<<1) If projecting future earnings is difficult, isn't projecting future cash flows more or less equally difficult?>
I alleged that cash flows are earnings stripped of accrual accounting manipulations. Therefore, it is actually more straightfoward to presict cash flows than it is to predict earnings.>
You make it sound so insidious! :-) Some accrual accounting adjustments are bogus, e.g. amortization of goodwill, depreciation (to a lesser extent), widely varying treatments according to whether a transaction is purchase or pooling, showing some securities at carrying value, etc. I take it you consider cash flows sufficiently estimable, or we wouldn't be carrying on this conversation.
<<2) Thus, when projecting far-off cash flows, should we use a higher discount rate for those cash flows?>
The discount rate used should relfect the individual entities cost of capital. That is the interest rate on debt (minus the tax shield), plus the cost of equity (the historical equity risk plus the companies indivudual beta), plus the business risk adjustment accorded to the use of debt and the cost reduction accorded by the use of said debt, plus the current long bond rate as of the date of calcualtion.>
I understand the first two components of your formula, but not the last three. Also, why not use a discount rate that resembles a yield curve, rather than a constant one through the whole life of a project?
<<3) I know that in finance, we are supposed to use weighted average projections of future cash flows rather than tamper with the discount rate. Is it even meaningful to be projecting flows on a 3 to 10 year horizon?>
The primary reason why business managers fail to accurately value a capital project is the fact that they are afraid to make long term projections. Unfortunately, that is the only way to get a meaningful result. When making long term projections, the proper way to avoid false assumptions is to create a scenario analysis which illustrates the delta of the major variables which influence value, ex. discounting rate (cost of capital), cash flows, business risks. My model details 54 different scenarios, concisely displayed in two (DCF and EVA) easy to read matrices.>
3^3 * 2? :-) Hence the use of the payback and discounted payback models! Do you find your valuation models to be more successful when applied to stable, mature industries, as opposed to high-risk, fluid areas? For instance, one can probably predict Gilette's or Coca Cola's cash flows out three to five years far better than one can predict Netscape's or, IMO, Microsoft's. Do you think this theory leads to systemic underpricing of technology stocks (or for that matter small stocks) which leads to the historical difference in returns and increased volatitlity?
<<4) Consider 3) in light of the fact that analysts have, I think, less than about a 50% chance of projecting earnings one quarter out within one sigma of actuals. Accurately projecting such earnings two years out is a virtual impossibility.>
The farther out you project cash flows, the less impact an error will have on the final discounted value. When you purchase a stock, you are not buying past "earnings," they are already consumed. What you are purchasing is the "earnings" POTENTIAL for an ongoing concern. >
Using a 20% discount rate, if I am 10% off on the t(2) cash flow estimate, I am still about 7% off on today's valuation. When dealing with growing companies, a 10% error (in EPS estimates) is not at all unusual; in fact, estimating within 10% is probably unusual. A 100% error four years off is still a 48% error in current dollars (at 20%).
<If you were to by a family business for your relatives, would you be be shopping for the money to be made over the next two quarters, or the earnings potential over the next 3 to 10 years. It is that potential, when discounted by the appropriate oppurtunity cost, that reveals the true value of the entity in question.>
Who could disagree with this? I guess it comes down to the original question: (1)Can we project cash flows with reasonable certainty (or with significantly better hopes than with EPS)? If the answer to that question is yes, we can rationally value an investment (given the use of sensitivity analysis). If the forecast is muddy, then we have problems, we act based on not much more than hope, and COMS drops 13 points in one day on bad news.
Regards,
Winston |
