Is it correct to say that to use a linear regression approach (10:47) to identify over/under-valued stocks, one is making an assumption that the market is on average more efficient than not?
@@elontusk9123 I think this is because the method that he was using is a statistical one in the regression example, so to quantify the risk he used the SD as the measure. Of course, there is different layers of risk but seeing as though he's using the relative valuation as a screening process to then perform IV, he probably thinks that it is good enough, and the other layers of risk will be accounted for in the IV.
Sir, how can you compare standard deviation as measure of risk for different companies whose size is different? Aren't we supposed to use Coefficient of Variation (SD/Mean) instead of comparing Standard Deviations directly?
In the regression approach you are calculating the predicted pbv ratio in comparison to the company's own financial variables like roe , beta , SD , payout ratio etc etc , and in the other approaches of taking median , you are directly comparing the company's pbv ratio to the average of other comparable firm's pbv ratios , and talking about the thing about regarding your doubt , comparable firms are those firms that have the same risk , growth profiles as your firm , so I don't know why you insist on taking coefficient of variation ?
That's why he said that it should be screening criteria not selection criteria he will study company as well that's the point you can find relatively cheap companies and that will build the foundation of stock selection