I want to start by noting, as mentioned in the article, that economists do not consider rent control as just an example of an undesirable price ceiling, but as a canonical example. There are many other potential examples, such as gas prices in the 1970s, or Nixon's wage and price ceilings, but economists haved picked rent control as the example as the best teaching tool to show that price ceilings are socially undesirable. Thus, we can look at the article first as criticizing this pedagogical choice. If removing rent control in typical, real-world situations does not typically result in increasing social welfare, but merely transfers surplus from consumers to suppliers, then it is a poor pedagogical tool, akin to trying to teach gravity by having students think about the behavior of airplanes (which do not, of course, fall when you drop them). In this regard, I think the authors succeed.
But I want to take a closer look at price ceilings in general. In general, economists talk about consumer and producer surplus. For consumers, this surplus is getting some good or service for less than they would have paid. I really love beer, for example. I would spend \$5 on the first beer of the day, but all my beer costs \$0.50. For my first beer, I'm getting \$4.50 of "surplus." For producers, the surplus is the difference between the marginal cost and the price. The brewery can make some of its beers for only \$0.45, so they're getting \$0.05 of surplus on those beers. (They sell a lot more beer than I drink, so that the \$0.05 adds up quickly.) We show this concept like this:
The shaded areas represent "surplus", a.k.a. "welfare", benefits that the consumers and producers get "for free". It is generally accepted that, absent externalities (people who gain or lose who are neither producers nor consumers of the good) and transaction costs (costs paid by the consumer that do not benefit the producer), the point where the (marginal) supply and demand curves meet, the market price, represents the price and quantity that maximize total welfare (consumer surplus + producer surplus).
Rent control is commonly modeled as a price ceiling, where the government arbitrarily sets a price for a good below the market price.
The canonical model of a price ceiling looks like this:
Here, the government has set a price ceiling, Pc. Because we don't force firms to produce at a loss, firms produce a smaller quantity if they can sell only at the price ceiling. Some of the welfare is redistributed from producers to consumers, but more importantly, some trade is just lost; welfare just vanishes. This lost welfare is labeled "Deadweight Loss" is the graph. At higher prices, producers are willing to produce goods, and consumers are willing to buy them, but the price ceiling prevents these welfare-generating trades. Economists see this loss of welfare as a Bad Thing, and thus argue against price ceilings.
Note too that at the restricted quantities, enough consumers are willing to pay a lot more, the "black market price" (Pb), than even the market price. This willingness can lead to illegality and corruption. Also, at the price ceiling, a lot more consumers are willing to buy (Qb) but cannot. Price ceilings can lead to a lot of problems all around.
But, like all economic graphs, the graphs above lie: specifically, they make assumptions about the supply and demand curves that are not always (indeed rarely) true, notably that they have a slope of about +/-1. I argue that in many cases, in the short term, the absolute value of slope of the supply and demand curves for rental housing is much larger than one, i.e. both are relatively inelastic.* The reason is that in the short term, i.e. holding the level of capital constant, we cannot build new buildings (since new buildings represent new capital). Thus, when there is increased demand, all we can do is use labor, a lot of labor, to rehabilitate the few older buildings back into economic use. Furthermore, there are cities, such as New York or San Francisco, where people will pay a lot to live there, especially at the margin (i.e. a new person coming to the city will pay a lot to move there). Hence, a small shift in the demand curve, such as when Google decides to add a thousand new engineers, whom they can pay enough to cover any rent increases, can have a dramatic impact on the market.
*Note that unless we're using a log scale, the slopes of curves do not directly represent elasticity, which is percentage change in the dependent variable per percentage change in the independent variable, which depends not only on the slope but the level. But slope is a pretty good proxy for elasticity.
I look at rental markets like this:
(Note that my graph is too small to show the short run quantity, which is between Qe and Qlr.)
Here, a small shift in the demand curve, from D to D' causes a large increase in the short run market price. I'm obviously oversimplifying here: e.g. not all rental housing is alike, but I think the graph gets the basic point across.
There is a perfectly good neoclassical economic argument that the government should do something here; the market, by itself, is insufficient to achieve the socially optimal outcome. The story includes the enormous transaction costs associated with moving, especially to a different city. The story goes like this: Google hires a thousand engineers, causing a shift in the demand curve. Rent goes to Psr, pricing a lot of people out of the rental market, people who could afford to pay the long run rent, Plr, and the city fills up with people who can afford to pay the short run rent. In the long run, rent decreases to the long run level, but people who could have paid that rent are long gone. Hence we incur wasteful transaction costs because of the short- and long-run disparity in rent. The government could prevent these wasteful transaction costs by imposing a price ceiling on existing rents (where no additional cost is necessary to bring the property into service), letting them rise to the long-run level, but not to the short-run level. New residents, who can afford it, pay the short run price in the short run, but eventually pay the long-run rent. This is, essentially, what most rent control regimes actually do.
Second, there's a political economy argument. We do not live in a laissez faire political economy; we live in a capitalist political economy, where people who own capital have political control. Furthermore, residential real estate is fundamentally different from an ordinary firm. The present value of any business is the discounted expected income stream of the business. For example, if my business runs a profit of \$100000 per year, and I discount future income at 2.5% per year, then the value of the business is $100000*∑↙{n=0}↖∞ (1-0.025)^n = $\$4 million. A normal business (like a sandwich shop or automobile manufacturer) has to actually be operating to have an expected income stream; the expected income of a closed sandwich shop is 0. However, real estate does not have to operate to have an expected income stream. An occupied and an unoccupied building have about the same value. A building owner can realize the value of his building just by borrowing against it (and paying minimal maintenance costs) rather than letting nasty tenants poop in the toilets. In a depressed economy, buying a building and leaving it empty is a viable investment scheme, as long as few enough people actually do it.
Building owners, then, have a double incentive to enact supply limits (shortages), both market-based and just by taking property off the market, to essentially make the long-run supply curve match the short-run supply curve, essentially creating a quota (which works the same as a price ceiling, except the producers expropriate surplus from the consumers).
Again, rent control can mitigate this market failure both by limiting the incentive to create shortages and by forcing buildings to be occupied to be valuable.
So... that, I think, is why rent control is a bad illustration of a quota, and why rent control is a Good Thing when both demand and supply are highly inelastic in a particular market.
