DOI QR코드

DOI QR Code

A Wilcoxon signed-rank test for random walk hypothesis based on slopes

기울기를 이용한 랜덤워크 윌콕슨 부호순위검정

  • Received : 2014.10.20
  • Accepted : 2014.11.14
  • Published : 2014.11.30

Abstract

Random walk is used for describing random phenomenon in various areas but tests for random walk developed so far are known to suffer from size distortion and low power. Kim et al. (2014) proposed a sign test for unit root (${\rho}=1$) hypothesis based on slopes. This article proposes a Wilcoxon signed rank test based on slopes for unit root hypothesis, and compares it with the augmented Dickey-Fuller test and the sign test by a simulation study. Our results confirm that the nonparametric tests are better than ADF test for small samples like n = 30. The results also show that the sign test is better than the Wilcoxon signed rank test and that for 0 < ${\rho}$ < 1 (-1 < ${\rho}$ < 0), the nonparametric tests suffer from power loss (improvement) as normal error changes to double exponential error.

랜덤워크는 다양한 분야에서 랜덤현상을 기술하는데 이용되고 있으나, 현재까지 개발된 랜덤워크 검정법에는 유의수준 왜곡과 낮은 검정력 등의 문제가 있는 것으로 알려져 있다. 이러한 문제점들을 개선하기 위해 Kim 등 (2014)은 부호검정에 기초한 랜덤워크 검정 (${\rho}=1$)방법을 제안하였다. 본 논문에서는 보다 개선된 랜덤워크 검정법을 제안하고자 부호검정보다 검정력이 우수한 것으로 알려진 윌콕슨 부호순위검정을 이용한 랜덤워크 검정법을 제안하고, 모의실험을 통해 부호검정, 윌콕슨 부호순위검정, 확장 Dickey-Fuller 검정의 성능을 비교하였다. 모의실험 결과 소표본에서 비모수 검정기법들이 ADF 검정보다 우월하다는 사실을 재확인하였다. 새롭게 밝혀진 사실은 부호검정이 윌콕슨 부호순위검정에 비해 높은 검정력을 가지며, 또한 비모수 검정기법들은 ${\rho}$가 양의 부호를 가지는 경우 (0 < ${\rho}$ < 1) 정규분포보다 이중지수분포에서 낮은 검정력을 가지게 되나, ${\rho}$가 음의 부호(-1 < ${\rho}$ < 0)를 갖는 경우에는 정규분포보다 이중지수분포에서 높은 검정력을 보인다는 사실이다.

Keywords

References

  1. Dickey, D. A. and Fuller, W. A. (1979). Distribution of the estimators for autoregressive time series with a unit root. Journal of the American Statistical Association, 74, 427-431.
  2. Hamilton, J. D. (1994). Time series analysis, Vol. 2, Princeton University Press, Princeton.
  3. Kwiatkowski, D., Phillips, P. C., Schmidt, P. and Shin, Y. (1992). Testing the null hypothesis of stationarity against the alternative of a unit root: How sure are we that economic time series have a unit root? Journal of Econometrics, 54, 159-178. https://doi.org/10.1016/0304-4076(92)90104-Y
  4. Kim, T. Y., Park, C., Kim, S. G., Kim, C. J., Kim, H., Yu, J. H., Jang, K. M. and Jang, Y. S. (2014). A sign test for random walk hypothesis based on slopes. Journal of the Korean Data & Information Science Society, 25, 385-292. https://doi.org/10.7465/jkdi.2014.25.2.385
  5. Maddala, G. S., and Kim, I-M. (1998). Unit roots, cointegration and structural change, Oxford University Press, Oxford.
  6. Malkiel, B. G. (1973). A random walk down Wall Street, 6th Ed., W. W. Norton & Company, Inc., New York.
  7. Pearson, K. (1905). The problem of the random walk. Nature, 72, 294.
  8. Phillips, P. C. and Perron, P. (1988). Testing for a unit root in time series regression. Biometrika, 75, 335-346. https://doi.org/10.1093/biomet/75.2.335
  9. Said, S. E. and Dickey, D. A. (1984). Testing for unit roots in autoregressive-moving average models of unknown order. Biometrika, 71, 599-607. https://doi.org/10.1093/biomet/71.3.599
  10. Schwert, G. W. (1987). Effects of model specification on tests for unit roots in macroeconomic data. Journal of Monetary Economics, 20, 73-103. https://doi.org/10.1016/0304-3932(87)90059-6
  11. Wilcoxon, F. (1945). Individual comparisons by ranking methods. Biometrics Bulletin, 80-83.
  12. Wilcoxon, F., Katti, S. K. and Wilcox, R. A. (1970). Critical values and probability levels for the Wilcoxon rank sum test and the Wilcoxon signed rank test. In Selected Tables in Mathematical Statistics, Vol. 1, edited by HL Harter and DB Owen, Markham, Chicago, 171-259.

Cited by

  1. A study on the slope sign test for explosive autoregressive models vol.26, pp.4, 2015, https://doi.org/10.7465/jkdi.2015.26.4.791