(R). Otherwise, the stock would have a negative value, a patently absurd
result. Theoreticians are not troubled. They say high growth rates are
necessarily temporary and that any growth rate used in this supposedly
perpetual model ought to be more permanently sustainable, and hence low
enough to avoid producing a negative denominator for the fraction.
Even if we could make reliable assumptions about what such a permanently
sustainable growth rate should be, real-world investors would
never be content to use it. There’s simply too much opportunity to prosper
by reacting to the wider variety of growth patterns, whether permanently
sustainable or not, that we see every day. Also, many high-quality companies
pay little or no dividends.
Hence it is essential that we modify our approach to dividend-based
valuation to accommodate these day-to-day realities. Such adaptations are
well described in Security Analysis on Wall Street: A Comprehensive
Guide to Today’s Valuation Methods by Jeffrey C. Hooke (John Wiley &
Sons, 1998) and Investment Valuation: Tools and Techniques for Determining
the Value of Any Asset by Aswath Damodaran (John Wiley & Sons,
1996; 2d ed. 2002). A popular theme is the multipart approach. Here’s how
a three-period variation would work.
• Assume in period one, the corporation reinvests all of its profits (i.e.,
paying no dividends) to facilitate rapid growth and a better stream of
dividends (than what would be the case if dividends were paid immediately)
that will start in the future. For the time being, there is nothing
we can directly use for purposes of computing a value. All we can
do is make assumptions about how fast earnings will grow. That will
give us a projected EPS level for the start of the second period. (If
the company is losing money now, we project revenue growth and
make assumptions that the company will break into the black at or
before the start of period two. At that time, estimate EPS by multiplying
an assumed net margin by the projected sales number and
then dividing by the number of shares.) For now, let’s assume period
one lasts five years.
• The corporation pays its first dividend in the sixth year, which is the
start of period two. The payout ratio (the dividend as a percent of
earnings) is not yet at what will ultimately be a long-term sustainable
level. So we are not going to plug this dividend into a DDM-like formula.
But we do recognize that the shareholder gets the dividend so
we factor it into stock valuation. We do that by using the present value
of this payment. Throughout period two, the payout ratio will gradually
rise to a “permanent” level. Also, the rate of earnings growth will
decelerate from the unusually rapid period-one pace to the long-term,
permanent, “mature rate.” The investor makes assumptions about how
this progress will occur year by year. For each year in this transition
period, calculate the dividend that is expected to be paid (multiply
projected profits by an assumed payout ratio), and add its present
value to the here-and-now estimate of stock valuation.
• We assume period three, which starts, say, in year 11, is the mature
phase characterized by a permanently stable rate of dividend growth
that is less than the required rate of return. Now, we can use the strict
DDM formula. The present value of this amount is added to the hereand-
now estimate of stock valuation.
That’s it. We have an objectively derived value for a company that is
growing rapidly at present, and may even be losing money. We stretched
the DDM mathematics as far as we’ll ever need. (If we’re looking at an initial
public offering for a company that’s barely out of, or even in, the development
phase, no problem; extend periods one or two to 10 years or 20
years or however many are needed to make the math work.)
Are you satisfied? I hope not.
Most investors know full well how hard it is to project sales and/or earnings
just one quarter into the future. Does anybody really want to commit
funds based on assumptions stretching 11 years and beyond? And is there
any rational way any investor can decide on the probable lengths of periods
one and two, not to mention crucial patterns of transition? We need one set
of assumptions regarding the pace at which the payout ratio makes a transition
from zero at the end of period one to a permanent level at the start of
period three. We also need to decide how quickly and smoothly earnings
growth decelerates from the rapid period-one pace to the mature periodthree
rate. It’s fun to make up spreadsheets and fiddle with things like
this. But if you do it, make sure you confine such efforts to entertainment.
Read More :Beyond the Dividend Discount Model
