ABSTRACT
The roles of probability in economics are most pronounced in econometrics, where estimation, stability, bounds for uncertainty, statistical inference and the validation of economic theories with empirical evidence are of primary interest. Economists quickly assimilated the basic notions of correlation and regression at the beginning of this century, and they actively sought cyclical and other patterns in economic time series with the periodogram and its successors after Fourier analysis was extended to stochastically disturbed data. This assimilation was directly recognized when the Econometric Society was established by prominent economists at the end of 1930, with its influential journal Econometrica appearing a few years later. By the end of the 1940s, economists were ready to analyse new collections of simultaneous equations and dynamic time series models, absorb the Neyman—Pearson framework for hypothesis testing, rely on linear stochastic processes and a few asymptotic properties of conventional estimators (including bias, consistency, efficiency and distributional convergence) within stationary and ergodic contexts with little concern for historical conditioning or special circumstances, and accept the informational requirements of maximum likelihood approaches.
