The equity premium puzzle

Corner Solution has raised a question: is the equity premium puzzle a justification for privatizing personalizing social security?

The puzzle is that equities have done far better than they should over the long-run. In theory – as Rajnish Mehra and Ed Prescott pointed out – equities should out-perform bonds by less than half a percentage point a year. But since 1872 they’ve out-performed by almost seven percentage points.

Now, I reckon the premium going forward won’t be anything like this big; and as Brad DeLong points out, if you think otherwise you’re in trouble.

But let’s assume, for the sake of argument, that the equity premium will be biggish. Does this justify investing in equities?

I’m not sure it does. It all depends on why the premium exists.

Clearly, it exists because some people don’t hold equities. Their lack of demand means prices are lower than they should be – which is another way of saying prospective returns are higher than they should be.

So, why do people avoid equities?

One possible reason is that they exaggerate the short-run riskiness of the market - myopic loss aversion in the jargon. Alternatively, they might be out of the market because they simply don’t save anything anyway, or are ignorant of the market’s prospects for other reasons.

Yet another possibility is that people avoid shares because a short-run loss would force them to restrain their spending, and being creatures of habit such cutbacks would hurt very hard.

And another possibility is that some people fear the market will fall at the same time that they lose their jobs, and so regard the market as especially risky.

None of these reasons, I think, justify avoiding equities as a way of saving for retirement. These, then, might justify privatizing social security.

But there’s another possibility, which is really nasty.

It’s that the US’s equity premium is due to luck. Will Goetzmann and Philippe Jorion have found that the average real return on global stock markets since the 1930s has been much closer to Mehra and Prescott’s predictions. This is because many of the markets that existed in the 1930s – like those in central Europe – subsequently shut down.

Looking at US history and inferring that equities are a good investment is therefore like looking at soldiers returning home and inferring that no-one died in the war – it’s the survivorship bias.

You might think this is a mere academic quibble. Surely, there’s no chance of the US market suffering such a calamity.

Well, try this for a theory.

In the new economy, the key asset is human capital, not physical capital. The power of the owners of physical capital – shareholders – is therefore declining at the expense of that of suppliers of human capital; Luigi Zingales has described this process, and its implications for corporate finance in “In search of new foundations.” You’d therefore expect it to become harder for shareholders to extract profits from a business.

There are two pieces of evidence that this is already happening. First, over the long-run, real the real earnings of S&P 500 companies have grown about one percentage point slower than real GDP: compare these figures with these figures.

Second, non-farm non-financial companies have generally seen a fall in return on tangible assets. At the peak of the boom in 1998, this return was 5.6 per cent - a full percentage point lower than the return in the cyclical trough of 1961; compare tables F.102 and B.102 here.

Now, what if the factors that have contributed to these trends accelerate in coming years? Global competition, more intense domestic competition (caused in part by the declining barriers to entry as a result of falling physical capital requirements) and rising importance of human capital are all threats to profits. And firms can’t innovate their way out of this squeeze; as William Nordhaus has shown, innovation is no way to make money.

And isn’t it quite possible that the sources of growth will increasing come from firms without equity owners – say, partnerships or private firms, with shareholders only able to get their hands on these, if they can at all, at the peak of their fortunes?

It’s conceivable, then, that the US stock market might die a slow death even if (indeed because) the rest of the economy thrives. Maybe the apparent equity premium is a reward for taking this risk.

And of course, if this happens, a shift from tax-financed pensions to equity-financed ones would prove disastrous.

Now, let me be clear. I’m not saying this will happen. And I’m certainly not telling Americans what to do about social security; some mad German cost me the right to a say-so in a bit of unpleasantness a few years back.

All I’m saying is that the tiny probability of a disaster should at least be considered before invoking the equity premium as a reason for privatizing social security.

Comments

[Chris Purnell]

All this seems to be saying is that the efficiencies in the market will go to the producers and not to the owners. This can be seen in the 'obscene' pay differentials of senior management and the return to shareholders. To build your pension plans into a scenario where there is effectively no control of the underpinning assets is insanity. The human capital can 'walk' and destroy your pension plan with it. There are no 'golden handcuffs' golden enough to retain value cf Robben of Chelsea: a prime asset until a severe tackle reduced his immediate value to zero.

[James]

I liked your analysis, but I think, also, there are other factors at work in the system. Perhaps return on equity has been bad on the US markets recently because of the dogma that retained earnings = prospective growth. Also, the tax system has traditionally rewarded capital gains over dividend income, which makes "growth" more desirable for the average private investor than income. Perhaps what is needed is a shakeup of these dogmas to revitalise the markets? Maybe applying the same tax regime to both capital gains and dividends? Why should a company's managers know better what to do with the stockholder's profit than the stockholder himself?

[Wat Tyler]

The prospective equity risk premium is a great topic- loads of historic data, and contributions from some of the brainiest people around, yet no real fix on whether the future is or isn’t going to be like the past. So we can all join in.

On survivorship bias, I thought we’d now got enough data to say that a cap weighted world equity index would have generated an historic premium broadly similar to that properly calculated for the US alone (eg see the Dimson and Marsh LBS work- the non-survivors were mostly small). But the range is 5-6% rather than the 7-8% often quoted.

On the human capital point- it may well be right, but of course major structural uncertainties of that kind are a key reason why equity investment continues to be risky, and ought to command a healthy expected return premium.

We should certainly be very concerned about arguments that say equities are pretty safe over the long haul. They’re not, neither in the context of US social security, nor our disintegrating final salary pension schemes here in the UK.