Can ECRI’s Weekly Leading Index be Trusted?

By Georg Vrba, P.E., Advisor Perspectives,

[Editor’s Note: In forecasting the economy, no proven dynamic forecasting model exists.  Econintersect has a benign opinion of leading indices – and views them like ground hogs – a concoction of logical and illogical data points which somehow provide forecasts better than opinion.]

I provide examples of indicators from two respected research institutions which provide a completely opposite outlook for the economy. I will look at an indicator derived from ECRI’s weekly leading index which signals almost a 100% probability of recession, and also one developed from the Conference Board Leading Economic Index which is well out of recession territory.

Finally there is an analysis as to why ECRI’s weekly leading index is signaling recession when most of the other recession indicators do not.

The percentage change of the ECRI’s Weekly Leading Index from its smoothed 10-week moving average one year earlier plus 3% – WLIyoy+3

The indicator is shown in the chart below. The amount by which the percentage change was increased to make zero the recession trigger line was found by optimizing for the highest score using my recession indicator evaluation system.

Superimposed on the indicator graph are the levels of the indicator three months before recession start and at recession start, the golden and red circles, respectively. Note, that all markers are above the current level of the indicator and only two of the red markers (the levels of the indicator at recession start) are below the recession trigger line. This indicates that historically recessions have always started before WLIyoy+3 reached a level corresponding to the current one.

Also shown is the linear regression line for the levels of the seven red markers. This line is about 2% above the recession trigger line, indicating that the recession trigger line should possibly have been placed at a higher level.

Likelihood Ratios were calculates as described in Likelihood Ratios and their use in Recession Indicators. Since I was interested to find the probabilities of a recession start and not particularly interested how well the indicator captures the recession length, the False Positives occurring immediately after recession ends (marked “G” in figure 1 of the above mentioned article) were omitted to calculate the likelihood ratios and probabilities for the indicator.

This indicator has been signaling recession since the middle of October 2011 when it crossed the trigger line to the downside. It has been indicating since then that there was a 99% chance that the US economy was within three month of a recession, and a 92% chance of the economy being in recession.

The high recession probabilities signaled by this indicator support ECRI’s repeated recession calls and their prognosis that the economy would enter recession not later than the end of June, 2012. See also the similar opinion expressed by “Mish” Shedlock in his recent commentary 12 Reasons US Recession Has Arrived (Or Will Shortly) and John Hussman in Enter, the Blindside Recession.

The smoothed 6-month annualized growth rate of the Conference Board Leading Economic Index – CBg+2.35

There are many other good indicators which are well out of recession territory, such as the Conference Board Leading Economic Index. Its smoothed six month annualized growth rate (CBg) is now about +2.5%. This indicator signals recession when the growth rate becomes -2.35%. Therefore its growth rate would have to decline considerably before the Conference Board LEI will signal a possible recession. The graph of CBg with 2.35% added to make zero the recession trigger line is shown below. It is plotted in real time – thus the May level appears when it was released, namely in the week ending June-22-2011.

Superimposed on the indicator graph are the levels of the indicator three months before recession start and at recession start, the golden and red circles, respectively. Note, that all red markers are below the 5% current level of CBg+2.35, and only one golden marker from 1973 is at the current level. This signifies that recessions have never started when this indicator was at a level of 5%, where it is now.

Also shown are the linear regression lines for the levels of the seven red markers and the seven golden markers. The trigger line is situated between the two regression lines, indicating that the trigger line was correctly placed when optimized for best position with my evaluation system.

Why are the current signals from these two indicators at odds?

In order to answer this question one has to look at the underlying components which make up the indices. The Conference Board provides comprehensive descriptions of all the components and many research articles in support of their index. ECRI’s methods are arcane; very little information is available from them on the composition of the WLI and thus their methods cannot be checked easily.

However, despite ECRI’s secrecy, we now have a very good idea how the WLI is assembled. Every week we are able to provide a Shadow WLI to the public before the official release of the WLI. Since we know the components of the WLI, we can also have a look at which of the components was mainly responsible for the big year-on-year decline of the WLI of almost 6%.

I found that since the beginning of this year the industrial metals component contributed on average 45% to the year-on-year change of the WLI, and for the last ten weeks a surprisingly high 60%. It can hardly be argued that the price of industrial metals is much influenced by US industrial activity, which has been on the wane for many years.  The metal prices are mainly influenced by China’s growth and have declined by about 30% since April 2011. It is not surprising then that the WLI is so much lower now than a year ago, but does the WLI reflect the state of the US economy correctly?

Industrial metals are not a component of the Conference Board Leading Economic Index, which explains the current opposing signals from these two indicators.

So which one shall we trust?  I think the answer is obvious.

About the Author:

Georg Vrba is a professional engineer who has been a consulting engineer for many years. In his opinion, mathematical models provide better guidance to market direction than financial “experts.” He has developed financial models for the stock market, the bond market and the yield curve, all published in Advisor Perspectives. The models are updated weekly. If you are interested to receive these updates at no cost, send email request to [email protected].