The conceptual development of population and variation as foundations of econometric analysis

Klein, Judy Lee (1986) The conceptual development of population and variation as foundations of econometric analysis. Doctoral thesis, City of London Polytechnic.

[img]
Preview
Text
371420.pdf - Published Version

Download (8MB) | Preview

Abstract / Description

Economics is a time-bound science. The analytical tools of statistical description and inference, however, were first developed for static comparisons of differences rather than formulation of processes of change. This thesis offers an historical perspective on the dichotomy of logical variation and temporal variation. I examine the interaction of statistical technique with the needs and concepts generated in the study of political arithmetic, observational errors. social physics, natural selection and economic motion.

Through these interactions the concept of statistical population changed. There was a shift in emphasis from the assumption of equivalence of constituents and from the mean as a manifestation of truth and divine order to the assumption of deviation and the mean as a typical value in motion. In Darwin's theory of natural selection, differences within a population were the source of evolutionary variation of a species. The quantitative techniques of correlation and regreSSion were developed to test theories of evolution and inheritance.

The problems of reconciling logical variation and temporal variation were most prominent in the application of correlation and regression to economic time series data. Differencing observations and calculations of deviations from moving averages were suggested as solutions. The most significant steps were taken in the the formulation of stochastic processes and in the development of errors-in-equations models. With the latter. the statistical properties of residuals rather than of series of observations became important. In building on some of these historical examples I suggest that acknowledgement of complementary statistical populations may enable us to further reconcile logical and temporal variations.

Item Type: Thesis (Doctoral)
Additional Information: uk.bl.ethos.371420
Uncontrolled Keywords: economics; estimation theory; econometric analysis; stochastic processes
Subjects: 300 Social sciences > 330 Economics
Department: Library Services and Special Collections
Depositing User: Mary Burslem
Date Deposited: 18 May 2022 13:40
Last Modified: 18 May 2022 13:40
URI: http://repository.londonmet.ac.uk/id/eprint/7658

Downloads

Downloads per month over past year



Downloads each year

Actions (login required)

View Item View Item