It seems as if, ever since the last recession ended in June 2009, doomsayers have been busy calling the start of the next one.
From the 'flash crash' of 2010 and that summer's growth slowdown to the eurozone 'smash crash' and debt ceiling debacle of summer 2011, the same recessionistas have emerged regularly (usually on the day of a stock market bottom) to regurgitate their familiar cry.
The bears gained more credibility when, in late 2011, the undisputed heavyweight champion of the recession-forecasting business stepped into the ring - namely ECRI, in the shape of its co-founder Laksman Achuthan.
ECRI, Sept. 2011: Not wrong, just early
Since inception this independent organization had accurately called every recession - 1990, 2001, 2007 - long before most people realized it was happening, and in the intervening years had not called a single 'false positive'. Considering the amount of conflicting data swirling around, the economic shocks during those intervening years and the abject failure of just about every other official body of economists to make accurate recession forecasts, that's a remarkable achievement.
Yet this time, the US economy has relieved itself all over ECRI's spotless record. It has wheezed along at not much more than stall-speed for nigh on two years and resolutely refused to crack.
Until now.
At least that is the conclusion you could be forgiven for reaching with the latest reading from one of the most widely-watched of economic indicators, the ISM Manufacturing Index. On Monday, ISM reported that the manufacturing index dropped below 50, the dividing line between expansion and contraction.
"So what?" you laugh, "it dropped below 50 a few months ago and there was absolutely no sign of a recession!"
Unfortunately, that's the problem.
ISM MOMENTUM -
A LOOK BACK THROUGH HISTORY
Like much economic data, the manufacturing sector's numbers move cyclically and their momentum can be measured quite easily. Just as it's possible to examine the acceleration or deceleration of a trend in stock prices to decide whether a major reversal is in the works, the great engine of US manufacturing has obvious periods of acceleration and deceleration which, when plotted on a chart, appear telling.
For clarity, let's split all the available data into three overlapping periods: 1948-1978, 1975-2000, and 1988-current.
1948 - 1978
 |
| Shaded grey bars indicate recessions |
Look closely and you find the data behaves with admirable consistency, completing a simple four-stage sequence.
- 1 Following a recession, manufacturing comes charging out of the gate, rising to a minimum reading of 55 (and usually well into the 60s). In this early stage the economy normally experiences its peak rate of expansion
- 2 The expansion begins to decelerate, sometimes precipitously, sometimes steadily, until the index dips below 50 (yellow arrows)
- 3 The first dip is almost always a headfake: the economy is merely taking a breather before another surge, even though that surge generally does not exceed the rate of expansion set in the first explosive period.*
- 4 The index - after failing to exceed its peak rate of expansion - dips below 50 once again (red circles). The economy is now in, or will shortly enter, a new recession.
* The only time since 1948 a recession has begun with the ISM data above 50 was in 1973/4, an outlier largely resulting from the OPEC oil price shock.
1975 - 2000
This simple way of looking at the data, even during the most powerful and sustained expansion in US history from 1980 to 2000, would have alerted you to all three recessions just as they were getting underway.
Unfortunately it would have also flagged up two of those dreaded 'false positives' (in green) which only ECRI and its black box of lead indicators have appeared smart enough to avoid. But that was an exceptional historical period, and those would be the only bad calls you'd make in 64 years of data.
1988 - 2012
Here, the first three signals overlap with the previous chart, so look to the right at the signals since 2000. This way of examining the data would have alerted you to the 2001 recession, though a touch earlier than normal, and would have rung warning bells as the economy began to contract in 2007.
So, what of 2012? In what looks like an even swifter version of the decline into 2007, again we see the classic signs of waning momentum: an initial surge from the prior recession is followed by a heavy drop off from the peak, a dip below 50, a pop higher which has failed to exceed the expansion high and now, in December 2012, a second drop below 50.
Of the 11 recessions since 1948, 10 recessions could have been identified in this way - and you would have done so either just as they were starting or, at most, a few months before they began.
Are there caveats to be applied here? Of course. This type of analysis is hardly scientific and the rules, such as they are, are essentially data-mined. But we do know beyond doubt that momentum studies, when applied to cyclical and market data, 'work'. If one believes that this data is indeed subject to momentum analysis - and if one believes that the manufacturing sector is still significant in determining US recessions (it was in 2007) - then even this simple process should have some validity.
Nevertheless, it is perfectly possible that this dip is merely another headfake and that manufacturing will re-accelerate due to some kind of catalyst, as it did in the mid-1980s and '90s on the back of the credit, consumer and tech booms. Such a move would reverse the currently plunging momentum trend in spectacular fashion, and would suggest our secular economic downtrend is well and truly over.
