Every house, every apartment building, every 2-flat in the village, measured against each other in the county assessor's 2026 valuation. Here's what the data actually says about living next door to multi-family housing.
SCROLL TO BEGIN ↓Ask around and you’ll hear it as settled fact: apartments drag down the value of the houses beside them.
It’s a testable claim. And Oak Park, dense and full of hundred-year-old multi-family housing, is a good place to test it.
It comes up at every zoning hearing, and it's stated as settled fact. Oak Park's own data lets us check whether it's true.
Oak Park has 2,041 multi-family buildings: 2-flats, courtyard apartments, condo buildings, townhomes. They stand on the same blocks as its houses, most of them for a century now. Each dot is one of them, plotted from the Cook County Assessor's records.
All 9,596 single-family homes in the village sit at some measurable distance from a multi-family building. If proximity really dragged values down, the pattern would be written all over this map. With the 2026 assessment of every property, we can go looking for it.
A house’s value is mostly its size, lot, age, and condition. So every comparison from here on holds those fixed: building and lot square footage, age, bedrooms, bathrooms, fireplaces, central air, garage, basement, construction quality, repair state, and neighborhood. What’s left is location. Does distance to multi-family move the number? There’s a problem with measuring that naively, and it’s visible on the map.
Color each building by its location and the red dots trace the arterials: Harlem, North Avenue, Madison, Roosevelt, Austin. A corridor building fronts a named arterial, sits on a corner within 150 feet of one, or stands within 300 feet of commercial property. That's 1,152 of the 2,041.
We measured this directly, with no multi-family terms in the model at all: houses within 100 feet of an arterial are assessed −3.7% against similar houses on quiet blocks, apartment or no apartment. Leave that out and the street's discount gets pinned on whatever happens to be built along it. So corridor buildings are set aside and estimated separately.
The blue dots are embedded: multi-family on quiet residential streets, away from the arterials and the commercial strips. This is exactly the kind of building zoning reform would allow more of. The question is whether these discount their neighbors.
Even within a neighborhood, blocks differ in ways no model fully captures. So the main analysis uses a stricter design: draw rings around each embedded building and compare only the houses that share it.
The design, in the paper ↗Houses within 100 feet of the building are compared to houses 400–800 feet from that same building: same micro-market, same schools, same street grid. Anything shared by the whole cluster cancels out. Only the nearness itself remains.
Run that comparison simultaneously around every embedded building with houses in its rings, with the full set of house controls on top. 155 of the 543 buildings have houses out in the reference band; those anchor the estimate. If living beside a 2-flat or a small apartment building costs you anything, this is where it shows up. That's the whole method. Now the results.
Houses within 100 feet of an embedded building are worth the same as houses a few hundred feet from that very building. The difference: +0.9%, statistically zero. And the method works: point it at the corridor buildings we set aside and it finds their street discount just fine. When there's a real effect, it shows up. Here, nothing does.
The ring result isn't a quirk of the design. Compare all 9,596 houses at once, with full controls, and distance to embedded multi-family still doesn't matter: +1.2% at 0–100 feet, statistically zero at every distance. It stayed zero under every alternative classification and check an independent adversarial audit of the pipeline demanded.
Count the embedded multi-family buildings within 400 feet of each house. Houses with one, two, or three or more are worth the same as houses with none. Every difference is a rounding error.
This is one suburb, measured once. But studies that look directly at what multi-family housing does to the single-family homes around it keep reaching the same conclusion: nothing.
MIT’s Center for Real Estate tracked 36,000 sales around seven large, deliberately controversial mixed-income rental developments. House prices next to them simply tracked the surrounding market.
Pollakowski, Ritchay & Weinrobe ↗After the 2040 Plan allowed triplexes on every lot, Kuhlmann found single-family parcels gaining multi-family rights rose modestly in value. They did not fall.
J. of the American Planning Assoc. ↗The ring method itself comes from Linden & Rockoff (2008), who used it to measure how home values respond to a genuinely unwelcome neighbor. The design finds real effects when they exist. That’s the point.
Houses beside Oak Park’s mid-block 2-flats, condo buildings, and townhomes are worth the same as identical houses two blocks away. At every distance, in every model, no matter how many are nearby. The buildings zoning reform would allow on residential streets are precisely the kind that show no effect at all.
One honest caveat: these are the assessor’s modeled market values, not closing prices, and a snapshot rather than an experiment. The full paper spells out what that does and doesn’t let us claim.