Last fall, the Economic Cycle Research Institute (ECRI) announced that the U.S. was “tipping into a recession.” This was a startlingly black-and-white call, one that, thus far, appears to be dead wrong. Instead of backing away from its call, however, ECRI has stood by it, even as its Weekly Leading Index has gotten better and better since the fall. Below, in a weekly column on ECRI, Doug Short of Advisor Perspectives analyses the latest:
The Weekly Leading Index (WLI) growth indicator of the Economic Cycle Research Institute (ECRI) came in at -1.4 in today’s public release of the data through March 9th. This is the ninth consecutive week of improvement (less negative) data for the Growth Index and the highest level (i.e., least negative) since August 5th of last year. The underlying WLI also improved, increasing from an adjusted 124.6 to 125.1 (see the fourth chart below).
The big news this week is the ECRI public commentary posted yesterday (March 15th) with the title Why Our Recession Call Stands.
For those of us who like to dig into the details, the most interesting revelation in the commentary involved a shift to the year-over-year WLI change from ECRI’s favoured, and rather arcane, method of calculating the WLI growth series from the underlying WLI (see note below).
Most data, both public and private, are seasonally adjusted. But the nature of the Great Recession seems to have had an unexpected impact on the statistical seasonal adjustment algorithms that are hard-wired to detect when the seasonal patterns evolve and change over the years. This is normally a good thing, but when the economy fell off a cliff in Q4/2008 and Q1/2009, it was partly interpreted by these procedures as a lasting change in seasonal patterns. So, according to these programs, data from Q4 and Q1 would be expected thereafter to be relatively weak, and therefore automatically adjusted upwards. Our due diligence on this subject indicates a widespread problem, resulting in many recent economic headlines being skewed to the upside.
However, we have no way to objectively measure the extent of these problems – either the upward bias for Q4 and Q1 or the downward bias for Q2 and Q3. Fortunately, year-over-year growth rates are naturally less susceptible to these seasonal issues because they involve comparisons to the same period a year earlier that is likely to be skewed the same way. In contrast, smoothed annualized growth rates, which we have traditionally preferred, presume proper seasonal adjustment. While the extent of the seasonal problem will be debated, monitoring year-over-year growth rates is a matter of simple prudence at this juncture not only for ECRI’s indexes but also for other economic data. The March 15th commentary includes a pair of charts: the WLI YoY and their US Coincident Index YoY. Here is my on version of the former. I’ve added dots to show the YoY value at the first months of recessions and a dotted line showing the current level.
As the chart above makes clear, the WLI YoY is currently at a lower level than at the month for five of the seven recessions during the published series. Of course, the same can be said for its interim YoY trough in 2010. In any case, the behaviour of this indicator over the next quarter or so will be especially interesting to watch.
The History of ECRI’s Recession Call
On September 30th, ECRI publicly announced that the U.S. is tipping into a recession, a call the Institute had announced to its private clients on September 21st. Here is an excerpt from the announcement:
ECRI’s recession call isn’t based on just one or two leading indexes, but on dozens of specialised leading indexes, including the U.S. Long Leading Index, which was the first to turn down — before the Arab Spring and Japanese earthquake — to be followed by downturns in the Weekly Leading Index and other shorter-leading indexes. In fact, the most reliable forward-looking indicators are now collectively behaving as they did on the cusp of full-blown recessions, not “soft landings.” (Read the report here.) For a close look at this movement of this index in recent months, here’s a snapshot of the data since 2000.
Now let’s step back and examine the complete series available to the public, which dates from 1967. ECRI’s WLI growth metric has had a respectable record for forecasting recessions and rebounds therefrom. The next chart shows the correlation between the WLI, GDP and recessions.
A significant decline in the WLI has been a leading indicator for six of the seven recessions since the 1960s. It lagged one recession (1981-1982) by nine weeks. The WLI did turned negative 17 times when no recession followed, but 14 of those declines were only slightly negative (-0.1 to -2.4) and most of them reversed after relatively brief periods.
Three other three negatives were deeper declines. The Crash of 1987 took the Index negative for 34 weeks with a trough of -6.8. The Financial Crisis of 1998, which included the collapse of Long Term Capital Management, took the Index negative for 23 weeks with a trough of -4.5.
The third significant negative came near the bottom of the bear market of 2000-2002, about nine months after the brief recession of 2001. At the time, the WLI seemed to be signaling a double-dip recession, but the economy and market accelerated in tandem in the spring of 2003, and a recession was avoided.
The question had been whether the WLI decline that began in Q4 of 2009 was a leading indicator of a recession. The published index has never dropped to the -11.0 level in July 2010 without the onset of a recession. The deepest decline without a recession onset was in the Crash of 1987, when the index slipped to -6.8. ECRI’s managing director correctly predicted that we would avoid a double dip. The positive GDP since the end of the last recession supports ECRI’s stance.
The Certainty and Dramatic Language of ECRI’s New Recession Call
What is particularly striking about ECRI’s current recession call is the fervor and certainty of the language in the public press release:
A Look at the Underlying index
With the Growth Index showing so little change over the past several weeks, let’s take a moment to look at the underlying Weekly Leading Index from which the Growth Index is calculated. The first chart below shows the index level.
For a better understanding of the relationship of the WLI level to recessions, the next chart shows the data series in terms of the per cent of the previous peak. In other words, a new weekly high registers at 100%, with subsequent declines plotted accordingly.
As the chart above illustrates, only once has a recession occurred without the index level achieving a new high — the two recessions, commonly referred to as a “double-dip,” in the early 1980s. Our current level is 14.5% off the most recent high, which was set 4.7 years ago. The longest stretch between highs was about 5.2 years from February 1973 to April 1978. But the index level rose steadily from the trough at the end of the 1973-1975 recession to reach its new high in 1978. The pattern we’re now witnessing is quite different.
Additional Recent Analysis
Here are some links to some useful articles for evaluating the substance of the ECRI’s controversial recession call:
- March 6: ECRI’s Recession Call: “Is There Something Magic About 2 per cent?” Georg Vrba digs into ECRI’s claims about the year-over-year growth rate of its indicators.
- March 3: The Elusive 2012 Recession: When Can We Expect It? Georg Vrba and Dwaine van Vuuren introduce the RecessionAlert.com website.
- February 21: Evaluating Popular Recession Indicators Georg Vrba offers a critique of three recession indicators, including ECRI’s WLI and WLI Growth.
- February 17: No Wonder ECRI’s Weekly Leading Index Data is Free Kishor Bhatia points out the remarkable similarity between the WLI and the S&P 500.
- February 14: Recession: Just How Much Warning is Useful Anyway? Dwaine van Vuuren questions the benefit of recession forecasts for portfolio management.
Earlier discussions of the ECRI recession call include these commentaries:
- January 18: Why the Average Investor Is Bamboozled by Recession Forecasts by Jeff Miller.
- January 17: Further Improving the Use of the ECRI WLI by Dwaine van Vuuren and Georg Vrba)
- January 10: Using the ECRI WLI to Flag Recessions by Dwaine van Vuuren
- January 3: US Recession – An Opposing View by Dwaine van Vuuren.
Here is Dwaine’s latest snapshot of the WLI, through last week’s data, which should be studied in the context of his January 17 article with Georg Vrba:
With regard to the ECRI shift to YoY, Dwaine comments:
We note with interest ECRI’s focus on co-incident indicators and 12-month growths versus their normal 6-months smoothed approach when in defence of their recession call. We agree with their assessment on seasonality factor risks (the reason they are now highlighting 12-month growths which show a more bearish outlook than their normal 6-month smoothed growths), that is why most of the components of the SuperIndex use 12-month growths in their respective recession probability models. Incidentally, the WLIg+1 in the below chart is unique to the other WLI growth variants in that it is a 52-week smoothed growth rate and theoretically devoid of seasonality risks. What makes it interesting is the fact it has only 1 false positive in the historical record versus the ECRI’s WLIg (6 months smoothed rate) which has 12 false positives to its name. RecessionAlert.com] Another Source for Recession Forecasts
But What About the 3.0 Q4 GDP?
At the end of February the BEA posted its Second Estimate of 3.0 for Q4 GDP, up from 2.8 in the Advance Estimate a month earlier. This is definitely above the conventional view of the economy on the threshold of a recession. ECRI itself did not offer a specific date for the start of the forecast recession. A general view is that ECRI’s headlights shine about six months into the future, which would make Q1 2012 GDP a critical number for evaluating ECRI’s stance. In his recent media appearances, ECRI’s Achuthan has said that his company’s date for the start of the recession extends though the first half of 2012.
The U.S. has had eleven recessions since the earliest quarterly GDP calculations, which began in 1948. In the month declared by the National Bureau of Economic Research (NBER) as the beginning of the recession, quarterly GDP for that month has only been negative four times. In fact, three of the six positive GDP recession starts were at GDP levels higher than the 3.0 of Q4 2011.
ECRI doesn’t provide the general public with the analytical details behind its calls, but the Hoisington Investment Management Q4 Report also makes a 2012 recession call accompanied by an extended economic analysis that includes several key topics:
- soaring debt-to-GDP
- contractionary fiscal and monetary policies
- anticipated decline in exports
- a weakened consumer
- capital spending cutbacks
See also the Advisor Perspectives interview with Lacy Hunt, Hoisington’s executive vice president, for a reassertion of their recession call with some additional explanation for their views.
Note: How to Calculate the Growth series from the Weekly Leading Index
ECRI’s weekly Excel spreadsheet includes the WLI and the Growth series, but the latter is a series of values without the underlying calculations. After a collaborative effort by Franz Lischka, Georg Vrba, Dwaine van Vuuren and Kishor Bhatia to model the calculation, Georg discovered the actual formula in a 1999 article published by Anirvan Banerji, the Chief Research Officer at ECRI: The three Ps: simple tools for monitoring economic cycles – pronounced, pervasive and persistent economic indicators.
Here is the formula:
“MA1” = 4 week moving average of the WLI
“MA2” = moving average of MA1 over the preceding 52 weeks
WLIg = [m*(MA1/MA2)^n] – m