The problems of the economics profession

Ronald Coase has an interesting new piece titled Saving economics from the economics profession. You may like to see What is wrong with Economics on this blog.

Last week, in the US, I heard that the number of Ph.D. graduates coming out vastly exceeds the number of academic job openings. Most economics Ph.Ds. are going to end up in non-academic jobs. In fields like Physics, the basic arithmetic became clear early on. Each academic in a research university produces 12 Ph.D. students, on average, over his or her life. In steady state, 11 of them have to go into non-academic lives. For some time, in Economics, this phenomenon was masked by the rise of business schools and schools of government, which recruited a lot of economists. With that transition largely behind us, the simple logic of 1-in-12 comes back to hit us.

I feel the profession is not doing enough to prepare the 11-of-12 economics Ph.D. students for a life in the real world. I am a sunny optimist on the importance of economics in the real world. Whether it is Google or a hedge fund or a consulting firm: I think a good economist has a lot to say. But what we do to Ph.D. students is pretty bad. The skills required to succeed in academic economics seem to be precisely unlike the skills required to engage with the world. I feel that fairness to the students requires turning this upside down. We should be primarily training Ph.D. students to gear up to be useful in the real world, for only a tiny fraction of them will go back into academics.

Academic economics in India suffers from one additional layer of trouble: the legacy of development economics. India has moved on. Only 15% of Indian GDP is agriculture; the labour force is moving away from agriculture; only 20% of India is below the poverty line. This implies that development economics is of little use in thinking about India. Whether it is P. Chidambaram or Mukesh Ambani, the decision makers of India are not too interested in development economics.

The early days of physics shows us a nice three-step story. First, the datasets fell into place, with Tycho Brahe. Then came the empirical regularities, with Kepler. Once Kepler’s laws were firmly established as hard facts of the data, you could curiously ask: Why might this be the case? And this gave us theory, in the hands of Newton. In economics, and particularly with economics in India, we are struggling with the first phase. We barely observe the economy.

When the physicists did not observe the world, the frontier lay in observation (Tycho Brahe), and not in the guys doing angels on pinheads. But in economics, in the early years, in the absence of data, the field got dominated by mathematicians analysing artificial worlds, the bulk of which was angels on pinheads exercises. Instead of looking at the world, we looked at blackboards and made up assumptions. Research papers got written by looking at other research papers, rather than looking at the world.

I am optimistic about where we will go from here, for the computer revolution is finally giving us datasets where there is high quality observation of the economy. E.g. retail stores are capturing scanner code data, financial exchanges see every order, massive databases of census or tax authorities are being prised open, google trends data is available, satellites measure illumination at night and give us estimates for the GDP of each square kilometre of the country every night, etc. The future of economics lies in data science. Just as astronomers are drowning in the data coming out of telescopes, we in economics will shake our heads in wonder, as we find our way around immense treasures of large datasets of high quality.

Yet, at present, most economists and economics Ph.D. students are focused on theory, or the old perspective where economics is seen as a part of axiomatic mathematics and not as an observational science. For most people in economics, there is a certain willingness to accept bad data and bad econometrics since all this is (in any case) just an excuse to get on with the thing that really matters, the model. Matters are made worse, in India, by the typical Western referee who does not ask questions about data quality. This gives the economist in India zero incentive to be careful about measurement, and gives us an equilibrium replete with garbage-in-garbage-out.

I don’t want to overstate the problem. Things have changed enormously when compared with the 1970s and 1980s, when economics was almost entirely dominated by theory. Today, the most important work in the profession is applied. Applied papers get more citations. The ship is turning. But as Ronald Coase is saying, it’s still far from where it needs to be.

Academic economics is a self-sustaining system, on the strength of the tuition fees paid by a large number of undergraduates who register for these course. There is relatively little pressure to change. The impetus for change will come from four directions:

  1. While wages for a small number of the superstars of the profession are sky high, most academic economists are not paid that well and are not experiencing real wage growth. This gives an incentive for some to engage with the world through consulting. Their work will matter.
  2. As Larry Summers has emphasised, a strength of the business school and the school of government (and the think tank) is that they engage with reality. They have incentives to look at the field with new eyes. The work done in these places will matter.
  3. The 11 of 12 freshly minted Ph.D.s who show up in the real world and puzzle over it matter a great deal. For the vast majority of them, the Economics Ph.D. will recede in their minds like a bad dream. A small fraction of them will do stuff that matters.
  4. The people with skills in data science will do unexpectedly cool things with the new datasets where we observe the economy. This stuff will matter.

Leave a Reply




You can use these HTML tags

<a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>