In its latest economic forecast, the IMF says something odd. It’s this: “Rich market valuations and very low volatility in an environment of high policy uncertainty raise the likelihood of a market correction.”
There are several problems with this.
One is that high policy uncertainty should already be reflected in share prices. In fact, if policy uncertainty increases downside risk it should be a source of a risk premium in equities – something that generates high returns for the investor brave enough to take it on.
History shows that this is the case. We have an index of global economic policy uncertainty compiled by Stanford University’s Nick Bloom and colleagues. Since January 1997 (when their data begins) the correlation between this index and subsequent annual changes in the MSCI world index has been slightly positive, at 0.2. (That between policy uncertainty and subsequent changes in the All-Share index has also been slightly positive, at 0.17.) This means that above-average policy uncertainty has been associated with equities being slightly more likely than not to do better than average in the following 12 months. Which is consistent with uncertainty being priced into markets and generating a risk premium. For example, uncertainty peaked in January of this year and since then global equities have risen almost 7 per cent.
There’s another problem. Low volatility is not a sign that investors are complacent. It’s a sign that they disagree: share prices are stable and volatility is low when sellers can easily find buyers and vice versa. Disagreement, however, should have no predictive power in itself for share prices.
We can test this. If the IMF is right, we’d expect to see a significant positive correlation between volatility and subsequent changes in global equities – so that low volatility leads to poor returns.
But this isn’t the case. Since January 1997 the correlation between the CBOE’s Vix index and subsequent annual changes in MSCI’s world index has been just 0.04 – essentially zero. Volatility, then, tells us nothing about future returns.This is consistent with low volatility being a sign of disagreement rather than of complacency.
But what about the combination of low volatility, policy uncertainty and “rich valuations”?
One obvious sign of “rich valuations” is the cyclically-adjusted price-earnings ratio (or caper) on the S&P 500, complied by Yale University’s Robert Shiller. It is now just over 30, which is almost twice its average since 1871. This is “rich”.
So, how do this, the Vix and policy uncertainty taken altogether predict returns? We can answer this simply by a regression equation linking annual changes in MSCI’s world index since 1997 to these three variables.
Such an equation tells us that the caper is indeed associated with lower subsequent returns. But policy uncertainty and the Vix have no statistically significant association with subsequent returns. Valuations tell us something, but volatility and policy uncertainty don’t.
This regression points to global equities rising slightly over the next 12 months, albeit with a significant chance of a fall – around a one-in-five chance of a drop of 10 per cent or more. Maybe this corroborates the IMF’s warning of a heightened chance of a correction. But if it does, all the work is being done by valuations: policy uncertainty and volatility are irrelevant.
Does this mean the IMF is wrong to warn of the risk of a correction?
Not at all. There is such a chance simply because there always is.
I suspect that what the IMF is doing here is trying to find a rational basis for what is in fact a hunch or gut feeling. Frankly, I share that hunch: in fact, I reduced my equity exposure in May. But it’s hard to fully justify such a hunch. And it might even be impossible: if there were obvious warnings of market falls, investors would sell as these warnings became clear and so there wouldn’t be a subsequent fall at all.
Not all hunches, however, reasonable, can be wholly rationalised.Not all useful economic statements are entirely scientific.
