Computational Statistics

, Volume 23, Issue 2, pp 317–335 | Cite as

Automatic selection of indicators in a fully saturated regression

  • Carlos Santos
  • David F. Hendry
  • Soren Johansen
Original Paper


We consider selecting a regression model, using a variant of the general-to-specific algorithm in PcGets, when there are more variables than observations. We look at the special case where the variables are single impulse dummies, one defined for each observation. We show that this setting is unproblematic if tackled appropriately, and obtain the asymptotic distribution of the mean and variance in a location-scale model, under the null that no impulses matter. Monte Carlo simulations confirm the null distributions and suggest extensions to highly non-normal cases.


Indicators Regression saturation Subset selection Model selection 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. Campos J, Hendry DF, Krolzig H-M (2003) Consistent model selection by an automatic gets approach. Oxf Bull Econ Stat 65:803–819CrossRefGoogle Scholar
  2. Campos J, Ericsson NR, Hendry DF (2005) Editor’s introduction. In: Campos J, Ericsson NR, Hendry DF (eds) Readings on general-to-specific modelling. Edward Elgar, Cheltenham (forthcoming)Google Scholar
  3. Doornik JA (2001) OX, an object oriented matrix programming language, 4th edn. Timberlake Consultants Press, LondonGoogle Scholar
  4. Foster DP, Stine RA (2004) Honest confidence intervals for the error variance in stepwise regression, Mimeo. Statistics Department, Wharton School, University of PennsylvaniaGoogle Scholar
  5. Granger CWJ, Hendry DF (2003) A dialogue concerning a new instrument for econometric modelling. Unpublished Paper, Department of Economics, University of OxfordGoogle Scholar
  6. Hendry DF (1995) Dynamic econometrics. Oxford University Press, OxfordzbMATHGoogle Scholar
  7. Hendry DF, Krolzig H-M (2001) Automatic econometric model selection using PcGets. Timberlake Consultants Press, LondonGoogle Scholar
  8. Hendry DF, Krolzig H-M (2003) New developments in automatic general-to-specific modelling. In: Stigum BP (ed) Econometrics and the philosophy of economics. Princeton University Press, Princeton, pp 379–419Google Scholar
  9. Hendry DF, Krolzig H-M (2004) Model selection with more variables than observations. Unpublished Paper, Department of Economics, University of OxfordGoogle Scholar
  10. Hendry DF, Krolzig H-M (2005) The properties of automatic GETS modelling. Econ J 115:C32–C61CrossRefGoogle Scholar
  11. Hendry DF, Santos C (2005) Regression models with data-based indicator variables. Oxf Bull Econ Stat 67:571–595CrossRefGoogle Scholar
  12. Hendry DF, Santos C (2006) Automatic tests for super exogeneity. Unpublished Paper, Department of Economics, University of OxfordGoogle Scholar
  13. Hoover KD, Perez SJ (1999) Data mining reconsidered: encompassing and the general-to-specific approach to specification search. Econometrics J 2:167–191CrossRefGoogle Scholar
  14. Hoover SJ, Perez SJ (2004) Truth and robustness in cross-country growth regressions. Oxf Bull Econ Stat 66(5):765–798CrossRefGoogle Scholar
  15. Krolzig H-M, Hendry DF (2001) Computer automation of general-to-specific model selection procedures. J Econ Dyn Control 25:831–866zbMATHCrossRefGoogle Scholar
  16. Salkever DS (1976) The use of dummy variables to compute predictions, prediction errors, and confidence intervals. J Econometrics 4:393–397zbMATHCrossRefGoogle Scholar
  17. White H (1984) Asymptotic theory for econometricians. Academic, LondonGoogle Scholar

Copyright information

© Springer-Verlag 2007

Authors and Affiliations

  • Carlos Santos
    • 1
  • David F. Hendry
    • 2
  • Soren Johansen
    • 3
  1. 1.Department of Economics and ManagementPortuguese Catholic UniversityPortoPortugal
  2. 2.Department of EconomicsUniversity of OxfordOxfordUK
  3. 3.Department of StatisticsUniversity of CopenhagenCopenhagenDenmark

Personalised recommendations