1. Berlin, Isaiah; Concepts and Categories; Princeton University Press (1978); p. 80.


  1. The first U.S. stock market regression model is appropriate when credit markets have seized up. The second error correction model of the stock market applies to a normal economy. Since it is highly tolerant of error, it can also identify financially significant historic events. We note in our essay, “The Nature of Financial Equilibrium” that the regression residuals, far from being extraneous annoyances, at times have serial correlation; they trace the arc of events like the OPEC crisis or the popularization of the Internet.


Isaiah Berlin is one of our favorite writers because he had an unmatched ability to explain complexity clearly. In his essay, “The Concept of Scientific History (1960),” he contrasts the opposed but complementary human demands, “one for unity and homogeneity, the other for diversity and heterogeneity.” The two regressions abstract historical events in the form of mathematical formulae – the first abstracts to a lesser degree than the second because it is conditional – but both are still abstract and restricted to a general (but useful) way of viewing things that is characteristic of some of the natural sciences, economics, and econometrics. Perhaps Berlin had studied econometrics. He writes, “At times some among these generalisations can be clearly stated, and combined into models; where this occurs, natural sciences arise. But the descriptive and explanatory language of historians...cannot…be reduced without residue (our emphasis) to…general formulae, still less to models and their applications.”


The second way of viewing known facts is historical. A historian’s view requires not abstraction, but verstehen, the ability to understand the lives that people who lived at a unique time and place and to place those lives in context of their societies. This is real; the above is abstract; and we think both have their places in the social sciences, noting each word. The real conditional sets the stage for possibly useful and definitely interesting abstraction.


Just don’t use one formula; the Gaussian copula blew up the entire financial system.


  1. Coming out of the trough between 1932 and 1936, a like calculated P/E increased to 18.6 . The multiplier of earnings per share rose to 12.54, but the ratio of the (rational) constant to the total remained about the same as before. Keynes (1953) emphasized short-term market behavior as opposed to long-term investor expectation.  



             The regression equation during that market phase was S&P = 6.06 + 12.54 (Earnings Per share).

                                                                                                       ρ2 = .98 out of 1