Wednesday, June 21, 2017

A market breadth model that works

Mid-week market update: Technical analysts monitor market breadth, as the theory goes, to see the underlying tone of the market. If the major market averages are rising, but breadth indicators are not confirming the advance, this can be described as the generals leading the charge, but the troops are not following. Such negative divergences are signs of technical weakness that may be a precursor to future market weakness.

That’s the theory.

I have been highly skeptical of breadth indicators as a technical analysis tool because breadth divergences can take a long time for the market to resolve, if at all. The chart below shows the SPX, the “generals”, along with several indicators of how the “troops” are behaving, namely the mid-cap and small cap indices, as well as the the NYSE Advance-Decline Line.

The behaviour of these indicators during and after the Tech Bubble was problematical. During the advance from 1998 to 2000, the NYSE A-D Line fell and flashed a negative divergence sell signal. At the same time, both the mid- and small-cap indices continued to rise and confirmed the market advance. Which divergence should investors believe?

After the market peaked in 2000, the mid and small caps traded sideways, which represented a non-confirmation of the bear market. Investors who followed this buy signal would have seen significant drawdowns during this bear market.

To be sure, there were periods when breadth indicators worked. The NYSE A-D Line traded sideways during the market decline of 2011, which was a buy signal. As well, it correctly warned of market weakness in 2015.

Putting it all together, the report card for these indicators can best be described as inconsistent and the timing of the signal uncertain. At a minimum, no investor would use breadth indicators in a trading system.

I believe that I discovered a solution to the problems that technicians struggle with when they use breadth indicators.

The full post can be found at our new site here.

No comments: