My last post, on urban geographic constraints and housing prices, led to an interesting discussion thread. The most common counterargument was that because dense cities are usually more expensive, density must cause high cost. But if this was true, cities would become cheaper as they became less dense.
Most American urbanized areas have become less dense, not more, over time due to suburban sprawl. Even where city populations have grown, much of that growth has been in areas that where undeveloped a century ago. Thus, the developed part of even growing cities were, I suspect, more dense in 1917 than than they are today: for example, Manhattan’s population peaked at 2.3 million in 1910, about 40 percent more than its current population. So rents should have come down. Did they?
Apparently not. The Census date has statewide data showing that rents rose pretty much everywhere in real terms over the late 20th century. In the District of Columbia, real rents increased from $346 to $612 in real dollars between 1950 and 1990, even as the city was losing population and the region was de-densifying. If Washington is typical, it appears that lower density and higher rent went hand-in-hand.