Failed generalizations

Physics doesn't lack for predictive theories of physical nature, but if it were evaluated on its ability to predict the weather, we'd call it a dismal science. Economics doesn't lack for predictive theories either, it's just that we evaluate it on its ability to predict reality. Reality is only partly predictable.

Looking at the history of the LES in the 19th century, you see a consistent pattern of quality of living space declining in inverse relation to density and rent. Demand at the bottom of the social scale was sticky -- choices were limited by the lack of convenient transit and work was concentrated between the downtown docks and downtown industry. Ghetto construction was structurally uniform for each decade. Outside the immigrant quarter, amenities chased big money as you'd expect in an elastic market: a broad gradient from middle-class town houses to immense mansions. 

Both trends are economically predictable, perfect fodder for economic theory. But there's a change in the LES that doesn't fit the pattern. After around 1910, virtually nothing is built, even though the American economy continues to grow. Not surprising, the reasons for the end of development are not economic. One was a technological innovation only marginally unrelated to economic production -- the subway system, which increased mobility and commerce but not so much production. It allowed labor to live far afield. Quotas on immigration in the 1920's turned the ghetto from a high-demand exploitative rent district to a low-demand, low rent district as residents left for better living spaces and no new immigrants replaced them. Less obvious was the third tenement house New Law act, which required so much courtyard space that landlords couldn't develop on single lots anymore, curtailing single-lot development. You can see it for yourself on 1st Avenue -- rows of four- and five-story Antebellum tenements on single lots. 

Disinvestment in the ghetto was not the consequence of a shift from capital. That would be a backward analysis. Disinvestment was the consequence of political policy (labor protectionism, anti-immigrant eugenicism paralleling growing isolationism), a technological advance in mobility, and an entirely unintended consequence of a progressive movement to improve labor housing (again, politics) with the New Law that made it harder to develop downtown. 

The historical lesson I get from it: politics, technology and public policies like zoning and housing laws have had more profound local consequences than constants like capital growth. Culture plays a role as well. Gentrification can no longer be described as a single economic phenomenon if the youth culture of Williamsburg drives upscale families to seek child-friendlier neighborhoods. That has consequences for construction, schools and commercial character. Until an economist comes up with a theory that explains why a fix-wheel bike with no brakes designed for race tracks with no inclines and no stop lights would be the trend among upscale youth in urban centers filled with lights and steeply inclined bridges, economics will be stuck with "60% chance of rain today."

Will the future be like the past? You can count on the constants, so maybe that's where policy should target, always bearing in mind the law of unintended consequences rules. The most depressing documents are the urban planning proposals of the past. In retrospect they look completely wrongheaded, as if their authors didn't have a clue.

That's why local community self-determination has promise. It's not urban planning from above; it's urban needs from below where life actually happens.