C. D. Howe: a bad study of vacancy taxes

Don't choose a control group in the same market!

Treatment group just west of Boundary, control group to the east

Ripple Effects: The Impact of an Empty-Homes Tax on the Housing Market. Gherardo Gennaro Caracciolo and Enrico Miglino, C. D. Howe, August 2024.

What Vancouver’s empty home tax really means for Canada’s housing supply. Gherardo Gennaro Caracciolo, The Hub, August 2024.

Like a lot of pro-housing people (“YIMBYs”), my understanding of the housing market mostly comes from economists. I like the C. D. Howe think-tank, which publishes economic studies aimed at journalists, policymakers, and the interested general public. They’re centrist and solidly grounded in economics. If you want to get a sense of what economists think would make good policy in a particular area, they’re an excellent reference.

But they’re not infallible. I was reading a study published by C. D. Howe recently, which looked at the impact of Vancouver’s vacancy tax on housing supply and rents. The estimate of the impact on housing supply looked okay. The estimate of the impact on rents looked completely wrong.

To back up a bit, if you see rents going up in Vancouver, or down, that doesn’t actually tell you whether the tax had any impact, because there’s lots of other things happening in the rental market. What you need is a control group: what’s happening in a separate rental market that isn’t affected by the tax? That is, what matters is the change in the difference between rents in the treatment group and the control group.

Problem is, they tried to do the study by comparing rents on the Vancouver side and the Burnaby side of Boundary Road, which doesn’t make any sense.

I wrote an email to Caracciolo.

I read your recent e-brief on vacancy taxes in Vancouver with great interest; it's good to have an estimate of the impact of the tax on the number of vacant homes.

However, as a layperson, I was extremely puzzled by the attempt to compare rents in the treatment group and the control group. This makes no sense to me.

The treatment group and the control group are part of a single rental market: someone looking for a place to live would be indifferent between living on the Vancouver side or the Burnaby side of the boundary. If rents on one side of the boundary were lower than rents on the other side, there'd be an arbitrage opportunity for renters to save money by choosing the lower-cost side.

To estimate the vacancy tax's impact on rents, I think you'd have to compare two separate rental markets.

More

• Demand-side policies. CMHC observed that 8,800 existing condos was added to Metro Vancouver’s long-term rental market in 2019, along with 2,300 new condos. This is a one-time boost, but it still helps. “A combination of market factors and housing policies designed to encourage long-term rental likely had an impact on the supply.” These policies included BC’s speculation and vacancy tax, Vancouver’s empty homes tax, and Vancouver’s restrictions on Airbnb.

• Noah Smith, Why you should not ignore economists. “It’s often very helpful to have smart thoughtful people who have the same goals as you, and who have thought deeply and rigorously about how to achieve them.”

• Paul Krugman, The Accidental Theorist. Slate, January 1997. An illustration of the pitfalls of trying to reason about the economy without an economic model: you have one, but you don't realize it! A couple decades ago, when I asked Krishna Pendakur, a friend who's an economist, what I should read to get a better understanding of economics, he recommended Krugman’s Peddling Prosperity. More articles by Krugman.