Market Breadth Primer: Hindenburg Omen
By Lawrence
Hindenburg Omen is a market breadth driven pattern. It is made famous when it has successfully forewarned several times in history on serious stock market sudden corrections. I am going to explain the concept with common sense so that you can make informed decision when you see this term showing up in the news.
Definition of Hindenburg Omen
I am not going to waste time on writing a full article on the subject. You can find a good description of the naming, history and criteria that defines the Hindenburg Omen from Wikipedia. My focus here is to clarify several important points about the concept.
Reasoning Behind
Any predictive forecast model designed to identify potential short term volatility (i.e. wild swings) are measuring some sort of excessive effort made in pushing the market in one direction (in terms of price movement) while other factors are not supporting the effort. Hindenburg Omen focus specifically on potential sudden drop or crash like move in the stock market. Thus it only measures excessive efforts made to push the stock market higher while the overall underlying conditions are not supporting it.
The underlying conditions Hindenburg Omen use are several market breadth statistics. By comparing the number of components demonstrating bullish behaviour against the number of components showing bearish signs, one can tell if the current strength in the up move of the index is justified. As the model is built from observing historical behaviour, the triggering threshold / event from the various statistics are discovered from history. Hence going forward, if the market breadth behaviour changes, it will affect the accuracy of the signal.
Think of Hindenburg Omen as the weather forecast warning that a potential thunderstorm is in the making. This will put you in the correct mindset to interpret the signal.
Like weather forecast, the signals get stronger and more accurate as more evidence present themselves. This, however, does not guarantee a sharp selloff in the market is coming. Using the thunderstorm metaphor, even if all the evidence like huge dark clouds covering the sky completely and that it is raining heavily already, there is still the chance of thunder and lightning not happening.
Great Track Record Coming From A Different Era
Ever since Quantitative Easing started back in 2009, majority of market breadth data has undergone significant changes. I have explained the changes in Advance / Decline Issues already in a separate article. These changes make the market breadth statistics more volatile. The mean reversion behaviour shortened from weeks to days at times.
With changes like this, Hindenburg Omen’s original design can no longer capture majority of the sudden moves in the stock market. The market breadth data it depends on are not sensitive enough. Hindenburg Omen has turned into a warning sign that works under specific conditions only. It is no longer very effective as a general market top warning tool. The main weaknesses is that it cannot produce a precise time window on the potential selloff for market timing purpose.
Going Forward More Sophisticated Analysis Will Do A Better Job
For weather forecasting, technological advancement continues to march forward while nature, the global weather system itself, is still pretty much the same dynamic system we have from hundreds of years ago. The improvement we made to our models on weather behaviour gets us closer to better forecasting ability, no matter how slow the progress has been.
For stock market crash warning system, it does not have the same luxury like our work in weather forecasting. The stock market has evolved significantly beyond what the old style broad based market breadth can offer to detect subtle changes within. The stock market grows bigger, more complex and more diversified everyday. Generic broad based market breadth paints a broad stroke on the stock market that is too broad to be useful.
A good example is the Russell 2000 Index. This index does not correlate much with the 3 major indices (Dow Jones Industrial Average, S&P 500 and Nasdaq 100). Most of the components within Russell 2000 are not even listed in NYSE. It is obvious Hindenburg Omen is of very limited value in forecasting potential market crash in Russell 2000.
Index specific breadth analysis will likely improve the signal quality and timeliness even if a similar framework like the Hindenburg Omen is used. The advantage of such models is that you can extract value from the work done directly. As a start, the predictive ability of the model can be translated into market timing model quite easily as you can target specific tradable market indices like the S&P 500. The disadvantage, of course, is that such targeted signal do not give you the broad market / overall financial system warning like the Hindenburg Omen.
A direction to look for better solution is to look at the transaction level behaviour to look for signs of instability. Work done by Didier Sornette shows us that there is indeed new way in identifying potential financial crisis on a global scale. The catch is that most stock market participants are not scientists thus it will be difficult to make progress in this direction unless more scientists are willing to work on this problem.
