Models · Model 004 · From Russell's research
Shrinking City
Demolition drains the tub. It doesn't turn off the tap.
In 2007 Buffalo set out to demolish 5,000 vacant buildings in five years and bring its vacancy rate from about 15 percent closer to 5. When Russell and Jason Knight studied the program, they found no link between the demolitions and the drop in vacancy that came around the same time. Here are sixteen blocks of a neighborhood that keeps losing people. You have the city's tools.
01 · The model
Sixteen blocks, your tools
Each square on the map is a lot, and one step is a month. Solid houses have someone living in them. Rust outlines are vacant houses, crossed out once they are past saving. Press Run, then try the city's tools as the years pass. The dashed line on each chart is the same neighborhood left alone, with the same owners and the same luck.
Year 1 · January
A new neighborhood. Press Run.
The city's tools
Housing market
About 10 households a year buy a house here, fewer than the houses that empty.
No demolitions.
No help for owners.
The city fixes up no houses.
Vacancy rate
This neighborhoodLeft aloneBuffalo's 15% and 5%
Households
HouseholdsLeft aloneHouses standing
Into and out of vacancy, each year
Owners moved outBought or fixed upTorn down or fell in
02 · The test
Twenty years, 9 policies
Each line is twenty years of one policy, averaged over 20 neighborhoods. Every policy starts from the same neighborhoods, with the same owners and the same luck. Choose the lines to compare.
20 neighborhoods for each policy, twenty years each.
The numbers
Every neighborhood starts with a vacancy rate of about 15%, about 280 households, and about 48 empty lots. Each number is the average over 20 neighborhoods; the last column is the year by which half of them had a vacancy rate of 5 percent or less.
| Policy | Vacancy rate | After 20 years | Demolitions | Under 5% by | |||
|---|---|---|---|---|---|---|---|
| Year 5 | Year 10 | Year 20 | Households | Empty lots | |||
| Nothing | 24% | 27% | 24% | 179 | 145 | 0 | Not in 20 years |
| Five in five | 15% | 22% | 23% | 182 | 145 | 44 | Not in 20 years |
| Keep demolishing | 18% | 16% | 10% | 186 | 175 | 105 | Not in 20 years |
| Help owners stay | 21% | 22% | 18% | 206 | 131 | 0 | Not in 20 years |
| Fix and sell | 19% | 17% | 12% | 231 | 119 | 0 | Not in 20 years |
| Keep up the lots | 19% | 18% | 13% | 229 | 120 | 0 | Not in 20 years |
| Demolish and keep up the lots | 13% | 8% | 3% | 242 | 132 | 74 | Year 14 |
| All of it | 10% | 3% | 0% | 285 | 98 | 34 | Year 9 |
| Buyers come back | 15% | 11% | 5% | 271 | 97 | 0 | Year 19 |
Left alone, the neighborhood keeps emptying. It loses more than a third of its households in twenty years, and its vacancy rate climbs from 15 percent to about 27 in the first decade. Then the rate eases a little, but not because anyone came back: houses past saving burn or fall in, and drop out of the count.
Demolition lowers the rate, and holds it down as long as the crews keep working. It doesn't keep households. Each policy that only demolishes ends twenty years within ten households of doing nothing. Russell and Jason Knight put the reason in one line: “demolition operates only on the outflow of existing vacant structures.” The crews take away houses that are already empty, and the owners who leave keep emptying more.
The tools that work on those flows keep households. With help for owners or kept-up lots, fewer owners leave; houses the city fixes up fill again. Each of the three adds from about 25 to about 50 households over twenty years. Together, with a little demolition, they hold the neighborhood at the size it started. That is the model's version of the study's call to fold demolition into “longer-term, multi-pronged, holistic approaches.”
The rate alone can mislead. In these runs it falls in two different ways, when people stay and when houses come down, and only the households tell them apart.
03 · How it works
Owners, buyers, and old houses
The blocks. Four streets cross four blocks each, with 384 lots of 30 by 110 feet. A face block is the 24 lots along one street between two corners, on both sides. When the clock starts, about one lot in eight is already empty, more on some blocks than others, and 15 percent of the houses still standing are vacant: the city's estimate for Buffalo when its program began.
Owners. On a face block with nothing empty, about 4 owners in 100 move out each year, each with a leaning of their own. Vacant houses and unkept lots on the face block raise the odds. Where half the other lots are empty, about one owner in ten leaves each year. This face-block pull is borrowed from Russell's study of property upkeep, where owners let their places slip more often as their face block slipped, and the pull depended on the wider neighborhood. Here it is an assumption about moving, and the model leaves out the wider neighborhood.
Buyers. About 10 households a year buy a house here, fewer than the houses that empty. Each looks at three vacant houses and takes the one on the steadiest block, so houses on the emptiest blocks wait longest.
Old houses. A vacant house is past saving after about four years, sooner or later depending on the house. Each year about one in eight houses past saving burns or falls in.
The count. The vacancy rate counts houses empty three months or more, out of the houses still standing, the way the postal data in Russell and Jason Knight's study count them. Empty lots aren't in it at all.
The tools. Crews take down houses past saving, the longest gone first. Help with repairs and taxes goes to owners on the emptiest blocks and halves their odds of leaving for five years. The city fixes up vacant houses on the steadiest blocks and sells them. Kept-up lots stop weighing on their blocks. Help for owners and kept-up lots are palliative in the sense of Russell's 2017 paper: they treat symptoms, aiming to “improve quality of life for existing residents in order to lower the probability of further decline.”
- Lots
- 384, on 16 face blocks of 24
- Step
- One month
- Owners
- About 4 in 100 leave a full block each year, more as it empties
- Buyers
- About 10 a year, or 16 in a recovering market
- Old houses
- Past saving after about four years empty
- Code
- About 200 lines of JavaScript, running in this page
Most ownersWith help
04 · What to try
Seven experiments
Each button sets up a new neighborhood with one policy and starts it running. Watch the dashed lines: they are the same blocks left alone. Every neighborhood is different, so run each one a few times.
-
Five in five
Twelve demolitions a year for five years, then none. While the crews work, the vacancy rate holds near 15 percent, where doing nothing would push it to about 24. Five years after they stop it is back above 20. The households fall just as fast as on the blocks left alone. The crews run short of work, too: houses don't pass saving fast enough to keep them busy, and they manage about 44 of their 60 demolitions.
-
Keep demolishing
Six demolitions a year, every year. After twenty years the vacancy rate is about 10 percent, against about 24 for the same blocks left alone. Yet the neighborhood keeps only a handful more households, and has about 30 more empty lots. The rate fell because there are fewer houses to count.
-
Help owners stay
Each year ten owners on the emptiest blocks get help with repairs and taxes, which halves their odds of leaving for five years. Fewer owners leave, and after twenty years the neighborhood has about 25 more households than if it had been left alone.
-
Fix and sell
Each year the city fixes up three vacant houses on the steadiest blocks and sells them: 60 houses in twenty years. The neighborhood ends with about 50 more households than if it had been left alone, and a vacancy rate of about 12 percent.
-
Mow what you clear
Six demolitions a year, and every empty lot mowed and cleaned. A kept-up lot no longer weighs on its block, so about 65 fewer owners leave over twenty years. The neighborhood keeps about 60 more households than if it had been left alone, and its vacancy rate usually ends below 5 percent.
-
All of it
Three demolitions a year, ten owners helped, three houses fixed and sold, and the lots kept up. After twenty years the neighborhood has as many households as when it started. Its vacancy rate falls below Buffalo's goal of 5 percent in about nine years.
-
Buyers come back
No program at all, but 16 buyers a year instead of 10. The neighborhood keeps nearly all its households, about 270 of 280, and its vacancy rate drifts down toward 5 percent. The strongest lever here is the one a city controls least.
Sources
Reading and code
- The studyRussell Weaver and Jason Knight, Can Shrinking Cities Demolish Vacancy? An Empirical Evaluation of a Demolition-First Approach to Vacancy Management in Buffalo, NY, USA, Urban Science 2(3): 69, 2018. Buffalo's 5 in 5 program, the postal vacancy data, and the stock-and-flow argument this model plays out.
- Palliative planningRussell Weaver, Palliative planning in an American shrinking city: Some thoughts and preliminary policy analysis, Community Development 48(3): 436–450, 2017
- The bookRussell Weaver, Sharmistha Bagchi-Sen, Jason Knight, and Amy Frazier, Shrinking Cities: Understanding Urban Decline in the United States, Routledge, 2016
- Face blocksRussell Weaver, The Tended Commons Within, Commoning Consciousness Sunday letter, October 11, 2026. Describes the property-upkeep study behind the face-block pull.
This is a teaching model, not a forecast for any real neighborhood, and it doesn't explain Buffalo's results. Its numbers were chosen to be plausible, not measured: how fast owners leave, how many buyers come, how long a house lasts empty, and how much an empty lot weighs on its block. That last choice matters most to how well kept-up lots work here. Every owner is also a homeowner; the model leaves out landlords, renters, tax foreclosure, and the rest of the city. The code is plain JavaScript and runs in this page.
Russell's research on vacancy and shrinking cities is one of the strands in his book in progress, Commoning Consciousness.