Thursday, May 4, 2017

Carl F. Christ, 1923-2017

Carl F. Christ (1923-2017) was first trained as a physicist at Colorado College and the University of Chicago. He later wrote: "Upon graduation in 1943 I went to work in Chicago on the Manhattan Project, the atom bomb project. Information circulated freely in internal seminars for the professional staff, and we soon found out what we were working on. At the time I was living in Concord Co-op House, founded and mainly occupied by pacifist Quakers. After  much thought I decided that I was not a pacifist, and that I would prefer the U.S. rather than Germany to be the first to develop the bomb. (It was known that they were working on it too.)." But after the war, he decided "to look for a
social science which I could use my mathematics."  He went to the University of Chicago for a PhD in economics, and spent most of his professional career at Johns Hopkins University. A substantial number of students learned their introductory econometrics from Christ's textbook, Econometric Models and Methods, first published in 1966. Here's an obituary from Johns Hopkins, and here's one from the Baltimore Sun.

Back in Spring 1990, Christ wrote a nice biographical essay for The American Economist called "A Philosophy of Life" (pp. 33-39, available through JSTOR, and the source of the quotation above).  He gives a sense of academia as it used to be:

"On the Lake Michigan beach in Michigan, where I spent summers as a child, and still do now, was an elderly gentleman in swim trunks, known to the children as Mr. Knight. I was quite surprised on entering graduate school at Chicago to find him, fully clothed, teaching economics. He had great skepticism of anyone who appeared to know all the answers. He told his classes, "As soon as a person gets a theory, he's lost. ...
"I owe a great debt as well to Leonid Hurwicz, who was not even on my committee. I had presented my paper at the famous Cowles seminar (where only "clarifying questions" were allowed until the discussion period arrived, and where "clarifying questions" from Arrow, Hurwicz, and Modigliani began to fly as soon as the seminar began). I thought I was finished with my thesis, until Hurwicz invited me to come down to the University of Illinois (where he had just moved) because he had some suggestions for me. I stayed in his World War II prefab hut with him and his family for two days while he went over my thesis with a fine-tooth comb. When I left I was very depressed. But soon I realized that he had done me an enormous favor, and that my thesis was much improved as a result."

And here's a comment about what econometrics can teach and what economists can know:

"I used to believe that it was possible to build and estimate an econometric model that would represent an invariant law of economic behavior, valid for many places and for long periods of time. I no longer believe this, because I have yet to see an econometric model that continues to describe new data with no change in its parameters. Instead, I believe that economic reality is so complex that the best we can expect of an econometric model is that it may approximately represent the relations among its variables for a limited place and time. Such an approximation may be very useful, and may permit us to make forecasts for short periods into the future. But until we have much more knowledge about human biology and its relation to economic, social, and political behavior I think we will not achieve econometric models that are invariant over wide reaches of space and time.
"What then do economists really know? I think we know a great deal. (Remember that this knowledge must be only tentatively accepted.) Most of our knowledge is about equilibrium situations and how they change, rather than about the path followed by the economy on the way to equilibrium. We know some rather simple things that can be stated in nontechnical terms. For example, we know that as incomes rise, a smaller fraction of income is spent for agricultural and extractive products, and a larger fraction for processed goods and for services. We know that increases in a country's output per person require either more effort per person, more capital per person, or better productive techniques. We know that price ceilings create shortages, and price floors create unsold surpluses. We know that sustained rapid growth in the stock of money is accompanied by rapid inflation and vice versa. We know that government spending must be financed by some combination of taxation, revenue from sales of product, borrowing from private or foreign sources, issuing high powered money, or depleting stocks of government-held assets. We know that excise taxes, tariffs, and quotas reduce economic welfare in the sense that without them the same total resource pool could produce more satisfaction for some people at no cost to others. We know that permitting individuals to own property and engage in transactions with each other freely will benefit the participants and harm no one, provided that everyone is well informed and acts in his own interest, no one has monopoly power, and there are no external diseconomies such as pollution and no external economies such as increases in the value of my neighbor's real estate if I improve mine. (These are very large provisos, and in some situations they are not even approximately satisfied.) We know that a system of private property and free contract leads to an unequal distribution of income and wealth. We know that social or political attempts to equalize the distribution of income and wealth have perverse incentive effects, so that equalization has a cost in the form of a reduction of total output, and we know a good deal about which kinds of policies are the most perverse in this respect (quotas and price controls) and which are the least perverse (income and inhelitance taxes at moderate rates, and good public education and health programs). We know how to look for long term as well as short term effects, and for indirect and well as direct effects. We also know a great many technical theorems that require mathematics to state clearly and to prove. ...
Perhaps the way to sum up this philosophy in the fewest possible words is to say, "Use your head. And use your heart, too."
Those interested in digging more specifically into the legacy of Christ's work on econometrics might usefully begin with the March-April 1998 issue of the Journal of Econometrics, which is a special issue devoted to "Studies in Econometrics in Honor of Carl F. Christ." The issue isn't freely available online, but for a taste, here are the paper title and authors:
  • "Editor's introduction: studies in econometrics in honor of Carl F. Christ," by Lawrence R Klein
  • "Econometric implications of the government budget constraint," by Christopher A Sims
  • "Impulse response and forecast error variance asymptotics in nonstationary VARs," by Peter C.B Phillips
  • "Business cycle analysis without much theory: A look at structural VARs," by Thomas F Cooley and Mark Dwyer
  • "Lending cycles," by Patrick K Asea and Brock Blomberg
  • "Quasi-rational expectations, an alternative to fully rational expectations: An application to US beef cattle supply," by Marc Nerlove and Ilaria Fornari
  • "Identification and Kullback information in the GLSEM," by Phoebus J Dhrymes
  • "The finite sample properties of simultaneous equations' estimates and estimators Bayesian and non-Bayesian approaches," by Arnold Zellner
  • "Model specification and endogeneity," by Alice Nakamura and Masao Nakamura
  • "Finite sample moments results for the quasi-FIML estimator of the reduced form: The linear case," by Michael D McCarthy
  • "Nonlinear and non-Gaussian state-space modeling with Monte Carlo simulations," by Hisashi Tanizaki and Roberto S Mariano
  • "Heterogeneous information arrival and option pricing," by Patrick K Asea and Mthuli Ncube
  • "The detection and estimation of long memory in stochastic volatility," by F.Jay Breidt, Nuno Crato, and Pedro de Lima
  • "Rational expectations, inflation and the nominal interest rate," by Jean A Crockett