A short note on the impact of the housing crisis on wages in tight labor submarkets, as a follow up to a Twitter conversation about the impact that moving to an expensive city has on disposable income.
In The Rent is Too Damn High, Matthew Yglesias argues that you can’t command higher wages to live in a high-cost city, even if your skills are relatively scarce, because employers roughly “pay workers according to how valuable their work is… nobody is going to offer you a higher salary to offset your high cost of living just to be nice.”
Anecdotally, I can tell you this is not true. Working in an industry where there is currently high demand for certain niche skills and limited supply of workers, we have to offer substantial pay raises to get workers to relocate to LA. My engineering skills are not particularly specialized, but you’d still have to pay me more if you wanted me to relocate to SF or NYC or some other place more expensive than LA. We’re not offering higher salaries to be nice; we’re doing it because the high cost of housing gives workers the ability to cut into our producer surplus.
The reason is that labor has a supply curve too, and the cost of housing is an input into the cost of supplying labor. If housing is more expensive, less people are going to be willing to relocate to the city. Obviously, the amount of money it would take to convince a worker to move will vary – some people don’t want to move to LA for reasons other than cost, some people really want to move to LA despite the cost – but overall, workers will demand higher compensation to relocate to higher cost cities.
The increase in pay that workers can demand depends on the number of potential workers and the demand for their skills. For jobs that many people can perform, modest increases in wages should bring about a decent increase in the number of workers. On the other hand, for niche skills, the supply curve might be pretty steep, and workers may be able to command substantially higher wages. On the demand side, if employers can easily automate work or move it to another city, they may not offer much in the way of wage increases. But if the work cannot easily be moved elsewhere, demand will be less sensitive to wages and workers can again demand substantially more money.
Employers almost always have some alternatives to raising wages. They can try to have the work performed remotely; the reduced productivity of remote work may be more than offset by wage savings. They can forgo the work entirely; in some cases it might be impossible to do the work profitably.
All of this implies that high housing costs will be borne jointly by workers and employers. Higher paid workers with niche skills will be able to force more of this cost onto their employers, though by no means all of it. The cost to employers becomes a deadweight loss, money that could have been otherwise invested in expanding a business and employing more people. Lower income workers are affected worse than higher income workers, making solving the housing crisis even more important.
That the realized set of workers in expensive cities is paid more than the realized set in less expensive cities doesn’t actually mean those workers got to demand they be paid more! It is just an equilibrium outcome: we only see engineers show up in $$ cities if a firm is productive enough to put them to good use.
Suppose a firm has a task that you can perform remotely so you can live wherever you’d like. They are going to pay you according to the value of the task, not according to where you choose to live: i.e., if you choose to live in SF they are not “going to offer you a higher salary to offset your high cost of living just to be nice.” I think this is what Yglesias is saying?
Anyway this is why math is useful because it makes clear what one means by “offer a higher salary just to be nice”!
Do they really pay the worker according to the value of the task, or do they pay according to the wage the worker will accept, with the difference being profit?
In some sense, is this a question of semantics? The equilibrium in $$ cities is different because of the cost of supplying labor, right?
It depends on why $$ cities are $$! To say something like “workers will demand higher compensation to relocate to higher cost cities” one needs a model for *why* those cities are higher cost. I can easily write one where they don’t (higher-cost cities have nice amenities => workers accept identical or even lower compensation in higher-cost cities). I think this was what I found objectionable in the first place..
But yes I think this is maybe mostly a question of semantics–and that’s why math is nice, clarifies the semantics!!
»(higher-cost cities have nice amenities => workers accept identical or even lower compensation in higher-cost cities)
Which basically means that the workers are paying for these amenities—not with an entrance fee like in Disneyland, but with higher housing costs. And paying for something requires that you are able to afford it. If you only earn a low wage, you have to think twice about what you can afford.
The most drastic outcome: You can’t sell below your own manufacturing costs; and the low-wage worker won’t work if housing costs more than (s)he earns.
I reject the premise that higher costs are explained by amenities!