The Case for Filtering and Moving Chains: Evidence from Honolulu
Does building new housing lead to lower rents in older housing? A new study out of Honolulu tests that claim by tracking “moving chains” in a single building, The Central, a 512-home tower in Honolulu, completed in 2021.
Each older home a resident leaves behind to live in The Central – which includes 202 market-rate homes and 310 deed-restricted units – becomes available to someone else. Each subsequent vacancy triggers another vacancy down the price scale, in a process economists call “filtering.”
In “The Downmarket Impact of New Multifamily Housing: Evidence from a Honolulu Condo Tower,” researchers Limin Fang, Emi Kim, and Justin Tyndall trace the moving chains set off by The Central’s first residents, comparing what happens when market-rate and income-restricted homes get new occupants.
Key Takeaways:
- Steep Price Drops: The homes vacated in the first round of moves to The Central cost 38% less per square foot; homes vacated in the next round cost 44% less.
- Vacancy Rate Split: New market-rate homes generate more moves and vacancies per home (0.43) than income-restricted homes (0.29).
- Moving up: The Central’s first residents moved from neighborhoods with 15% lower median incomes, freeing up less-costly housing stock while also demonstrating the demand in higher-income neighborhoods.
To trace these chains, the researchers combine four sources: the tower’s unit-by-unit pricing plan, county property assessments, statewide deed-transfer records, and a private address-history database covering Hawai’i from 2017 to 2024. Starting with the 322 residents identified in the tower after it opened in 2021, they trace each mover’s prior address, then whoever fills that vacancy next, following the chain through four rounds. That database captures only about 46% of income-restricted households and 42% of market-rate ones, so 180 confirmed vacancies are treated as a floor, not a full count. The sample thins fast, too: by round two, only about 20 properties remain in the analysis, so those figures carry more uncertainty. The authors frame this as a single case study. Honolulu’s housing market has unusual features: a large share of out-of-state vacation-home buyers and unusually high rates of household overcrowding—which may shorten the moving chains observed here relative to what a similar building might produce in another city.
New Homes Trigger a Price Cascade: The tower’s own homes average $1,152 per square foot. What people move out of, though, is typically much older housing. The average vacated home was built around 1985, versus 2021 for the tower, and it’s often larger. Despite that, those vacated homes are cheaper: $715 per square foot in round one (i.e., the homes that Central residents vacated) and $642 by round two (i.e., the homes vacated by the people who moved into the Central residents’ original homes), 38% and 44% below the tower. Income-restricted homes filter into an even lower price tier. Households that move out of them tend to start from a cheaper point, pushing the first-round drop to 43% per square foot, versus 30% for market-rate movers.
Same Building, Different Vacancy Math: The 202 market-rate homes triggered 87 downstream vacancies, while the 310 income-restricted homes triggered 90. That near-even split hides a gap once you account for the fact that the building has far more income-restricted units. Per home built, market-rate homes trigger a vacancy almost 50% more often than income-restricted ones (0.43 versus 0.29). Many movers into income-restricted homes are people leaving a family member’s home or a roommate situation to move into their own home. If even one member of the original household stays behind, the study doesn’t count it as a full vacancy — though the authors note it may still ease overcrowding in the home left behind. Market-rate homes generate more vacancies per home built, but vacancies in income-restricted homes reach further down the price scale because those homes start at a lower price point.
A Neighborhood Upgrade for the First Movers: The Central’s own census tract has a median household income of about $106,000 — close to the Honolulu County median — but a far higher share of college graduates: 71%, compared with 38% countywide. Households that moved into the tower came from tracts with 15% lower median income and 34% lower college-education rates than the tower’s tract, meaning the move itself was a real neighborhood upgrade for the people who made it.
The paper’s authors draw two policy lessons. First, they credit Hawai’i’s 201H program with making the project viable, and argue similar streamlined pathways can speed other projects. Second, they note that requiring developers to sell some homes below market price, as 201H did, can channel the same filtering benefit as government-built public housing, with the developer absorbing the discount instead of taxpayers. But citing other research, they caution that if such a requirement is set too high, it can discourage a project from getting built at all.
The authors can’t say for certain that any single vacancy wouldn’t have happened without The Central — some of those movers might have relocated anyway. But in that scenario, they would have taken an existing vacant home instead of a new one, leaving one fewer option for whoever came next. Over three years, this one building is tied to 180 confirmed vacancies — and, by the researchers’ own estimate, likely several hundred more — that benefited households who never lived in it.
Photo by Jess Loiterton via Pexels