Yes, history tells us that valuations matter. Since 1988 there's been a link between the non-financial PE ratio and subsequent one- or three-yearly returns on the All-Share index. These relationships now predict decent returns - of just over 10 per cent a year - simply because the PE, at 16.2, is still slightly below its post-1988 average of 16.5.
However, although they are statistically significant, these links aren't very strong. The PE ratio explains less than one-fifth of the variation in subsequent three-year returns, and less than 10 per cent of the variation in annual returns.
What's more, these relationships are averages. They conceal the fact that there have been quite long periods when valuations didn't predict near-term returns at all. For example, in the late 1990s tech bubble 'expensive' markets became more expensive. And in the 2008-09 crisis, 'cheap' shares became even cheaper.
Recent months have been one of these periods. During the last three years, the correlation between the PE ratio and subsequent monthly returns has been plus 0.06 - implying that the PE ratio has told us nothing useful about subsequent returns at all.
You might object that we shouldn't expect a short-term relationship between valuations and returns. But there was just such a relationship in much of the 1990s and 2000s. It has vanished now.
In this sense, we are in a 'post-valuation' era, in which valuations don't matter.
There's a reason for this. The PE ratio depends upon investors' risk appetite and expectations for growth; the higher these are, the higher will be the ratio. In ordinary times, it was reasonable to suppose that these would mean-revert. When optimism is high, we can expect it to fall; and when it's low, we can expect it to rise. When this is the case, the PE ratio will predict returns. It's only when sentiment doesn't mean-revert, that the PE ratio will fail to do so: this is what happened in the tech bubble (bullish sentiment became more bullish) and in the crisis (pessimism led to more pessimism).
However, since the crisis, growth expectations have come into question. We don't know for sure whether the crisis has simply reduced the level of GDP and hence profits, or their growth rates. If it's the former, PE ratios should be around their long-run average, albeit at a lower level of earnings and share prices. If it's the latter, they should be lower than they were before the crisis, because investors should anticipate a slower future growth rate of earnings.
But we can't say confidently which it is. We cannot therefore say whether the PE ratios we've seen have been too low or too high. As a result, they've lost their ability to predict returns. Maybe this ability will return, when our ideas about future growth rates are less uncertain (which is a different thing from them being correct). Until then, though, valuations won't tell us much.
Talk of valuation has always been in part an illusory comfort, which investors have used to give themselves the confidence to buy or sell which wouldn't otherwise be warranted. This is even more true now than before.
