Spurious Regressions and Near-Multicollinearity, with an Application to Aid, Policies and Growth

Chatelain, Jean-Bernard and Ralf, Kirsten (2012): Spurious Regressions and Near-Multicollinearity, with an Application to Aid, Policies and Growth.

Abstract

In multiple regressions, explanatory variables with simple correlation coefficients with the dependent variable below 0.1 in absolute value (such as aid with economic growth) may have very large and statistically significant estimated parameters which are unfortunately �"outliers driven" and spurious. This is obtained by including another regressor which is highly correlated with the initial regressor, such as a lag, a square or interaction terms of this regressor. The analysis is applied on the "�Botswana outliers driven" Burnside and Dollar [2000] article which found that aid had an effect on growth only for countries achieving �good macroeconomic policies.

• Item Type: MPRA Paper
• Language: English
• Keywords: Near-Multicollinearity, Student t-Statistic, Spurious regressions, Ceteris paribus, Parameter In�flation Factor, Growth, Foreign Aid
• Subjects: F - International Economics > F3 - International Finance > F35 - Foreign Aid; C - Mathematical and Quantitative Methods > C5 - Econometric Modeling > C52 - Model Evaluation, Validation, and Selection; C - Mathematical and Quantitative Methods > C1 - Econometric and Statistical Methods and Methodology: General > C12 - Hypothesis Testing: General; P - Economic Systems > P4 - Other Economic Systems > P45 - International Trade, Finance, Investment, and Aid

References

Belsley [1991]. Conditioning Diagnostics: Collinearity and Weak Data in Regression. Wiley, New York.

Burnside C. and Dollar D. [2000]. Aid, policies and growth. American Economic Review, 90, 847-68.

Chatelain J.B. and Ralf K. [2010]. �Inference on Time-Invariant Variables using Panel Data: A Pre-Test Estimator with an Application to the Returns to Schooling," PSE Working Papers hal-00492039, HAL.

Cohen J. [1988]. Statistical Power Analysis for the Behavioral Sciences. 2nd edition. Lawrence Erlbaum Associates, Hillsdale.

Cohen J., Cohen, P., West S.G., Aiken L.S. [2003]. Applied multiple regression/correlation for the behavioral sciences. Wiley, New York.

Christensen, R. [2006]. Comment on Friedman and Wall (2005). Letter to the Editor. The American Statistician, 60, 101-102.

Doucouliagos H. and Paldam M. [2008]. Aid effectiveness on growth: A meta study. European Journal of Political Economy, 24, 1-24.

Doucouliagos H. and PaldamM. [2009]. Conditional aid effectiveness: A meta study. Journal of International Development, 22, 391-410.

Easterly, W., R. Levine, and D. Roodman. [2004]. New Data, New Doubts: A Comment on Burnside and Dollars Aid, Policies, and Growth (2000). American Economic Review, 94, 774-780.

Freedman D. (1997). From Association to Causation via Regression. in V.R. McKim and S.P. Turner (eds), Causality in Crisis?, University of Notre Dame Press, Notre Dame, 113-61.

Friedman, L. and Wall, M. [2005]. Graphical views of suppression and multicollinearity in multiple linear regression. The American Statistician, 59, 127-136.

Fisher R. [1921]. On the �probable error �of a coefficient of correlation deduced from a small sample. Metron, 1, 3-32.

Fisher R. [1924]. The Distribution of the Partial Correlation Coefficient. Metron, 3, 329�332.

Fisher R. [1925]. Applications of �Student�s� distribution. Metron, 5, 90-104.

Hill and Adkins [2001]. Collinearity. In: Baltagi B.H. (Ed.). A Companion to Theoretical Econometrics, Blackwell, Oxford (Chapter 12).

Hoover K.D. [2001]. Causality in Macroeconomics. Cambrige University Press, Cambridge.

Horst P. [1941]. The role of prediction variables which are independent of the criterion. in The prediction of personal adjustment. Social Sciences Research. Council, New York, pp. 431-436.

Ioannidis J.P.A. [2005]. Contradicted and initially stronger effects in highly cited clinical research. Journal of the American Medical Association. 294, 218-228.

Ioannidis J.P.A., Trikalinos T.A. [2006]. Early extreme contradictory estimates may appear in published research: The Proteus phenomenon in molecular genetics research and randomized trials. Journal of Clinical Epidemiology, 58, 543-549.

Ioannidis J.P.A. [2008].Why Most Discovered True Associations Are In�flated. Epidemiology, 19, 640-648.

Mason C.H. and Perreault W.D. [1991]. Collinearity, Power and Interpretation of Multiple Regression Analysis, Journal of Marketing Research, 28, 268-80.

Pearl J. [2009]. Causality: Models, Reasoning and Inference (2nd edition). Cambrige University Press, Cambridge.

Roodman D. [2008]. Through the Looking Glass, and What OLS Found There: On Growth, Foreign Aid, and Reverse Causality. Center for Global Development Working Paper Number 137.

Spanos A. and McGuirk A. [2002]. The Problem of high correlation among regressors revisited: erratic vs systematic volatility. Journal of Econometrics, 108, 365-393.

Spirtes P., Glymour C.N. and Scheines R. [2000]. Causation, Prediction, and Search (2nd edition). Cambrige University Press, Cambridge.

Silvey [1969]. Multicollinearity and Imprecise Estimation. Journal of the Royal Statistical Society B, 31, 539-552.

Stanley, T.D., [2005]. Beyond publication bias. Journal of Economic Surveys. 19, 309-45.

Yule G.U. [1897]. On the theory of correlation. Journal of the Royal Statistical Society, 60, 812-854. 21