Brian Kelsey provides the provocation for this post, "78704 population didn't change much 2000-2012 but % w/ bachelor's degree or higher increased 37% to 50%." Translating character-restricted Twitter lingo, "78704" is an Austin, Texas, zip code. For that part of the metro, residents didn't become more numerous so much as they became more educated. Brian was discussing housing affordability in Austin (hence the hash tag "#atxaffordability"). Real estate appreciates without population gains or more restrictive regulations on building.
If population growth isn't driving demand for housing in 78704, then will greater supply drive down prices? That's not the matter of debate, but it should be. I can subscribe to basic supply and demand theory while questioning the efficacy of increasing housing units as a solution to the affordability crisis. Define demand.
Most conversations about housing affordability define demand in terms of population change. No one defines housing demand in terms of income, save the scholars worried about gentrification. Regarding gentrification, higher incomes displace lower incomes. The population in a neighborhood could go down. Tenured residents are still forced out of their homes. Vacancy rates can increase.
I wouldn't put much stock in population numbers. I learned my lesson from studying Rust Belt population decline, particularly the recent good fortune of the shrinking city of Pittsburgh. You'll find a similar story in St. Louis:
American families have changed dramatically over the past half century. The average household size in St. Louis in 1950 was 3.1 and in 2010, 2.2. With every other factor held constant, the decrease in city population would have been 248K or 29%. This means that with the same number of homes, the same number of apartments, and the same number of families as resided in the city in 1950, the decrease in average household size could account for 46% of the city’s population loss.
I do notice that a shrinking household does not account for all of the population decline in St. Louis. Still, a drop of almost 250,000 people would constitute a substantial dent in demand. And yet, the same number of houses would be occupied. Population decreased by a quarter of a million residents and demand for housing remained constant. What a lousy way to measure demand.
Instead of population, I recommend using income to measure demand. The income variable has warts, too. It's still better than using population change. A household earning $60,000/year eyes another household in a different neighborhood making $30,000/year. The wealthier household is currently paying 50 percent of its income on shelter to be near the city center. This household could pay a lot more than the poorer household for the same home a bit further from the city center and still have more disposable income. Demand goes up without any population change.
Just because I have redefined demand in terms of income doesn't mean that increased housing supply can't help balance the real estate market. The policy geography for increasing supply is different. Birth of a gentrifier:
In the opposite scenario where housing and office development remains static, the resulting paucity of workers to meet labour demand leads to an increasingly tighter labour market, sparking a bidding war, which in turn leads to wages spiralling higher, which finally leads to still higher housing costs and eventually an out-migration of precisely the kinds of workers that are needed.
I don't expect the feedback loop described would price such highly skilled talent out of the regional labor market. But it could push workers into other parts of the city. Hello gentrification. Instead of building more in the destination neighborhood, the target would be the source (of the gentrifier) neighborhood. Note how greater supply is addressing the upward pressure on wages, not greater numbers of people piling into the same place. This is the Financial Times talking and the prescription for the problem is to liberalize land use and building regulations. The author believes in supply and demand theory while linking higher wages with higher housing prices. Build baby build.
I wish the solution were so simple. Let migration theory explain. Most people, when they move, move short distances. Long-distance relocation is rare. Why? Long-distance relocation is scary. You leave your social network behind. The new place is, well, new. It is filled with known unknowns and unknown unknowns. Thus, the new-to-town cluster in the same neighborhoods where they are less likely to encounter townies who hate them. The new-to-town make more money—much more—than the average household in the area. Gateway neighborhoods for outsiders tend to sport some of the most expensive real estate in the metro. Meanwhile, locals moving around the region (most migration) are seeking some sort of return on their intimacy with place, "Can I have it all at half the price?" Eventually, newcomers transform into tenured residents and join the hunt. And the erstwhile newcomers enjoy a greater return for their labor. They ruin everything. They ruined Austin, Texas.
Places such as Austin are importing higher labor costs which will eventually diffuse. Places such as San Francisco not only import higher labor costs, they also offer attractive investment opportunities for the wealthy living and making money in other countries. Effectively, San Francisco's real estate market imports higher wages without offering such jobs locally. The price for shelter has long been disconnected from local population change. Increasingly, the price for shelter is disconnected from local wages. Supply-side solutions to the housing affordability problem have a woefully anachronistic definition of demand.