Template-Type: ReDIF-Article 1.0
Author-Name: Oseledets V.
Author-Email: oseled@gmail.com
Author-Workplace-Name: Budget and Treasury Academy of the Ministry of Finance of the Russian Federation
Author-Name: Oseledets I.
Author-Email: ivan.oseledets@gmail.com
Author-Workplace-Name: Budget and Treasury Academy of the Ministry of Finance of the Russian Federation
Author-Name: Pospelov D.
Author-Email: d.a.pospelov@gmail.com
Author-Workplace-Name: Budget and Treasury Academy of the Ministry of Finance of the Russian Federation
Title: Predictions Made for Two or More Financial Series via the Diagonal Correlations Made to Their Common Past History
Abstract: The article authors offer an approach to be used when predictions are made of the future stock price provided that there is some data on the stock market price history. Estimates are made of the probability figures in the forecasts made of the dynamics of the stock price formation, and the criterion of quality of the method is the extent to which the estimates are correct. The authors show that if the diagonal (or cross)-correlation-based method is used in order to make the predictions for the two shares’ price formation dynamics in the period of high volatility (that is, in the period of a crisis), the direction in which the change curve goes are correct in the 75% of the total number of cases. However, the limitations of the method are also identified. The forecasts are correct in not more than 50% of all cases in the calm market periods. The method can be easily adjusted to be used to make forecasts for a larger number of shares.
Classification-JEL:
Keywords: financial series, stock price forecast, diagonal correlation
Journal: Finansovyj žhurnal — Financial Journal
Pages: 105-112
Issue: 2
Year: 2010
Month: April
File-URL: http://www.finjournal-nifi.ru/images/FILES/Journal/Archive/2010/2/statii/100210.pdf
File-Format: Application/pdf 
Handle: RePEc:fru:finjrn:100210:p:105-112