Chart:  S&P 500 Positive Earnings Surprises

A subscriber passed along a note from the Hussman Funds that discussed various measures related to earnings.

One of the studies looked at a Bloomberg calculation for the percentage of firms in the S&P 500 whose earnings report beat an average of analysts' estimates.  The article suggested that buying the S&P when the 12-month average crossed over the 24-month average, and selling when the opposite occurred, was a good timing signal.

And for the most part, it was.  But it could be a bit better.

Using the Hussman system, since 1993 you would have generated +1298 points in the S&P during the buy signals and you would have avoided -411 points in losses during the sell signals.

If we tweak the parameters a bit, though, it gets better.  By using 6-month and 36-month average, you would have generated +1521 points during the buys and avoided -634 points during the sells.

Here are the buy signals using that system:

Pretty impressive.  The "Max Loss" column shows the worst loss you would have endured during the signal (using monthly closing prices only), and five of them didn't have any loss at all.

It took an average of 9 months for the signals to reverse, though there was a wide disparity among them.  All of them showed a gain of at least +12% sometime during the signal, and on average the S&P was a whopping +21.3% higher by the time a sell signal hit.

And here are the sells:

This isn't nearly as pretty.  On average, the S&P dropped more than -3% at its worst point during the sells, with a couple of them extremely large.  The maximum gains were very modest, with all but one under +10%.  By the time they reversed to a buy signal, only one showed an ending gain more than +5%.

What's the point of showing this now?

Both systems just gave sell signals.