S&P 500 And Its 10-Day Moving Average

The momentum in the major equity averages over the past month has been remarkable, and over the past couple of weeks in particular.

A couple of weeks ago, we started seeing quite a few studies that were very consistent in preceding weakening prices in the past, but those have all failed so far and are close to expiring - most of them had a time frame of 2-3 weeks.

During this span, the S&P 500 has held above its 10-day moving average almost the whole time, and at no point has it closed more than 0.5% below that average.  Yesterday marked the 38th straight day its has held that line.

Such a feat has been rare over the past decade.  We did see it just this past April, but other than that there have only been four other such streaks since 2000.

After the streak hit 38 days in April, the S&P managed to continue higher and didn't start to top out for another couple of weeks.  Let's go back to 1928 and look for every other time the S&P managed to go 38 straight days without closing more than 0.5% below its 10-day average, and see how it performed going forward.

The "Days 'Til High" column shows how many trading days it took before the S&P formed an intermediate-term peak with a loss of at least -5.0%.

Out of the 44 historical occurrences, 7 of them ended right at 38 days, and 16 ended before the streak reached two months.  15 went on longer than 50 days, and 5 longer than three months.  3 of them lasted an amazing 80 days or more (1/16/43 lasted 80 days, 2/17/64 lasted 84 days and 9/21/65 lasted 80 days).

So in just under half the cases, the streak was very close to ending.  But it took a median of 52 trading days - about two and a half months - before the S&P actually formed a peak of any significance.  There is an exceptionally wide variance in those figures, though.  The longest before a top was 330 trading days (almost a year and a half!) while 6 of them topped out within two weeks.

During the past decade, the five instances were a bit more uniform, and it took a median of 19 trading days before the S&P peaked.  Three of the five peaked out before a month went by.

Overall, the S&P's raw performance going forward was OK, but nothing spectacular.  In fact, it was almost exactly in line with random across all of the time frames.  The momentum required for such a remarkable streak didn't seem to have any predictive power over how the index did going forward.

Still, the fact that we've rolled right over so many bearish studies from the past couple of weeks should be respected, as there are apparently forces at work that are overpowering what we normally see.  Perhaps that will change after the election and FOMC meeting next week, but that's just idle speculation.