Saturday, March 17, 2012

An Irish Economic Statistician

Today, of all days, it seems appropriate to mention an outstanding Irish statistician - one who made seminal contributions to "economic statistics", as well as to mathematical statistics. I'm referring to Robert ("Roy") Charles Geary.

First, let me quote, without apology, from Wikipedia:

"Dr. Robert (Roy) Charles Geary (April 11, 1896 – February 8, 1983) was an Irish statistician and founder of both the Central Statistical Office and the Economic and Social Research Institute. He held degrees from University College Dublin and the Sorbonne. He lectured in mathematics at University College Southampton (1922–23) and in applied economics at Cambridge University (1946–47). He was a statistician in the Department of Industry and Commerce between 1923 and 1957. He was the founding director of the Central Statistics Office (Ireland) (in 1949). He was head of the National Accounts Branch of the United Nations in New York from 1957 to 1960. He was the founding director of the Economic and Social Research Institute (in 1960) where he stayed till his retirement in 1966. He was an honorary fellow of the American Statistical Association and the Institute of Mathematical Statistics. In 1981, he won the Boyle Medal. Roy Geary is known for Geary's C, the Geary–Khamis dollar, the Stone–Geary utility function, and Geary's Theorem, which has that if the sample mean is distributed independently of the sample variance, then the population is distributed normally."

It's worth noting that Geary's time in the Department of  Applied Economics at Cambridge was spent under the direction of Sir Richard Stone, about whom I've posted previously. Maximization of the Stone-Geary (direct) utility function leads to the set of demand equations called the "Linear Expenditure System". This was the "work-horse" of empirical demand analysis up to the introduction of the "Almost Ideal Demand System" by Deaton and Muellbauer (1980).

Also in the department at Cambridge in that period were Donald Cochrane and Guy Orcutt  - no explanation needed! What is less generally known, though, is that Don Cochrane also worked on the early development of national accounts, also with the U.N. in New York, before returning to Australia and becoming the foundation Dean of Economics and Politics at Monash University.

Geary's Theorem is a result that econometrician's use, perhaps without even being aware of it, on a daily basis. Although it was already known that the sample mean and variance are statistically independent if the population is Normal, Geary (1936) proved the converse result, thus establishing Normality as a necessary and sufficient condition for this independence.

Fittingly, we use this result whenever we conduct a t-test.

So, if you're enjoying a pint of Guinness in honour of St. Patrick today, you might spare a few moments to raise a toast not only to their previous brewer,"Student", but also to another of our own -Roy Geary.


Deaton, A. and J. Muellbauer (1980). An almost ideal demand system. American Economic Review, 70, 312-326.

Geary, R. C. (1936). The distribution of the Student's ratio for the non-normal samples. Supplement to the Journal of the Royal Statistical Society, 3, 178-184.

@copy; 2012, David E. Giles


  1. There is now a institute named in his honour at UCD

    1. Kevin - I wasn't aware of that. That's excellent!