Saturday, May 20, 2017

The NIMBY challenge


The other day I wrote a Bloomberg View post about how California is waking up to the problem of NIMBYism - development restrictions that limit economic activity and make cities less affordable. Ground Zero for this struggle is the Bay Area, San Francisco in particular. The pro-development activists known as the YIMBYs have been at the forefront of the fight. Economists have also been weighing in. Enrico Moretti & Chang-Tai Hsieh and Andrii Parkomenko have both come out with new theory papers showing negative impacts of housing development restrictions. Ed Glaeser and Joe Gyourko have a paper reaching similar conclusions after looking at data, theory, and institutional and legal details. And Richard Florida has a whole new book about the problem.

But the YIMBYs have faced a great deal of intellectual pushback from certain folks in the Bay Area. Even as I was writing my post, physicist Phil Price was writing an impassioned attack on YIMBYism over at Andrew Gelman's blog. He followed it up with a second post three days later, after getting a great deal of pushback in the comments. The commenters have made most of the points I would make in rebuttal to Price, but I think his posts are worth a close look, because they reveal a lot about the way NIMBYs think about the housing market. In order to understand and meet the YIMBY challenge, pro-housing activists should familiarize themselves with the arguments Price makes.

The first thing to note is that NIMBYs think that a house's price is defined when it's built - almost as if the price is built into the walls. Price writes:
[N]ew high-rise apartments are going in that have hundreds of apartments each, typically with a rent of $4000 – $8000 per month. If you let a developer build “market rate” apartments, that’s what they’ll build.
These numbers are a bit exaggerated, but that's not the point. What Price seems to ignore is the impact of construction on all the non-new units. Here's an example. I live in SF, in a market-rate apartment (though not one quite *that* pricey). But when my apartment was built, it didn't have the high rent it now has. It's a small, older apartment, once occupied by working-class families. The rent changed over time, turning an affordable home into a luxury home for a member of the upper middle class. In fact, when I moved into this apartment, I increased demand in this neighborhood, putting increased pressure on any working-class people who still happen to live here. What if, when I moved to SF, instead of moving into this apartment, I had moved into a nice fancy new "market-rate" unit in one of those towers that Price decries? I would not have increased demand in this neighborhood, and would not have put upward pressure on the rents of the families living nearby. 

Later, Price repeats the fixed-price idea when he writes:
Sorry, no. If the ‘market rate’ for newly developed apartments is substantially higher than the median rent of existing apartments, then building more market-rate apartments will make median rents go up, not down.
That sounds like simple math. And if the price of an apartment was somehow built into its walls and floors, it would just be simple math. In fact, though, it's wrong. Here's why. Suppose there are 2000 people in a city, living in 2000 apartments. One quarter of the people are rich, and rent apartments for $4000 apiece. Three quarters are poor, and rent apartments for $1000 apiece. The median rent is therefore $1000. Now build 400 fancy new luxury apartments that rent for $5000 each. And suppose no new people move to the city. All 500 rich people move into the fancy new $5000 places, leaving their old $4000 places vacant. The previously-$4000 apartments fall in price to $2000, and 500 poor people move into them, leaving 500 of their apartments vacant. These are used as second apartments, storage, or whatever. The rent of the 1500 apartments that used to all cost $1000 falls to $900 because of this drop in demand for low-end apartments. The median rent of the city's 2400 apartments is now $900, down from $1000 before.

So the "simple math" is not necessarily correct.

NIMBYs do seem to recognize this on some level. So they intuitively turn to a phenomenon called "induced demand" (though they may not realize it's called that). The theory is that if you build more housing in SF, you encourage people to move into SF, preventing prices from going down, or even pushing them up. Price espouses a version of this theory when he writes:
Tens of thousands of high-income people who would like to live in San Francisco are living in Oakland and Fremont and Berkeley and Orinda because of lower rents in those places. As market rate housing is built in San Francisco, those people move into it...There is a cascade: some people move from Berkeley and Oakland to San Francisco, which allows replacements to move from Richmond and El Cerrito into Berkeley and Oakland, and so on. Ultimately, rents in San Francisco go up, and rents in some outlying communities go down. Yes, the increased supply of housing lead to decreased housing prices on average but they’ve gone up, not down, in San Francisco itself.
It's perfectly possible in theory that this happens. In fact, this is even possible in the simplest, Econ 101 type supply-and-demand theory - it's just the case where supply is infinitely price-elastic.

Is this realistic, though? Price cites Manhattan as a counterexample - a very dense place where rents are still high. I'm not sure this counterexample applies - I see a lot more poor Black people living in Manhattan than in SF, for example. But anyway, a counter-counterexample is Tokyo, where construction seems to have been successful in keeping rents low.

The question is what would happen to SF. As I wrote in a Bloomberg View post last December, there's OK evidence that more housing would ease the city's affordability crisis:
In 1987, economists Lawrence Katz and Kenneth Rosen looked at San Francisco communities that put development restrictions in place. They found that housing prices were higher in these places than in communities that let developers build... 
[Recently, blogger Eric] Fischer collected more than 30 years of data on San Francisco rents. He modeled them as a function of supply -- based on the number of available housing units -- and demand, measured by total employment and average wages. His model fit the historical curve quite nicely. 
Recent experience fits right in with this prediction. In response to the housing crisis, San Francisco recently allowed a small increase in market-rate housing. Lo and behold, rents in the city dropped slightly
Admittedly, this data is not decisive. More SF construction might have pushed rents down a bit this year, but a big construction boom might suddenly induce a flood of rich people to decide to move to the city. It's not possible to know.

But if NIMBY theorists like Price really believe that induced demand determines SF rents, they should do the following thought experiment: Imagine destroying a bunch of luxury apartments in SF. Just find the most expensive apartment buildings you can and demolish them. 

What would happen to rents in SF if you did this? Would rents fall? Would rich people decide that SF hates them, and head for Seattle or the East Bay or Austin? Maybe. But maybe they would stay in SF, and go bid high prices for apartments currently occupied by the beleaguered working class. The landlords of those apartments, smelling profit, would find a way around anti-eviction laws, kick out the working-class people, and rent to the recently displaced rich. Those newly-displaced working-class people, having nowhere to live in SF, would move out of the city themselves, incurring all the costs and disruptions and stress of doing so. 

If you think that demolishing luxury apartments would have this latter result, then you should also think that building more luxury apartments would do the opposite. Price should think long and hard about what would happen if SF started demolishing luxury apartments. 

In any case, I think Price's posts have the following lessons for YIMBYs:

1. Econ 101 supply-and-demand theory is helpful in discussing these issues, but don't rely on it exclusively. Instead, use a mix of data, simple theory, thought experiments, and references to more complex theories.

2. Always remind people that the price of an apartment is not fixed, and doesn't come built into its walls and floors.

3. Remind NIMBYs to think about the effect of new housing on whole regions, states, and the country itself, instead of just on one city or one neighborhood. If NIMBYs say they only care about one city or neighborhood, ask them why.

4. Ask NIMBYs what they think would be the result of destroying rich people's current residences.

5. Acknowledge that induced demand is a real thing, and think seriously about how new housing supply within a city changes the location decisions of people not currently living in that city.

6. NIMBYs care about the character of a city, so it's good to be able to paint a positive, enticing picture of what a city would look and feel like with more development.

I believe the YIMBY viewpoint has the weight of evidence and theory on its side. But the NIMBY challenge is not one of simple ignorance. Nor is it purely driven by the selfishness of incumbent homeowners trying to feather their own nests, or by white people trying to exclude poor minorities from their communities while still appearing liberal (two allegations I often hear). NIMBYism is a flawed but serious package of ideas, deserving of serious argument.

Monday, May 15, 2017

Vast literatures as mud moats


I don't know why academic literatures are so often referred to as "vast" (the phrase goes back well over a century), but it seems like no matter what topic you talk about, someone is always popping up to inform you that there is a "vast literature" on the topic already. This often serves to shut down debate, because it amounts to a demand that before you talk about something, you need to go read voluminous amounts of what others have already written about it. Since vast literatures take many, many hours to read, this represents a significant demand of time and effort. If the vast literature comprises 40 papers, each of which takes an hour to read, that's one week of full-time work equivalent that people are demanding as a cost of entry just to participate in a debate! So the question is: Is it worth it?

Often, reading the literature seems like an eminently reasonable demand. Suppose I were to think about minimum wage for the first time, knowing nothing about all the research economists have done on the topic. I might very confidently say some very silly things. I would be unaware of the relevant empirical evidence. There would probably be theoretical considerations I hadn't yet considered. Reading the vast literature would make me aware of many of these. In fact, I think the minimum wage debate does suffer from a lack of knowledge of the literature.

But the demand to "go read the vast literature" could also be eminently unreasonable. Just because a lot of papers have been written about something doesn't mean that anyone knows anything about it. There's no law of the Universe stating that a PDF with an abstract and a standard LaTeX font contains any new knowledge, any unique knowledge, or, in fact, any knowledge whatsoever. So the same is true of 100 such PDFs, or 1000. 

There are actual examples of vast literatures that contain zero knowledge: Astrology, for instance. People have written so much about astrology that I bet you could spend decades reading what they've written and not even come close to the end. But at the end of the day, the only thing you'd know more about is the mindset of people who write about astrology. Because astrology is total and utter bunk.

But astrology generally isn't worth talking or thinking about, either. The real question is whether there are interesting, worthwhile topics where reading the vast literature would be counterproductive - in other words, where the vast literature actually contains more misinformation than information.

There are areas where I suspect this might be the case. Let's take the obvious example that everyone loves: Business cycles. Business cycles are obviously something worth talking about and worth knowing about. But suppose you were to go read all the stuff that economists had written about business cycles in the 1960s. A huge amount of it would be subject to the Lucas Critique. Everyone agrees now that a lot of that old stuff, probably most of it, has major flaws. It probably contains some real knowledge, but it contains so much wrong stuff that if you were to read it thinking "This vast literature contains a lot of useful information that I should know," you'd probably come out less informed than you went in. 

Of course, many would say the exact same about the business cycle theory literature that emerged in response to the Lucas Critique and continues to this day. But if so, that just makes my point stronger. The point is, a bunch of smart people can get very big things wrong for a very long period of time, and that period of time may include the present.

I have personally encountered situations where I felt that reading the vast literature didn't improve my knowledge of the real thing that the literature was about. For example, I read a lot of the macro models that came out in the years following the 2008 financial crisis. Obviously, the financial sector is very important for the macroeconomy (as more people should have realized before 2008, but which almost everyone realizes now). But the ways that macro papers have modeled financial frictions are pretty unsatisfying. They are hard to estimate, the mechanisms are often implausible, and I bet that most or all will have glaring inconsistencies with micro data. I could be wrong about this, of course, but I felt like reading this vast literature was setting me on the wrong track. I'm not the only one who feels this way, either.

The next question is: Can a misinformative vast literature be used intentionally as a tactic to win political debates? It seems to me that in principle it could. Suppose you and your friends wanted to push a weak argument for political purposes. You could all write a bunch of papers about it, with abstracts and numbered sections and bibliographies and everything. You could cite each other's papers. If you wanted to, you could even create a journal, and have a peer review system where you give positive reviews to each other's B.S. papers. Voila - a peer-reviewed literature chock full of misinformation.

In practice, I doubt anyone ever does this intentionally. It takes too much coordination and long-term planning. But I wonder if this sometimes happens by accident, due to the evolutionary pressures of the political, intellectual, and academic worlds. The academic world gives people an incentive to write lots of papers. The political world gives people an incentive to use papers to push their arguments. So if there's a fundamentally bad argument that many people embrace for political reasons, there's an incentive for academics (or would-be academics) to contribute to a vast literature that is used to push that bad argument.

And in the world of intellectual debate, this vast literature can function as a mud moat. That is a term I just made up, sticking with the metaphor of political arguments as medieval castles requiring a defense. A mud moat is just a big pit of mud surrounding your castle, causing an attacking army to get trapped in the mud while you pepper them with arrows.

If you and your buddies have a political argument, a vast literature can help you defend your argument even if it's filled with vague theory, sloppy bad empirics, arguments from authority, and other crap. If someone smart comes along and tries to tell you you're wrong about something, just demand huffily that she go read the vast literature before she presumes to get involved in the debate. Chances are she'll just quit the argument and go home, unwilling to pay the effort cost of wading through dozens of crappy papers. And if she persists in the argument without reading the vast literature, you can just denounce her as uninformed and willfully ignorant. Even if she does decide to pay the cost and read the crappy vast literature, you have extra time to make your arguments while she's so occupied. And you can also bog her down in arguments over the minute details of this or that crappy paper while you continue to advance your overall thesis to the masses.

So when I want to talk and think and argue about an issue, and someone says "How about you go read the vast literature on this topic first?", I'm presented with a dilemma. On one hand, reading the vast literature might in fact improve my knowledge. On the other hand, it might be a waste of time. And even worse, it might be a trap - I might be charging headlong into a rhetoritician's mud moat. But choosing not to read the vast literature keeps me vulnerable to charges of ignorance. And I'll never really be able to dismiss those charges.

My solution to this problem is what I call the Two Paper Rule. If you want me to read the vast literature, cite me two papers that are exemplars and paragons of that literature. Foundational papers, key recent innovations - whatever you like (but no review papers or summaries). Just two. I will read them. 

If these two papers are full of mistakes and bad reasoning, I will feel free to skip the rest of the vast literature. Because if that's the best you can do, I've seen enough.

If these two papers contain little or no original work, and merely link to other papers, I will also feel free to skip the rest of the vast literature. Because you could have just referred me to the papers cited, instead of making me go through an extra layer, I will assume your vast literature is likely to be a mud moat.

And if you can't cite two papers that serve as paragons or exemplars of the vast literature, it means that the knowledge contained in that vast literature must be very diffuse and sparse. Which means it has a high likelihood of being a mud moat.

The Two Paper Rule is therefore an effective counter to the mud moat defense. Castle defenders will of course protest "But he only read two papers, and now he thinks he knows everything!". But that protest will ring hollow, because if you can show bystanders why the two exemplar papers are bad, few bystanders will expect you to read further.

If it proves to be as effective as I think, the Two Paper Rule, if widely implemented, could make for much more productive public debate. The mud moat defense would be almost entirely neutralized, dramatically reducing the incentive for the production of vast low-quality literatures for political ends. It could allow educated outsiders and smart laypeople access to debates previously dominated by vested insiders. In other words, it could shine the light of reason on a lot of dark, unexplored corners of the intellectual universe.


Update

Some people seem to misunderstand the purpose of the Two Paper Rule. The Two Paper Rule is not about summarizing the literature's findings - for that, you'd want a survey paper or meta-analysis. It's about evaluating the quality of the literature's methodology.

Sometimes a lit review will reveal pervasive methodological weakness - for example, if a literature is mostly just a bunch of correlation studies with no attention to causal effects. But often, it won't. For example, if the literature has a lot of mathematical theory in it, a lit review will generally contain at most one stripped-down partial model. But that doesn't give you nearly as much info about the quality of the fully specified models as you'll get from looking at one or two flagship theory papers. Or suppose a literature consists mostly of literary theorizing; the quality of the best papers will depend on the clarity of the writing, which a lit review is unlikely to be able to reproduce. Sometimes, lit reviews simply report results, without paying attention to what turn out to be glaring methodological flaws.

In other words, if you suspect that a literature functions mainly as a mud moat, what you need to assess quickly is not what the literature claims to find, but whether those claims are generally credible. And that is why you need to see the best examples the literature has to offer. Hence the Two Paper Rule.

Meanwhile, Paul Krugman endorses the Two Paper Rule. Tyler Cowen is more skeptical of its universality.  

Sunday, May 14, 2017

Actually good Silicon Valley critiques?


Scott Alexander has a post with some pretty spectacular smackdowns of Silicon Valley's more exuberant critics. Some excerpts:
While Deadspin was busy calling Silicon Valley “awful nightmare trash parasites”, my girlfriend in Silicon Valley was working for a company developing a structured-light optical engine to manipulate single cells and speed up high-precision biological research. 
While FastCoDesign was busy calling Juicero “a symbol of the Silicon Valley class designing for its own, insular problems,” a bunch of my friends in Silicon Valley were working for Wave, a company that helps immigrants send remittances to their families in East Africa.. 
While Gizmodo was busy writing that this “is not an isolated quirk” because Silicon Valley investors “don’t care that they do not solve problems [and] exist to temporarily excite the affluent into spending money”, Silicon Valley investors were investing $35 million into an artificial pancreas for diabetics. 
While Freddie deBoer was busy arguing that Silicon Valley companies “siphon money from the desperate throngs back to the employers who will use them up and throw them aside like a discarded Juicero bag and, of course, to themselves and their shareholders. That’s it. That’s all they are. That’s all they do”, Silicon Valley companies were busy inventing cultured meat products that could end factory farming and save millions of animals from horrendous suffering while also helping the environment.
Alexander then goes on to look at a bunch of venture-funded startups, and concludes that most of them are either run-of-the-mill computer-related businesses, or idealistic save-the-world type of stuff, not goofy overpriced trinkets.

Alexander's post is an entertaining and timely reminder that most of the tech industry's more ardent critics are probably just using the Valley as a misplaced whipping boy for their general frustration at the larger problems of the American economy. When you live in daily fear of losing your $40,000/year job, skate on the verge of bankruptcy from overpriced medical bills, lose your house to a bailed-out bank, and realize every day that you make less than your parents did, seeing some high-flying computer whiz getting handed $10 million or $50 million seems to just rub salt on the wounds. (Also, Silicon Valley badboy Peter Thiel sued Gawker out of existence, so Gawker-derived outlets like Gizmodo and Deadspin may have a bit of a chip on their shoulder about that.)

Alexander's post asks the right question, which is "What could Silicon Valley be doing to make the world better, that it's not currently doing?" The answer is: Probably not a lot. There are a few excesses here and there, but by and large, these are just people doing the best they can, trying to both make a buck and make the world a better place. They just happen to have gotten luckier than most over the last few years.

The American public, unlike the writers Alexander spotlights, believes in the tech industry. Gallup tracks the favorability ratings of U.S. industries. The "computer industry" is the second most favorable (behind restaurants), with an enormous positive-minus-negative gap of 53 points, and the "internet industry" comes in at #7 with a positive-minus-negative of 29 points. This stands in stark contrast to pharmaceuticals and health care, both of which garner significantly negative ratings.

In other words, angry Gawker writers and pugilistic lefty bloggers to the contrary, most Americans love the heck out of tech. But OK, just to play devil's advocate, suppose we did want to criticize Silicon Valley and not end up looking foolish. What would actually non-silly criticisms look like? Here are some candidates:


1. Silicon Valley culture is still too sexist.

I haven't worked in the tech industry, so I can't speak to this personally. Female friends' anecdotes range from "Oh my God, it's so sexist" to "I don't really see any sexism". And my male friends in the industry - who are, of course, a highly selected set - almost all try to go out of their way to create a supportive, inclusive, fair environment for women. But it's hard to deny that at some companies like Uber, there is a pervasive culture of sexism. And surveys say that sexism is still fairly common in the industry, though not overwhelmingly so. Since employment at the less sexist companies is limited, that means if you're a woman looking to work in tech, there's a good chance you're going to be forced to take a job at one of the more sexist ones, and endure unwanted sexual advances, lack of promotion, casual slights regarding your technical competence, etc. Female founders also probably have extra difficulty getting funding.

How could Silicon Valley mitigate this problem? One way is for non-sexist tech industry leaders to speak out more aggressively against workplace sexism. Just let everyone know it isn't OK. Another way is for venture capitalists to hire more women (some are doing this already). But I suspect that a lot of the change will have to come from big established companies like Google, Apple and Amazon. Big institutions are probably better at creating female-friendly cultures, are more susceptible to public pressure, and have a larger profit cushion to allow them to make deep organizational changes. 10-person startups are too busy trying to survive the month to examine their gender attitudes, and closely held behemoths like Uber are generally less transparent and accountable than their public counterparts. So I expect the big public companies to lead the way.


2. Silicon Valley is late coming out with the Next Big Thing.

Tech venture financing is a hit-driven business - a few big wins make up all of the returns. VC funding is tiny compared to private equity or hedge funds - a few tens of billions per year - but three out of five big companies got their start with venture financing.

But it's noticeable that really huge successes, at least as measured by stock performance, seem rare in recent years. Facebook seems like the last really successful behemoth to come out of the Valley and it went public five years ago. Twitter's stock is down by half since its IPO more than 3 years ago, it hasn't managed to branch out beyond its core product, and it remains plagued by Nazis and bots. LinkedIn did well in the markets but was acquired last year by Microsoft. Stuff like Groupon and Zynga flamed out.

Meanwhile, the current crop of behemoth candidates seem like they could be on shaky ground. Uber, despite being valued in the tens of billions in private markets, still hasn't made any money, and if its pricing power doesn't improve it might never turn a profit; meanwhile, it's plagued by scandals, dysfunction, and an exodus of talent. Snap doesn't seem to be very ambitious as a company, and might already be getting outcompeted by Facebook even as it continues to lose money. The only real contenders for post-Facebook behemoths seem to be Netflix, which actually went public long before FB and only recently became an entertainment giant, and Tesla, whose long-term success remains to be seen.

This doesn't mean VCs themselves aren't making good returns. How much money venture funds are making depends on how you measure it, and is often something that isn't known til years after the fact. A VC fund's return on any company depends not just on where the company ends up, but on how much the VC paid for it.

But it's possible that the frontiers of technology are shifting toward capital-intensive things that favor large established players with deep pockets and long investment horizons. Machine learning, for example, might favor big companies with lots of data over plucky startups in their garages. If that's true, it would mean that tech is becoming a more mature, sedate industry - at least for now.

Anyway, I guess this only sort of counts as a "criticism", since if this is true there's not much anyone can do. Also, it was 8 years between Google's IPO and Facebook's, and 7 years between Amazon's and Google's, so I'd give it at least a few more years before we get impatient.


3. Peter Thiel is an evil man.

In the tech industry, there's a culture of not criticizing anyone publicly. I like that culture, but I'm not part of it, so I'm free to say that Silicon Valley badboy Peter Thiel looks like a bad guy. I'm kind of neutral on the Thiel vs. Gawker war - Gawker definitely had it coming, but having rich people be able to sue newspapers out of existence due to personal feuds seems like a scary precedent. But Thiel's support of Trump, his habit of making a buck off of government surveillance, and his promotion of nasty political ideas combine to make him the closest thing America has to a comic-book evil mastermind. Thiel's sort-of-reactionary ideas are confined to a small minority of techies, but the Valley's friendly culture means that even those who disagree with him are out there publicly singing his praises. I certainly wouldn't mind if tech industry people got more vocal about disagreeing with Thiel's values.


4. Silicon Valley is too blase about disruption.

Economists are rapidly learning they were wrong about something big - the economy is not as flexible and dynamic as many had assumed. People who lose their careers, to globalization or automation or regulation or whatever, often never find anything as good. Retraining has proven to be much harder than economists had hoped. Americans are moving less, too. Basically, a lot of people who lose their career jobs have crappy lives forever after.

It's not clear what Silicon Valley companies could do about this problem. If the frontiers of technology are shifting from things that complement human skills to things that substitute for human skills, or if technology is widening the skills gap and making inequality worse, it's not clear whether the boardroom decisions of Google, Amazon, etc. can do anything to alter that trend. No one really knows how much of tech progress is intentional, and how much just sort of happens automatically.

But it would be nice to see big tech companies actually worrying about this out loud. So far - and this is based on anecdote - there seems to be a general presumption in the tech industry that displaced workers will just find something better to do. It would be nice to see tech execs grapple more explicitly with the emerging realization that many displaced workers will in fact not find something better to do, but will sink into a lifetime of low-paid service work, government welfare, and unhealthy behavior.

Even if tech companies can't actually do anything about this problem, it would be nice to see more acknowledgment that the problem is real and significant.


5. Tech might be in the middle of a bust.

Venture financing is falling, indicating that some of the enthusiasm of 2013-2015 might have been overdone. But even if this continue, and tech has a bust, this is just not that worrying. Even if tech startups turn out to be overvalued, it represents almost zero danger to the U.S. economy or financial system. The dollar amounts are small, and stock in closely held tech companies is not a big percentage of any normal person's wealth. If a bunch of unicorns go bust, the vast bulk of the pain will be felt by tech workers and investors themselves, not by the broader public. So this criticism, while potentially true, is just not that big a deal.


There, you have my list of potentially valid critiques of Silicon Valley. Notice that this is pretty weak tea. #2 might not be anything the Valley could change at all, #5 is no biggie, and #3 and #4 are mostly just a matter of optics. Only #1 represents anything real and substantive that the tech industry could definitely be doing differently. All in all, Silicon Valley represents one of the least objectionable, most rightfully respected institutions in America today.

Saturday, May 13, 2017

How should theory and evidence relate to each other?


In response to the empirical revolution in econ, and especially to the rise of quasi-experimental methods, a lot of older economists are naturally sticking up for theory. For example, here's Dan Hamermesh, reviewing Dani Rodrik's "Economics Rules":
Economics Rules notes with some approbation the rise of concern among applied economists, and especially labor economists, about causality. It fails, though, to observe that this newfound concentration has been accompanied, as Jeff Biddle and I show (History of Political Economy, forthcoming 2017), by diminished attention to model-building and to the use of models, which Rodrik rightly views as the centerpiece of economic research. He recognizes, however, that the “causation ├╝ber alles” approach (my term, not Rodrik’s) has made research in labor economics increasingly time- and place-specific. To a greater extent than in model-based research, our findings are likely to be less broadly applicable than those in the areas that Rodrik warns about. Implicit in his views is the notion that the work of labor and applied micro-economists might be more broadly relevant if the concern with causation were couched in economic modeling. If we thought a bit more about the “how” rather than paying attention solely to the “what,” the geographical and temporal applicability of our research might be enhanced... 
In the end, the basic idea of the book—that models are our stock in trade—is one that we need to pay more attention to in our research, our teaching, and our public professional personae. Without economic modeling, labor and other applied economists differ little from sociologists who are adept at using STATA.
Oooh, Hamermesh used the s-word! Harsh, man. Harsh.

Anyway, it's easy to dismiss rhetoric like this as old guys defending the value of their own human capital. If you came up in the 80s when an economist's main job was proving Propositions 1 and 2, and now all the kids want to do is diff-in-diff-in-diff, it's understandable that you could feel a bit displaced.

But Hamermesh does make one very good point here. Without a structural model, empirical results are only locally valid. And you don't really know how local "local" is. If you find that raising the minimum wage from $10 to $12 doesn't reduce employment much in Seattle, what does that really tell you about what would happen if you raised it from $10 to $15 in Baltimore?

That's a good reason to want a good structural model. With a good structural model, you can predict the effects of policies far away from the current state of the world.

In lots of sciences, it seems like that's exactly how structural models get used. If you want to predict how the climate will respond to an increase in CO2, you use a structural, microfounded climate model based on physics, not a simple linear model based on some quasi-experiment like a volcanic eruption. If you want to predict how fish populations will respond to an increase in pollutants, you use a structural, microfounded model based on ecology, biology, and chemistry, not a simple linear model based on some quasi-experiment like a past pollution episode.

That doesn't mean you don't do the quasi-experimental studies, of course. You do them in order to check to make sure your structural models are good. If the structural climate model gets a volcanic eruption wrong, you know you have to go back and reexamine the model. If the structural ecological model gets a pollution episode wrong, you know you have to rethink the model's assumptions. And so on.

If you want, you could call this approach "falsification", though really it's about finding good models as much as it's about killing bad ones.

Economics could, in principle, do the exact same thing. Suppose you want to predict the effects of labor policies like minimum wages, liberalization of migration, overtime rules, etc. You could make structural models, with things like search, general equilibrium, on-the-job learning, job ladders, consumption-leisure complementarities, wage bargaining, or whatever you like. Then you could check to make sure that the models agreed with the results of quasi-experimental studies - in other words, that they correctly predicted the results of minimum wage hikes, new overtime rules, or surges of immigration. Those structural models that failed to get the natural experiments wrong would be considered unfit for use, while those that got the natural experiments right would stay on the list of usable models. As time goes on, more and more natural experiments will shrink the set of usable models, while methodological innovations enlarges the set.

But in practice, I think what often happens in econ is more like the following:

1. Some papers make structural models, observe that these models can fit (or sort-of fit) a couple of stylized facts, and call it a day. Economists who like these theories (based on intuition, plausibility, or the fact that their dissertation adviser made the model) then use them for policy predictions forever after, without ever checking them rigorously against empirical evidence.

2. Other papers do purely empirical work, using simple linear models. Economists then use these linear models to make policy predictions ("Minimum wages don't have significant disemployment effects").

3. A third group of papers do empirical work, observe the results, and then make one structural model per paper to "explain" the empirical result they just found. These models are generally never used or seen again.

A lot of young, smart economists trying to make it in the academic world these days seem to write papers that fall into Group 3. This seems true in macro, at least, as Ricardo Reis shows in a recent essay. Reis worries that many of the theory sections that young smart economists are tacking on to the end of fundamentally empirical papers are actually pointless:
[I have a] decade-long frustration dealing with editors and journals that insist that one needs a model to look at data, which is only true in a redundant and meaningless way and leads to the dismissal of too many interesting statistics while wasting time on irrelevant theories.
It's easy to see this pro-forma model-making as a sort of conformity signaling - young, empirically-minded economists going the extra mile to prove that they don't think the work of the older "theory generation" (who are now their advisers, reviewers, editors and senior colleagues) was for naught.

But what is the result of all this pro-forma model-making? To some degree it's just a waste of time and effort, generating models that will never actually be used for anything. It might also contribute to the "chameleon" problem, by giving policy advisers an effectively infinite set of models to pick and choose from.

And most worryingly, it might block smart young empirically-minded economists from using structural models the way other scientists do - i.e., from trying to make models with consistently good out-of-sample predictive power. If model-making becomes a pro-forma exercise you do at the end of your empirical paper, models eventually become a joke. Ironically, old folks' insistence on constant use of theory could end up devaluing it.

Paul Romer worries about this in his "mathiness" essay:
[T]he new equilibrium: empirical work is science; theory is entertainment. Presenting a model is like doing a card trick. Everybody knows that there will be some sleight of hand. There is no intent to deceive because no one takes it seriously. 
In addition, there are also paper groups 1 and 2 to think about - the purely theoretical and purely empirical papers. There seems to be a disconnect between these two. Pure theory papers seem to rarely get checked against data, leading to an accumulation on the shelves of models that support any and every conclusion. Meanwhile, pure empirical papers don't often get used as guides to finding good structural models, but are simply linearly extrapolated.

In other words, econ seems too focused on "theory vs. evidence" instead of using the two in conjunction. And when they do get used in conjunction, it's often in a tacked-on, pro-forma sort of way, without a real meaningful interplay between the two. Of course, this is just my own limited experience, and there are whole fields - industrial organization, environmental economics, trade - that I have relatively limited contact with. So I could be over-generalizing. Nevertheless, I see very few economists explicitly calling for the kind of "combined approach" to modeling that exists in other sciences - i.e., using evidence to continuously restrict the set of usable models.