How To Solve Housing With Capitalism

My grandfather, Conrad Pat Harness, came back from the Sicily campaign of World War II to Washington DC and found himself a real estate journalist for the Washington Post. This led to a long career in home building, and ultimately a position on the board of NAHB. My father, Chris Harness, has been home building for my entire life. By our family’s count, his footprint is over 20,000 new units across a long career.

Generations of Harnesses have focused on housing. It's in my blood, on my mind when I wake up. It's a passionate obsession.

We have a massive housing gap in the United States. I won't hang this essay on one national number: every estimate of the shortage depends on what it counts — households that never formed, the vacancy rate you decide is healthy, homes that wore out — and the estimates disagree. What I will defend is what I have watched from the building side my whole life: where people most want to live, homes get approved and built more slowly than people arrive, and rents and prices do the rationing. The question is whether the tools of capitalism — speed, incentive, competition — can be pointed at that gap deliberately enough to close it.

Start with the incentive, because it is the part people get wrong. Capitalism does not build housing because housing is needed. It builds housing when a deal pencils, and a deal pencils one home at a time. Here is an illustration — round numbers I chose, not observed prices. An acre costs $1,000,000. Construction runs $150,000 a home. If the zoning on that acre allows 24 homes, the land costs $1,000,000 ÷ 24 = $41,667 per home, and each home costs $191,667 before soft costs, interest or profit. Let the same acre hold 72 homes and the land falls to $13,889 per home; the total falls to $163,889. Same dirt, same contractor, same capitalist: $27,778 less per home, about 14.5% ($27,778 ÷ $191,667), just by changing the denominator.

What happens next is a prediction, not an observation. A lower cost per home lets a developer accept a lower rent and still clear the return their capital demands, or build on a site that never penciled before. Rents only come down if enough of those deals actually get built. And the illustration leaves out everything that can eat the savings: denser buildings can cost more per home to build, parking takes land, a real site may not fit 72 homes, and capital runs out first. A $10,000,000 budget funds at most ($10,000,000 − $1,000,000) ÷ $150,000 = 60 of those 72 homes. The law can allow homes that nobody can afford to finance.

Speed is the other tool, and it is the one TestFit is built for. Every week a site sits between “interesting” and “decided” costs money — carrying the land, the option, the team’s time — and most sites, in my experience on the software side, get looked at and dropped. The faster a team can see the per-home math on a site, the more sites it can test before one gets bought. That is what TestFit is trying to do: make the economic case for building housing so fast and so clear that developers can't afford not to act. That is a bet, not a result. Software doesn't pour foundations.

So here is the judgment. Capitalism will build housing the moment the per-home number works, and that number is set less by greed than by arithmetic: what the land costs divided by how many homes the law lets onto it, plus how long it takes anyone to find out. If you want more housing, fight about the denominator and the clock.