Bad Market News From Junk Forecasting Models
The mistake is comparing a recovery from a financial crisis with the rebound from a normal recession.
Bloomberg, August 23, 2017
Today’s column is a necessary evil. It is one of those things where I point out stupid things said by smart people with track records of saying things that turn out to be wrong.
The latest: the idea that the business cycle has peaked and that markets are bound to follow.
No, this won’t be another harangue about the folly of forecasting 1 , although obviously it would be warranted. Regardless of whether I agree or disagree with this particular prediction, my beef is that the various methodologies used seem suspect. What’s more, the resulting track records are even worse.
Just to be clear, I’m not saying U.S. stocks are not pricey (they are). Nor am I saying that bonds are not overvalued (for all I know, they may be). I’m not even saying that the bull market isn’t long in the tooth (that’s simply not relevant). Nor am I arguing that the markets gained too much too fast (they didn’t), or that we’re not late in the cycle (perhaps). 2 These things have all been true for a long time now; they will continue until they stop, at which point the people who have been wrong for many years and missed an enormous run up in markets will gleefully say, “I told you so.”
This is how it has always been and how it always will be.
• Correlations: Morgan Stanley’s model “shows assets across the world are the least correlated in almost a decade” — and of course, we know what happened in 2007. But 10 years ago, the crisis came about because of reckless lending (by many of these same banks), securitized subprime loans and derivatives. Why asset-class correlations are relevant is pretty murky. And let’s think back to the financial crisis — there was almost perfect correlation across almost all asset classes, as everything but the safest holdings (e.g. Treasuries) collapsed.
• Earnings: Bank America Merrill Lynch notes that “For the first time since the mid-2000s, companies that outperformed analysts’ profit and sales estimates saw no reward from investors.” I find this statement a bit bizarre. In the past 12 months, the Nasdaq Composite Index and Standard & Poor’s 500 Index have gained 19 percent and 12 percent, respectively. Are we going to blithely declare that these gains were not at least, in part, in anticipation of better earnings? Adherents of the efficient market hypothesiswould like to have a few words with you.
• Slowing Economy: The thinking here seems to be that the U.S. economy is at full employment and now will experience a slowdown in momentum. This should lead to “a decline in corporate profit margins,” according to Societe Generale SA. Maybe that is true; or maybe that is the nature of post-credit-crisis recoveries. Indeed, the entire recovery fits that pattern — a long slow deleveraging and gradual improvement that doesn’t look like the usual snap-back from a normal recession. Just as an aside, how many models like this also predicted the financial crisis? The answer is few, if any.
I am among the first to admit that these are unprecedented times: The U.S. is still healing from a huge financial crisis that culminated in a generational market crash. To stimulate the economy, the Federal Reserve pushed down interest rates to near zero for the better part of a decade. These elements make it almost impossible to compare this cycle with earlier ones that lack these extremely unusual characteristics.
Hence, I believe GMO’s Jeremy Grantham — who has a longer history and a better track record than almost everyone else on this subject — has the right take. He suggests that although “mean reversion in margins and price-earnings ratios is still probable, the speed of regression has slowed way down and become sticky.” In other words, markets always return to more rational pricing and may be doing so now, but they can take a very long time to get there.
In light of that, I suggest being wary of acting on models used by financial analysts that suggest the business cycle has peaked.
1. For some of my earlier tirades about predictions and forecasting, see this.
Originally: Bad Market News From Junk Forecasting Models