Two new papers on housing restrictions are noteworthy, Housing Constraints and Spatial Misallocation by Chang-Tai Hsieh and Enrico Moretti, and The Economic Implications of Housing Supply by Ed Glaeser and Joe Gyourko.
Readers of this blog will not be surprised at the idea that zoning and other restrictions drive up the cost of housing, and that this has many bad consequences on economic growth and inequality. The papers are especially noteworthy for much deeper implications.
Hsieh and Moretti:
...high productivity cities like New York and the San Francisco Bay Area have adopted stringent re- strictions to new housing supply, effectively limiting the number of workers who have access to such high productivity. Using a spatial equilibrium model and data from 220 metropolitan areas we find that these constraints lowered aggregate US growth by more than 50% from 1964 to 2009.1) The costs of regulation. The biggest problem in economics right now (yes, I mean that) is, How do we measure the growth consequences of regulation? Looking at the Western world's sclerotically slow growth rate, and listening to many anecdotes, it seems at least plausible that productive innovation is being strangled by byzantine bureacracy, captured by rent-seeking and anti-competitive forces. (Your other choices are, we just ran out of ideas, or some sort of endless "lack of demand.")
But how do we move past anecdote? How to we come up with "regulation is costing the economy x percentage points of growth?" Our statistical measurement system, GDP, unemployment, inflation, and so on, was beautifully designed in the 1940s to measure very Keynesian demand concepts. It isn't designed to answer the question of our time, how much growth is regulation costing us? We are flying in the dark. And Europe, perpetually in an Augustinian moment -- Lord, grant me structural reform, just not yet--is also.
Well, Hsieh and Moretti are doing it, and by doing so showing one path to answering the larger question.
Half of all US growth for a half century is an astounding amount. 1964: $3,734 trillion; 2009: $14,419 Trillion. Growth = 3.05% per year. At 6.1% per year, $3734 x (1.061)^(2009-1964)=$53.6 trillion dollars!
OK, maybe that's too huge. Well, read the paper and see how they came up with the number. If you don't like their assumptions make different ones. More important than this number is how they are coming up with answers to this, the most important question of economics.
2) Models and micro vs. macro
So how do they make the calculation? Roughly, they measure productivity in cities. They assume that people get higher wages in San Francisco because there are some very high productivity activities that have to be done here. They assume that business could expand and form here, and workers could move here and join in those high productivity activities, both earning higher wages and making more and better stuff for the rest of us. But those workers can't move, and businesses can't expand and form, because housing supply is restricted.
You can see grounds for objection.
Housing costs are high in places like Carmel (I think) because retired rich people like to live there, together with restricted supply -- the amenity theory of housing cost. Workers moving there would not earn a lot more. Maybe people live in tiny apartments in San Francisco because they prefer it, so restricted housing supply is not restricting productive activity. Well come up with your own model, and check it out in the micro data as Hsieh and Moretti do. Every model has assumptions. To calculate a counterfactual -- how much would people earn and businesses make if they could move to San Francisco -- you have to do that. It takes a better model to beat a model.
But there is a deep lesson in their style of modeling: Heterogeneity. Misallocation. Dispersion. Inequality. The key lesson is not that "regulation is killing US firms on average." The US as a whole is doing badly because firms are in the wrong place -- misallocation. Each individual firm may feel it's doing fine. It might consider moving to San Francisco, but say "well, we might be more productive there, but wages are much higher because you have to pay people enough to buy a house, so we wouldn't make any more money if we were there." More to the point, a new business that would embody the higher productivity, get workers to move, and put the old unproductive business out of business can't start.
Macroeconomics and our numbers are designed around the "representative firm" and the "representative worker." But you are seeing here the macroeconomic effects of microeconomic distortion, and only visible in the amazingly large, widening and persistent differences in productivities, wages, and incomes across areas and companies.
(Random additional examples: The Allocation of Talent and U.S. Economic Growth, Chang-Tai Hsieh Chad Jones Erik Hurst and Pete Klenow; "About one-quarter of growth in aggregate output per person [from 1960 to 2010] can be explained by the improved allocation of talent," talented african-americans and women moving in to occupations from which they had been excluded. See above eye popping numbers.; Chang Tai Shieh and Pete Klenow, Misallocation and Manufacturing in China and India, or keep browsing all these authors' websites.)
3) Regulation again. There is a deeper lesson in this story for my call to measure the cost of regulations. The usual measures -- hours spent filling out forms, costs of activities required by regulators, wages of compliance employees and lawyers -- are completely meaningless. Suppose that getting a building permit in the Bay Area took 5 minutes. (It doesn't. It's an amazingly wasteful and time consuming process. But suppose it did.) But they just say no. Well, the cost of regulation measured that way would be zero. Businesses would answer surveys, "Regulations? They aren't a problem. It's just too darn expensive here." Economics is always about the unseen.
The cost of regulation is in the higher prices it imposes, the businesses that don't get started, the people who can't move here to earn the high wages that high productivity brings, the higher tax rates we all must pay to bail out social security medicare and pensions because the tax base is half as large as it should be, the personal social and government costs of the islands of poverty we have created away from the productive cities, and so on.
Glaeser and Gyourko is a great readable summary of broader issues in housing economics. This is going on a bit, so I'll give you one quote and save the rest for another day
In some parts of America, there has been a revolution in the regulation of home building over the past 50 years (Glaeser, Gyourko and Saks, 2005). For most of U.S. history, local economic booms were met with local building booms, so labor could follow shocks to local productivity. However, between the 1960s and the 1990s, it became far more difficult to build in the nation’s most desirable locations, especially those along the coasts. Higher economic productivity in San Francisco now leads to higher prices, not more homes and more workers (Ganong and Shoag, 2013). This change has both led to a transfer of wealth to a few lucky homeowners and to a distorted labor market where people move to regions such as the Sunbelt that make it particularly easy to build (Glaeser and Tobio, 2008).(YIMBY stands for yes in my backyard, a new movement that even here in the Bay area recognizes that letting people build houses and apartments might lower housing costs.)
Update: See also "Tarnishing the Golden State: Regulations and the US Slowdown" by Kyle Herkenho , Lee Ohanian and Ed Prescott. From the abstract, the finding,
Deregulating existing urban land from 2014 restriction levels back to 2000 restriction levels would increase US GDP growth by nearly .5% per annum from 2000 to 2014, bringing output and TFP growth roughly in line with their historical trends. The most significant expanding regions from these hypothetical deregulations are California, New York, and the Mid-Atlantic.A half percentage point of growth is still huge. What's the logic? As in Hsieh and Moratti, counterfactuals must come from a model
We use a variety of state-level data sources, including the USDA, the Census and the BEA to develop a general equilibrium spatial model of the US states
... general equilibrium congestion forces in the market for housing and land offset some of the gains from deregulationI get an inkling here that there is a production function with decreasing returns to scale, so not everyone can move tomorrow to SF (if they had a place to live) and earn currently high salaries. Just how many more people could benefit from whatever is the local magic is of course the key question.
Thanks to a correspondent for the link.