Joe Fish

thanks for walking me through this, btw!

when you say "the ROI looks good", i assume you mean something along the lines of "if the builder were forced to hold the property in perpetuity, would the baseline rents cover costs plus some target profit margin"?

so the ability to exit the market is not meaningfully in the pro forma?

i might be misunderstanding how these deals work (if so, lmk!).

im just trying to think of a world where my cash-out price is a multiple of my rental income and that multiple depends (implicitly) on future price increases. are you saying that cash-out price is marginal compared to 1-year rents?

i think the question wrt to how long the exemption should be is to what extent the buyers of the buildings price in possible above-inflation rent increases and to what extent they price in above inflation cost increases

The median housing unit in California is twenty years older than in texas. 10% of California's housing stock was built before WWII compared to 3% in Texas. Likewise, only 18% is built after 2000, compared to 39% in Texas. CA housing is really old!

how much of this is data centers crowding out residential structures? seems plausible AI capital competes with, say, multifamily, which has dropped off a lot since it's peak in 2023

(although from your later tweets, maybe the answer is the capex boom is a lot of relabelling?)

nobody knows how to set expectations like a central banker

as someone who loves (and is from) san francisco, no political organization or thought should escape a strict containment zone without thorough vetting.

media blackout on the six people at the billionaires march, the phrase "dean preston", and anything that has ever been discussed at a polycule.

More vacant homes give tenants relatively more bargaining power over landlords, in the same way more open jobs gives workers relatively more bargaining power over employers.

I got 2 free months rent because of an apartment surplus and high vacancy rates; i want that for everyone

the ratio of vacant homes to homeless people should be a lot higher because we should have far more vacant homes and far fewer homeless people

In many ways, a headline called "Why Building Alone Won’t Solve the Housing Crisis" is progress; people are willing to admit that supply does help

Lord willing, we'll fix subheadings next: "An imbalance in the kind of housing getting built..." The problem is not imbalance. The problem is inadequacy

i think we've gotten to the point where most of the people writing these kinds of articles will concede that changes in quantity tell you little about much if you press them.

but then they'll slip in headings like "As Supply Grew, the Share of Units Serving Lower-Income Renters Decreased"

i don't want to be an annoying economist, but i will be: Looking at places that built more housing as evidence for whether more housing helps affordability is like asking why when people buy umbrellas they tend to get *more* wet. Changes in quantity don't tell you if the demand or supply curve moved

are you conflating "not everyone can earn an above average income", which is true, with "not everyone's income can increase", which very much is not true?

"how do hedonics work with heterogeneous valuations" is an interesting topic that I, as a housing person, should really know more about tbh. my guess is that they work badly in theory and okay in practice, but i do not know for sure!

So is the argument the 2022 era multi family boom was mostly pulling units forward? Plus a few places that did do supply reforms?

>If building more lowered housing prices, then Tokyo & Manhattan would be cheap.

"If umbrellas stopped rain, the places with the most umbrellas wouldn't have the wettest people". Manhattan is high density *because* people want to live there, but if more housing were built, it would be cheaper.

idl if any SF housing people follow me, but the sf "pipeline" numbers are basically fake, no? I'm looking at the data and a huge percent of the oft-cited "Sf is sitting on 72K units" are

1. stale projects
2. projects that aren't approved in any meaningful sense (2700 sloat, 469 stevenson)

San Francisco has built about net 60,000 units in the past *twenty* years. And san francisco only builds expensive housing because SF makes it expensive to build!

The 72K units also includes a bunch of projects that are not really "approved for building" in any meaningful sense, e.g., 2700 sloat

oh yeah, i should've phrased this better: i think the percentage is a better number to lead with for public communication

again, love all the work you all do!

(I say all of this out of love, since i look forward to the state of the nation's housing report every year)

also JCHS if youre reading this, i think the residual income charts need to be fixed, too

bsky.app/profile/sadb...

i hope JCHS switches to talking about percentage of cost burdened renters (also rising!). the number going to "record highs" is not always super informative when population growth is happening. same reason why the unemployment rate is preferred over the number of unemployed people

of course population didn't increase much, san francisco barely built housing! if they had built more housing, population would have gone up more, and prices would be lower

Even if you take the MMT line that inflation was all supply shocks, that implies that, like WWII, you would’ve needed to do non price rationing.

And to come full stancil, if you want to do that, you need a media ecosystem that will be supportive, which very much does not exist lol

Maybe “haul a bunch of grocery ceos in front of congress” kind of helps?

Idk, I think people would’ve been upset either way. If you had done price controls you would’ve gotten a million stories about shortages (as w/ inflation, this is true regardless of severity).

It’s not like there was a big media ecosystem of people ready to ride or die for biden

how should i square your substack with food prices having grown more slowly than wages? Eg., Grocery-inflation outpaced wage growth during 2022-3, but it's been stable since.

the timing kind of, but kind of doesn't line up with the changes in your chart since the pandemic

food and clothing are two categories that have gotten removed from the bucket of "essentials" when talking about affordability concerns precisely because they are now so much more affordable. i hope this happens with other goods!

ive been overhauling a bunch of data pipelines with it (plus chatgpt) and it's great at writing much better boilerplate code with like logging, testing, etc.

which is code im often too lazy to write and so that's a big win. but it's still decidedly meh at making substantive decisions