The broader area of econometrics revolve around regression and regression residuals, for long term investment portfolios, the strategy to mitigate risks inherently is a critical component. One such aspect is to understand whether the issues, like Normality of residuals can provide some empirical justification to the ever-challenging area of long term portfolio risk reduction.
The present paper is therefore touching on this very aspect; it is to refer to portfolio risk minimization with long term perspective, with fundamental financial statement information. This will create a two-fold study, first to gather relevant financial ratios for regression purposes (mainly in terms of multiple regression), and then utilizing the residuals, creating relevant portfolio combination and simulating thereby using portfolio attributes (weights) for risk reduction. This will be precisely covering only the cement sector, and this sector-specific selection implies testing the empirical soundness of the topic concerned. The result strongly differentiates the performance of “non-normal & normal” residual portfolio risks values (as referred several times in paper as impurecombination) with that of purely normal residual portfolio risk values(as referred several times as pure residual combinations). And, the author in the last stress on using Normal residuals as a better tool for portfolio risk minimization.
OLS, Normality, Heteroskedasiticity, Autocorrelation
