zero mean error

abbreviated thoughts on economics, policy, and tall buildings. views my own.

[johnricco]

Externalities, again

I've written before about how the notion of "externalities" is fundamentally empty. To reiterate: in order to label something an externality, we need to have a clear idea about who is entitled to what. Only then can we meaningfully say something about a third party being "affected" by a transaction. 

I'm returning to this because a group of homeowners in Lanier Heights, DC wants to downzone the neighborhood. Among other things, they worry that newly developed condos will (a) block sunlight and (b) destroy neighborhood character. Now, let's say I live in a rowhome on Lanier Place and a developer buys my neighbor's property and builds a pop-up. Maybe it blocks some light and maybe it looks like shit. This sucks. I've been made manifestly worse off by that transaction -- no way around it.

BUT, if I cry "externalitity!" and run to the ANC board asking for a zoning change, I've begged the question. I've assumed that I'm entitled to the area over which certain rays of light travel and a pleasant view when I walk out my door. And in calling it an externality, I've subtly compelled people to think there's a market failure here correctable by government regulation. Not only is this a nontransparent way of making an argument, it's incompatible with the post-restrictive covenant tradition of property rights in America. When you buy a home, you end up owning some land, wood, bricks, etc. You don't own your neighbors' land. This is why people use rent-seeking and not tort law to influence neighborhood development.

So I think we'd be better off if we lessened our dependence on the externality as a tool in policy discussion. Instead, let's slow down and discuss who is entitled to what and why -- legally, morally, etc -- and make sure we're not baking our conclusions into our assumptions.

[johnricco]

Let's play a game: Umair Haque or random inspirational quotes twitter account?

Today we're going to play a game where you guess who authored the following tweets. There are two possibilities. The first is Umair Haque, a self-styled economist (er, MBA) who thinks the world is going to shit because of like bankers and tech and inequality and whatever else seems appropriate for a 140 character Serious Thought lacking any sort of nuance. The other possibility is @Inspire_Us, the first result that came up when I googled "inspirational quotes twitter". The former writes for Harvard Business Review and the latter is one of those lowest-common-denominator accounts that profits from reposting misattributed platitudes. Let's see if we can tell their tweets apart.

1. The haters will hate you either way. So you might as well go ahead and live a life you love.

2. When you say yes to others, make sure you are not saying no to yourself.

3. If you don't love someone for their imperfections, you don't love at all.

4. Accept no one's definition of your life; define yourself.

5. Happiness doesn't result from what we get, but from what we give.

6. Good things happen better when you leave bad things behind.

7. You aren't wealthy until you have something money can't buy.

8. We do small things out of fear. We do great things only out of love.

9. Give people a piece of your heart rather than your mind.

10. The cynics will always tell you you can't. So you might as well go ahead and live a life that does.

[johnricco]

Streetcars and economic development

Streetcars are a pretty inefficient means of moving people around. Even most support-all-transit-at-any-cost fanatics will admit that. So instead of selling streetcars on actual transportation benefits, advocates claim they spur development. This is a point that I have mixed feelings about – I don’t think it’s obvious whether A) that’s fully true, and B) if it is true, that localized development is an unambiguously good thing, as is implicitly assumed. I’ll save point B for another day though.

So, point A: Do streetcars spur development? Probably, but it’s not that simple. When we observe development and a streetcar occur simultaneously occurring, three things could be happening:

The streetcar caused development. Again, this is widely accepted and is probably true. I see three problems with accepting this one its face, though. One, the empirical argument used to defend this is weak: just because we observe development happening after a streetcar opens does not mean the streetcar caused the development. Correlation, causation, etc. Two, the theoretical argument – that streetcars’ permanency relative to other modes is the mechanism by which they cause development – is unconvincing. Bus transit can be permanent too if done well. Here’s Alexandria’s brand new BRT line (photocred goes to Dan Mauloff):

Note the center-running dedicated lanes, separated by curb, with pleasant, sheltered stops that are more like stations. This looks more permanent to me than rails in the middle of a street. It’s probably not a coincidence that lots of dense housing is going up around there. And third, it might not be the streetcar itself that causes development, but instead the expectation of a streetcar. DC’s streetcar is a great example of this: the streetcar has seen delay after delay and hasn't opened yet, but in spite of this the H st corridor experienced tons of development anyway. Citylab wrote a good article on this a little while back. This is actually very similar to how central bankers try to work their way out of deflation – the Fed chair may not always be able to spur inflation now, but if they can credibly promise to leave interest rates low for a long time, they create the expectation of inflation which has the same effect.

The development caused the streetcar. That is, planners saw a neighborhood becoming popular and decided it would be a good place for a streetcar. I think this is an underrated possibility. Surely this is true to some degree; we don’t run streetcars into a random field and expect a vibrant neighborhood to pop up. Instead, the decision of where to route new lines is always influenced by current trends in development, capacity for increased density, expected ridership, and so on. So it’s totally reasonable to think that development itself reverse-causes streetcars in part.

A third factor caused both the streetcar and development simultaneously. Another underappreciated possibility. The most obvious example of a confounding variable is what we might call neighborhood characteristics, broadly speaking. For example, neighborhoods containing a lot of young hipsters might have an inclination for both new development and streetcars. A neighborhood of NIMBYs would have the opposite preferences. These are important fixed effects that need to be controlled for if we want to have a clear understanding of the relationship between streetcars and development. 

Just thinking out loud here really. But my point is that there are reasons to doubt that the assumed streetcar -> development line of causality isn't theoretically or empirically obvious, and that advocates shouldn't treat it as such. 

[johnricco]

The World Bank HQ is home to one of like two well-done spires in all of DC.

[johnricco]

Land use restrictions have national implications too

A new NBER working paper looks at how land use restrictions are hurting the US as a whole. I'll quote at length here because there's a lot of good stuff:

There are vast differences in nominal wages across US metropolitan areas. These differences are robust to controlling for worker characteristics and do not disappear over time – indeed, today they are much larger than in the past. In particular, the geographical dispersion in average wage conditional on worker observable characteristics in 2009 is twice as wide as the dispersion in 1964.

People's salaries differ across geographical areas. This should be obvious: If I moved to like, rural Idaho, I'd certainly be less productive than I am in DC. What's not obvious a priori is that this effect is growing or why it's growing. But more on that in a second. 

Using this insight and data on a longitudinal sample of 220 cities in 1964 and 2009, we calculate the effect of the increased dispersion in the nominal wage across US cities under a counterfactual distribution of labor supply across US cities. In the counterfactual, we set the gap between the nominal wage and the mean wage of each city in 2009 equal to the gap of the same city in 1964, holding the distribution of local TFP fixed at its 2009 level. Intuitively, this scenario involves setting the determinants of labor supply in each city – amenities and housing supply – at their 1964 level, while keeping labor demand constant at its 2009 level. We show that under this scenario, aggregate output would increase by 13.7%, or $1.7 trillion. This amounts to a wage increase of $8,900 for the average worker.

If workers could freely move to places where their marginal product would be higher then our country would be 1.7 trillion dollars richer. (This is the same type of estimate that pro-immigration economists make when they claim that open borders could double global GDP.) That's a massive number. So why don't more people move to places where they could actually use their talents? Because it's illegal:

More specifically, we need to understand whether the increased geographical dispersion in wages between 1964 and 2009 is driven by changes in amenities or changes in housing supply policies. If the relative increase in nominal wages in high TFP cities such as San Francisco and New York is due to increased restrictions to housing supply, then the aggregate output loss due to differences in the marginal product of labor also imply welfare losses. In this case, removing constraint to housing supply in cities like San Francisco and New York would allow more workers to move there and take advantage of their higher productivity, pushing down local wages and increasing both aggregate output and welfare.

And...

We conclude that the aggregate gain in output and in welfare from reducing constraints to housing supply in high wage cities are likely to be substantial in the US. The effect of stringent land use regulations on local housing prices is well documented (Glaeser and Gyourko and Saks, 2005 and 2006; Gyourko and Glaeser, 2005; Saiz, 2010), and our contribution is to highlight the aggregate negative impacts of such regulations. 

So not only do zoning laws have local effects, they hurt the US as a whole, too. The authors recommend that this be used a rational to move land use lawmaking up the ladder of federalism to kill the vested-interest political power of localized NIMBYs. (Development capital of the world Japan's system of national land use laws is a good example of how this works.)

Again, this stuff is jaw-droppingly important. There's a $1.7 trillion dollar free lunch just sitting out there and we somehow lack the political will to take it. 

[johnricco]

When data-driven journalism is neither data-driven nor journalism

A friend directed me to this awful BloombergBusinessweek article (?) yesterday in which the authors claim that there is no link between CEO pay and company performance. This is a wholly reasonable conjecture; it's possible if not likely that corporate executives' salaries are decided by a process that has nothing to do with marginal revenue productivity. (Piketty's thesis, whereby top rates determine whether an additional after-tax dollar of CEO pay is worth the public outrage, strikes me as sensible.) And better yet, it's a testable hypothesis -- one that has been the focus of empirical economic work for quite some time.

So how did these guys approach this topic? Well, they did what any sincere researchers would do: made a scatterplot, drew a line through the middle, and called it a day.

Certainly this One Chart provides substantial evidence to "prove" their claim despite a number of glaring issues, omitted variables being first on the list -- you know, the many other things that determine stock return. You can also find fault in their use of rankings instead of raw figures, which probably results in lost information via aggregation. Or their decision not to consider other measures of firm performance. Or their use of R^2 as a metric for hypothesis testing (an avant garde approach, I must say!). Or how they butcher the interpretation of R^2. Or their juvenile error in putting the dependent variable on the x axis. You get the idea: this is very sloppy work.

Complex social questions require complex toolkits. It's not a coincidence that the economists who have produced the body of work on CEO compensation have done so with high-level regression and financial simulations. It's also not a coincidence that this literature has been in the works for twenty-five years and counting. But in the Vox.com era of two-minute explainers (oh yeah, this got picked up by them too), this sort of thing is applauded for being hard-hitting analysis based on The Numbers, despite being downright bad. This is a horrible use of data, and anyone who says otherwise is probably just looking for pageviews. 

[johnricco]

[johnricco]

afloweroutofstone:

Let’s take your example of river pollution. The polluted party claims that their land is harmed, while the polluting party rejects the concept of land ownership outright, and, thus rejects the complaints of the polluted party. But if there’s a clearly established framework of property ownership, what the polluter’s personal beliefs are about land ownership are irrelevant, because (presumably) land ownership is an established fact. For example, if I were a communist who did not believe in the private ownership of the means of production, that still wouldn’t give me any justification to do what I like with machines or factories that I don’t own, as long as there is an established structure of private property in place.

Thus, personal definitions of “affect” are made irrelevant when property rights are agreed upon. It’s true that, if we take things to multiple degrees of impact, any action could “affect” anyone, but direct harm isn’t as difficult to define. While a private property framework is in place, externalities can be clearly defined.

Perhaps the pollution exercise wasn't the best example because it's unlikely you'll encounter a Marxist factory owner any time soon. Generally, people aren't going to deny your claim to land ownership, but reasonable people both working from a private property framework can disagree about the nature of that framework -- and consequently, externalities are hard to define sometimes.

Take this 2006 paper about gun ownership as a more nuanced example. The authors write:

But widespread gun ownership could also lead to increased risk of various sorts, including the possibility that guns will be misused by their owners or transferred to dangerous people through theft or unregulated sale. Whether the social costs of gun ownership are positive or negative is arguably the most fundamental question for regulation of firearms in the United States.

They ultimately find the effect to be negative, indicating that gun ownership imposes negative externalities and that a pigouvian tax as high as $1800 would help internalize social costs. But a lot of gun owners would be outraged, claiming that they themselves have never harmed anyone, and just because there's a risk of something bad occurring doesn't mean that they should be subject to high licensing fees. Note that both sides are working from a baseline of private property; they just disagree on how risk fits into that framework. So here, "externality" is only as useful as the underlying theory of entitlement -- whether we think that carrying the risk of harming others is morally equivalent to harming others.

And this is tough stuff! We ban some activities that run the risk of harming innocent third parties (like drunk driving) but not others (like flying a plane over a city). My point is that the line we draw is often arbitrary and values-laden, and when economists talk about externalities without acknowledging this, their policy advice is presented as being technical and positive rather than fundamentally normative. The paper above is a perfect example of that. 

[johnricco]

"Externality" as a question-begging concept

Maybe it’s because I’ve been reading old Unlearning Economics posts recently, but I’ve been thinking a lot about the implicit assumptions that economists make — and finding more and more things in economic models that seem shaky. One example: the concept of externalities as a whole seems suspiciously circular to me. Let me try to explain why; I hope that writing about it will clarify my thoughts. 

Externalities are usually defined as something like “instances in which an economic transaction affects a third, uninvolved party”. Seems reasonable and divorced from ideology. But on its own, this definition can’t tell us anything meaningful until we define “affect”. Think about it. Before we can claim that a third party is “affected”, we need to establish an underlying, comprehensive theory of justice that answers questions like “what sort of things can people own?” and “what constitutes aggressive violence?”. Only then can we say anything about who’s being harmed or helped. 

So really, when economists talk about externalities, they’re working from a pre-decided idea about entitlement — which rarely, if ever, is acknowledged explicitly. And that theory usually takes a libertarian form, quietly pushing this supposedly objective framework to the right, as a lot of other ideas in economics do.

Pollution is a good example here. Suppose I own a factory and I dump a bunch of garbage into a river, which floats downstream and washes up in your backyard. You claim that this is a negative externality; clearly I’ve harmed you by inconveniencing your land. But what if I rebut by claiming that it’s not your land because God made the world for everyone? I might have affected the earth but I didn’t affect you. “Externality” has no practical meaning in our dispute because we disagree on the underlying theory of entitlement. You’ve implicitly assumed “affect” to mean one thing and I’ve assumed it to mean another. Irrelevant to our argument is the notion of externalities because we’ve begged the question.

Economists apply this framework widely when it comes to policy justification (usually under the guise of providing “positive” and not “normative” analysis). Single-payer will limit financial externalities! We need to implement congestion pricing to curb traffic externalities! Subsidize college so that students can internalize the benefits of a more-informed population! If these things are true, they’re true independent of whether we call them “externalities”. Doing so literally adds nothing to any argument and encourages fuzzy thinking about important issues. 

[johnricco]

Listen to these albums from 2014

This year in music has been unimpressive, but there's been a handful of good records. These are my five favorite albums from the first half of 2014 (with links to some good songs to check out). 

5. Benji, Sun Kil Moon

Sixty-two minutes of otherwise unspectacular acoustic guitar made great by a Sufjanian resolve to tell stories that incorporate both extreme specificity and universal human experience, like death.

4. I Never Learn, Lykke Li

All the makings of a good pop album are here: love, loss, brevity, and reverb-soaked power ballads.

3. What is this Heart?, How To Dress Well

Imagine if the Weeknd were a lanky white dude who wrote lyrics with actual substance and didn't do so many drugs.

2. Salad Days, Mac DeMarco

Maybe the key to making effective guitar-based music in an era where electronic artists do all the inventing is to be a total piece-of-shit goofball. 

1. St. Vincent, St. Vincent

Annie Clark is the most talented songwriter around and it's not really close. 

[johnricco]

Want rowhomes instead of condos? Build condos

DC planners are proposing a ban on carving up R-4 rowhomes into condos. The idea is that developers are buying single family townhouses and turning them into small apartments for yuppies, which displaces working class families. This is a shortsighted approach and like most other attempts at curbing gentrification, it loses sight of the actual problem.

One place where this is especially relevant is my own neighborhood, Columbia Heights. CoHi is a rapidly gentrifying area that white people like because it's walkable and close to nightlife. It's also primarily zoned as R-4, meaning that most of the housing consists of 3-story attached rowhomes instead of larger apartment buildings.

Now, there are a few areas where large apartment buildings are allowed. But because this land is at such a premium, it's been developed into luxury units that require something like an $80,000 salary to afford. So when well-to-do-but-not-rich people like me move into the neighborhood, we find the next best thing which is attached housing with one or two BR apartments -- places that were initially occupied by families. 

If we want neighborhoods to have cultural and economic diversity, we need to focus on accommodating the influx of new residents. For CoHi, this means letting more multifamily housing go up -- something that is all but impossible under existing municipal laws. (That developers' most profitable option is to take townhomes and add one or two units says a lot.) Columbia Heights is an awesome place, so we should build more of it. 

[pgdigs]

December 1959: The hotel that transformed the Point

For some time, the Digs staff has been partaking in a Pittsburgh tradition: Watching construction work at the city’s most visible hotel. We’re thrilled that renovations at the Wyndham Grand are finally complete.

More than half a century ago, the hotel was known as the Hilton. And as it slowly rose over the Point, one steel beam at a time, thousands of Pittsburghers stood on sidewalks and craned their necks to watch the progress.  One was a Pittsburgh Press reporter named Gilbert Love.

This was in the late 1950s — an exciting time in the city. Love and every other Pittsburgher had fresh memories of the Point as a place of crumbling warehouses, dilapidated brick dwellings, saloons and narrow cobblestone streets.

Then it changed — in dramatic fashion. The Point was cleared, the old buildings demolished. Silver towers named Gateway Center 1, 2 and 3 rose skyward, looking spiffy and modern. Surrounding them were trees and shrubs. Greenery, on the Point! It was barely believable.

Further up Liberty Avenue stood Four Gateway Center, 22 stories of green glass — a mirror, so the new and emerging city could preen itself.

The Hilton would be the gem of the Point. “We intend to build a beautiful structure worthy of its beautiful site,” said the hotel’s director. When was the last time anyone had used the word “beautiful” and Pittsburgh in the same sentence?

But there it was, 24-stories of anodized gold, an exclamation point on the Golden Triangle. Workers wrapped it in a quarter mile of royal blue ribbon — “the world’s largest ribbon and bow,” proclaimed The Pittsburgh Press.

Tanned and dapper Conrad Hilton showed up to cut the five-foot wide strip of cloth. Trailing behind him were a bunch of his Hollywood friends. After opening ceremonies, the sexes parted in typical ’50s style: Women filed inside to attend a fashion show while men sat down to a luncheon.

The city shivered with anticipation of the next day’s dinner and dance, a $75 per person affair open to 1,000 Pittsburghers and 200 special guests of Hilton. “The doors of the largest ballroom in the eastern part of the nation will be thrown open to reveal what promises to be one of the most spectacular scenes of elegance and conviviality ever witnessed here,” gushed the Press.

It was a smashing event. A few days later, Love looked up at Mount Washington and imagined what the new city would look like in the years to come. He could visualize “the crest of the hill lined with restaurants with glass walls overlooking the triangle, and apartment houses with balconies or decks to take advantage of the view. Below them, the hillside would be a hanging garden of blooming plants.”

He imagined pleasure boats on the river, a fountain sending a geyser of water into the sky, walkways along the river.

Visitors, he wrote, “will wonder why this place was ever called ‘The Smoky City.’”

—Steve Mellon

Top photo: The Hilton Hotel, under construction in October 1958, framed by one of the tiers leading to the new Fort Pitt Bridge. (Photo credit unknown)

[johnricco]

Day 3 of living car-free: I love my new toy.

[johnricco]

Warming up to One57 now that those penthouses are lit up. Pretty. (h/t flickr user Bjo - ao)

[johnricco]

In case you thought the future wasn't here. (h/t flickr user JBB | MK00)

[johnricco]

Information asymmetry and predatory colleges

I graduated college last week. Over the last four years, I've spent a lot of time thinking about how American universities prepare students for the workforce. This is because some of my college friends, all of whom are capable, smart kids, are alarmingly ill-prepared for real life. They have no skills. They face no job prospects and are moving back home with their parents -- or worse, going to a second-rate law school. Most are saddled with a considerable student debt burden.

Cyclical labor market slack certainly plays a role here. But every 2014 graduate I know who can't find a job happened to study a regrettable major: broadcast journalism, public relations, English, jazz performance, and so on. Spending four years and $100,000 for these degrees is demonstrably stupid. So why do people do it?

The market for higher education suffers from serious information problems. First, prospective students are usually 17 years old when they begin applying for schools. These kids -- children, legally -- are expected to have an idea of what they want to do with their lives, which is of course ridiculous. Second, kids (and their well-meaning parents) typically have an all-too-optimistic view of their abilities and subsequent chances of finding work after graduation. Part of this is because humans are bad at what economists call intertemporal optimization (i.e. what happens when an art student isn't exposed to the full costs of art school until four years after enrolling).

What's most worrying, though, is that universities exasperate these problems by preying on low-information consumers. Many students don't have a clear understanding of their post-graduation employment prospects. This is in part due to student laziness, but it's also because colleges can get away with providing blatant misinformation to students. The data show that this is rampant in law school recruitment. I've also experienced this firsthand: I was once a freshman music technology major who was swindled into a $120,000 investment because the program advertised a 98 percent placement rate. As it turns out, any employment -- unpaid internships and low-paying music gigs included -- counts as "career placement".

This kind of stuff borders on fraud and needs to be treated as such, especially given the onerous state of student debt in 2014. A light dose of regulatory paternalism would be a wonderful start. Here's a great idea I saw in a twitter exchange today:

Mandatory, transparent databases that track employment and earnings statistics would do a lot to help students make informed decisions. You'd see more students choosing legitimate career paths and less student debt disaster stories.

I'd even go so far as to say that this could reduce wealth inequality. Information asymmetry in higher education almost certainly harms poor people disproportionately. For rich kids, graduating with a useless degree is an inconvenience, but for poor kids, there's a greater chance it results in poverty. Here's Matt Bruenig on how screwing up in school is devastating for the poor but not for the rich:

Taking risks that will harm conventional indices of merit (grades, scores, activities, whatever) is probably fine enough for a rich kid for whom the negative consequences are not so great. Like we saw before with Ezra Klein, you can screw up in high school and come out with a 2.2 GPA and somehow get into an elite top 100 university, if you come from the right background.

However, the costs of screwing up without all of that background assistance are massive. Taking risks that result in “failure” for a poor kid is more likely to mean joblessness, poverty, or prison than it is to mean having to go to the dreaded backup college.

So sure, go ahead and let your rich kids fail. The joys of inequality are such that they can’t really fail in any deeply harmful way, not with all of the second chances their position in the economic hierarchy provides them. 

So at the margin, a better-informed student body should make for fewer underprivileged 22-year-olds graduating with enormous negative net wealth. College is the most important investment many people will make in their lifetime, so it's only fair that students are well-equipped to make the right educational decisions. 

[johnricco]

Yikes. NYC's new $4 billion "transportation hub" (read: subway station) is coming together. This thing sucks on a few levels: it's ugly (fish spine or ladyparts?), it's exceeding its budget by about 100%, and it's a perfect example of how US transit systems systematically misallocate funds. (Photo from the WTC Progress twitter page.)

[johnricco]

Coincidental correlation isn't an actual problem

Everyone's been having fun with this site that shows you correlations between obviously unrelated variables. It's being sold as a neat way to demonstrate a cliched truism in statistics: correlation does not imply causation. This is correct, of course. Nick Cage doesn't cause drowning:

and eating less margarine doesn't cause divorce rates to plummet: 

The problem is that this isn't a good example of why relationships aren't always causal. Spurious correlations can result from either random chance, as is the case here, or a third confounding variable. The latter is more insidious because it's both easier to mistake as causal and is far more common. For instance, an econometrician just isn't going to include margarine consumption in a model where the dependent variable is divorce rate. But when trying to specify all the factors that cause divorce, it's alarmingly easy to omit (or not have data for) relevant variables, meaning that we've failed to control for that third confounding variable. This is the biggest obstacle to empirical research giving us *anything* substantive in the social sciences. While coincidental correlation is cute, endogeneity is the real enemy.