iiinova — where your digital self keeps evolving

Shane Phillips @shanedphillips.bsky.social · 05/10/2026

🎵Democrat for Senate, California🎵

Shane Phillips @shanedphillips.bsky.social · 05/10/2026

Another entry for the "impact fee nexus studies are bullshit" file! Student enrollment at LA Unified has fallen for 24 years straight, ~48%. And yet! We have an impact fee because: "As new development occurs, additional school facilities will be needed to house the growth in the student population."

A screenshot from the Los Angeles Unified School District 2026 School Impact Fee Justification Study. The text reads: "Education Code Section 17620 allows school districts to assess fees on new residential and
commercial construction within their respective boundaries. These fees can be collected without
special city or county approval, to fund the construction of school facilities necessitated by the
impact of residential and commercial development activity. In addition, these fees can also be used
to fund the reconstruction of school facilities to accommodate students generated from new
development projects. Fees are collected immediately prior to the time of the issuance of a building
permit by the City or the County.
As new development occurs, additional school facilities will be needed to house the growth in the
student population. Because of the high cost associated with constructing school facilities and the
District’s limited budget, outside funding sources are required for future school construction. State
and local funding sources for the construction and/or reconstruction of school facilities are limited."

Source: https://4.files.edl.io/0bc1/05/22/26/161250-26fd9501-6360-44fb-a158-3bb37a543fd9.pdf

A chart showing TK-12 enrollment in LAUSD declining in every year since 2002-03, from 724,000 in that year to 408,000 in 24-25, and a projected 376,000 in 27-28.
Source: https://gpsnla.org/wp-content/uploads/2025/08/Facing-the-Decline-Building-A-Resilient-LAUSD_FINAL.pdf

Current school impact fee rates effective August 11, 2026, per square foot by construction type:
Residential: $5.38
Commercial/industrial: $0.87
Rental Self Storage: $0.32
Parking Structure: $0.44

512920

Shane Phillips @shanedphillips.bsky.social · 26/09/2026

Maybe I'm just a partisan hack and this happens all the time with both parties, but it seems like putting the flag of a major geopolitical rival literally on top of the Oval Office is kind of insane???

0102

Shane Phillips @shanedphillips.bsky.social · 23/09/2026

We cover a few dashboards built to help people understand what determines the cost and feasibility of new housing, and a nice little tour through new French suburbs with Eric Kronberg

110

Shane Phillips @shanedphillips.bsky.social · 23/09/2026

We also cover Pew Charitable Trusts' report on overly restrictive lending standards. Lots of great stuff in here, and a call-back to our previous Housing Voice episodes with @kerdmann.bsky.social

100

Shane Phillips @shanedphillips.bsky.social · 23/09/2026

Our research overview is on "How Housing Supply Expansions Reshape Cities" by Marco Giacoletti, Matthijs Korevaar, and Franklin Qian (2026). It's about what happened when the Netherlands expanded the areas open to development around its major Randstad cities in the 1990s. (Hint: good stuff.)

100

Shane Phillips @shanedphillips.bsky.social · 23/09/2026

This one was wild to me: Only 63% of people who voted for Bass in the primary know they're voting for her in the general, compared to 89% for Raman. Raman is also supported by likely voters who didn't vote in the primary by more than 2-to-1, though with lots still undecided.

1221

Shane Phillips @shanedphillips.bsky.social · 11/09/2026

Amazing time capsule of a photo from Sept 11, 2001 in Los Angeles, from this @lamaterial.com article: lamaterial.com/p/alex-padil...

020

Shane Phillips @shanedphillips.bsky.social · 10/09/2026

I've got various other figures in there too. Here are bar charts for median home size, average bedrooms, and average bathrooms by lot size band. This is partly intended to address the opposite critique, that small lots will only produce "tiny homes," i.e.: www.aei.org/commentary/d...

110

Shane Phillips @shanedphillips.bsky.social · 10/09/2026

Figure 4 is further evidence that land values are determining home sizes. On the y-axis, I replace lot size with land value, and I remove the tract-level income restriction—it's all Dallas County. The relationship is very clear, with higher land values consistently associated with larger new homes.

121

Shane Phillips @shanedphillips.bsky.social · 10/09/2026

Figure 3 is the same as #1, but for n'hoods with median hh income *below* the countywide median. The clear relationship between lot size and home size breaks down, which I argue is because land values are lower: they're not exerting pressure to build bigger and maximize your investment in land.

120

Shane Phillips @shanedphillips.bsky.social · 10/09/2026

This figure shows that home size *does* matter a great deal for the price of housing, even when other characteristics are held roughly constant. So when lot sizes are smaller, we're not just benefiting from lower land cost, but lower improvement/construction cost as well.

120

Shane Phillips @shanedphillips.bsky.social · 10/09/2026

Figure 2 shows relative values, inclusive of land, for homes of increasing size in five major metro areas including Dallas, holding constant home type (single-family detached), tenure (owner-occupied), lot size (1/8 to 1/4 acre), and building age (1980–1999).

110

Shane Phillips @shanedphillips.bsky.social · 10/09/2026

Very clearly, smaller lots are associated with smaller new homes. For example, for parcels 1,500-2,999 sf. in size, the most common new home size was 2,000-2,249 sf. For parcels 20,000 sf (about half an acre) and above, the typical new home is over 6,000 sf It's a smooth progression

131

Shane Phillips @shanedphillips.bsky.social · 10/09/2026

This figure shows the frequency of different lot size and home size combinations in Dallas County, restricted to single-family houses built 2010 or later in neighborhoods with median incomes above the countywide median. Darker squares represent the more common home sizes within a given lot size band

130

Shane Phillips @shanedphillips.bsky.social · 04/09/2026

Hail from WA, moved to CA. Checks out.

030

Shane Phillips @shanedphillips.bsky.social · 26/08/2026

And from a study shared by @stephenjacobsmith.com: are high-rises a threat to birds? No they are not, according to researchers in this study published by the American Ornithological Society. Low-rises and houses cause >99% of bird deaths from buildings, and cats cause 4x more than that.

1165

Shane Phillips @shanedphillips.bsky.social · 26/08/2026

And check out this detailed estimate of cost reductions and revenue gains from different building code reforms (in the Ontario study)!

280

Shane Phillips @shanedphillips.bsky.social · 26/08/2026

We also cover two great reports on building code reform, one looking at US policy (specifically Boston) and the other at Canada (specifically Ontario). The Boston one has a great chart that I adapted into the one below uclahousingvoice.substack.com/p/ucla-housi...

130

Shane Phillips @shanedphillips.bsky.social · 18/08/2026

Life expectancy by world region, 1770-2023.

010

Shane Phillips @shanedphillips.bsky.social · 17/08/2026

Anyone know what's going on with LIHTC development in Colorado? This report got a lot of attention for high costs in CA and high LIHTC costs relative to market-rate in CA and TX, but I don't recall anyone talking about how CO is building LIHTC cheaper than market-rate. www.rand.org/pubs/researc...

4183

Shane Phillips @shanedphillips.bsky.social · 08/08/2026

I look at this map of bike paths and lanes in LA County and I want to say something like "thanks, fragmented governance among 88 different cities," but a huge share of this is the city of LA and it doesn't look any better. If your trip is more than a few blocks then you're probably dodging cars.

127111

Shane Phillips @shanedphillips.bsky.social · 31/07/2026

I'm playing around looking at median family income data and noticed that Washington DC has a way higher median than other large metros (chart 1) but a pretty run-of-the-mill 80% MFI (chart 2). Also, note how 80% MFI increased a lot faster than MFI in LA and NYC (chart 3 + 1). What's the deal?

110

Shane Phillips @shanedphillips.bsky.social · 29/07/2026

Unintentionally well-targeted spam

010

Shane Phillips @shanedphillips.bsky.social · 25/07/2026

Ontario, what is going on with these very straight long roads? Across the border in Michigan and New York at the same scale for comparison. Is it topographical (fewer mountains and hills requiring indirect routing)?

220

Shane Phillips @shanedphillips.bsky.social · 25/07/2026

I had Claude give this a go using Census and ACS data. I also added unmarried couples, which I think makes sense? They only started tracking that in 1960.

1101

Shane Phillips @shanedphillips.bsky.social · 01/07/2026

Good use of land within a couple blocks of one of the best public universities in the world in the largest city in a state with an economy bigger than every country's except the US, China, and Germany? Who's to say.

1282

Shane Phillips @shanedphillips.bsky.social · 01/07/2026

Here are the prices of all the recently sold homes within a half-mile of my office at UCLA.

87612

Shane Phillips @shanedphillips.bsky.social · 08/06/2026

- These migration patterns are at the heart of the moving chain research by Mast (2023) and others. Mast found that roughly 60–70% of people moving into new market-rate apartment buildings came from the same metro area, mirroring the 30–40% coming from outside the metro area seen in this dashboard.

130

Shane Phillips @shanedphillips.bsky.social · 05/06/2026

Updates based on today's vote drop. Of the 12 scenarios I posted yesterday, Raman beat Pratt in 6. After the 6/4 vote drop, she wins in 8. If turnout meets projections, she needs an ~11.4 point margin over Pratt. Her margin was 10.4 in the 6/3 drop and 13.3 today.

Estimates showing Raman needs a margin of 11.43 over Pratt to win if turnout meets projections, 13.39 if turnout is 5% below projections, 9.97 if it's 5% above, and 8.84 if it's 10% above.

3435

Shane Phillips @shanedphillips.bsky.social · 04/06/2026

I was curious about how the primary for Los Angeles mayor might turn out, so I modeled some scenarios with different turnout levels and vote splits. Upshot: Raman still has a decent chance, esp. if turnout exceeds projections. 2nd place scenarios marked in green. Notes and caveats in the replies.

A table showing vote shares for Bass, Pratt, and Raman under different scenarios. The scenarios are: 20% of the remaining vote goes to Pratt and 30% to Raman, 20% Pratt + 35% Raman, and 25% Pratt + 35% Raman. Results for each scenario are shown for four turnout levels: Total turnout matches projections, falls 5% short of projections, exceeds projections by 5%, or exceeds projections by 10%. Raman wins second place in the 20% Pratt + 35% Raman scenario for all turnout levels, and wins second place in all scenarios if turnout is 10% above projections.

1313

Shane Phillips @shanedphillips.bsky.social · 06/04/2026

Interesting chart from @kevinerdmann.bsky.social, with this even more interesting commentary: "Since [the 1990s], the growth of owner and renter households have moved in opposition to each other because the housing shortage means that more of one requires less of the other."

193

Shane Phillips @shanedphillips.bsky.social · 26/03/2026

Much appreciation for this abundance of seating

0191

Shane Phillips @shanedphillips.bsky.social · 06/02/2026

This is such pathetic petty tyrant bullshit. People should point and laugh at this guy every time he goes out in public.

090

Shane Phillips @shanedphillips.bsky.social · 26/01/2026

@michaelwiebe.bsky.social we didn't have time for my question about this in the interview, but one caveat to note is that credit scores have increased across the board pretty significantly over the past two decades. Got some pushback over LinkedIn over this www.federalreserve.gov/econres/note...

140

Shane Phillips @shanedphillips.bsky.social · 25/01/2026

Where we're headed, very very quickly

"Under the North Korean system, if one person in your family does something wrong or commits a crime against the regime, then the entire family can be sent away to a prison camp.

The reason the regime has this kind of system in place is because it wants to instill fear in North Koreans. Kim Il-Sung once said that any person who has anti-government sentiments should be wiped out entirely and up to three generations of their family should be punished."

050

Shane Phillips @shanedphillips.bsky.social · 21/01/2026

My favorite part of the tariffs is how, in addition to the tax itself, I also get to pay an even larger fee to the broker. More than doubling the additional cost to me, the consumer.

180

Shane Phillips @shanedphillips.bsky.social · 21/01/2026

...that drove up rents, and that disproportionately hurt lower and middle-income households and communities. (The black line in this chart is price-to-income in a low-income Atlanta census tract; the red line is a high-income tract.)

130

Shane Phillips @shanedphillips.bsky.social · 21/01/2026

The result was a post-2008 crash that was much deeper and longer-lasting than it had to be, that decimated the construction industry even as it kept growing in peer countries like Canada and Australia, that dramatically shrank the market for new construction...

140

Shane Phillips @shanedphillips.bsky.social · 21/01/2026

Tightening of lending standards leading up to and following the crash ended up shutting out roughly a third of the conventional mortgage borrowers — mostly people with good credit scores and low borrower risk.

150

Shane Phillips @shanedphillips.bsky.social · 08/01/2026

Very strange that LADWP's rebate program for high-efficiencly appliances, thermostats, etc. also applies to TVs. I have to assume almost no one actually claims it since it's so small, but in that case why even have it?

120

Shane Phillips @shanedphillips.bsky.social · 18/12/2025

These figures really capture the problem — a problem that poorly design construction defect liability laws contribute to, but by no means fully explain. The report that our conversation focuses on is here: ternercenter.berkeley.edu/wp-content/u...

Chart showing average unit value for single-family homes and units in 5+ unit buildings in LA/Orange and San Francisco Counties. Units in 5+ unit buildings (condos) are 40-50% less expensive on average.

Chart showing condo units as a share of new multifamily production in projects with 5+ units for 2011-2021. 3% of units in California are condos, nearly 40% in Canada are condos.

Charts showing condo production in LA/Orange and San Diego Counties. Production peaked in 2005/2006 and fell roughly 70-90% in the next few years and never recovered.

241

Shane Phillips @shanedphillips.bsky.social · 08/12/2025

For those interested, here are the model inputs for the different financial models, with "Ideal Loan Product SPR" being the preferred, and the one referenced in results throughout the report. (This table is on page 65 of the report.) The second table (pg. 80) are assumptions for the owner model.

110

Shane Phillips @shanedphillips.bsky.social · 08/12/2025

I also share this chart because unlike the others, it shows SPR returns only *during* the tenancy—which is why they appear lower than elsewhere. But the comparison to homeownership also shows that they're quite high, especially when we make reasonable assumptions about ownership.

Graph with "returns on rent or equity" on the y-axis and "years of tenancy or ownership" from 0 to 30 on the x-axis.

The line representing homeownership with a 3% discount rate starts at nearly -10% in year 1, then increases sharply to 30% by year 4-5 and then more slowly in later years, hitting nearly 60% in year 30.

The line representing homeownership with a 5% discount rate also starts near -10% in year 1 and is also near 30% by year 4-5, but it increases more slowly from here, peaking around 44% around year 15 and then dipping slightly over the next 15 years to around 41%.

The line representing SPR with a 3% discount rate starts at 0% and rises in a sawtooth pattern with large jumps every 7 years and slower, smaller declines in the intervening years. The peaks are around 14% in year 3, 19% in year 10, 25% in year 17, 32% in year 24, and 40% in year 31.

110