Two months ago, I blogged about the com­mon argu­ment that cities should­n’t allow new hous­ing because new res­i­dents might over­load their infra­struc­ture. I raised a vari­ety of coun­ter­ar­gu­ments, but it seems to me that one point deserves more atten­tion than I gave it.

In par­tic­u­lar: there is almost no place in the Unit­ed States where new res­i­dents don’t affect infra­struc­ture needs to some extent. For exam­ple, in a city with decent pub­lic tran­sit, new res­i­dents will mean more peo­ple on the sub­ways and bus­es. So anti-hous­ing activists can plau­si­bly com­plain that new res­i­dents will over­load the trains and bus­es. Even though it is not tech­ni­cal­ly cor­rect (even in New York City) that trains and bus­es have as many pas­sen­gers as they can han­dle, it is true that more rid­ers mean more peo­ple sit­ting next to you, or maybe even a high­er like­li­hood that you may have to stand on the train or bus. 

But in a city or sub­urb that is high­ly depen­dent on roads and high­ways, new res­i­dents might mean more cars on the roads and thus more traf­fic con­ges­tion. So there­fore, it could be plau­si­bly argued that noth­ing should also be built in auto­mo­bile-depen­dent sub­urbs.

And we can’t avoid the prob­lem by build­ing in rur­al land. In fact, new hous­ing in rur­al areas might require even more infra­struc­tur­al improve­ments than new hous­ing in exist­ing sub­urbs, because rur­al areas often lack sew­ers, and often have two-lane roads instead of major high­ways. So new sub­di­vi­sions in rur­al areas might require even more sig­nif­i­cant infra­struc­ture improve­ments than new sub­di­vi­sions in sub­urbs. 

Trans­porta­tion infra­struc­ture is just one type of infra­struc­ture. Schools are anoth­er type of infra­struc­ture that is affect­ed heav­i­ly by growth; new res­i­dents mean more chil­dren in the schools, and thus risk mak­ing schools over­crowd­ed if they aren’t over­crowd­ed already.*

So the “infra­struc­ture” argu­ment against new hous­ing is one that sug­gests no hous­ing should be built anywhere—obviously an absurd result. 

It seems to me that a fair­ly sig­nif­i­cant body of com­men­tary sug­gests that sprawl cre­ates more need for new infra­struc­ture than does devel­op­ment in exist­ing areas: for exam­ple, if gov­ern­ment per­mits a new low-den­si­ty sub­urb in the coun­try, it might build new roads and sew­er pipes that might not be nec­es­sary if hous­ing is built in exist­ing areas. So if our only goal was to lim­it the need for new infra­struc­ture, we would lim­it new hous­ing to already-built-up areas.

But the cur­rent zon­ing sys­tem pro­motes the oppo­site result. Here’s why: if neigh­bors are giv­en veto pow­er over new hous­ing (as is often the case), we may actu­al­ly need more infra­struc­ture than if the free mar­ket held sway. Why? Because if new hous­ing is built in low-den­si­ty rur­al areas, there will be few­er neigh­bors around to object. Thus, a region where neigh­bors are giv­en veto pow­er is one which has more hous­ing in corn­fields and less hous­ing in cities and exist­ing sub­urbs, which in turn means that soci­ety’s col­lec­tive infra­struc­ture bill will be larg­er. 

In sum, anti-hous­ing poli­cies nev­er real­ly reduce soci­ety-wide infra­struc­ture costs; at best, they shift those costs to more per­mis­sive juris­dic­tions. 

*If anti-hous­ing activists real­ly believed that school over­crowd­ing was the only rea­son to oppose new hous­ing, they would sup­port hous­ing tar­get­ed towards sin­gle peo­ple instead of families—for exam­ple, small apart­ments. But in gen­er­al, apart­ments are even more polit­i­cal­ly radioac­tive than hous­es. So I sus­pect that con­cerns over schools are either a) not real­ly a major rea­son for anti-hous­ing activism or b) insin­cere.