Being of Prussian descent, I’ll admit to a certain fascination for all things Teutonic, especially in the field of technical analysis. One of these is the Hindenburg Omen (HO), a fairly obscure indicator named after the German zeppelin that crashed in a fiery explosion in May 1937. The urban myth surrounding the HO is that it accurately predicts market crashes.
The creation of the HO dates to the 1990s and is generally credited to a blind former physics teacher named Jim Miekka who died in 2014. When the HO flashes a signal it is often front page news in major financial newspapers.
Miekka claimed in a 2011 interview that the Omen has appeared ahead of every market crash in the U.S. from 1987 but he also admitted that not every Omen signal means the market will crash. As he put it, “not every tropical storm turns into a hurricane.”
The theory behind the Omen is that at major turning points, divergences begin to develop within the market. Generally, in a strongly uptrending market, one would expect to see many more new 52 week highs than lows and in a strongly downtrending market more new 52 week lows than highs. The Omen looks for periods when large numbers of both 52 week highs and lows are occurring at the same time.
The traditional definition of the Hindenburg Omen contains four criteria listed below, and is based on NYSE data. Over the years various authors have deviated from the original strict definition by adding or subtracting criteria depending on their own research.
1. The daily number of new NYSE 52 week highs and the daily number of new NYSE 52 week lows must both be greater than 2.2 % of total advancing or declining issues traded that day. Some analysts now use 2.8%. This criterion shows the divergence that is occurring within the market.
2. The NYSE 10 week moving average is rising. This criterion tells us that the market has generally been rising in the intermediate term. More recently, technicians have altered this to require instead that the NYSE is simply trading at a higher level than it was fifty days ago.
3. A technical indicator called the McClellan Oscillator is negative on the same day. This indicator is a market breadth indicator based on advance/decline data. It is calculated by taking the difference between the 19 day and 39 day exponential moving averages of advancing minus declining issues. A negative reading implies that the number of declining issues, on average, is growing faster than the number of advancers, a potentially bearish condition.
4. The new 52 week highs can’t number more than twice the new 52 week lows, although it is fine if the 52 week lows are more than twice the 52 week highs.
A HO signal is considered valid for the next 30 (some say 40) day period. Some authors feel that one signal occurring in isolation is not as predictive as seeing multiple signals within a thirty day period.
The next obvious question, of course, is whether the HO is at all accurate or even worth considering at all. The answer, it seems, depends on who you ask.
Jason Goepfert of Sundial Capital Research examined data from 1965 to 2013. He looked at the performance of the S& and P 500 at various intervals after the HO triggered at least 5 times during a two week period. He found that three months after the signals, the S and P was up only two times out of eleven, with the average decline being 2.1%. Heeding the Omen saved you some pain both in February 1980 and November 2007 when the market decline over the subsequent three months was 10.5 % and 9.8% respectively.
Tom McClellan, the son of the couple who created the McClellan Oscillator, wrote in 2013 that the Hindenburg Omen “has not failed to fire off a signal ahead of all of the major price declines of the last thirty years.” He also pointed out, though, that not every signal leads to a decline.
Analyst Darryl Guppy, on the other hand, in an article published on CNBC in December 2014, warned about “confusing co-incidence with correlation.” He went on to point out that the market was up more than 20% over the nine months after the Omen triggered in August 2010 and up over 48% after the June 2013 signal. The Omen also flashed in mid-December 2014 and three months later the market was making new highs.
So where does all this leave us? Perhaps it is better to think of the Hindenburg Omen as a warning, rather than a specific trading signal. Consider using the Omen to put other technical signals in context.
Here are a couple of examples of what I mean. One market timing method is to get more bearish when the 50 day moving average crosses below the 200 day. Some traders sell a portion of their holdings when this occurs. If this signal occurs after a cluster of HO signals, you might sell a higher percentage of your holdings than if there was no Omen.
Another market timing method is to count distribution days. A distribution day is defined as a more than a 0.2% decline of a major index on greater volume than the previous day.
One guideline some traders use is to get more bearish if there are between six and eight distribution days over a 25 trading day period. Maybe you might start taking action after the sixth day as opposed to the eighth if they occurred after an HO signal.
The bottom line, I think, is that it is a wiser strategy to use Hindenburg Omen to provide a contextual background for further analysis rather than using it as a trading strategy. Seeing clouds in the sky on a slate gray German morning doesn’t mean it is going to rain, but it might pay to bring your umbrella.
Lee Bohl can be contacted at Charles Schwab UK Ltd