Harold Hotelling

Harold Hotelling (1895–1973) was an American mathematical statistician and economist whose work established the foundations of multivariate statistical analysis. His 1933 paper on "Analysis of a Complex of Statistical Variables into Principal Components" gave the modern formulation of principal component analysis as an eigenvalue problem, transforming Pearson's geometric intuition into rigorous linear algebra. He also invented canonical correlation analysis, a method for finding relationships between two sets of variables that remains central to multivariate statistics, and developed the T-squared statistic for testing hypotheses about multivariate means.

Hotelling's influence extends beyond statistics into economics. His 1929 paper on the economics of exhaustible resources — the "Hotelling rule" — established that the optimal price of a non-renewable resource should rise at the rate of interest. This result, derived from a simple arbitrage argument, became the cornerstone of resource economics and environmental policy. The same mind that formalized dimensionality reduction in data also formalized intertemporal resource allocation, suggesting that the mathematical structures of statistical inference and economic optimization are more closely related than the disciplinary boundaries of the twentieth century admitted.

Hotelling is a reminder that the most important intellectual contributions often cross boundaries that institutions later harden. His principal components and his resource rule were published within four years of each other, by the same person, working at the same university. The separation of statistics and economics into distinct departments was not a natural fact; it was an administrative decision that obscured the unity of mathematical reasoning about constrained optimization under uncertainty.