Arnold Kling  

The MIT Connection

PRINT
We Are All Austrians Now?... No Comment...

Bloomberg News has an article on the many MIT economics Ph.D's currently in top policy slots at central banks around the world. The article paints a rosy picture of grad school life during the period when we were there, including:


Its emphasis on solving policy problems instead of perfecting theories... "We turned out students who were actually interested in macroeconomic policy and understanding daily events and not in showing off." [quote from Robert Solow]... the emphasis at the university was on trying to understand "real phenomena and how the world works" rather than seeking to come up with elaborate theories or better techniques.

Oh, barf. Or, to put it more politely, these passages struck a false note. In reality, the MIT training was all about the math. All about the math. Abstract mathematical methods were known as "tools" and graduate education was about "adding tools to your tool kit."

I think that Solow really was focused on useful economics as opposed to pure math, but he was already being shoved aside by the young mathematical theorists. The main message you learned in graduate school was that more difficult math equated to better economics.

I failed to accumulate an impressive tool kit. Writing my dissertation under Solow was not a good career move, because Dornbusch and Fischer controlled the placement of macro students.

With no attractive academic offer, I ended up spending my career in organizations. Along the way, I learned some things that you don't learn in graduate school. In an actual business, you are not given a demand curve and a cost function; instead, you grope. The internal alignment of an organization cannot be taken for granted; instead, a lot of time and effort goes into just trying to keep people focused on common goals. Day-to-day life in a organization is a soap opera, with individuals and departments often working at cross-purposes. No one, including the CEO, has full knowledge or control.

When I came to think of every organization as a dysfunctional family, it affected my mental model of markets and government. I don't assume that organizational units know what they are doing. Instead, I ask: what institutional pressures exist that ensure that more effective units survive and less effective units disappear? That in turn leads me to be relatively pessimistic about government as an institution, because I see the tools of voice (elections and representative democracy) as less effective than the tools of exit (consumer choice, leading to profit and loss).

MIT's contribution to producing technocrats was what it did not teach. It did not teach humility. It did not teach that the world is too complex for technocrats to control.


Comments and Sharing


CATEGORIES: Economic Education



COMMENTS (14 to date)
david writes:
That in turn leads me to be relatively pessimistic about government as an institution, because I see the tools of voice (elections and representative democracy) as less effective than the tools of exit (consumer choice, leading to profit and loss).

I do wish you would attempt to formalize this; the conditions under which competitive elimination lead to good outcomes are no less restrictive than the conditions under which elections lead to good outcomes. It is easy to wave the flag of consumer selection with the full artifice of neoclassical theory under your feet - but you jettison those, so what do you have?

The First Welfare Theorem is no more tolerant of institutional vagaries than public-choice theories of elections are, and at least the latter has legitimacy of outcomes built-in.

Practice is no more encouraging than theory: historically, the state decisions to cease to enforce contracts gauged as abusive in some way was an inherently political judgment; it did not spring from the ether of natural law fully-formed. While we might grumble about the Feds proactively investigating Microsoft, we should readily agree that it would be a Far Worse Thing if Microsoft could enforce collusion via contract law.

RGV writes:

What do you mean by controlling placements? How does the econ PhD market work? I got my PhD in CS and no one was controlling anything. You had four choices of advisor type: a) Famous but won't do a darn thing when you gradaute b) Famous but will rally his students or make phone calls c) not famous but work as much as he/she can to get you a job d) neither famous nor useful. Seems like even a powerful voice like Solow can't do much in the econ market?

PrometheeFeu writes:

This reminds me of the "lean startup" methodolody. The basic idea of the methodology is that certain types of businesses or organizations operate in a highly uncertain environment. When that is the case, all your time and effort should be devoted to learning about that uncertain environment. I somehow doubt that mathematical economists are able to model such firm behavior...

Jack writes:

Back in the late 70s and early 80s (and maybe still today), MIT was the top economics dept; their grads got the top academic jobs, the most prestige, and since top policy jobs usually go to the most prestigious academics, ergo MIT grads. But Dr. Kling is right that reading the research output of MIT grads after 1980, there is nothing particularly empirical or policy-oriented. On the contrary, a charge made by some dissatisfied economists is that after the 1980s, every econ dept tried to become a second- or third-rate MIT, meaning all theory. In the last 15 years, MIT and Harvard non-macro grads have been more empirical. But I see no difference among the macro PhDs. It is still DSGE, dynamic prog, contraction mapping thms, etc.

@RGV: In economics, at least in the top-50 departments, job placements are heavily influenced by Dept decisions of where to place their students. A job market candidate will be said to be "top-20 dept material" or "top-100 dept material but not top-20", etc. This assignment is no guarantee, of course, but it is the first hurdle, and if you are assigned "top-100 but not top-20", you will not get an invitation to a top-20 dept. I assume your paper will not even be considered. Funny that economics is about competitive markets, etc, and yet we have such a top-down controlled economy of job placements...

Mark Michael writes:

I second the "oh barf" comment. The WSJ ran a front page article on the Fed's 2006 minutes of the FOMC last week. In 2006, we were coming to the near peak of the housing bubble, yet, the brilliant economists advising the FOMC were hardly ringing the fire alarms, waving the red flags of the coming bubble bursting. A little humility from the git-go would have been more comforting.

We'll have to wait another year to read the minutes from 2007. I wonder if they'll reinforce Dr. Kling's wish that we could see a little more humility in the Fed's internal discussions of the economy.

Tangent. I wonder if the (apparently) easy way the Fed and the economy handled the dot.com bubble and its bursting in 2000-03 gave those lacking-in-adequate-humility Fed types too much confidence they could handle the likely housing bubble bursting?

"Gee, we got through that one okay. A housing 'correction' won't be any tougher to handle." Hence, the hubris.

Consider the period 1920 to 1935. There was a sharp depression between 1920 and 1922. The GDP and employment dropped at a faster rate than 1930 to 1932, believe it or not. Yet, the Fed under Benjamin Short and the Harding administration handed that depression okay. Unemployment peaked around 11% and then fell as the economy recovered and the Roaring Twenties started.

That may have given the 1920s Fed the hubris to not worry about the obvious "bubbles" occurring in the economy in the late 1920s. In fact, I recall reading that Benjamin Short talked about giving the economy a little goose around 1927 or 1928 by lowering interest rates some more. He did, and the stock market rose sharply. He was proud of himself. I believe there was more buying stock on margin back then than in the dot.com bubble. I think debt levels in the late 1920s were also very high, similar to what we saw in 2006-2008.

Short died in 1928 or 1929 and his successor did not have the reputation he did. The Fed did a bunch of wrong things that Friedman & Schwartz documented in their famous work. Some of the Fed governors were pretty ignorant of banking and economics, I've read, at that time. Political hacks appointed as favors to powerful pols.

Anyway, maybe the 1920-1922 depression created the same kind of hubris that the dot.com bubble did at the Fed. The very high levels of debt in the Great Depression are like those the Fed faced in 2006 & following. At least Bernanke pumped money into the system this time, unlike the Fed in the 1930s. But, I don't see how we get back to reasonable levels of debt on our current path, other than austerity.

Eric Rasmusen writes:

The secret of MIT's success around 1980 was that it occupied a middle position in math. MIT profs used a little math and theory in everything, including economic history, but there was hardly anybody (one assistant prof who later changed his mind, in 1980) who was a pure math economist, or even a pure math econometrician. The MIT style was to build small simple partial equilibrium models to illustrate particular effects one at a time. That's useful for policy work. Now, it's standard economics--- the non-math economists have mostly died out, general equilibrium has lost its glamor, and fancy math has lost its power to impress.

kerokan writes:

"because I see the tools of voice (elections and representative democracy) as less effective than the tools of exit (consumer choice, leading to profit and loss)."

I always thought of elections as tools of exit because voters can threaten the parties by voting for another party, which is a form of exit. What am I missing?

andy writes:

@david: I'm not sure if that was what Mr.Kling meant, but I would certainly see evolution as a better way than voting about the structure of the DNA of our children. The possibility to try many solutions and eliminate the worst seems to me significantly more effective than electing and settling on one solution.

John Bailey writes:

Wow, very insightful.

Jeremy, Alabama writes:

Hi Kerokan,

We each get a vote, but we are all subject to the winning faction's 'product' regardless of how our vote was 'spent'. It is not a form of exit.

Rings true. What size organisations are you talking about? Because I think the experience differs greatly with size (esp logarithmically).

srp writes:

Let me second Eric Rasmussen's comments and append a few:

1) Dornbusch and Fisher were intensely interested in policy questions. That's why Fisher ended up in government service, and why their macro course was all about the strengths and weaknesses of each of the zillions of models they covered for drawing policy conclusions.

2) MIT leaned on students to do empirical work. They really discouraged pure theorists and took pride in the applied econometric skills of their graduates. Of course, lots of them ignored that advice...

3) MITers saw themselves as the sensible middle-of-the-roaders in many areas, not like those extremist types at Chicago and Minnesota. That applied both methodologically and substantively. Remember, at that point the high-tech macro was coming out of Lucas, Sargent, etc., and it was taught at MIT more in terms of self-defense than out of conviction. You had to know what those bomb-throwers were saying and be able to manipulate their models, but you weren't supposed to take their theoretical ideas that seriously. Then they figured out how to use the same tools and tweak the assumptions, and the rest is Neo-Keynesian history.

david writes:

@andy: evolution, as I'm sure you are aware, places zero weight on the welfare of the vehicles carrying the units of selection.

It has no problem opting for a solution where there are regularly many poorly-resourced individuals, most of whom die, for example...

Likewise, even an optimistic competitive selection a la Friedman has one answer for the individual who regularly makes bad choices: you run out of income and starve, and your impact on wider equilibrium goes to zero. All this to eliminate some given behavior of making bad choices? Surely we might not regard this as desirable. And competitive selection is not generally so optimistic (bad equilibria, etc.).

Greg writes:

Beautiful, Arnold. Thank you. If only they would listen

Comments for this entry have been closed
Return to top