Early in my apprenticeship at Eberstadt, when I was splitting my time between being a political economist and being a trainee investment banker, I discovered computers. That is to say, I discovered why computers are interesting. This came about as an unexpected result of the collapse of the post-Second World War golden age. From the autumn of 1973, under the impact of the oil embargo and energy crisis triggered by the Yom Kippur War, both political and market processes broke down, nowhere more definitively than in the United States. Making sense of the new economic environment in which the financial markets were functioning was as challenging as it was necessary.

By the early 1970s, the macroeconomics of Samuelson’s neoclassical synthesis, universally and misleadingly termed Keynesian, had come to be intimately associated with large-scale econometric models. Otto Eckstein’s DRI (Data Resources, Inc.) Model, based on his research at Harvard, led the field, with competition in the commercial world from Michael Evans’s Chase Econometrics and Lawrence Klein’s Wharton Model. Every major central bank had its own version, as did the Treasury Department. Derived from the work that had won Jan Tinbergen his share of the first Nobel Prize in Economics, these models all deployed a statistical methodology intended to define consistent relationships between variables, using the correlations between time series to establish predictable patterns of systemic behavior.