Chart:  Coppock Curve

On Thursday, Albert Edwards from Societe Generale came out with a research report touching on a technical indicator called the Coppock Curve.  Apparently that measure just carved out a "killer wave".

Mr. Edwards's note was making the rounds on Wall Street all day, particularly because the implication was that the S&P declined an average of -40% during the next 20 months after one of these developments.

Before anyone can dismiss this as a curve-fitted example of confirmation bias, we should at least objectively look at its history and see if there may be something to it.

The Coppock Curve is calculated by summing up a 14-month and 11-month rate-of-change in an index, then taking a 10-month weighted moving average of that sum.  The "killer wave" occurs when the Coppock Curve reaches a high point, then declines but doesn't dip below the zero line.  It rises again, then begins to fall yet again, forming a kind of double top.

According to the report, there have been 8 of these waves in the past 83 years.  If we clearly define the parameters, then we're forced to use the signals as they would have occurred at the time.  Often, that will change the outcome from what we get when we just eyeball a chart.

In this case, there wasn't too much of a difference, at least in terms of the number of occurrences.  I came up with 9 since 1928, which are highlighted on the chart below with red arrows.

The table below shows the performance in the S&P 500 after the Coppock Curve first turned down from a double top.  Because our calculations and the identification of the waves may be slightly different, I'm not sure if the signals are the exact same ones that Mr. Edwards highlighted in his report.

If the dates are mostly the same, then I can't confirm the dire prediction from the "killer wave".  The S&P's performance going forward was sub-par, certainly, but I don't see the "average -40% decline during the next 20 months".

When we look at the next 20 months, the maximum decline in the S&P was a median -13.9%.  That's definitely higher than what we see during any random 20-month stretch, but way off from what was mentioned in the report.

Also, the maximum gain during the next 20 months was actually quite a bit higher than that, and in fact it was even higher than random.  The last six "killer wave" signals, in fact, led to quite bullish outcomes, excepting the one from November 2007.

Overall, I don't see much about the indicator that would have me overly worried about stocks' long-term prospects.