Does “filtering” make housing more affordable?
Councilperson Warwick and Councilperson Gross have both expressed skepticism on filtering in housing. It's worth taking a couple paragraphs to explain what filtering is, what empirical evidence we have for its existence, and why it does not appear to have provided cheap housing in Pittsburgh.
What is Filtering?
When new market rate homes are built, households move into them, reducing demand and prices for the homes they leave behind. Other households now move into these cheaper homes, leaving behind still more homes.
An easy illustration of this uses hermit crabs:
If you line up a bunch of hermit crabs, too big for their shells, and place a large, empty shell among them, the largest crab moves into the new shell. The second smallest crab then moves into the shell the largest crab left behind. And so on and so forth, until even the smallest crab has a new, larger shell.
The same process happens in housing. When supply increases faster than demand, the wealthiest households move into the newest, most desirable homes. Less wealthy families move into the homes the wealthiest households left behind, even less wealthy families move in the homes the less wealthy families left, so on and so forth, until even the poorest households have affordable housing. (Some people have derisively referred to this as "trickle down" - that is an entirely different concept. Trickle down is the idea being that wealthier people will use an increase in income to increase demand, thus generating higher wages for lower income. households. Filtering, on the other hand, is the mechanism of how an increase in supply results in lower costs).
Empirical Evidence
There have been multiple papers studying filtering since the 1970s (see Boddy & Gray, 1979, for a review of these early papers). Since then many more have come out(Bratu et al, 2023; Weicher et al, 2018, and Kindström and Liang, 2023).
I will focus here on only one: Evan Mast's 2023 paper "The effect of new market rate housing construction on the low income housing market", published in the Journal of Urban Economics.
In this paper, Mast looks at the construction of 686 new, large, market-rate buildings in 12 major US cities (Atlanta, Boston, Chicago, Dallas, Denver, Houston, Minneapolis, New York City, Philadelphia, Seattle, San Francisco, and Washington, DC). Then he traces all the households that moved into those buildings (over 52,000 individuals), the buildings they moved out of, the households that moved into those buildings, etc - for a total of 6 such moves.He finds that for every 100 new market rate units built, the resulting migration chain creates 45-70 new housing units for below median income households, including 19-39 new housing units for bottom quintile income households.
We have lots of new housing in Pittsburgh - why don't we see filtering and why are rents still going up?
First, we don't see filtering because we're not specifically looking for it - we're not looking at individual households, asking where they moved from, who moved into their previous homes, etc. But more importantly, we don't see filtering and we don't see rents going down because demand for housing in Pittsburgh is increasing faster than supply. See this recent Pro-Housing Pittsburgh blogpost explaining why in detail.
But in short: if 20,000 households want to move into the city, and we've only built 6,000 housing units, they can all only move in if 14,000 households move out. And if those 20,000 households are willing to pay more than the poorest 14,000 households can, then the poorest 14,000 will be pushed out. The only solution, the only way out, is to build more housing, of all sorts, as quickly as possible.
Every delay, every friction, every unit canceled means another family forced out of housing in this city.
Sources:
Boddy, M. and Gray, F., 1979. Filtering theory, housing policy and the legitimation of inequality. Policy & Politics, 7(1), pp.39-54.
https://bristoluniversitypressdigital.com/view/journals/pp/7/1/article-p39.xml
Bratu, C., Harjunen, O. and Saarimaa, T., 2023. JUE Insight: City-wide effects of new housing supply: Evidence from moving chains. Journal of Urban Economics, 133, p.103528.
https://www.sciencedirect.com/science/article/pii/S0094119022001048
Kindström, G., and Liang, C. 2023. Does new housing for the rich benefit the poor? Working paper.
https://tinyurl.com/KindstromLiang
Mast, E., 2023. JUE Insight: The effect of new market-rate housing construction on the low-income housing market. Journal of Urban Economics, 133, p.103383.
https://www.sciencedirect.com/science/article/pii/S0094119021000656
Weicher, J.C., Eggers, F.J. and Moumen, F., 2018. The long-term dynamics of affordable rental housing: creating and using a new database. Cityscape, 20(2), pp.235-244.
https://www.jstor.org/stable/26472176
For more papers on housing, I highly recommend Stephen Hoskins' excellent literature review:
https://stephenhoskins.notion.site/YIMBY-Lit-Review-27ae7791bab141058b82d94875ca98f3
Here is another popular press article from Strong Towns explaining filtering:
https://www.strongtowns.org/journal/2024/1/3/the-best-evidence-yet-for-the-housing-musical-chairs-theory
By Jack Billings.
