Quantile regression and independent variables
Posted: Thu Aug 08, 2013 4:11 am
Hello
I need to estimate volatility response (dependent v.) to several independent variables in a certain point. (a time series analysis). I have 10 quintiles (0.1 to 1.0). To my understanding, eviews 7 estimates, say q=0.4 of the dependent variable, using whole time series of the independent variable (in my case stock returns) to estimate the volatility response. However, I would need to estimate the effect of 10 different return categories (by size, from small to large) to volatility. So, I'm interested of return size effect to volatility. I.e. at 0.4q the fourth smallest return category effect on volatility and so on. Eviews would need to estimate the response variable in regarding the return size in 10 separate instances. Is there a function to do it, or do I need to manually separate the returns into categories and just use countless numbers of OLS regressions?
Thanks.
I need to estimate volatility response (dependent v.) to several independent variables in a certain point. (a time series analysis). I have 10 quintiles (0.1 to 1.0). To my understanding, eviews 7 estimates, say q=0.4 of the dependent variable, using whole time series of the independent variable (in my case stock returns) to estimate the volatility response. However, I would need to estimate the effect of 10 different return categories (by size, from small to large) to volatility. So, I'm interested of return size effect to volatility. I.e. at 0.4q the fourth smallest return category effect on volatility and so on. Eviews would need to estimate the response variable in regarding the return size in 10 separate instances. Is there a function to do it, or do I need to manually separate the returns into categories and just use countless numbers of OLS regressions?
Thanks.