Brian Keegan: Rethinking Boulder’s growth debate — with data, not nostalgia

Brian Keegan is a regular opinion columnist for Boulder Reporting Lab. His “Charting Boulder” column uses public data to make sense of how the city is changing — from housing and politics to income and population — with clear explanations and a focus on equity.

One of my goals in writing this column is to raise the bar for how our community makes — and evaluates — arguments. In a world where political and media claims are often repeated until they become “facts,” we owe it to ourselves to examine these assumptions closely. Are these beliefs still true? Were they ever?

Another part of this mission is accountability. Today, we have more data and better methods for understanding our community than ever before, but fewer ways to turn facts into change. Politicians are more insulated, reporters are stretched more thinly, and scientists face greater pressure to avoid public interest topics. Yet science, at its core, belongs to all of us. That means scientists should engage the public, and the public should hold scientists to high standards.

To do my part, I make two promises to you:

First, I will use public datasets and share all my data and code openly. Our governments (federal, state, county and city) release an abundance of data, sometimes in accessible formats like CSV or JSON, sometimes buried in PDFs or audio files. Using these data resources is not just about transparency or curiosity; it’s essential for keeping them public, as there are very real efforts to enclose or destroy open data. You can always find the data and code for these columns (usually Jupyter Notebooks in Python) on my GitHub repository. Pull requests are welcome!

Second, I invite your scrutiny and feedback. I am human — mistakes will happen, and I am committed to correcting them. An already promising part of writing this column is hearing from readers. Your questions spark important conversations that go far beyond what fits in one column. Many of your insights will — and should — shape future contributions.

My June 29 column challenged the widely held belief that Boulder is “full,” using decades of population data to show how slow-growth policies have stalled growth, shifted development to neighboring communities and weakened Boulder’s long-term vitality.

Here’s a look at some of the feedback it received, particularly around Boulder’s plateauing population growth:

CU student population

Many of you pointed out that my last column did not examine the impacts of CU Boulder’s enrollment, often with the assumption that the university disproportionately burdens the city. (Full disclosure: I am faculty at CU Boulder.) Conflicts between a university and its neighboring community — often called “town-gown tensions” — are hardly unique to Boulder, and several readers asked how CU’s student numbers shape our city’s population.

Combining CU’s historical enrollment data with the State Demography Office’s annual municipal population estimates since 1980, the figure below plots CU enrollment as a percentage of the city’s population. CU’s student population in 2023 was at its highest fraction since 1980. 

But there are some caveats to consider. First, not all CU Boulder students reside in Boulder. CU’s estimates of where students live since 2017 show approximately 15% of undergraduates and 25% of graduate students live outside of Boulder. 

Second, CU’s enrollment hasn’t grown steadily, and its future is in flux. There were several significant declines in the late 1990s and early 2010s. And while the Prime Effect is real, the consequences of Trump 2.0 policies and the impending “demographic cliff” are just beginning to affect CU enrollment. 

Finally, research universities are essential shock absorbers for cities and regions, helping to cushion against economic downturns, policy shifts and demographic changes. More students means more customers in local businesses, more tax dollars for our governments, more skilled graduates, more progressive citizens and a stronger university, especially as TABOR continues to strangle higher education funding in Colorado.

Jobs-housing balance

Another recurring critique argues the number of jobs within the city is greater than the amount of housing. The argument goes that the city’s encouragement of job growth without adding enough housing creates a mismatch that drives up housing costs and contributes to sprawl, traffic and pollution. In other words: Because Boulder is already full, send new jobs somewhere else to stop attracting new people.

While historical data on employment are readily available for the county, numbers about jobs within the city are harder to find.

Then I came across a gem of a Census dataset with the pithy name “Longitudinal Employer-Household Dynamics Origin-Destination Employment Statistics” (LODES) that describes employment “where workers are employed and where they live” with data from 2002 to 2022. Dividing the number of jobs from the “workplace area characteristics” by the SDO’s municipal housing data (only available back to 2010), the figure below is one way of capturing the number of jobs per housing unit in Boulder since 2010.

This jobs-housing ratio increased modestly from 2010 to 2019 before crashing during the Covid-19 pandemic and bouncing back to pre-pandemic levels in 2022 (the last data available). While the trend is upward, the change in absolute terms is modest: from 1.86 jobs per house in 2010 to 1.93 in 2022. It is certainly not a flood of Silicon Valley refugees commuting in from Louisville, like some overheated rhetoric would have us believe.

Suburbanization

One of my favorite critiques came from a Redditor who piquantly argued that my column purposefully misunderstood “the most basic contours of suburbanization.” They claimed that the tapering growth of the city relative to the county “happened to every single city” and should not be attributed to Boulder’s adoption of slow-growth policies in the 1970s.

This is a genuinely interesting question. To understand whether Boulder was just a normal suburbanizing metropolis, I created a “basket” of 24 mid-sized peer cities that also contain major universities as a comparison. Using historical data from the National Historical Geographic Information System, I retrieved the diennial population for each of these cities and their counties since 1970 (no place-level data is available before then). The population growth of the cities from 1970 to 2020 was divided by the growth of their counties: Did university towns keep up with their counties? The figure below ranks the ratios of these city-to-county growth rates.

Boulder still comes in as the third slowest-growing university city relative to its county in this basket, which is right alongside similar growth-anxious communities like Berkeley, Santa Barbara and Mountain View. However, other college towns like Eugene, Iowa City and College Station grew more quickly than their surrounding counties. It turns out different cities have different historical patterns of suburbanization.

A single column cannot answer every question, nor is it the final word. But your feedback and the discovery of other local data resources (Downtown Boulder Partnership has fascinating pedestrian count data!) could sharpen our debates. As Boulder’s demographics, politics and economy change, so should the way we use local data resources to guide community decisions — like the Boulder Valley Comprehensive Plan that will shape development for decades to come.

That’s the promise of public interest data science: not just to mythbust, but to make arguments about our community more accountable and actionable. By openly sharing data, code, precedent and other evidence, we can create more — dare I say — enlightened civic values where empirical scrutiny is the norm and hazy nostalgia and lazy ad hominems are exceptions. 

So keep the feedback coming, and let’s hold each other to higher standards in our arguments.

The data and code for replicating these analyses can be found on GitHub.