Effect of outliers on forecasting temporally aggregated flow variables

Abstract

Economic time series are of two types: stock and flows, and may be available at different levels of aggregation (for instance, monthly or quarterly). The economist, in many situations, is interested in forecasting the aggregated observations. The forecast function, in this case, can be based either on the disaggregated series or the aggregated series. The forecasts based on the disaggregated data are at least as efficient, in terms of mean squared forecast errors, as the forecasts based on temporally aggregated observations when the data generating process (DGP) is a known ARIMA process. However, the effect of outliers on both forecast functions is not known. In this paper, we consider the effect of additive and innovation outliers on forecasting aggregated values based on aggregated and disaggregated models when the DGP is a known ARIMA process and the presence of the outliers is ignored. Results when the model is not known and tests applied for the detection of outliers are derived through simulation.

This is a preview of subscription content, access via your institution.

References

  1. Abraham, B., andBox, G. (1979). Bayesian analysis of some outlier problems in time series analysis.Biometrika, 66:229–236.

    MATH  Article  MathSciNet  Google Scholar 

  2. Ahsanullah, M., andWei, W. (1984). Effects of temporal aggregation on forecasting in an ARMA(1,1)process. InProceedings of the Business and Economic Statistics Section, pp. 297–302. American Statistical Association, Alexandria, VA.

    Google Scholar 

  3. Box, G., andJenkins, G. (1976).Time Series Analysis, Forecasting and Control. Holden Day, San Francisco, 2nd ed.

    MATH  Google Scholar 

  4. Chang, I., Tiao, G., andChen, C. (1988). Estimation of time series parameters in the presence of outliers.Technometrics, 30:193–204.

    Article  MathSciNet  Google Scholar 

  5. Chen, C., andLiu, L.-M. (1993a). Forecasting time series with outliers.Journal of Forecasting, 12:13–35.

    Google Scholar 

  6. Chen, C., andLiu, L.-M. (1993b). Joint estimation of model parameters and outlier effects in time series.Journal of the American Statistical Association, 88:284–297.

    MATH  Article  Google Scholar 

  7. De Jong, P., andPenzer, J. (1998). Diagnosing shocks in time series.Journal of the American Statistical Association, 93:796–806.

    MATH  Article  MathSciNet  Google Scholar 

  8. Fox, A. (1972). Outliers in time series.Journal of the Royal Statistical Society Ser. B, 34:350–363.

    MATH  Google Scholar 

  9. González, P. (1992). Temporal aggregation and systematic sampling in structural time series model.Journal of Forecasting, 11:271–281.

    Google Scholar 

  10. Harvey, A., andKoopman, S. (1992). Diagnostic checking of unobserved component time series models.Journal of Business and Economic Statitics, 10:377–389.

    Article  Google Scholar 

  11. Hillmer, S. (1984). Monitoring and adjusting forecasts in the presence of additive outliers.Journal of Forecasting, 3:205–215.

    Google Scholar 

  12. Hotta, L. (1993). The effect of additive outlier on the estimates from the aggregated and disaggregated ARIMA models.International Journal of Forecasting, 9:85–93.

    Article  Google Scholar 

  13. Hotta, L., andCardoso Neto, J. (1993). The effect of aggregation on prediction in autoregressive integrated moving-average models.Journal of Time Series Analysis, 14:261–269.

    MATH  Google Scholar 

  14. Kleiner, B., Martin, R., andR. J. T. (1979). Robust estimation of power spectra.Journal of the Royal Statistical Society, Ser. B, 41:313–353.

    MATH  Google Scholar 

  15. Ledolter, J. (1987). The effect of outliers on the estimates in the forecasts from ARIMA time series models. InProceedings of the Business and Economic Statistics Section, pp. 435–458. American Statistical Association, Alexandria, VA.

    Google Scholar 

  16. Ledolter, J. (1989). The effect of the additive outliers on the forecasts from ARIMA models.International Journal of Forecasting, 5:231–240.

    Article  Google Scholar 

  17. Luceño, A. (1998). Detecting possibly non-consecutive outliers in industrial time series.Journal of the Royal Statistical Society, Ser. B, 60:295–310.

    MATH  Article  Google Scholar 

  18. Lutkepöhl, H. (1986). Forecasting temporally aggregated vector ARMA processes,Journal of Forecasting, 5:85–95.

    Google Scholar 

  19. Lutkepöhl, H. (1987).Forecasting Aggregated Vector ARMA Processes. Lecture Notes in Economics and Mathematical System. Springer Verlag, New York.

    Google Scholar 

  20. Pino, F., Morettin, P., andMentz, R. (1987). Modelling and forecasting linear combinations of time series.International Statistical Review, 55:295–314.

    MATH  MathSciNet  Article  Google Scholar 

  21. Schmid, W. (1996). An outlier test for time series based on a two sided predictor.Journal of Time Series Analysis, 17:497–510.

    MATH  MathSciNet  Google Scholar 

  22. Schwarz, G. (1978). Estimating the dimension of a model.Annals of Statistics, 6:461–464.

    MATH  MathSciNet  Google Scholar 

  23. Stram, D., andWei, W. (1986). Temporal aggregation in the ARIMA process,Journal of Time Series Analysis, 7:279–292.

    MATH  MathSciNet  Google Scholar 

  24. Wu, L. andHosking, J. (1993). Reallocation outliers in time series.Applied Statistics, 42:301–313.

    MATH  Article  MathSciNet  Google Scholar 

Download references

Author information

Affiliations

Authors

Corresponding author

Correspondence to Luiz K. Hotta.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Hotta, L.K., Pereira, P.L.V. & Ota, R. Effect of outliers on forecasting temporally aggregated flow variables. Test 13, 371–402 (2004). https://doi.org/10.1007/BF02595778

Download citation

Key Words

  • Additive outliers
  • innovation outliers
  • forecasting
  • temporal aggregation

AMS subject classification

  • 62M10
  • 62M20