Mandelbrot presents the cotton study as evidence for a non-Gaussian stable scaling model, and also recounts later evidence both for and against strict scaling that the initial model could not reconcile.

groundsUntitled source

Power laws, the spread of rich and poor, and the mathematics of the exceptional, stable probability distributions. One is a general way of looking at the world, the next a practical example of it in economics, and the last a set of mathematical considerations no one viewed as useful. How do they come together in the cotton mystery? The IBM Computing Center chewed through the thousands of cotton prices as I had asked. It quickly confirmed Houthaker's view. The price changes from one day to the next, one week or month or year to the next, did not behave as the Baccalier model assumed. The variance misbehaved. Each time I added an extra price change to the dataset, my estimate of the variance changed. It never settled down to one simple number, say, 1% volatility. Instead, it roamed erratically from about 0.4% to 3%, nearly a tenfold difference. That was surprising, considering that the quality of the data itself could not be challenged.

groundsUntitled source

Markowitz's portfolio theory, sharp asset pricing, and the Buckelier market model were spreading. And the next year, Black and Scholes published their influential options pricing formulae. Modern finance was the official religion. My hypothesis contradicted it, and I was about as welcome in the established Church of Economics as a heretical Aryan at the Council of Nicene. In 1972, a University of Chicago grad student, Robert R. Officer, crystallized many of the qualms of the economics establishment. His Ph.D. thesis found evidence both for and against strict scaling in the same set of data. Other seemingly contradictory reports appeared, But the critics could not simply explain away the evidence supporting it, and I preferred just to forget about them. I found many of the criticisms were based on statistical tests inappropriate to the data, often a problem in statistics. But at the time, no one could devise a theory to reconcile all the conflicting data.