Slowdown ahead, but not recession
In hindsight, that forecast was spot on. But it was a very lonely position to take at the time – in part because well-known indexes of leading economic indicators (LEI) were still falling.
Given that their LEI hadn’t seen a single monthly uptick since June 2008, the Conference Board spokesman declared last April that “There's no reason to think that this recession is going to end any time this spring or this summer.” Their LEI wouldn’t show even a single month’s uptick until its May 2009 release – but that wasn’t enough, of course, to predict an economic recovery.
The OECD’s U.S. leading index, which is revised significantly each month like its international counterparts, didn’t register its first monthly uptick until its June 2009 release. Again, that was hardly enough to forecast economic recovery.
So how could ECRI predict a recovery last April? Sure, ECRI’s Weekly Leading Index (WLI) had already been in a cyclical upturn for six weeks, but, in itself, that wasn’t enough to make such a bold call.
The key was the U.S. Long Leading Index (USLLI), which goes back to 1919, spanning crises and depressions. While it isn’t publicly distributed, the USLLI has the longest lead time among all of ECRI’s U.S. leading indexes – certainly longer than the WLI and the conventional leading indexes from the Conference Board and the OECD, which include short leading indicators like stock prices.
The USLLI doesn’t include the stock market, which doesn’t qualify as a long leading indicator. Basically, unlike standard leading indicators, the USLLI is more prescient than the markets about the business cycle.
We noted the first monthly uptick in the USLLI in January 2009, four months before the Conference Board LEI and five months before the OECD leading index showed such one-month upticks. Therefore, when the clear cyclical upturn in the USLLI was corroborated by a six-week upturn in the WLI, we publicly predicted a business cycle recovery – a full year ago.
But we anticipated more than just an economic recovery. Quoting words written 90 years ago by an economist familiar with financial crises, we wrote, “The error of optimism dies in the crisis but in dying it 'gives birth to an error of pessimism. This new error is born, not an infant, but a giant.’”
Of course, that’s exactly what happened. Even as the economic recovery took hold and stock prices soared, most remained in the grip of that “giant error of pessimism,” as people had done in the wake of past crises. It’s only recently that they’ve begun to recognize the end of recession in their rear-view mirrors.
The funny thing is, most people who, a year ago, didn’t expect the recession to end so soon (which is to say, most people) don’t remember it that way anymore. Rather, many of them think they correctly anticipated the upturn. Such revision of memories in our own favor stems from what psychologists call hindsight bias – the tendency to believe that our forecasts were more accurate than they really were.
Research demonstrates (Goodwin, P., “Why Hindsight Can Damage Foresight,” Foresight, Spring 2010) that those who show greater hindsight bias typically fare worse as forecasters. In fact, finance professionals who exhibit greater hindsight bias usually make less money – including smaller bonuses!
The bottom line is that, to improve our future investment decisions, we need to reduce our hindsight bias, acknowledging our errors, but, most importantly, thinking through why the unexpected actually happened. That should teach us how to do better next time.
To check our own hindsight biases, we routinely go over what we had actually written or said at the time. People could also review their investment decisions, which ought to tell them something about their actual expectations at the time.
Looking back, we find that, last summer, based on a surge in ECRI’s leading indexes, we declared that there was no double dip anywhere on the horizon, while predicting that this recovery would be more rapid than the last two. Again, this is precisely what’s happened, and not just in terms of GDP growth.
Of course, the leap in long-term jobless rate, which won’t fall to prerecession levels for many years, is a huge structural problem that a business cycle recovery just can’t solve. But please note that the jobless rate already peaked in October 2009, only four months after the recession likely ended – compared to lags of 15 and 19 months after the prior two recessions.
We were virtually alone last summer in predicting such a revival amidst an ocean of gloom. Jim Grant, writing in Grant’s Interest Rate Observer at the time, tagged us as “the exception to this predictive consensus,” and quoted us as follows: “It’s as if you drop a ball and it has a very big drop, then it also shows a big bounce, but it’s the bounciness of the ball that has been going down over the decades since World War II. In other words, sure – it’s less bouncy, but a big drop in economic activity still is followed by a relatively large rebound. What these leading indexes are saying is not that following the worst recession since the Great Depression you will get the biggest rebound since the Great Depression, merely that … it’s going to be stronger than the last two recoveries. In that context, (ECRI’s forecast) is not that audacious.”
Where does the economy go from here? Following the initial post-recession spurt, economic growth always throttles back at some point, since it can’t keep accelerating forever.
So what are ECRI’s leading indexes saying about the timing of this pullback? Quite simply, that in line with the first quarter’s downshift in GDP growth from a six-year high, income and sales growth will also start easing in the next few months.
The good news is that there’s still no “double dip” in sight. The first indication of the next recession is likely to be a downturn in the level of the USLLI. That would be bad news indeed, because recessions often go hand in hand with major bear markets.