<For example, if Intels earnings estimate is $1.00 as reported but $1.15 if the depreciation schedule is dragged out a few years and we go by cash earnings to adjust for any non-cash deal related amortization charges, because that is how AMD happens to be doing it, then one needs adjust the multiples for the earnings differential. This can usually be estimated (albeit with errors).>
If you start adjusting for non-cash related charges, then you are now not dealing with accouting earnings, but reconciled income. We are saying the same thing, but I don't think you are differentiating between reported accounting earnings and actual operating profit.
<My point is that a study of the lack of correlation between printed estimates (or reported numbers) relative to prices and implied market caps does not necessarily mean that stocks do not trade relative to earnings expectations. >
I have already explained that there is a lack of significant correlation between the 12 month consensus estimate and market value as well.
<My view rests on something which one is not able to prove or disprove; that investors do care about sales and earnings but normalize the valuations through the P/E and PSR ratios.>
You are erroneously (in my humble opinion of course) lumping sales and earnings in the same category. My study shows sales to have a high correlation to market value in most cases. Yahoo and AMZN sales have a high correlation between market value to sales yet a negative correlation between earnigns and market value). The reason is most likely due to the fact that these companies reivest so much of thier monies back into the business before it gets to the net income available to common line item.
If your asertions concerning P/E's are true, should the P/E have a high correlation to market value and the consensus estimate? According to the cursory research that I have done they don't.
<If one company has higher Return On Invested Capital than another, that too will be factored into the multiple.>
Again, we have an apples and oranges situation here. ROIC is impossible to determine when dealing with accounting earnings because of the ambiguity that I complained about as well as the lack of cash transactions. Even if one assumes that these accounting earnings are reconciled to produce a number through which one can produce a credible ROIC number, the ROIC methodology itself is incomplete. If one does not capitalize R&D, marketing and advertising expenses, ROIC will value AMZN, YHOO and most mining companies and many other high growth companies considerably lower than the market has valued them for several years running. The reason is that many companies invest adn reinvest capital at a breakclip rate to sustain growth. The recently popularized method of calculation ROIC, as marketed by companies such as Stern Steward and McKinsey & Co., as well as many hip B schools, do not take into consideration the reinvestment requirements of high tech/high growth, therefore calculate the company as having a lower return on invested capital instead of compensating for the need for the high reinvestement rate. The tendency to use invested capital anchored to a specific period in lieu of averaged out over the entire fiscal year or several years also injects additional distortions. I feel the market is more than smart enough to figure this out (hey, if I can do it, then anybody else can as well). That is why you never see many high growth technology, mining, or biotech companies (as well as other hyper-growth non-cash cow companies) rank high on popular listings that use the EVA style methodology.
I hope I have gotten my point across without wandering too far, it is late. |
