Leading Indicators Have Turned South

I have been following leading indicators for a while and today David Rosenberg ran some forecasts from his favorite leading indicator measure, the Economic Cycle Research Institutes’s (ECRI) Weekly Leading Index (WLI). Their chief economist, Lakshman Achuthan, is frequently seen on CNBC’s programs. His WLI is turning down and has been for a while.

Rosenberg said:

The growth rate on the ECRI leading index did it again! It sank further into negative terrain, now at -9.8% during the week ending July 9, down from -9.1% the prior week. This was the tenth deterioration in a row and the growth index is now negative for six straight weeks. We have never failed to have a recession with the ECRI at current levels but there is also inherent volatility in the index that requires acknowledgment. Our reckoning is that in the past few weeks, the index has gone from pricing in even-odds of a double-dip to two-in-three odds. It may take a while, but Mr. Market will figure it out before long.

I’ve been forecasting a downturn for some time but for different, more fundamental reasons, but the data seem to be bearing me out. Rosenberg does make an excellent point in his analysis in that his detractors point out that such “double-dips” have been rather rare events in a recovery (my emphasis added):

These “extremely rare” events have been the norm for the past 24 months: negative nominal GDP growth; negative operating earnings; a massive contraction in credit; a 30% slide in home prices (these same economists — Bernanke too — were telling everyone that home prices never deflate over a 12-month time span … but they did this time!); a record-high duration of unemployment.

The past 24 months have given us a lifetime of “extremely rare” events, but as we suggested last week, these are only “rare” from the perspective of an analyst that sees the past 24 months as a typical post-WW2 recession.

Economists always look backward and then interpret data. They rely on historical data to make and justify their positions. This is what almost all economists do and they generally don’t have a very good track record. The reason is, is that the data they choose is either incorrect, insufficient, or based on the wrong economic theories.

My position is that while historical data may have some value in interpreting past events, it is not very useful in predicting future economic events. You can’t look back to other cycles and say, “since X happened in 1982, and the charts look similar today, that X will happen in 2010.” The best way to look at the economy is with a thorough understanding of business cycle theory and try to interpret today’s events in that light. It’s not easy, and, as Dr. Athreya, the Fed economist who trashed bloggers said, “one has to think hard about many, many things.”

I’m not saying I don’t look at data; I do. What I’m saying is that the data doesn’t drive my ideas. At best data is only good for broad brush analysis which is what I try to do. Data can prove me or anyone wrong however. By the way, Rosenberg may complain about his critics, but he does the same thing as they do (leads his analysis with data), only his “hunches” have been better than theirs. Also he dares to interpret data to predict markets behavior and he’s pretty good at it I believe. I don’t do this because I have no clue what the market will do tomorrow.

I recently began following the Consumer Metrics Institute’s data (they were kind enough to send it to me). They have their own proprietary data to look at leading indicators, but generally they are based on actual consumer purchases, primarily of durable goods. They ignore things like food and gasoline, because those are things we have to buy, and are not discretionary. Their growth indicator has been declining since October, 2009 and went negative in January 2010.

What this shows is that they lead the next GDP decline by about three months. Note that  BEA data is slow in coming, and is often revised two times over several months, whereas Consumer Metrics claims that their data is current.

Look at this chart they did on cyclical contractions:

They say:

One measure of the true severity of an economic slowdown is the ‘area under the curve’ (or ‘above’ the curve in this case) swept out by the ‘Daily Growth Index’ over time. This area is just the average magnitude of the decline times the duration of the contraction event. During the 2006 slowdown this area was about 136 percentage-days of contraction, while the 2008 event was much more severe at 793 percentage-days. The 2010 event has now reached 288 percentage-days, over twice the severity of 2006 and well over a third of 2008 ‘Great Recession’ — and it is still growing.

The key point to notice in the above chart is that if the current 2010 curve continues its current course, in about 20 days the 2010 slowdown will be more severe on a day-to-day basis than the 2008 ‘Great Recession’ was at the same point in its respective evolution. Unless the economy begins to pick up quickly, a double dip is likely — with the second round milder but lingering longer than the first.

Of course, the important word in the above quote is “if.”