Usual caveat: I am not a supply skeptic. I am a “don’t assume simple solutions to complex problems” person. Supply and demand occur in contexts, and those have an unfortunately large effect on how things turn out.

Modern economists, even those who take Ricardo seriously like Thomas Piketty, tend to be a bit too hasty to conclude that Ricardo was wrong about the scarcity principle of land. In 1817 On the Principles of Political Economy and Taxation, he tackled the problem he saw about the increasing scarcity of agricultural land vis-a-vis growing populations. He was not, unlike Malthus or Smith, willing to conclude that more laborers were the problem. He suggested that increasing the number of people who needed ag land, along with the decreasing marginal productivity of extra land units, meant that adding additional units into production would still, depending on the rate of demand increase, result in increasing scarcity, rents, wealth and market dominance among the land-owning class.

Sound familiar? Fans of Henry George might find a lot to like in Ricardo.

We should point out, as Piketty does, that Ricardo wasn’t the only person observing the economic problems looming in continental Europe, notably France, with growing labor pools driving down wages and driving up rents. Arthur Young wrote about similar themes in his essay, drawing similar conclusions.

Why do economists generally think Ricardo was wrong? Piketty says as much in Capital. I’m not sure why, as they have to know that Ricardo’s situation was pretty exceptional. Not long after Ricardo, Young, and Malthus, ag land suffered a severe decline. But I doubt it was because there was some sort of market correction where landowners decided to start supplying more accessible units to people: early 1800s, ag tech and practices are changing, informed by Enlightenment science (for both good and ill) and better land is becoming more productive so that marginal ag land becomes excess supply, losing value quickly. I suspect prime ag land had a small dip in value for a bit and then continued to rise, and that the overall decline was an average effect. (Also see: more ag prodo getting shipped farther as the century moves on). The point being, we just began needing far less ag land as time wore on, something that continues today.

Ricardo’s point about increasing scarcity even with increasing supply matters to our discussion because it’s one reason why I think short-term effects of more housing supply are just about impossible to predict without better empirical models that consider the possibilities. Ricardo in Principles allowed that when there are extreme shortages of something, a new increment of supply would not drive prices down, but up. The problem is, like a lot of smart fellows back then, he didn’t have reviewers on his ass to explain his thinking there, so he doesn’t. Not very much anyway. But I think there are some key ideas in the theory, especially given Freemark’s counterintuitive findings in Chicago. (Lots of us didn’t feel like they were counterintuitive, but that’s ok.)

Here’s my thought experiment for the day: under what conditions are we likely to see Ricardo’s odd price responses to supply?

When there is population growth in a place, accommodating that population growth any number of things. I’ll collapse it to two scenarios for simplicity in the argument:

Scenario 1, the good marginal context for new supply. In this scenario, the shortage is not that bad or nonexistent, so at the margin, new supply disciplines landlords and we see slower rent increases, if not softening rents entirely.

In this instance, the shortage isn’t holding back any real urban productivity effects we get from including more people, and the increment in new people isn’t having any measurable deleterious impact on wages. These are relatively simple moves up and down demand curves.

I suspect this is why we see so little action in Freemark’s study in Chicago. Chicago’s hole isn’t as deep as the coastal markets, and there are all sorts of reasons why even there, rents and land values might not go down in ways easy to measure to social science studies. For one, prices are only one way landlords can compete. Small marginal changes in local demand might prompt landlords to market more, add a new service, or just stop acting like jerks about things like taking key money or any of the side hustles landlords with sufficient market power can grab.

Scenario 2: The more dicey context for new supply. In this scenario, which I’d say quite a few coastal cities are likely in, the shortage is profound. New housing supply is, as they say, a drop in the bucket. Housing shortages are cutting into the overall productivity of the metro area because they are dampening labor productivity by keeping productive laborers out. New supply is likely to go in at a market rate, which is expected at or even above where local landlords are, and they take their cues from the higher rents, not the possibility that their unit is going to sit empty. Why not? There is still no real supply competition because the shortage is so grave that any real price competition is years out, if ever. If anything, they have a motivation to grind harder now in anticipation that supply competition may be coming, but not for some time yet.

I’ve written before that infill housing is likely to increase local amenities and local landlord’s expectations about what they might charge. Various people have argued with me that the new supply effects should easily outweigh new amenity effects and that new supply should discipline landlords. That is entirely possible in the first scenario. In the second scenario, that is not necessarily true. Moreover, productivity effects are not amenity effects. Productivity effects are the bump land prices get because of getting spillovers from their location in an economically productive region. Amenity effects are the bump in price parcels get from being located reasonably proximate to nice things. Both can and likely do happen with new supply, and their size relative to any possible supply competition effects strike me as much more significant in places where landlords are secure price setters.

In the first scenario, by contrast, I could easily see supply competition cancel out amenity effects. I have less confidence it would cancel productivity effects in the long term across business cycles. That’s another essay for another day.

Laid atop these differences are submarket problems. This is another layer of complexity, but suffice it to say, we don’t have great information on the degree to which locations are substitutes in residential location decisions. At all. Again, added complexity for another day.

And again, this is not “don’t build.” That just makes the hole bigger. Instead, I’m of the “help low-wage laborers across the board with rental supports, unions, and a robust welfare state” argumentation.