Higher Education, Higher Rent Burdens?
Getting to the Bottom of the Student Household Distortion in Bozeman
Off-campus student households make up a significant fraction of renter households in Bozeman.
Income data for student households provides a very misleading account of their true financial condition.
Levels of rent burden need to be adjusted for this distortion to understand the rental housing situation.
In a previous post, we urged the City to do the analytical work to make a proper diagnosis of the causes of the housing cost stress that Bozemanites have been experiencing before embarking on treatment that presupposes the cause is overly restrictive zoning. A prerequisite to a proper diagnosis is a deep understanding of the symptoms. Aggregate numbers on rent burdens, for example, may suggest one story, while breakdowns by household type may require a reconsideration of that story.
One complication that faces anyone trying to understand the housing market in Bozeman arises from the presence of Montana State University. Student housing represents a very distinct submarket with its own dynamics. Student households, moreover, exhibit very different financial characteristics than non-student households. Mixing all households together will paint a misleading picture of the housing market and housing cost burdens. Yet, the City’s work has mostly failed to consider this complication.
The Census Bureau, on the other hand, has in recent years done illuminating work on this very topic. In 2018, it published a research piece (1) on the importance of considering the student distortion when looking at poverty statistics in college towns. The Bureau’s analysis mentions Bozeman directly as one of the cities where the poverty rate was most heavily distorted by the presence of off-campus student households. The crux of the issue results from the fact that most student households’ cash flow is not tightly connected to reported income. Statistics will consider a couple of students living in an apartment together as a “household”, even if they all remain financially dependent on their families and show almost no income. Students also may rely on student loans (to be paid back out of future earnings boosted by their degrees) for paying expenses during college or graduate school. Loan proceeds do not count as income. Therefore, most student households show extremely low incomes, often below the poverty line, and high rent-to-income ratios, frequently in excess of 100%, even though these households may not be truly suffering poverty or a high housing cost burden relative to available resources. Here you can see the impact on Bozeman’s poverty statistics, as presented by the Census Bureau’s 2018 study.
American Community Survey results across the US available here.
Excluding off-campus student households cut the poverty rate by over 40%! During the study period, 16.4% of households in Bozeman consisted of off-campus students, and those households had a 54% reported poverty rate. In case a reader might be tempted to believe that 54% of student households genuinely fell below the poverty rate, we adduce data from the Department of Education (2) that show that only 19% of MSU students received an income-tested Pell Grant and only 33% of MSU students receive any federal loans. MSU enrolls what appears to be a student body with below-average financial need. Clearly, we are dealing with a large statistical artifact that will distort all poverty and housing stress statistics.
The obvious follow-up question is: how large a distortion does the presence of this student household “artifact” cause in Bozeman’s metrics of housing affordability and prevalence of rent-burdened households?
Unfortunately, regular Census cross-tabulation does not allow any easy way to exclude student households. Instead, we built an analysis using Census Public Use Microdata (PUMs). We could look at data in a much more granular way, with the one downside being that PUM areas (PUMAs) do not match up exactly with cities or counties. However, the PUMA that contains Bozeman is a pretty good proxy — Bozeman city contains in excess of 60% of the rental housing units in the entire PUMA.. We were able to go from the “headline” rent-stress numbers to numbers reflecting subpopulations that might be of greater interest when considering whether we have a rent affordability problem. That is, market rent will always appear unaffordable to students showing no or little income. Rent-stress caused by high rents will show up in high levels of rent-stress among the employed population.
The chart below illustrates what we found in Bozeman, using the latest (2022) PUMS data. We have included aggregate numbers for renter households in the US as a whole as a point of reference. First, we look at the median gross-rent-to-income ratio of renter households of various types: all renter households, student renter households (defined as households where the primary householder is a higher-education student and at least 75% of the residents of the unit are students), non-student renter households (renter households that do not fall within the student household definition), households with at least one full-time (35+ hours per week of work) worker, and households with a gross income equivalent to what two full-time workers would earn at $15/hr jobs.
The median gross-rent-to-income in Bozeman comes out as below the US average, even with students in the mix. But student households have an absurdly high rent-to-income ratio, for the reasons we explained above – it is a statistical artifact. For non-student households, the median rent-to-income falls even further below the nationwide median.
The picture resolves even more clearly when we consider another often-cited statistic related to housing stress, namely the percentage of “rent-burdened” households. The thresholds most often cited are a gross-rent-to-income ratio of 30% (“rent-burdened”) and of 50% (“severely rent-burdened”). As an aside, we call the reader’s attention to the fact that the 30% threshold level comes from the Department of Housing and Urban Development without much concrete justification. In fact, 30% is just about the median gross-rent-to-income level in the US in the past few decades. In any case, using those thresholds, we report in the chart below the percentage of rent-burdened households in each category of household.
Even without adjusting for the student distortion, Bozeman PUMA has a lower percentage of renter households above either the “burdened” or “severely burdened” threshold than does the US as a whole. But student households cause a very large distortion. Nearly 80% of student renter households count as “burdened” and well over half as “severely burdened.” Filtering out students makes a very meaningful difference, particularly in the “severely burdened” category. About half the severely burdened households in Bozeman are student households, whereas student households only represent around 20% of renter households in the Bozeman PUMA! It is also very interesting to see that for households in the area that earn the equivalent of what two full-time workers making $15/hr bring in, severe rent-burden is basically non-existent.
Bozeman PUMA’s rank among all US PUMAs looks very different when we consider the entire population, just non-students, and finally students. Where affordability concerns appear the most acute (student households) is the very segment of the market where the standard affordability statistics prove least reliable. And even if they did accurately represent reality, does solving the housing challenges of one institution’s customers (i.e. MSU’s students) rightly fall to the city? MSU is an institution that serves the entire state and beyond.
All that may come as a surprise to readers who hear a lot about rent burdens in Bozeman. Yet, the academic literature provides a clue to this conundrum. The places that have the highest percentages of severe rent-burden are not, as one might naively suppose, expensive, high-opportunity cities. They feature high rents, yes, but not always the highest percentages of rent-burdened households. In fact, the places with the highest fraction of burdened households tend to be places with cheap rents. The problem is that these places also have extremely high levels of single-parent households and households disconnected from the labor market. According to one detailed study, “even substantially relaxing land use restrictions and regulatory barriers in development is unlikely to reduce the cost of construction enough to significantly reduce housing affordability problems for many single-parent households. In most of the country, the problem for these renters is not one of insufficient supply of housing but rather one of insufficient incomes.” (3) Rent burden represents a ratio. The numerator contains the level of rent, and the denominator income. The problem of severe burdens can result from the numerator, the denominator or both. A careful policy maker will analyze how much of a rent-burden problem relates to problems in the housing market and how much due to problems in the labor market. Have Bozeman’s policymakers done so?
A final note on the rent-burden numbers. The PUMs data comes from 2022. Does that make it stale? Unfortunately, it will be some time before the Census Bureau releases the 2024 numbers. They publish with a long lag. We can, however, see how rents and wages have evolved in the Bozeman area at a high level since 2022, because ApartmentList and the Bureau of Labor Statistics publish higher-frequency data on each, respectively (ApartmentList covers Bozeman City specifically, BLS tracks Gallatin County).
Wages have outpaced rents since 2022 in Bozeman, and by a larger margin than has obtained in the US as a whole. It would be reasonable to conclude that rent burden statistics will show improvement from the 2022 numbers when more recent data becomes available.
We have written before that an effective treatment plan requires an accurate diagnosis. An accurate diagnosis, in turn, presupposes a detailed understanding of the presenting symptoms. In a market with a heavy presence of student housing, the symptom of aggregated elevated housing-stress statistics is not highly informative. A careful analyst must try to abstract away the powerful distortionary effect of student households on all metrics of housing stress. In Bozeman’s case, we can see that a great deal of the measured stress in the rental market comes from student households, whose reported household income does not represent their true situation in most cases. The city must likewise account for the student housing distortion in formulating goals for alleviating housing stress. WIth nearly 20% of rental households consisting of students, and nearly all of those households exhibiting “statistical” rent stress, Bozeman will never achieve a report level of “rent-burdened” households of less than 10%, even if rents were to fall to the point that no non-student households experienced rent stress.
Author Benjamin Heller, Bozeman Resident
(1) Bishaw, Alemayehu and Benson, Craig, Small and Large College Towns See Higher Poverty Rates (2018). Linked here
(2) Data from US College Scorecard (https://collegescorecard.ed.gov/school/?180461-Montana-State-University)
(3) Ghent, Andra C. and Leather, David, Is America’s Housing Affordability Problem a Housing Problem? (August 25, 2021). Available at SSRN: https://ssrn.com/abstract=3761305 or http://dx.doi.org/10.2139/ssrn.3761305