The Unimportance of the 50-day Moving Average and Other Long-term Observations

For a sixty-year period between 1940 and 2000, there was one market timing
strategy that trumped everything, and it was the ultimate in simplicity. Buy
when the market closes higher and shift to the short side when the market
closes lower. As incredible as it sounds, this simplistic approach racked up
astounding profits from 1940to 2000. So astounding you wouldn't believe the
figures. Unfortunately, times have changed, and in recent years this one-day
approach has been completely turned on its head, severely underperforming a
buy&hold approach. Testing the one-day system only since 2000 rather than 1940
reveals it would be down over 75%. The phenomenal results in the sixty years
prior to 2000 illustrates how radically different the markets are now. Prior
to 2000, the S&P exhibited a definitive tendency to trend in one direction for
a sustained period of time, allowing a profit to be made just by latching on
to whichever direction the market happened to be moving. Since 2000, that
entire concept has been turned on its head, as buying the market (on a closing
basis) when it starts rallying and shorting when it starts to fall has been
disastrous. It's clear that the market's tendency to move in one direction for
more than a day or two has been completely wiped out. This tends to suggest
that we recently (from a long-term sense) entered into a completely different
environment in which an up day should generally be looked upon as a sell
signal and a down day a buy signal. Indeed, running the same one-day strategy
test since 2000 with the rules reversed (sell when the S&P closes higher, buy
when the S&P closes lower) has produced a profit of 250% since 2000 vs. a 7%
loss for buy&hold. While I'm not advocating the implementation of such a
strategy, it does provide insight regarding the overall market environment,
particularly for short-term traders.

I'm surprised by the number of columns exclaiming the importance of the S&P500
(and other indexes) closing above its 50-day moving average. As if this were
some sort of 'all clear' for further gains. I've discussed this in the past,
but I think it bears repeating - historical testing reveals that the break of
a 50-day average (or most averages for that matter) means very little. Anyone
drawing broad conclusions from this development might want to take another
look at the historical data.

Prior to 1990, this used to be a viable strategy. Pull up this equity chart of
going long the S&P on an upside break of the 50-day average and selling short
on a downside break of the average. In the fifty years from 1940 to 1990, this
strategy racked up impressive gains. But like other moving average-based
systems, things really fell apart starting around 1990. Since then, trading
penetrations of the 50-day moving average has produced ZERO profit. That's
eighteen years, and a lot of trading, with nothing to show for it.

If you like the idea of this concept, studies show the most consistent returns
with very long-term moving averages. Replace the 50-day average with a 200-day
moving average, make it exponential rather than simple and you've got a moving
average-based strategy that actually works. While historical testing
invariably turns up certain averages that perform better, the 200-day often
ranks right near the top of the list. While it stumbled badly in 2000, as did
all averages tested, it's made back that loss in recent years (see equity
chart.) One could also eliminate short sales from the approach with little
change in performance. Trading the 200-day penetration from a long-only
perspective has proven to be a solid combination in terms of performance and
limited drawdowns. While you're only winning approximately 33% of the time,
the average winner has been nearly ten times the average loser, leading to
consistent gains over the last seventy years (see equity chart). This strategy
has returned more than twice the performance of buy & hold since 1930.

Keep in mind that exponential moving averages perform far better than simple
moving averages, regardless of the type of strategy implemented. For the
purposes of this column, we examined moving average crossover strategies
(which don't work), strategies involving the slope of the moving average
(again, doesn't work at all) as well as moving average penetrations. In
virtually every test, exponential averages, which give more weight to recent
activity, proved to be most profitable. As of Tuesday's close, the 200-day
exponential moving average for the S&P500 stood at 1419.

Categories : Archived EOD Commentary, End-of-Day Commentary

 Print This Post