Sticky Luck
Essays on Stock Market Patterns and Expected Returns
Mark O'Reilly, FIA, ASA, MAAA
oreilly
Twelve:
Other Market Math
Twelve:
Other Market Math
Though it looks awfully simple, most big market models depend way too much on this crude equation...
Consider the simple investor-decision model:
p = ∑et * (1 + jt)^t * (1 + it)^-t for t =1,∞
therefore P ~e /( i – j)
Where
- et are earnings assigned to the unit of stock p;
- i and j are weighted-average discount rate and earnings growth rate over the series it and jt
Also
- Set i = risk-free rate, so we must adjust fully for all risk in setting j. (Models investor choice of equity versus risk-free alternative.)
P is so sensitive to the relative values of i and j that even rational valuations (so-called fundamental values) can be measured only by convention. That convention can shift over time. It moves by sentiment following short-term changes in actual it and estimated jt for short t, despite their negligible impacts upon i and j.
All such market models are capable of working brilliantly for many years when chosen (i-j) coincides with future market-estimated (i-j), and then crashing in flames when the relationship breaks.
Mark O'Reilly, FIA, ASA, MAAA
oreilly
