Pur Autre Vie

I'm not wrong, I'm just an asshole

Wednesday, September 18, 2019

Long Sigh

I am perhaps unreasonably angry about two Kevin Drum blog posts (first, second). In the first post, he writes:
Let me put this in concrete terms. If you truly believe that climate change will broil the planet in the next 50 years or so, the very least you should do is immediately get rid of your car and adopt a vegan diet. How many of you have done that? How many of you have even considered it? Virtually none of you.

In the second post, he writes:
Climate change isn’t a game, and trying to make people feel bad about living their lives isn’t going to increase support for the kinds of things that really make a difference.
Okay.

Now in fairness to Drum, his consistent position appears to be that no one should feel obliged to change behavior to fight climate change, and no one should support policies such as carbon taxes that might address climate change by altering the behavior of large numbers of people. What is called for is the invention of technologies that will allow us (meaning affluent Americans) to keep our current lifestyles while limiting global temperature increases to manageable levels.

Still, if that's what you believe, I think it really behooves you not to write troll-bait like the passage I quoted from Drum's first post. I suppose his (honest) response would be something like, "For a lot of people, if they don't hate-read me they won't read me at all," but it is really very unhelpful to the underlying substantive argument. (Which, by the way, is stupid as hell.)

If We Had But World Enough and Time

From time to time I am asked to interview prospective hires at the law firm where I am employed. This is usually depressing because the interviewees are young and bright and interesting, and they would generally be happier and more productive doing something other than practicing law. But you've got to feed the monkey.

Anyway in one particularly depressing interview, the interviewee explained the biscuit conditional to me. The biscuit conditional is a linguistic concept illustrated by the following Demetri Martin joke (which I am paraphrasing):

I went to shop for clothes and the woman in the store said, "If you need anything my name is Kim." This startled me because I had never met anyone with a conditional identity before.

The point of the biscuit conditional is that an "if" statement sometimes pertains to the relevance of the information, not the information itself. The traditional example is: "If you're hungry there's a biscuit on the table." Of course there's a biscuit whether or not you are hungry, but the information is only relevant to you if you're hungry. Similarly Demetri Martin only needed to know Kim's name if he needed to ask for help.

Anyway my point today is that biscuit conditionals are a kind of exception that proves the rule, the rule being that statements contain an implicit "and this information is relevant to you" appended to them. To put this another way, at any point you can choose to pay attention to any of the nearly infinitely many streams of information available to you. Each of them demands a certain amount of your attention, and you can't conceivably pay attention to all of them. Of course in some circumstances you are cut off from most of them, e.g. your phone dies in the middle of an intercity train ride and you are left with only a single magazine to read. But this example is revealing precisely because it is unusual—in our ordinary circumstances we face no such scarcity and instead it is attention that is scarce.

So how do we allocate attention? I'll write some more posts about it, but in short I think we do it haphazardly or driven by a desire for entertainment or comfort. A few quick implications:

  • This is the reason that propaganda needs to entertain us. Fox News and Donald Trump are very good at this. You also see this on Twitter where ridiculous arguments couched in humorous terms often go viral.
  • Because attention is so scarce it is often sufficient for an idea to be superficially plausible or attractive. Few people have the attention spans to delve deeper. (Again, this is clearly visible in propaganda and viral tweets.)
  • It is unreasonable to go through life without consciously allocating attention, and yet this is how most of us behave in our personal lives. (Professionally there are usually few opportunities to pay attention to random things rather than the task that's been assigned.) Of course allocating attention well may be a very hard problem, but simply going with the flow is almost certainly worse and exposes people to many of the well-known hazards of modern life (polarization, in-group thinking, manipulation by advertisers, and so forth).

I think the right way to go through life (one that I seldom practice, admittedly) is to consciously ask, "Is this worth my time?" before committing any serious amount of attention to a subject. Moreover, I think it's important to maintain a kind of awareness of uncertainty for most things. It's okay to form a provisional opinion based on very limited information (e.g. the FDA says pasteurization is good, so I'll provisionally take the view that milk should be pasteurized), but it's also important to remember the degree of confidence that you attach to your beliefs and to behave accordingly.

Twitter is terrible for all of this. I've been thinking about how to use Twitter responsibly and I am not sure it can be done. More on that later.

Tuesday, September 17, 2019

City Dwellers Pay the Price

Just a few quick observations.

First, cities are generally the greenest place for people to live. "Greenest" in the sense that an urban lifestyle (small house, relatively little car use) is generally less polluting than a suburban or rural one. I think this is pretty well understood these days, at least among the audience for this blog. (It is worth noting briefly that (A) big cities tend to have big, polluting suburbs, so if you take the entire metro area into account I'm not sure big cities are actually all that green, (B) poor people tend to have low-carbon lifestyles wherever they live, and rich people tend to have high-carbon lifestyles wherever they live, and (C) with better public policy smaller cities and suburbs could be made to be pretty green. These are all topics for a different post.)

Anyway my real point has to do with my second observation, which is that cities tend not to be at all green in the sense of having healthy environments. This used to be a severe problem, now in the developed world it is not nearly as bad (more on that in a minute), and it's even improved quite a bit in many big cities in developing countries (Mexico City for instance). But it's still pretty bad to live in a city. This is extremely unfair in my view. If you live in a leafy green suburb you really will live a longer, healthier life than if you live in a city. In particular as I've started to ride a bicycle around the city more and more, I've noticed that I'm breathing a lot of really terrible air. (This post was motivated by this article, but there's plenty more evidence that small particulate pollution is very damaging. By the way, that article was published in Britain, which is why there are so many misspellings.) It kills me to think that I'm trading away my health in this way. (I know, probably the exercise benefits outweigh the costs to my heart and lungs, but still, it's a shitty tradeoff.)

My third point is that despite what I've said, cities' environmental conditions seem to have improved a lot relative to the rest of the country in the last 50-100 years. This may be a partial explanation for the high prices that houses in some urban areas command. Emphasis on partial! I'm not trying to deny the supply constrains that appear to be the core problem. But I suspect demand would be far weaker if pollution were as bad today as it was in the 1950s. In some ways, what is happening is that the product is improving dramatically and the price is responding as you would expect.

I'll write a separate post on how to think about where to live.

[Update: I think I should link to a recent piece on the conservative movement's abandonment of cost-benefit analysis. It's a well-written piece that I think manages to explain the problem without glorifying CBA itself. I will return to this topic too, but for now I'll just observe that it appears that reducing small particulate matter pollution is a hugely beneficial policy.]

Friday, September 06, 2019

Build More Housing for the Sake of Housing

A quick point about urban zoning. I am stealing this point from Matt Yglesias, but his decision to delete all his old tweets has freed me from the obligation to dig up the specific tweets that I'm thinking of.

The point is simply that the main virtue of allowing more housing construction in urban areas is that there will be more housing in urban areas. I'm not convinced that housing will be significantly cheaper in the long run as a result of upzoning. If upzoning does drive prices down in the long run, it's probably because it makes the city less livable, not because of supply and demand effects.

But all of that is speculative and mostly beside the point. It is simply good for more people to live in dense, walkable cities. Adding housing to dense, walkable cities means that more people will live there. And allowing supply to expand when people live there means that choosing to live there doesn't simply displace someone else.

In a perfect world, lots of cities would be dense and walkable, and existing cities would become more dense and walkable over time. In practice this is difficult and maybe impossible in a lot of places (not that we should stop trying). But we do have cities that are dense and walkable, and they are capable of housing millions more people than they presently do, and we should try to make that happen.

Wednesday, July 24, 2019

Regulation and Cheap Stuff

A quick followup to my last post. One question that libertarians sometimes ask is why the government should be involved in product safety regulation. Let's say you've got n manufacturers of cheese, with the safety of their products evenly spaced along a spectrum. And let's say it's readily apparent to the public how safe each cheese is. Isn't it reasonable that someone with limited resources might choose one of the cheaper, less safe cheeses? Maybe that's better than no cheese at all. If the government steps in and imposes minimum safety regulations, then all that has happened is certain manufacturers have been put out of business and certain poor people have been denied cheese.

This is of course very stupid, but it's worth thinking about why it's stupid. To me the answer is that in reality people are very bad at figuring out how safe cheese is. (By contrast people are pretty good at figuring out how tasty cheese is, and there's no need for tastiness regulations.) If you end up with a world in which safety is strongly correlated with price, but people have limited information, then bargain-hunting becomes a very difficult thing to do. You never know if this half-price-off cheese is cheap because the retailer accidentally ordered too much, or because listeria was found in the warehouse where the cheese was stored. You can't shop around for cheap cheese without putting your family's health at risk.

By contrast, if all cheese is required to be reasonably safe, then bargain-hunting becomes a totally reasonable thing to do. This means that manufacturers and retailers can compete on price without signaling that their product is unsafe. This, in turn, drives down prices (at least at the low end) and facilitates a market in economical cheese, which might not exist in an unregulated system.

Now you can imagine all of this happening without government involvement. You can imagine a certifying organization that runs safety checks on a voluntary basis and controls a label that only reasonably safe cheese is allowed to use. Then people would shop for bargains within the universe of safety-labeled cheese, and truly desperate people would buy the unlabeled stuff. This is the kind of model libertarians hang their hat on, and it's not totally crazy. (In fact, when it comes to product quality, we use this kind of system all the time. I believe it is also how compliance with religious dietary restrictions are monitored/communicated. By contrast allergen labels are mandatory.)

Anyway it's a thought experiment. But much the same logic applies to Amazon's counterfeit goods problem. In a world where Amazon reliably delivers the product that you order, you are free to look for the lowest price possible. Unfortunately that's not the world we're living in. You can never tell whether you've found an amazing price on an LCD monitor because of a temporary oversupply or because someone has just manufactured a run of counterfeit Samsung monitors.

Of course with Amazon the rot goes deeper—even if you go out of your way to buy from a reputable retailer on the platform, foregoing suspiciously low prices from scammy sellers, your order may be fulfilled from a scammy seller anyway. So in a way my point doesn't apply to the Amazon situation. But in general maintaining standards helps people by making it safe to buy an economical option.

The Corruption of the Platform

Yesterday I read two stories about different businesses tied together by a common theme. The first was a stunning piece in the New York Times about how easily Jeffrey Epstein was able to use mainstream media outlets to restore his reputation after his negotiated guilty plea for soliciting a minor. The implicated outlets are Forbes.com, the National Review, and Huffington Post. These sites basically published a press release as a straight opinion item, with no disclosure of its source. (Read the Times piece for the details.)

The second story was a Financial Times piece from a few months ago (which only came to my attention yesterday), which described how the "commingling" of goods sold by Amazon mixes counterfeit goods with genuine ones. I'm going to spend a little bit of space describing the article's revelations because I think they are shocking, and after discussions with friends I have come to believe they are not obvious to the casual reader.

Commingling is actually only half the problem, and in a sense it's the less important half. The initial problem is that when a seller submits a product to Amazon for sale on its website, Amazon puts very little effort into ensuring that the product is what it purports to be. So when you buy, say, Apple EarPods, you may end up receiving counterfeit goods of much lower quality. I want to pause here for a moment and highlight how profitable this scam can be. In general, you can make money by manufacturing a high-quality product (product A) and selling it for a premium price, or you can make money by manufacturing a lower-quality product (product B) and selling it for a modest price. But both of these activities are highly competitive, and you have to be really good at what you do (good at making a quality product, or good at keeping costs to a minimum, or both) to remain profitable. But this assumes that the rules of the game require you to label your product accurately. If you can just make product B and sell it as product A, then fat profit margins are yours for the taking. And that's what Amazon is permitting people to do.

In my view this is shocking—why would you ever trust a retailer that charges you for product A and ships you inferior product B? The answer is that Amazon is both a retailer and a platform for other retailers, and it is a very low-margin business. Amazon and its sellers would not be able to sell products so cheaply if the goods were genuine. First, genuine products are expensive (at least relative to counterfeit ones). But second, verifying that sellers aren't fraudulent requires spending serious resources on an ongoing basis. This is trivial for the retail business itself (just communicate with the manufacturer, who has no reason to tolerate counterfeits), but it's more difficult when it comes to products supplied by the manufacturer.

So anyway the result is that a lot of counterfeit goods get sold on Amazon. But you can imagine a business model where some sellers (maybe including Amazon itself, when it's acting as a retailer) are scrupulously honest, develop a good reputation, and sell their products for a premium. Buyers willing to pay that premium will get the genuine product, while buyers looking for a bargain will get counterfeit goods. It's not ideal but it makes a certain amount of sense. (More on this in a separate post.)

But that's where commingling comes in. Commingling is driven by the fact that Amazon maintains multiple warehouses around the world (I guess they call them "fulfillment centers"), and it is helpful to ship products from the warehouse closest to the customer. To facilitate this, under certain circumstances Amazon treats all products bearing the same manufacturer bar code as a fungible bulk. In other words, if Seller A and Seller B have each delivered a product with the same bar code to Amazon, and those products are held in different locations, then Amazon fulfills each seller's orders from the warehouse that is closest to the customer, regardless of whether the goods in that warehouse actually came from Seller A or Seller B. If Amazon also sells the same product on a proprietary basis, then it may also fulfill its own orders with products provided by Seller A and Seller B, again based on the location of the warehouse and the location of the customer.

Now I hope the problem is apparent. Amazon's platform lacks integrity, because policing its sellers would be expensive. And you can't avoid the problem by buying from honest sellers, because Amazon doesn't silo its sales by seller, it freely substitutes counterfeit goods for genuine ones. The platform's integrity is defined by its least honest seller. It's like a fast food chain that buys good quality produce and then mixes it with produce contaminated with salmonella or listeria or whatever.

There is probably also a Gresham's Law effect where the counterfeit products chase out the good, because they are cheaper to produce. On the other hand, Amazon will go out of business very quickly if the public concludes that its products can't be trusted, so Amazon is probably at least somewhat motivated to cut off ties to fraudulent sellers. In equilibrium there will be lots of genuine goods and lots of counterfeit goods, and customers roll the dice every time they place an order. Speaking for myself, that's not a gamble I care to make.

But putting all that aside, I hope it's obvious why I opened this post with the corruption of Forbes.com, National Review, and the Huffington Post. It is a remarkably similar story. If you want to persuade people of something, you can always place an ad, but ads are relatively unpersuasive because everyone understands that their claims are self-interested. If I read an ad that says that a movie is great, I discount it to zero. If I read a review that says a movie is great, I give it more weight.  But the problem is that no one is going to write an honest opinion piece saying that Jeffrey Epstein is a great guy. (Actually I'm probably wrong, you could probably arrange that without paying the writer, but it would not be easy.)

So the trick is to pay for an ad but present it as a review. And that's basically what was going on with the corrupt publications, except that (as with Amazon) I don't think the illicit profits were paid directly to the publications. Instead they benefited from the extremely cheap content that their shoddy business practices made possible. But as with Amazon, I don't see why anyone would trust a publication with such low standards.

Monday, June 17, 2019

Kinds of Rent Control

In my previous post I argued that rent control is generally undesirable. Affordable housing is important, but it is not the kind of thing that can be done by fiat, at least not without serious negative consequences. In general if you want cheaper housing you need to (A) allow more of it to be built, and more densely, or (B) improve transit so that the supply of housing within x minutes of the city center is increased (the Tokyo approach). Or, ideally, both.

However, not all rent control is created equal. The principal harm of rent control is that it pushes rents far below market prices, causing serious distortions as described in my previous post. Small deviations from market prices aren't nearly as big a problem. Here are a few policies that I think could alleviate a lot of tenants' worries without causing too many problems.

1. Encourage (or mandate) "rolling" leases that don't terminate on any specified date, but instead are perpetual until they are terminated on x months' notice. The landlord also can't raise rents except on x months' notice. By itself this isn't really rent control, since the amount of the rent increase needn't be regulated, this simply means that tenants will always have x months to make new housing arrangements if the landlord either decides to allow the lease to expire or to raise the rents dramatically. (By the way, I think x = 8 is pretty good, but I could see going all the way to x = 12.) The idea is that tenants shouldn't be caught in a situation where they have to secure new housing in a hurry or face steep rent increases. It's kind of crazy that how protected your living situation is fluctuates every year between 1 and 12 months. Much better to have a constant, modestly high level of protection throughout the year.

A landlord who truly needs a hard deadline on a lease (for instance, because he wants to move back into the unit) would have to make that abundantly clear at the outset. Basically any lease shorter than x months would have to be for special circumstances only (e.g. a student leasing an apartment for a summer internship). This creates a bit of a line-drawing problem, but not an insurmountable one.

(By the way, this regulation would only apply to the landlord. The tenant would still be able to terminate a lease on shorter notice, maybe 2 months or something.)

Admittedly it's a little puzzling to me that this solution doesn't seem to be very common in the marketplace. I would have thought it would be very popular with tenants and only mildly annoying to landlords, and therefore would be bargained for. This is maybe a warning sign that this would be worse for landlords than it appears, although there's also just a tremendous amount of stasis in the way people do things, and quite possibly landlords wouldn't mind much.

2. Allow rents to increase by no more than y% per year. Here, y needs to be low enough to offer real protection to tenants, but high enough that rents don't get too far below market prices. To me, the high single digits seems good. The idea is not to mimic traditional rent regulation, where an apartment remains cheap for generations, eventually becoming comically cheap relative to market rents. The idea is to protect tenants from rapid shocks that totally upend entire neighborhoods at once. If a neighborhood becomes vastly more desirable, because of a drop in crime or a new subway station or whatever, under this system rents can still double within about 10 years, ensuring that they don't remain below market for too long. On the other hand, tenants have plenty of time to make new arrangements, never facing a terribly steep rent increase in any given year.

Now there are three important caveats here. One is that if you adopt this policy, you also have to make leases perpetually renewable at the tenant's option. Otherwise the landlord will simply terminate the lease in lieu of letting the tenant stay at below-market rents. (Obviously a tenant who is violating the lease can still be evicted, but a tenant who is paying rent timely and not wrecking the place has to be able to renew as of right.) Alternatively I suppose you could lock in the rent even if a new tenant moves into the apartment, so the landlord can decide to get rid of an existing tenant he doesn't like but can't evict him for profit.

The second caveat is that you probably have to make an exception for big improvements that the landlord pays for. If the landlord does something like install a new laundry room so that tenants don't have to walk outside to do laundry, the landlord should probably be allowed to recoup the costs through a one-time rent increase in excess of the legal limit. You have to be careful to make sure these improvements aren't pretextual ("I installed some new patio furniture on the roof, now your rent is 50% higher!"), but legitimate improvements should probably be grounds for increased rent.

The third caveat is that landlords may front-load their rent increases. Let's say a landlord is leasing an apartment at market rent. If the landlord anticipates that market rents might rise by more than y% for the next several years, he might raise rents by y% this year so as not to fall behind. That means that temporarily the rent for this apartment is above market rent, a perverse outcome. I don't think a landlord could get away with this for very long (the tenant will just move to a market-rate apartment), but you can't rule out the possibility that this would become a problem in some situations. In particular, landlords would probably start raising rents at the legal limit well in advance of a major improvement to a neighborhood like the addition of a new subway station. I'm not sure what you could do about that, maybe nothing.

By the way, I think these two policies go really well together. Every year your landlord would tell you, "Hey, in x months your rent is going up by (z where z <= y) percent, or you can move, your choice." You would be able to stay in the apartment at a non-ruinous price, or you would have x months to find a new place. Meanwhile an apartment would generally only stay below market rent for a few years at a time, so the distortions caused by the policy would be pretty minor.


Wednesday, June 12, 2019

Thinking About Rent Control

It appears the New York State legislature is moving toward an expansion of rent control. I want to lay out some basic thoughts, possibly in preparation for a more detailed post later, we'll see. Throughout this post I use the term "rent control" to mean a legal prohibition on charging more than a specified amount for leasing a residential unit (so in other words, I use the term to include "rent stabilization"). I also assume that tenants in rent-controlled apartments have perpetual renewal options—that is, the landlord cannot choose to rent the unit to anyone else or to move into the unit at the termination of the lease. Instead the landlord has to continue to rent the unit to the tenant for as long as the tenant is in compliance with the lease. This is not a necessary feature of rent control, but I believe it is true in New York.

Rent control is sometimes presented as a left-right issue, the assumption being that leftists will favor rent control while conservatives will oppose it. I think it's important to resist that framing for a variety of reasons, the most important being that the winners and losers from rent control don't neatly correspond to "the rich" and "the poor" or whatever.

Instead I think the right way to think about rent control is to consider (A) its direct consequences, (B) its political consequences, and (C) basic principles of fairness.

The main pragmatic problem with rent control is that it leads to under-provision of housing and leads to inefficient use of existing housing. Taking the second issue first, in general rent control will encourage people to consume more housing than they need where it is scarcest. Here are two examples:

1. A family lives in a 4-bedroom apartment in NYC with a maximum legal rent of $1,400. The children move away, and the parents would prefer to downsize to a 1-bedroom apartment in NYC. But they can't afford to—the rent for a 1-bedroom apartment would cost at least $2,500. As a result, an apartment that could house an entire family is used by two people. If they get divorced, maybe the apartment will end up being occupied by one person, a tremendous waste of space.

2. An individual lives in a 1-bedroom apartment in NYC with a maximum legal rent of $1,000 (on the free market, the apartment would cost $2,500). This individual is offered a job in Chicago that pays just as well as her job in New York, and she is inclined to take the job because she prefers Chicago to NYC. Also, the cost of living in Chicago is lower than in NYC (there is less of a housing shortage there). However, the market price of a 1-bedroom apartment in Chicago is $1,700, far more than she is paying in NYC. As a result, she stays in NYC, unwilling to take what would amount to a $700/month pay cut.

In both cases, people had preferences that, if acted on, would have freed up housing in NYC (where it is especially scarce), but rent control prevented it. Now in fairness, rent control can cut the other way—imagine a couple that has a baby and wants to move to a bigger apartment, but can't afford to give up the rent-controlled studio that they are occupying. But that's not great either, even if (unlike our earlier examples) it reduces observed demand. In an ideal world, the price of housing would reflect its social cost, and people would make decisions according to their needs. In a lot of cases, that would involve economizing on space or moving to a different city, which would increase the available supply and lower market costs.

Rent control also tends to suppress supply, although there is a tricky question here involving how landlords and developers form expectations about the future. One thing you can theoretically do is impose draconian rent control on existing units, while leaving all units built in the future totally unregulated. In that case, if developers believe that rent control will never be imposed again, their incentives to build new units are not necessarily reduced. However, I think it's somewhat unrealistic for expectations to work this way, particularly if the legislature periodically imposes new rent controls. Certainly if the topic being debated is "Should we impose new rent controls on top of the ones that already exist?" it seems as though developers should not expect the pro-rent-control side of the debate to win just this once and then lose forever after.

To the extent developers expect rent control to be imposed in the future, it obviously affects the buildings that are built today. In particular, no one has ever suggested imposing rent controls on owner-occupied units, so developers will tend to build condos rather than rental units in expectation that rental units will lose lots of value in the future. Or developers will focus on luxury units that are likely to have tenants who are not likely targets of the legislature's sympathy. All of this will decrease the construction of affordable rental units, exacerbating the shortage of rental housing in the city. (To the extent developers simply build condos one-for-one instead of rental units, possibly the housing shortage wouldn't be worsened too much, but of course a lot of people can't afford a down payment and so the affordability of the city's housing would still take a hit.)

Zoning politics are complicated, but I think it's fair to say that people are more likely to support construction of new housing units when they will personally benefit. Of course the main beneficiaries of housing construction are (A) people who don't live in the city yet but would like to move here (but who definitely don't vote here yet), and (B) renters who live in the city and would like rents to fall. Obviously people who don't vote in the city aren't going to exert very much electoral power, but market renters are a natural constituency for more building. Expanding rent control shrinks this constituency and makes it less likely for pro-building policies to be enacted. This, in turn, exacerbates the housing shortage.

Finally, fairness is obviously debatable, but rent control has a few features that are particularly obnoxious. It amounts to a tax imposed on some landlords and not others, rather than a broad-based, equitable tax. Property taxes in New York are shockingly low for some kinds of housing (mostly single-family houses, I believe), and it would be possible to raise taxes and use the revenues for housing vouchers targeted at low-income people. Instead we picked random landlords and stripped them of most/all of the value of their property and bestowed it on random renters (not necessarily low-income). Those incumbent renters are now paying vastly less rent than poorer people who happened to move to the city more recently.

So this is my basic framework for thinking about rent control. Rent control reduces the supply of housing, increases effective demand, reduces support for upzoning, and is much less fair than a broad-based policy that benefits people proportionately to their current circumstances. But all of this has been fairly broad-brush, and I hope to drill down on a few variants of rent control in future posts.

Wednesday, June 05, 2019

ugh