Project quacks like student housing, and city schools will pay the price.

by Joe Fitzgerald

It took me a long time to figure out why Harrisonburg’s Community Development Department doesn’t use one of the clearest metrics available to predict growth in the city’s schools. It’s similar to the reason no local news media use it. In each case, under the rules of their operations, they don’t have permission to use it.

The metric in question is the correlation between growth of JMU student housing and growth of Harrisonburg City Public Schools’ student population. Since Sunchase opened in 1999, three bedrooms of new student housing generate one new K-12 student as the older housing is gradually repurposed as family housing. If you put new bedrooms next to new students in a spreadsheet and ask the program if they rise together, the program will tell you that they do 88 percent of the time. If you ask the program to build a three-year lag into its calculations, it will tell you that the numbers line up 97 percent of the time. For shorthand purposes, let’s call this the Sunchase correlation.

Saying that one asks the program is an old man’s phrasing. (I recently turned 70, which means that in Biblical terms, I’m in overtime.) Today, one does simply ask the computer that’s running the program. Asking in Biblical times, when Alan Turing, Grace Hopper, and dinosaurs walked the earth, meant writing a procedure to step through the numbers one by one and run a formula. Those of little faith would test the formula with a pencil, and a sheaf of paper, and a cup of strong coffee, and perhaps a slide rule. Later generations used a calculator and an energy drink, which didn’t affect the numbers.

References to numbers rising together or lining up is called correlation in math terms. Correlation is often used badly. For instance, the underside of leaves being visible generally only happens when the wind is blowing. This could lead to the fallacy that leaves turning upwards cause the wind to blow. So the sort of people who want to dismiss a statistic might say that correlation isn’t causation.

They’re right. Sometimes two numbers rise and fall together because a third factor is affecting both of them. Cigarette smoking correlates with lung cancer, but my father, who didn’t want to quit smoking, maintained that we needed a stronger breed of lab rats, or that the rats were born with cancer. This is an extreme example. Perhaps there is a third factor in the Sunchase correlation that drives both student housing and K12. I don’t know what it would be. The growth of student housing is driven by financial opportunity or, in Biblical terms, greed. The growth of K12 is driven by how many people move into the city or are born here.

Weldon Cooper barely mentioned student housing in a $75,000 study of housing and K12 growth performed for the city. They did not mention the Sunchase correlation. Possibly, that means there was no one to tell them about it. They did attempt to figure out how many K12 students reside in areas originally built as student housing. They did this by using the addresses of JMU students, and then admitting they didn’t trust the addresses. HCPS, by the way, used the addresses of K12 students to find out where K12 students live.

One could then make an argument that Weldon Cooper went at it ass-backwards, but more generously and more generalized, they don’t adapt their student projection figures to an individual locality in any meaningful way. They know Virginia and they know statistics but they don’t know, just for instance, Harrisonburg. It they have formulas for projecting or forecasting numbers for Virginia, and college housing doesn’t drive K-12 enrollment in the state, then they’re going to get Harrisonburg wrong or, generously, they’re not going to get it right.

If Weldon Cooper had used the Sunchase correlation in its flawed $75,000 report, this would have given Community Development permission to use it. Community Development does not obviously or apparently develop numbers on its own. They might calculate numbers, but only if the calculations are accepted norms in their field. If they need original numbers, they use studies. The flawed Weldon Cooper study is one such. Downtown 2040 is another.

Downtown 2040 was cited by some planning commissioners in their justifications for recommending approval of The Link. Downtown 2040 does not mention student housing. Taking the less generous route again, a report that does not mention student housing was used as a justification for allowing student housing.

Trigger warning: If you hate numbers and arithmetic, take my word that The Link is a bad idea, quit reading now, and get on with your life.

The developers of The Link originally advertised the project as student housing, and then didn’t. It still is. If a 4-bedroom apartment is assumed, based on every metric and anecdote known to God and man, to be aimed at students, then the maximum of 85 4BR apartments promised by the developer means that student housing constitutes 34 percent of the 250 apartments. The minimum 110 apartments that will be one bedroom, aimed at young professionals, constitute 44 percent of the apartments. If you think that means more young professionals than students, then you might be a planning commissioner.

The single-bedroom apartments would constitute about 20 percent of the proposed 555 bedrooms in the building. The 4BR apartments, 320 bedrooms total, would constitute 61 percent of the bedrooms. The remaining 55 apartments, with 105 bedrooms in them, would, according to Community Development, be two or three bedrooms each. If that’s true, then the average number of bedrooms in each apartment would be between 2 and 3. But putting 105 bedrooms into 55 apartments yields an average of 1.9 bedrooms per apartment.

So let’s go back to the single-bedroom apartments, and remember that 110 of them is the minimum. They could go as high as 140, which is not a randomly chosen number, but one chosen to show what might be coming if this brick-plastered hulk is approved. That’s 30 more single-bedroom apartments, leaving 75 bedrooms and 25 apartments. If you divide 75 by 25, or pay Weldon Cooper $75,000 to do it for you, you get three, which is often the number of bedrooms in a student apartment that doesn’t have four, either because you can’t fit the extra one, you think you can sell it with three, or you’ve made a proffer to Community Development. If those 75 hypothetical bedrooms are added to the 340 in the 4BR apartments, the total of 415 bedrooms constitutes 75 percent of the 555 in the complex.

So now we have a building that is 75 percent student housing. If the Sunchase correlation holds, those 415 student bedrooms will in three years with a 97 percent likelihood, result in 138 new K12 students. The proposed city school budget seeks $48 million in local funding for an estimated 6,350 students, or $7,594 per students. If 138 students are added at that cost per student, that’s $1.05 million more in local funding. That number is not precise. Some costs, mainly personnel, go up when you add students. Some, such as the debt service on the six new schools we’re still paying for, are stable. But while the number may not quite reach $1.05 million, it is going to be substantially more than the $600,000 the developers say The Link will pay in local taxes. That’s a deficit of $400,000 or about two-thirds of a penny in real estate tax, but assume they’ll round up to a full cent. Even the 340 JMU students the developers admit to would add $860,000 in K12 costs, a deficit of a quarter million dollars.

Community Development has not done this math or, if they did, they didn’t tell the Planning Commission. It’s not the kind of projection they do. They depend on, just for instance, Weldon Cooper. They frequently refer to doing the minimum they’re required to do, which may mean they fear lawsuits if they go beyond that, or that no one has given them permission. And traditional media do not do this kind of math. They ask government for the numbers. They have to get permission from the government, if you want to be ungenerous, and it’s probably too late to stop now.

Maybe these numbers are wrong, or should be tempered with some other factors. But they were not mentioned at all during the recent Planning Commission hearing on The Link. They will likely not be mentioned in next month’s City Council hearing on The Link, but someone may surprise us.

Recapping the numbers, The Link could have up to 415 of its 555 bedrooms rented to JMU students, which could add 138 K12 students, which could add a million dollars to the city budget. These numbers will likely only be mentioned by opponents of the project. Maybe they’re the only ones who have permission.

Joe Fitzgerald is a former mayor of Harrisonburg. This column is republished with permission from his blog Still Not Sleeping.