Cointegration and Error Correaction Modeling for BSE and NSE Stock Prices Time Series Data
Econometrician always have been observed that most of the economics time series are non-stationary. The ordinary least square technique could be applied to estimate the model parameters only if the variables have been found to be stationary, i.e., they do not have unit roots. Otherwise, an alternative approach has to be followed, which is ‘co-integration’. Co-integration is a method of finding out the long-term relationship between economic variables under consideration. In the present study, it is necessary to know whether National Stock Exchange (NSE) and Bombay Stock Exchange (BSE) daily closing stock prices time séries data will have a long-term movements or relationship between them for which the co-integration analysis is being proposed. The error correction model (ECM) captures the impact of long-run equilibrium on the short run dynamics, provided that the variables are co-integrated. The ECM developed by Engle and Granger reconciles the short run behavior of an economic variable with its log run behavior. The ECM results showed that the coefficient of the error correction term ONELAGU, of about -0.020 suggests that only about 2% of the discrepancy between long-term and short-term LOGNSE is corrected, suggesting a slow rate of adjustment to equilibrium. Further, the validity of ECM showed that all three tests conducted through spurious of the model and serial correlation of residual and normality of residual are in favour of the model. Since, all tests in favour of ECM, the results of the model can be used for policy decisions.
Keywords:Stationarity, co-integration, error correction model, breusch-godfrey test, granger representation theorem
Cite this Article
Rajarathinam A, Balamurugan D. Cointegration and Error Correaction Modeling for BSE and NSE Stock Prices Time Series Data. Research & Reviews: Discrete Mathematical Structures. 2017; 4(3): 30–41p.
- There are currently no refbacks.