3 ms·

Rent-control policies likely 'fueled' SF gentrification, Stanford economists say

15 points by MilnerRoute 9y ago

wahern 9y ago
Ah, more economists making predictions based on very general theory without bothering to validate against the empirical evidence. Here's an awesome article posted on HN last year, where the author laboriously compiled advertised rental rates going back to 1948. https://experimental-geography.blogspot.com/2016/05/employment-construction-and-cost-of-san.html https://experimental-geography.blogspot.com/2016/05/employme... His data clearly shows that the long-term rate of increase in rents continued unchanged after rent control was imposed in 1979. It's basically been steady since the mid 1950s. Year-to-year variations are best explained by variations in income and number of employed persons. Moreover, growth in housing stock didn't appreciably drop after rent control. It has slowed slightly since the 1960s but it would be hard to pin that all on rent control as the regulatory environment has become much more strict. EDIT: Here's a more succinct conclusion from someone who tried to reproduce the results from the same data: Using those methods and Mr. Fischer's data, we find that median rents in San Francisco are best predicted by average per capita annual income and total employment. We further find that there *IS NOT sufficient evidence that total housing unit inventory has a significant effect on the median rent in San Francisco*. http://moreuseful.blogspot.com/2016/05/another-look-at-factors-driving-rents.html http://moreuseful.blogspot.com/2016/05/another-look-at-facto... EDIT EDIT: The Stanford authors did use Fischer's data for median rents. See page 10.
- pzone 9y ago
  Ah, more amateur data analysts running a few time series regressions and thinking they're doing economics. Check out the big warning at the bottom of the moreuseful.com post with a very forthcoming "We have not established causality!" The entire crux of Diamond, McQuade and Qian is that there was a pseudorandom assignment of rent control benefits to certain tenants but not others in the 1994 law. They use this variation to identify their models. This gives them a defensible causal story: here are two similar groups, with and without rent control, we'll study how their behavior differs. You can see all this starting in section 4 of their paper.
- closeparen 9y ago
  Supply and demand. No one claims that housing supply growth alone would predict the rent increase in a given year. We claim that housing supply growth per job, per migrant, etc. would. YIMBYs are prone to leave off the denominator on the assumption (right or wrong) that there's not much housing policy can do about it, but the argument is fundamentally that the price comes from the relationship between the population seeking housing and the available inventory. It's much easier to invalidate this strawman by invoking areas where nothing is being built but prices are flat or declining. As the article mentions, housing inventory has not been allowed to vary much, while the labor market has been through huge boom-bust cycles. So yes, the labor market should be the best predictor of housing prices.

15 points by MilnerRoute 9y ago

Ah, more economists making predictions based on very general theory without bothering to validate against the empirical evidence. Here's an awesome article posted on HN last year, where the author laboriously compiled advertised rental rates going back to 1948. https://experimental-geography.blogspot.com/2016/05/employment-construction-and-cost-of-san.html https://experimental-geography.blogspot.com/2016/05/employme... His data clearly shows that the long-term rate of increase in rents continued unchanged after rent control was imposed in 1979. It's basically been steady since the mid 1950s. Year-to-year variations are best explained by variations in income and number of employed persons. Moreover, growth in housing stock didn't appreciably drop after rent control. It has slowed slightly since the 1960s but it would be hard to pin that all on rent control as the regulatory environment has become much more strict. EDIT: Here's a more succinct conclusion from someone who tried to reproduce the results from the same data: Using those methods and Mr. Fischer's data, we find that median rents in San Francisco are best predicted by average per capita annual income and total employment. We further find that there *IS NOT sufficient evidence that total housing unit inventory has a significant effect on the median rent in San Francisco*. http://moreuseful.blogspot.com/2016/05/another-look-at-factors-driving-rents.html http://moreuseful.blogspot.com/2016/05/another-look-at-facto... EDIT EDIT: The Stanford authors did use Fischer's data for median rents. See page 10.

pzone 9y ago
Ah, more amateur data analysts running a few time series regressions and thinking they're doing economics. Check out the big warning at the bottom of the moreuseful.com post with a very forthcoming "We have not established causality!" The entire crux of Diamond, McQuade and Qian is that there was a pseudorandom assignment of rent control benefits to certain tenants but not others in the 1994 law. They use this variation to identify their models. This gives them a defensible causal story: here are two similar groups, with and without rent control, we'll study how their behavior differs. You can see all this starting in section 4 of their paper.
closeparen 9y ago
Supply and demand. No one claims that housing supply growth alone would predict the rent increase in a given year. We claim that housing supply growth per job, per migrant, etc. would. YIMBYs are prone to leave off the denominator on the assumption (right or wrong) that there's not much housing policy can do about it, but the argument is fundamentally that the price comes from the relationship between the population seeking housing and the available inventory. It's much easier to invalidate this strawman by invoking areas where nothing is being built but prices are flat or declining. As the article mentions, housing inventory has not been allowed to vary much, while the labor market has been through huge boom-bust cycles. So yes, the labor market should be the best predictor of housing prices.