On the Accuracy and Efficiency of GMM Estimators: A Monte Carlo Study
GMM estimators are now widely used in econometric and financial analysis. Their asymptotic properties are well known, but we have little knowledge of their small sample properties or their rate of convergence to their limiting distribution. This paper reports small sample Monte Carlo evidence which helps discriminate between the many GMM estimators proposed in the literature. We add a new GMM estimator which delivers better finite sample properties. We also test whether biases in the parameter estimates are either significant or significantly different between estimators. We conclude that they are, with both relative and absolute biases depending on sample size, fitting criterion, non-normality of disturbances, and parameter size.
KeywordsMaximum Likelihood Estimator Beta Distribution Significant Bias Small Sample Property Small Sample Bias
Unable to display preview. Download preview PDF.
- Duffie, D. and Singleton K.J. (1989) Simulated Moments Estimation of Markov Models of Asset Prices, Stanford University Discussion Paper, Stanford, CA.Google Scholar
- Hughes Hallett, A.J. (1992) Stabilising earnings in a volatile market, paper presented in the Royal Economics Society Conference, London (April).Google Scholar
- Kendall, M.G. and Stewart, A. (1973) The Advanced Theory of Statistics, Vol. 2, Third Edition, Griffen & Co., London.Google Scholar
- Mood, A. F. Graybill and D. Boes (1974) Introduction to the Theory ofStatistics, McGraw-Hill, New York.Google Scholar
- Smith, G. and Spencer M. (1991) Estimation and testing in models of exchange rate target zones and process switching, in P. Krugman and M. Miller (eds), Exchange rate targets and currency bands, Cambridge University Press, Cambridge and New York.Google Scholar
- Tauchen, G. (1986) Statistical Properties of Generalised Method of Moments Estimators of Structural Parameters Obtained from Financial Market Data, Journal of Business and Economic Statistics, 4, pp.397–425.Google Scholar