Macro theory vs. string theory

After reading this cool article about the debates surrounding string theory, someone on Twitter asked me to do a post comparing it to macro theory. Well, I did that a long time ago back when I was a snarky young lad, but as someone at Bloomberg once said to me, "If you know a good story, tell it from time to time," so here goes.

String theory and macro (business cycle and growth theory) have one big problem in common: They're not easily testable. They are not untestable. There's a big difference. With string theory, if you could build a really really really huge powerful particle collider, you could probe particles down to the Planck scale, and you could establish whether or not particles are shaped like tiny strings. With macro, if you could observe a whole bunch of planets with no mutual trade between them, or if you could establish a world government and get it to do random policy experiments, you could definitely test growth and business cycle theories.

The problem is, we can't do any of these things. String theory deals with ultra high-energy phenomena, and macro deals with aggregate time series, and those are just very difficult things to observe. In both cases, one big problem is the lack of experimental controls. High-energy phenomena can sometimes be indirectly observed by looking at black holes or the echoes of the Big Bang, and the effects of time-series macro phenomena can be observed every quarter. But you can't really put these things in a lab (well, not real versions of them, anyway), so you can't poke and prod them and explore counterfactuals and deduce the underlying structure, etc.

(Update: This was a little confusing. Yes, you can test theories by looking at non-controllable phenomena. If a macroeconomic theory says "There will be a huge recession next year" and there is no recession, the theory has major problems. But because you can't do controlled experiments, you often have to deal with multiple competing theories that fit the same data. Did they U.S. economy recover from the Great Recession because of easy monetary policy, or in spite of it? And so on. In physics, this is especially problematic when looking at data from the Big Bang, of which there is only one. As you might expect, there are tons of theories floating around about how the Universe began, some of which incorporate string theory.)

So that's one similarity. A second is that studying both has led to useful spinoffs. Studying macro has led to time-series techniques like GARCH models, and arguably to much of microeconomics itself. Studying string theory has led to some cool mathematics, and to some other useful modeling techniques that have been applied in other areas of physics.

Now for the big difference. String theory is not really relevant to the real world. Although string theory techniques have found real-world applications, the theory itself has not and probably never will. No one in our lifetimes or our posthuman grandchildren's lifetimes is likely to build a machine that runs on string theory. Even if the theory - in all its myriad offshoots - turns out to be total crap, it doesn't really matter whether anyone believes in it or not. You might complain that we're wasting our society's brightest minds by paying them to study an untestable theory, but you'd then have to make the same complaint about much of mathematics itself.

In that sense, string theory is "safe" for the world. Macro theory might not be. If we pick the wrong macro theories, we could enact policies that cause real human disasters for millions. In fact, this probably happens quite often.

When data don't give you a guide to policy, major policy failures are inevitable. But what if macro is not totally untestable, but just mostly untestable? What if the evidence gives us a very weak but nonzero signal about which theories are good and which are bad?

This is the danger. If macro policy is hugely important and the signal from data is very weak, then small sociological phenomena within the econ profession - conformity, political bias, etc. - might cause great harm to real people. And bad scientific techniques - ignoring data entirely while placing too much faith in plausibility - might also cause cause real-world harm.

To sum up: If string theory happens to be complete B.S., there is essentially no loss to society. If prevailing macro theory (New Keynesian models now, RBC in the 80s, etc.) is complete B.S., there might be no loss, but there might very well be a big loss. This is probably why the public gets a lot madder about macro debates than about high-energy physics debates.

Academic B.S. as artificial barriers to entry

"Stratos has never sampled the full terror stalking the stars." - David Brin, Glory Season


Paul Romer complains of "mathiness" in macroeconomics. Paul Pfleiderer talks about "chameleon" models. Ricardo Caballero says macroeconomists encourage the "pretense of knowledge". Everywhere, people complain about economists' fetish for pointless model-making. And of course, some folks accuse the economics profession of being a front for laissez-faire ideology.

But we should remember that compared to other disciplines, econ is in great shape. A friend just sent me a paper by Ananya Roy, a professor of urban planning at UCLA, entitled "What is urban about critical urban theory?" Here is an excerpt from the abstract:

This essay discusses how the “urban” is currently being conceptualized in various worlds of urban studies and what this might mean for the urban question of the current historical conjuncture. Launched from places on the map that are forms of urban government but that have distinctive agrarian histories and rural presents, the essay foregrounds the undecidability of the urban, be it geographies of urbanization or urban politics. What is at stake is a critical urban theory attentive to historical difference as a fundamental constituting process of global political economy and deconstruction as a methodology of generalization and theorization.

And here is the spectacular final paragraph:

I conclude then with the invitation to read the urban from the standpoint of absence, absence not as negation or even antonym but as the undecidable. I conclude too with the provocation that theory, including a theory of the urban, can be made from the tealcolored building at the edge of the world that is the Dankuni municipality, a panchayat office repurposed for urban government. But in a gesture befitting the task of provincializing the urban, I note that the dedication plaque for the panchayat building references a fin de siècle poet, Jibanananda Das and his writings on “rupasi bangla,” or beautiful Bengal, envisioned as rural and verdant. But Das is also the first urban poet of Bengal, with a set of starkly neo-urban poems that are now etched into the region’s self-imagination of urban modernity. The plaque can thus be read as a serendipitous anticipation and premonition of the urban yet to come but its rurality cannot be effaced or erased (Figure 4). The sign of a constitutionally demarcated urban local body, it is the undecidability of the urban.

To many readers not steeped in critical theory, this may sound like a broken fire hydrant of nonsense. One may be tempted to reach for a copy of Pennycook et al.'s paper, "On the reception and detection of pseudo-profound bullshit."

But I don't think critical theory is simply the academic equivalent of meaningless auto-generated guru wisdom. My guess is that it's actually something else: Obscurantism.

Here's what I kind of suspect is going on.

For a given level of demand, supply restrictions generally push up price. You don't want to have any old dork walk in off the street and get a full professorship in urban studies. That would send salaries crashing, and prestige as well.

But what if urban studies is just inherently a really easy field? (I'm not saying this is true, I'm just being hypothetical!) What if all the remaining big truths could be uncovered by running a few regressions in Stata? In that case, the supply of potential urban studies profs would be really big. Danger!

If existing urban studies profs can form a cartel, they can artificially raise the barriers to entry and bring supply back down again. Cartelization in academia doesn't seem that hard, since admissions, hiring, and tenure committees are already cartels, and since the barriers to creating new universities and new top journals are very high.

The barriers to entry will probably be some combination of A) a psychometric test, and B) an ideological loyalty test. These tests are relatively easy to administer. They also take advantage of natural supply restrictions - very high-ability people (along whatever dimension you want to measure) are relatively rare, and ideological buy-in is limited by the diversity of ideas in society. For example, there are just not that many people who are both A) really good at parsing dense paragraphs of text, and B) deeply committed to a quasi-Marxist lefty ideology.

Artificial entry barriers provide a tidy explanation for the rise of "critical theory" in humanities, urban planning, anthropology, and sociology departments. Critical theory is basically just the practice of taking lefty social criticism - of the type you might find in any college dorm - and dressing it up with a bunch of neologisms and excess verbiage. Stephen Katz of Trent University explains this in an essay entitled "How to Speak and Write Postmodern". He gives some hypothetical examples:

For example, let’s imagine you want to say something like, “We should listen to the views of people outside of Western society in order to learn about the cultural biases that affect us”. This is honest but dull...[Instead]  say, “We should listen to the intertextual, multivocalities of postcolonial others outside of Western culture in order to learn about the phallogocentric biases that mediate our identities”... 

You want to say or write something like, “Contemporary buildings are alienating”. This is a good thought, but, of course, a non-starter. You wouldn’t even get offered a second round of crackers and cheese at a conference reception with such a line. In fact, after saying this, you might get asked to stay and clean up the crackers and cheese after the reception...[Instead, say] “Pre/post/spacialities of counter-architectural hyper-contemporaneity (re)commits us to an ambivalent recurrentiality of antisociality/seductivity, one enunciated in a de/gendered-Baudrillardian discourse of granulated subjectivity”.

As multivocalities on the internet say: LOL.

The supply of super-lefty people who are able to parse artificially dense text is limited, but not so limited that it's hard to find people who are willing to do it for $70,000 a year. Meanwhile, student demand for humanities, anthropology, urban studies, and sociology majors is probably pretty inelastic, so university demand for professors in these areas is probably inelastic. Hence, for departments and journals in these fields to make "critical theory" a soft requirement for professors is probably an effective way of keeping their salaries (and job perks) as high as they are.

It seems pretty obvious that humanities departments have been almost entirely consumed by this sort of thing. Any semblance of objectivity (whatever that would even mean in the humanities!) is gone, replaced by pervasive quasi-Marxist doofiness. And humanities scholars' research work now appears to largely consist of parsing and writing artificially dense text. As for the social sciences, anthro seems to have taken some big steps in this direction, and sociology more modest steps.

How about economics?

The econ profession as a whole is shifting toward empirics, and possibly even toward reduced-form empirics. Since the flood of data is a recent thing, most of the new insights in the field over the next decade or two will probably be gained from doing this kind of work - as John Cochrane says, "The stars in their 30s are scraping data off the internet."

Scraping data off the internet is cognitively demanding work, but not that demanding. Without artificial entry barriers, the data flood would probably increase the supply of people qualified to be economics professors. To preserve their high salaries and high levels of intellectual prestige, it therefore behooves the economics profession to create some artificial barriers.

Econ isn't going for ideological tests - or at least, not very much. There's too much pressure from the world at large to stay objective. So the entry barriers rely mostly on psychometric tests. Mathematical theory, of the type economists do, is hard to do - much harder, for most people, than parsing dense paragraphs of woo-woo "critical theory".

This could explain the pressure on empirical economists to include a structural theory section that has little relation to the reduced-form empirical analysis that forms the core of a paper. It also might explain why even though the economics literature is more and more filled with reduced-form empirical studies, theory papers are still very common on the job market.


Updates

Brad DeLong is annoyed with me for picking on Ananya Roy, and links to an essay of hers that he implies is much better than the paper I cited (and which is written in more-or-less plain English). But I didn't quote one of Roy's papers because I think she's a bad urban studies researcher (how would I even know if that were true?). My conjecture was that many successful urban studies profs will have written at least one or two papers like this - much like almost any top macroeconomist will have at least one or two theory papers with unrealistic assumptions and complex math that can't really be tested against data. I'm not trying to single out any individual for criticism.

Some commenters have been suggesting that critical theory is actually on the wane in the humanities. That is interesting; I don't know many humanities people, so I'm pretty uninformed about recent trends. I knew a bunch of Michigan humanities PhD students back in grad school, and they were all very into critical theory. I also met an incoming literature prof at Stony Brook who said she does "theory", and when I asked her "What kind of theory?", she blinked in surprise and asked "Are there multiple kinds?". So anecdotes suggest it ain't dead yet...

A commenter points out that, as usual, Feynman did this snark way before I did.

Efficiency in growth's clothing? (reply to John Cochrane)

In response to a John Cochrane policy paper on growth policies, I wrote a post chiding Cochrane for selling efficiency policies as long-term growth policies. Cochrane now has a long and rather testy reply to Yours Truly. The basic message of the reply is: "If level effects are really, really big, they tend to look like long-term growth effects."

Yep. That's right.

Cochrane's original paper described the impact of growth policies by drawing an analogy with the period from 1950-2000:

If the US economy had grown at 2% rather than 3.5% since 1950, income per person by 2000 would have been $23,000 not $50,000. That’s a huge difference. Nowhere in economic policy are we even talking about events that will double, or halve, the average American’s living standards in the next generation. 

To get a policy change that had that effect, we'd need:

A) A doubling of the level of detrended steady-state GDP, and

B) Frictions in the economy and/or the policy-making process that smoothed that change over 50 years.

Cochrane, in his new post, asserts that (A) is possible (I'll get to that later). He doesn't mention (B). How much of the growth we enjoyed from 1950 to 2000 a result of policy improvement? Cochrane certainly believes strongly that Reagan's policies caused a lot of the growth in the 80s and 90s, but how about the equally impressive growth from 1950-1980?

When I look at U.S. history I see a pretty smooth growth trend (I'll use a David Andolfatto tweet instead of just grabbing FRED data, because it's funnier!):




See the big growth takeoffs from policy liberalization? Neither do I.

Of course, at any time we could have chosen to become North Korea, at which point the line would have crashed and burned, so in some sense the whole upward trend was due to policy. But I think that it's hard to look at this steady upward march and see anything other than the steady improvement of technology. If you look at other countries, you see that they did just about as well as us (or better) over this time period, and that our growth and theirs was highly correlated. To me, that says that it was technology (and trade), not policy changes, that drove most of the global growth trend.

But OK, OK, if we DID engineer policy changes that doubled our income, would it matter to us if it was abrupt or spread out over decades? No, it would not. So Cochrane is right - a "level effect" that huge really would be party time.

So maybe my Bloomberg piece didn't really manage to express what actually annoyed me about Cochrane's paper. I guess it wasn't just about "growth effects" vs. "level effects". It was about proof versus conjecture.

Cochrane, by citing the growth we enjoyed from 1950-2000, and then telling us that we can enjoy similar growth if we do his preferred policies, seems to imply that this is something we've done before and therefore something we can do again. To me, that doesn't seem to fit the facts, even if you give Reagan as much credit as Cochrane gives him. To me, it seems pretty obvious that liberalizing policy changes produced, in the past, at best small bumps in the trend of steady technological progress.

But OK, OK, John obviously really believes that his preferred policies would engineer a HUGE change - a doubling or tripling - of our potential GDP. I may disagree with that prediction, but I guess I shouldn't dismiss it out of hand. I definitely think that pointing to the growth from 1950-2000 as an example of what we could achieve with deregulation is misleading. But ultimately that springs from the fact that my priors about how the world works are just very different from John's.

Anyway, on to John's other points:


Cochrane Point 1: If China did it, why not us?

Because China started off very far away from the technological frontier. If you liberalize your economy in ways that allow you to start importing and applying foreign technology, you will be able to grow fast. Some of China's growth is simple Solow capital catch-up, but some of it is a sudden and dramatic influx of foreign technology.

The U.S. is at or near the technological frontier, so I assume it would be a lot harder for us to do what China did. I think this again illustrates a difference in the way John and I think about productivity - he seems to instinctively think of policy, I always instinctively think of technology.


Cochrane Point 2: We could triple or octuple our GDP by making it easier to do business!

I was pretty skeptical of this argument. Reverse causation is a much bigger problem than Cochrane seems to think; he dismisses it with some anecdotes, but I don't think it can be dismissed. Another problem is the poor fit of the regression line Cochrane shows - see Kevin Grier for more on this point.

A third problem is that the World Bank's "Ease of Doing Business" indicators are constructed from surveys, which have all kinds of huge methodological issues (which Cochrane himself has pointed out in other contexts). They are not objective measures of the ease of doing business.

What this means is that A) the hypothetical "frontier" that the World Bank and Cochrane construct may not exist, and B) attempts to reach that frontier might actually hurt rather than help growth/efficiency. Alternatively, engineering improvements in the rankings might help a lot for poor countries but not help much for rich countries.

If we look at real-world examples, only one single country - Singapore - has both A) substantially higher GDP than the U.S., and B) better performance on the World Bank rankings. Singapore's GDP is about 60% higher than ours in PPP terms, so if we could reach that level it would indeed be great. They are #1 in the World Bank's rankings. The U.S. is #7. Countries #2-#6 are actually all a bit poorer than the U.S.

But maybe we can emulate Singapore.


Cochrane Point 3: Permanent growth effects might not actually exist.

Yep, true. Cochrane points to a Chad Jones paper showing this, which I actually already knew. It's a good paper. In fact, if you just use a simple Solow-type exogenous growth model you get a similar conclusion - there's nothing you can do to boost growth in the very long run.

Whether or not this is true, it is orthogonal to my point. I was talking about the overselling of efficiency-based policies by appealing to the history of long-term growth. I did not intend to claim that there are other clearly identifiable policies we could take to boost the growth trend for 50 years.


Cochrane Point 4: I should not have used the word "conservative".

I think this word was appropriate. First of all, it's one that the public intuitively understands. Second of all, it accurately communicates Cochrane's apparent love for Republican politicians, especially Reagan. In an earlier post (also a rebuttal to Yours Truly), Cochrane wrote:

In 1980 Ronald Reagan announced some pretty radical growth-oriented policies, at least by the standards of the time. (Not much new since Adam Smith, of course.) The standard liberal commentators made the standard objections: voodoo economics, numbers don't add up, it will take generations of unemployment to lower inflation, the debt will explode, and so forth. (Plus, the Soviet Union will be there forever, we might as well get along.)  Reagan offered optimism; won, malaise ended, we won the cold war, and there was an economic boom.

I would note that:

A)  John uses the word "liberal" to describe people who disagree with his desired policies, and most people use "conservative" to mean the opposite of "liberal".

B) Reagan's policies included more tax cuts, while the big deregulations came under Carter. Deregulation is the centerpiece of Cochrane's current growth proposals, so it's interesting that he credits Reagan 100% and Carter 0%. If the World Bank's Ease of Doing Business rankings had been around at the time, I think Carter's reforms would have resulted in a lot more improvement than Reagan's, but Cochrane gives Reagan all the credit. That sort of feels like a politically "conservative" view of things.

So I don't think that any harm was done by the use of "conservative".