This article is part of Morningstar's "Perspectives" series, written by third-party contributors. Here, Kevin Gardiner, chief investment officer for Barclays says that a market correction is more likely to be event driven than based on valuations.
The signal-to-noise ratio in financial commentary is low. One rule of thumb is that some valuation metrics for stocks, namely ‘Tobin’s q’ and long-term cyclically-adjusted price earnings (or ‘CAPE’) ratios, are of limited value and can be largely ignored. Here we explain our scepticism. Apologies for being bit geekier than usual – and for temporarily adding to that noise.
Tobin’s q (for quotient) compares the market value of companies with estimates of the cost of (re)building them. It seems intuitive – why pay more than replacement cost to buy a business?
Alas, the real world is not quite that simple. Data on aggregate replacement costs are patchy and unconvincing – the Fed provides some, but it bundles quoted and non-quoted sectors together, and offers little detail (and no financial sector or non-US data). More importantly, replicating a business as a going concern is not just a matter of re-assembling its assets.
Microsoft, Apple or GSK’s value does not reside in their premises, plant and equipment; Tesco’s worth is not in its shelves; financial sector assets themselves are particularly hard to value. An inaccessible and flawed valuation tool is of little use.
The idea of smoothing corporate earnings in CAPE ratios to eliminate short-term – cyclical – noise is a good one, and pre-dates q. But CAPE proponents often use US data stretching back to the late nineteenth century to argue that today’s stock market is expensive relative to (say) its 100-year moving average. Those claims are not supported. Prior to the great Depression there weren't any well-established accounting standards, and earnings data from that period wouldn’t pass muster today. Stock market indices themselves have changed drastically – from around a dozen or so railroad companies in the 1880s to today’s multi-sectoral baskets with hundreds of names. Information is cheaper, and liquidity greater, nowadays. And while ultralong - term historical charts are seductive, they are no less vulnerable than shorter-term ones to the fallacy of induction: across most of recorded time, stock markets didn’t exist at all.
None of this deters the advocates of the two ratios, whose latest claim is that both metrics unexpectedly move together, and so must be onto something. Wrong on both counts. Each measure compares a volatile, stock market numerator with a more slowly-moving denominator, so of course they move similarly: most of that movement comes from the common component, stock prices. If they are indeed “onto something”, it is simply that spikes in stock prices are often reversed. This is where most of their alleged predictive power comes from, not the fancy denominators. Chartists tell us the same thing more quickly and less patronisingly. Finally, you may read that these ratios are historically proven, but they aren’t: they were popularised long after the events they supposedly predicted.
There is no short-cut to valuing stocks: it is an unavoidably subjective and constantly-evolving quest, and in our view the simpler and more transparent the statistical input, the better. To us, conventional PEs and the earnings outlook, dividend yields, price⁄ book values and implied risk premia all suggest that stocks are not especially expensive. for sure, developed markets have gone some 15 months without even a 10$ correction, and a setback in the weeks ahead would not be surprising – even though tapering seems largely priced-in, and an intervention in Syria seems less imminent. But such a setback in our view will likely reflect not valuations but events.
