On the Use of Limit Theorem Arguments in Economic Statistics
- 1 May 1977
- journal article
- research article
- Published by Taylor & Francis in The American Statistician
- Vol. 31 (2) , 85-87
- https://doi.org/10.1080/00031305.1977.10479202
Abstract
The heavy reliance on the normal distribution in the teaching of economic statistics is usually justified by appealing to the central limit theorem. Limit theorem arguments can, however, also lead to the nonnormal stable distributions. It is argued that the conditions required for nonnormal stable limits are, in fact, frequently satisfied by economic variables and in particular by disturbance terms in regressions. Normal theory results should, therefore, be used with caution, and it might be wise for economic statisticians to devote more attention to robust statistical methods.Keywords
This publication has 10 references indexed in Scilit:
- Generating non- normal stable variates using limit theorem propertiesJournal of Statistical Computation and Simulation, 1978
- A Simple Method for Robust RegressionJournal of the American Statistical Association, 1975
- Stable Distributions in Statistical Inference: 1. Symmetric Stable Distributions Compared to other Symmetric Long-Tailed DistributionsJournal of the American Statistical Association, 1973
- The Distribution of Stock ReturnsJournal of the American Statistical Association, 1972
- Convolutions of Distributions Attracted to Stable LawsThe Annals of Mathematical Statistics, 1968
- The Variation of Some Other Speculative PricesThe Journal of Business, 1967
- The Behavior of Stock-Market PricesThe Journal of Business, 1965
- New Methods in Statistical EconomicsJournal of Political Economy, 1963
- The Variation of Certain Speculative PricesThe Journal of Business, 1963
- Application of Least Squares Regression to Relationships Containing Auto-Correlated Error TermsJournal of the American Statistical Association, 1949