Warp Drive Valuations
OK, we know that bubbles sometimes form in markets and we also know that sometimes the price of a stock doesn’t so much drift away from a realistic valuation as engage the Warp drive and disappear off into the Outer Limits. There’s also modelling research which also suggests that the way markets behave depends on the balance between short-term chartists and long-term fundamentalists.
Which is nice and all that, but doesn’t really explain how we should actually go about investing. After all, why is one method better than another and what should we actually invest in? To which the answer, it seems, is all tied up with dividends.
The Miller-Modigliani dividend irrelevance proposition states, unsurprisingly, that dividends are irrelevant to investors. Apparently we shouldn’t really care how we get rewarded for investing in shares. On this basis stock buybacks are as good as dividends (see: Buyback Brouhaha). Moreover, capital gains obtained by companies retaining their earnings are likely as good as anything else: in fact much of the idea behind investing in growth companies is that they have so much opportunity to grow by reinvesting their earnings that they’d be positively barmy to give these away in the form of dividends.
So, given this, it seems a bit strange that investors should want their companies to pay dividends at all, especially as experts reckon shareholders should prefer buybacks as these are more tax-efficient. Douglas Skinner suggests in The Evolving Relation between Earnings, Dividends and Stock Repurchases that firms are increasingly moving towards buybacks:
"Overall, the evidence suggests that corporate earnings now drive total firm payouts – dividends and repurchases – and that repurchases play an increasingly important role, which helps to explain the disappearance of dividends."
It’s likely that income loving investors will continue to prefer dividends anyway: this desire for a nice income tends to cause companies to behave in slightly irrational ways – as we saw in The Psychology of Dividends – managements are extremely disinclined to raise dividends too much in case they have to cut them in future. Nonetheless there’s not much doubt that many investors prefer dividends on the basis that cash in their pocket is worth more than money in a firm’s bank account.
In fact, over long periods of time dividend income is rather critical to investors. As recorded in Global Investing: The Professional's Guide to the World Capital Markets for the US stock market over the period from 1798 to 1990:
"Note that nearly all of the real return over the period comes from the reinvestment of dividends."
So the growth in investor portfolios can almost wholly be explained by dividend income, rather than capital growth. This is one reason why experts reckon that reinvesting dividends is a critical part of any investor’s long-term strategy, since capital gains and losses are likely to even out over long enough periods.
Because of this one of the ways in which long-term valuations are often derived is through discounted cash flow (DCF) models. This type of model derives a so-called “fundamental value” for a company by modelling the discounted value of future dividends. Which is problematic because the future is essentially unknown and unknowable so any model based on long-term predictions is likely to be flawed.
Moats and Invisible Hands
There is, however, one situation where such models may well be of use to investors: where the firm in question has a strong defensive moat meaning that they are able to hold off the dreaded invisible hand of the market (see: Moats, Unbundled). If a company can prevent its margins being eroded by competition then there’s a decent chance that a DCF model will provide some reasonable estimate of fundamental value. These types of models work by a process known as backwards induction. By starting from a rational, albeit hypothetical, future dividend stream we can backward induct our way to a current valuation.
Now analysis has shown that backward induction can theoretically fail if investors aren’t sufficiently rational or if the trajectory of dividends is itself uncertain. Since both of these conditions are likely to be true we might stop at this point and argue that the whole idea is stupid. However, the alternative argument is that over long enough periods markets behave as though investors are rational and dividends will, for that subset of companies able to protect their earnings, return to their long-term trends.
So if we assume that there are longer term investors out there making sensible and realistic assumptions about future dividends and current valuations it would still leave open the question about how bubbles form and burst. A neat little experiment by Shinichi Hirota and Shyam Sunder, Price Bubbles sans dividend anchors, suggests a reason.
The researchers hypothesised that although the market may contain long term investors who are making rational estimates about fundamental valuations by backwards induction it may also contain shorter term traders who aren’t using dividends as anchors in their valuations. Rather they’re using share price which is, of course, a self-fulfilling prophesy because if you base your investing on share prices, which are affected by your investing you run into positive feedback. Investing on the basis of dividends, however, means you’re anchoring on something you, as an investor, can’t directly affect.
As the researchers put it:
“The results are consistent with the proposition that when they are unable to backward induct from dividend anchors, investors tend to form their expectations of future prices by forward induction using first-order adaptive or trend processes. In these markets, allocative efficiency is unpredictable, and the cross-sectional dispersion of wealth increases with the deviation of prices from fundamentals. In contrast, prices in markets populated by long-horizon investors tend to converge to the fundamentals.”
Or to put it another way: short-term investors with no interest in long-term fundamentals make their investment decisions based on the existing share price or the trends in share prices. Longer-term investors tend to base their investing decisions more on whether or not the security is trading above or below a fundamental value.
Additionally the research shows that where future dividends are uncertain there’s an increased possibility of a share’s price getting out of whack with its fundamental valuation. This would explain why companies that don’t pay a dividend – which either tend to be high growth or speculative – can end up with share prices which don’t really reflect a sensible valuation.
This finding is in line with Arnott’s research into clairvoyant value, which shows that glamour stocks tend to grow at twice the rate of value stocks but that investors are willing to pay four times the price (see: Don't Overpay for Growth). The net result of this is that a value investing style outperforms a growth investing style, despite the better underlying performance of companies in the latter case.
Which, for longer term investors disinclined to trade a lot suggests that focusing on companies that are susceptible to backward induction from a predictable dividend anchor would be a sensible approach. Or, to put it in English, look for companies with defensible moats on reasonable valuations, paying a decent dividend and then hold a long time.
Which is a long-winded way of stating the bleeding obvious, but it’s always nice to have some evidence to support one’s intuitions. Meanwhile investors with short-term horizons will continue to focus on momentum, because that’s what you do when you’re adrift without a dividend anchor.