbers of households will have to be displaced each year, and many households are willing to pay a lot to avoid being one of those families. The effects on rents and prices of that process are much larger than the effects of regional amenities for high-income newcomers and local amenities for low-income locals.In Figure 2, the dashed lines represent Phoenix and the solid lines represent New York City. Using price/income ratios within and between Phoenix and New York City to estimate the scale of the price effect, (1) is the scale of regional amenities (rich NYC vs. rich Phoenix), (2) is the scale of cyclical inflation (2005 rich vs. 1999 or 2022 rich), (3) is the scale of local amenities (dense NYC poor vs. non-dense NYC poor), and (4) is the scale of perpetual shortage (non-dense poor vs. non-dense rich).(These don’t correspond to the 4 points above. They are 2 separate lists, only partially related.)By 2022, the resistance to displacement from a perennial shortage was pushing up prices in poor neighborhoods in both Phoenix and New York City. That was never a factor in Phoenix before 2008. And, in Phoenix, as long as we don’t prohibit single-family rental housing in order to finally block the adequate creation of every single form of housing, the condition will be temporary. If nothing else, rental homes will be constructed in the exurbs, and slowly, pent up household formation will commence and interrupted migration to Phoenix will return.In the meantime, Americans, on average, will be spending an uncomfortable portion of our incomes on housing.Will More Homes Increase the Aggregate Value of Residential Real Estate?No.Oddly, some housing advocates for more dense urban housing and some supply skeptics agree to be wrong about this. And, oddly, they both seem to be making the same error - underestimating the cost of the shortage. The skeptics have talked themselves into the ridiculous idea that housing expenditures aren’t that different than they have been historically. That’s just ahistorical bullshit. We all came together with our pitchforks out in 2008 to demand the most disruptive financial crisis in generations because everybody knew that the Case-Shiller index had suddenly gone haywire because of a lending bubble. Everyone was wrong about that. A lending boom, which was a subset of the factors that created the cyclical inflation, was part of note number 2 in Figure 2. The Case-Shiller was out of whack because of note number 4 (and the cyclical impulse in cities like Phoenix because of so many families giving up on note number 4 in Los Angeles and accepting regional displacement to places like Phoenix where new homes could be permitted).Discovering how important that error was has been the raison d’etre of my entire project. But, for a bunch of sophists to come along these days, after everyone went through that, and pronounce, “aktchuallly, there’s nothing much going on here. Cities have always been expensive” is chutzpah I can’t abide. This isn’t the introduction to a debate. This is an admission that debate will be useless.At the other end of the spectrum, some housing advocates understand that when development in cities is related to higher valuations, that’s a reflection of progress and value. But, they underestimate the scale of the shortage premium, and so they have stretched out gap number 1 in Figure 2 to explain gap 4.That blind spot leaves them neutered to respond to the sophists because they accept the sophists’ anti-empirical claim - that building more will increase aggregate real estate values.Larger, more dense cities are related to higher housing costs, higher incomes, and even more spending relative to those incomes. The point is true on all counts. It’s just not the only thing that is true. Thirty years ago, to a rough approximation, it was the only thing that was true. Today, it’s a small fraction of what is true. I hope that Figure 2 can help to think through that.Figure 14 is from my recent paper, “We are not as wealthy as we thought we were.” The green line is an estimate of aggregate home prices/incomes. The dark line is the portion of home prices that pays for the construction of the buildings, amenities related to location or region, and cyclical fluctuations. You can see the big cyclical bump before 2008 - it really was a large cycle (although some of the “cyclical” impulse was the displacement of families out of the Northeast and coastal California, which created a demand boom in Arizona and Florida).To an extent, after adjusting for cycles that are mean-reverting, the dark line is very stable under a number of circumstances at near 3x. When families purchase locational or regional amenities, they mostly trade off structural value. The choice they make isn’t between a 3,000 square foot home in Naperville and a 3,000 square foot condo in the Loop. It’s between the 3,000 square foot house and a 1,000 square foot condo. On net, families spend a little more in some contexts when they trade off structure and space for public and locational amenities, but it is mostly traded off against other factors.Of the 3x income American families would be paying for housing under non-shortage conditions, maybe 0.5x of that could be described as some sort of locational amenity - either luxury condos in downtown St. Petersburg, or row houses in the Bronx near a subway line. If American cities make more locational value available, and we double that, total real estate value won’t rise to 3.5x incomes. Maybe it will rise to 3.1-3.2x. This is why the Case-Shiller index was so flat for a century. Families adjust on the margin to keep nominal housing expenditures level.The yellow line is stacked on top of the dark line. It is a measure of the recent phenomenon of a scarcity premium - note number 4, above (in both lists). Supply conditions (first in the coastal metropolitan areas, then across the country) that are so bad that families pay ransom rents in order to keep the endowments of location - the personal reasons they refuse to move in a housing market that insists on making them feel poorer until some of them do.The yellow line is actually an understatement of that problem, because it is the net of two factors. (1) Housing inflation that is due to the supply problem that has basically doubled the value of the average urban American home and (2) a discount of about half that scale because families that would choose to buy those homes rather than rent them are not able to convince Uncle Sam to let them.My estimate of the shortage is unusually high. I put it at north of 15 million units, which amounts to about 10% of the existing housing stock.The key to understanding the scale of all these factors is recognizing the peculiarity of the source of elevated home values in Figure 14. The difference between the yellow line and the dark line is from the right panel of Figure 2. Mathematically, it is the estimate of the total value across cities of the land rents households have to pay for the privilege of living in the poorest neighborhoods - the peculiar development that the dark red line in Figure 2 for every city across the country has moved higher. It is from poor families paying more than rich families for housing for the right to live tomorrow where they lived yesterday. It is more expensive to be poor when there aren’t enough homes.Let’s say we dropped all zoning rules everywhere and 15 million new homes were constructed in exactly the places they should be, and many of those new homes unlocked new locational value. The first-order effect of that would be a return to the left panel of Figure 2, when families were generally able to live where they wanted to live, which in most cases was where they were already. A second order effect would be that families would more easily be able to move to where they want to live.My contention is that a free-for-all would lead to 15 million extra units. Regardless of where the free-for-all would place those new homes, it would return us from the right panel of Figure 2 back to the left panel.That would involve increasing the stock of homes by about 10%. If every single one of those 15 million homes landed in the closed access cities (NYC, LA, Boston, San Francisco, and San Diego) it would roughly double the sizes of those cities.When Los Angeles doubled in population from 1950 to 1980, did housing get a lot more expensive? How about when New York City doubled in population from 1900 to 1930? What did the Case-Shiller chart look like?Meanwhile, in that extreme context, the rest of the country would definitely be back to 1999 affordability, after losing more than 10 million households.What if the 15 million units were more evenly distributed? Gap #4 goes away. Is 10% population growth really going to make up the difference in amenity value? There are only 2 ways to make the scale plausible. Either my estimate of the 15 million units required to get rid of gap #4 is a huge underestimate, or there is some other explanation for gap #4. What reason, other than the supply crisis, explains why poor families in non-dense, low amenity neighborhoods are living in homes worth 8x their incomes in Phoenix and 12x their incomes in New York City? I don’t think either of those explanations can be plausible.The thing is, there is no existing literature (that I am aware of) that even noticed that pattern. That’s understandable, because it’s a new phenomenon. (The “super city” amenity value camp might point to some Rosen-Roback friendly story about how housing for low-income families can become expensive because of option value or because it is near to places that have increased amenities. This is clearly not the cause of the systematic and large scale of gap #4, and it shouldn’t take much attention to the matter to realize that. Rents are higher now in poor neighborhoods. Not because of some mysterious 3rd thing that affects some of them. The straight correlation between income and price/income ratios is very regular and quite strong.) It’s the year 2026, and it’s past time to understand it. I hope the descriptions and figures above are helpful for you to understand it.Original Post