Good schools do not grow NOI, and the reason is a basis problem.

Columbia Business School's Milstein Center published a study on August 3, 2026, by Michael Steinberg of MetLife Investment Management and Christopher Munsell of Columbia, that has been circulating through commercial real estate newsletters since mid-August. They tested one of the most durable assumptions in multifamily investing, that a property assigned to better schools should produce better net operating income (NOI) growth. I want to walk through what they found and why it changes how I would underwrite a school boundary story.

They pulled loan-level multifamily records from CRED iQ and kept every property they could observe in both 2015 and 2024. They matched each one to its assigned school attendance zone using the School Attendance Boundary Survey (SABS) for 2015 and ATTOM's 2024 update, then ranked each school on SchoolDigger's in-state percentile, converted so 2015 and 2024 are comparable. Whenever a property's assigned school changed between the two years, they counted it as a redistricting event. They found 795 of them.

Properties redistricted into worse schools grew NOI about 38% cumulatively over the period. The non-redistricted control group grew about 37%. Properties redistricted into better schools grew only 32%. Run through a regression with metro fixed effects, favorable redistricting carries a statistically significant NOI growth penalty of 6 to 7 percentage points. That result holds up, and it gets slightly worse, not better, when the sample is narrowed to elementary and middle schools, exactly where the existing rental literature says the school effect should be strongest. But before anyone rewrites their underwriting model, look at the effect size.

In plain English, redistricting explains under half a percent of the total variation in NOI growth. It is statistically real and economically irrelevant.

This matters because the prior academic work on this question, Gabe, Robinson, and Sanderford (2022), matched roughly 100,000 asking rent observations to Great Schools ratings across 45 metros and found that renters do pay up for school quality, about half of what buyers pay per point of quality, call it 1 to 2% of rent per point on a 10-point scale. That earlier study measured a level at a point in time. It never asked what happens to a specific property's NOI over a hold period when its assigned school changes. This new study is the first to ask that question directly, and the answer is that whatever premium exists never clears the noise of nine years of tax reassessment, insurance escalation, and operating expense growth.

Here is the basis problem for anyone underwriting an acquisition. If you pay a premium price (i.e., a lower cap rate) because a property feeds into a strong school district, you are underwriting a rent and NOI outcome that this data says does not show up in the cash flow within a normal hold period. Pay for the school anyway, and you have impaired your basis relative to what the asset can actually produce.

Not to get too technical, but the authors' own explanation for the negative coefficient is endogeneity. School boundaries do not move at random, they shift when enrollment shifts, usually because new supply is being delivered or a neighborhood is turning over. A favorable redistricting event may just be picking up new competitive supply pressure on rents, not a school premium at all. That reframes redistricting as a due diligence flag rather than a value creation lever. It tells you something in the submarket is changing, and you need to find out what before you assume it works in your favor.

​The link is here if you want to go to the source.

In other news, I'm still offering on-demand financial modeling courses that are tailored to the commercial real estate industry. Interested? Check out my "Everything" package; I've distilled years of financial modeling experience into easy-to-follow videos with supporting spreadsheets.