December 24, 2012
A New Business Cycle Indicator
Interpreting the endless stream of macro data points in search of context for analyzing the business cycle confuses and confounds more than a few investors, and even some economists. The Capital Spectator Trend Index (CS-ETI) is one attempt at a solution, or a least a partial solution. The regular updates in recent months on CS-ETI, in fact, have done a decent job of telling us how the economy is faring based on a broad read of the numbers. But there are several ways of interpreting the data, and CS-ETI is only one approach. How can we stress test its signals? One answer is reviewing the same data through a different statistical lens. Enter The Capital Spectator Economic Momentum Index (CS-EMI), a companion to CS-ETI.
As a brief recap, the older CS-ETI is a diffusion index that measures the percentage of 14 leading and coincident indicators that are trending positive for economic growth. Higher readings equate with more indicators signaling that the economy is expanding; lower readings imply a weaker economy, with levels below 50% signaling a high probability that a recession has started.
A diffusion index, however, doesn’t offer much qualitative information about the degree of strength or weakness in the data. As a remedy, CS-EMI looks at the numbers and quantifies the degree of positive or negative momentum for the overall data set. The methodology for CS-EMI is simply the monthly median percentage change for the 14 indicators that comprise the diffusion index that is CS-ETI. Why use the median rather than an average (i.e., the mean?). It’s well known that the mean is subject to outlier numbers, which can mislead us as to the true average. The median, by contrast, is immune to extremes because it reflects the middle point for a batch of numbers.
Here’s how CS-EMI compares through the decades. Note that near-zero and below-zero readings accompany recessions, usually with a slight lead time relative to the start dates, as determined by NBER. Whereas an NBER dating announcement arrives months, even a year, after the fact, CS-EMI is updated much closer to real time, subject only to the timeliness of the economic reports.
For instance, the current update is November 2012, which is nearly complete in terms of initial reports. Last month's CS-EMI reading is a strong 9.4%. The only missing number is real manufacturing and wholesale trade data.
Considering that November was generally a healthy month for growth in most of the economic indicators, it’s likely that the manufacturing and wholesale trade numbers will follow suit. Meantime, the current November profile looks robust in terms of the growth trend--the macro momentum is positive and strong.
In sum, the odds are low that NBER will declare November 2012 as the start of a new recession. That’s also been the message all along in CS-ETI, and it’s a message that’s corroborated in CS-EMI.
Posted by jp at December 24, 2012 9:12 AM
Much appreciated JP.
Posted by: David at December 26, 2012 11:57 AM
I do track vintage data and the signals still look pretty good, albeit not quite as clean as the revised numbers. But remember, since I'm looking primarily at year-over-year data, along with several market numbers, which aren't revised, the revisions aren't as critical as they would be if we were looking only at month-over-month changes for indicators sans financial indexes. In any case, I'll have a post about vintage data soon.
Posted by: JP at December 24, 2012 9:58 PM
Impressive. However, could you test your model on vintage data?
Posted by: David at December 24, 2012 1:25 PM