ECRI Weekly Update: Beating The Recession Drum

by: Doug Short

The Weekly Leading Index (WLI) of the Economic Cycle Research Institute (ECRI) rose slightly in the latest public data. It is now at 126.3, up from 125.4 in the previous week. See the WLI chart in the Appendix below. The WLI annualized growth indicator (WLIg) declined to 3.4, down from last week's 3.6. WLIg has been in expansion territory since August 17th, although it is now at a six-week low, with the high at 6.0 on October 12th.

But ECRI's indicators appear to have diminished in significance. Thursday, ECRI's Lakshman Achuthan made the rounds on TV interviews reaffirming his company's recession call and now pinpointing July as the business cycle peak. That puts us in the fourth month of recession.

Here are four November 29th video links.

  • Bloomberg: Defends recession call (link)
  • Bloomberg: Highlights four indicators (link)
  • CNN International: Interview with Ali Velshi (link)
  • Yahoo! Daily Ticker: Ignore GDP and the Fiscal Cliff, U.S. Is Already in Recession (link)

The most interesting part of the new ECRI media blitz is an article on the website with the intriguing title The Tell-Tale Chart. Here is the chart in question.

With this chart now upstaging ECRI's indexes, we see the company's dependence on the same Big Four Economic Indicators I've been tracking for the past several months.

What Are the Big Four Economic Indicators Telling Us?

Since last week's ECRI update, there has been one new update to Big Four Economic Indicators, the October Real Personal Income Less Transfer Payments. Here is one of my Big Four charts with ECRI's 2011 recession call annotated. Note especially the table below.

There is, however, one curious anomaly between ECRI's chart of the Four Indicators and the series as I track it. I'm referring to Sales, which ECRI shows peaking in July. Exactly what Sales data series they are using is a mystery. I track sales using the same formula as the Federal Reserve economists (see PDF file), Retail and Food Services Sales (RSAFS) deflated by the Consumer Price Index for All Urban Consumers: All Items (CPIAUCSL). This metric, sales continued to increase until October, the data for which was significantly impacted by Hurricane Sandy.

The NBER is also on record as using Real Manufacturing and Trade Sales. The problem with this series is that the data is less timely. Also, it didn't correlate well with the onset of the last recession (peaking seven months after the recession started). Here is a side-by-side comparison of the two using a percent-off-peak chart to illustrate how they behaved in advance of the last recession up to the present.

Are we really in a recession that started four months ago? Thursday, my friend Mish Shedlock announced his agreement with ECRI in his commentary ECRI Sticks With Recession Call.

My personal view is that we are walking an economic tightrope with the average of the Big Four indicators flatlining. However, a post-Sandy rebound, good holiday sales and favorably received outcome to Fiscal Cliff negotiations could easily put the economy into indisputable expansion mode.

As for the recent data, of course they are subject to revision, so we must view these numbers accordingly.

Appendix: A Closer Look at the ECRI Index

The first chart below shows the history of the Weekly Leading Index and highlights its current 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 percent 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 11.9% off the most recent high, which was set over five years ago in June 2007. We're now tied with the previously longest stretch between highs, which was 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 in ECRI's indicator is quite different, and this has no doubt been a key factor in their business cycle analysis.

The WLIg Metric

The best known of ECRI's indexes is their growth calculation on the WLI. For a close look at 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 WLIg metric has had a respectable record for forecasting recessions and rebounds therefrom. The next chart shows the correlation between the WLI, GDP and recessions.

The History of ECRI's Latest Recession Call

ECRI's weekly leading index has become a major focus and source of controversy ever since September 30th of last year, when 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:

Early last week, ECRI notified clients that the U.S. economy is indeed tipping into a new recession. And there's nothing that policy makers can do to head it off.

ECRI's recession call isn't based on just one or two leading indexes, but on dozens of specialized 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.)

Year-over-Year Growth in the WLI

Triggered by another ECRI commentary, Why Our Recession Call Stands, I now include a snapshot of the year-over-year growth of the WLI rather than ECRI's previously favored method of calculating the WLIg series from the underlying WLI (see the endnote below). Specifically the chart immediately below is the year-over-year change in the 4-week moving average of the WLI. The red dots highlight the YoY value for the month when recessions began.

As the chart above makes clear, the WLI YoY, now at 3.2%, down from 3.7% the previous week. Nevertheless, this is higher than at the onset of all but one of the seven recessions in the chart timeframe. The second half of the early 1980s double dip, which was to some extent an engineered recession to break the back of inflation, is a conspicuous outlier in this series, starting with a WLI YoY at 4.1%.

Additional Sources for Recession Forecasts

Dwaine van Vuuren, CEO of, and his collaborators, including Georg Vrba and Franz Lischka, have developed a powerful recession forecasting methodology that shows promise of making forecasts with fewer false positives, which I take to include excessively long lead times, such as ECRI's September 2011 recession call.

Here is today's update of Georg Vrba's analysis, which is explained in more detail in this article.

Earlier Video Chronology of ECRI's Recession Call

  • September 30, 2011: Recession Is "Inescapable" (link)
  • September 30, 2011: Tipping into a New Recession (link)
  • February 24, 2012: GDP Data Signals U.S. Recession (link)
  • May 9, 2012: Renewed U.S. Recession Call (link)
  • July 10, 2012: "We're in Recession Already" (link)
  • September 13, 2012: "U.S. Economy Is in a Recession" (link)

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
"n"= 52/26.5
"m"= 100

WLIg = [m*(MA1/MA2)^n] - m