"The whole idea that the stock market reflects fundamentals is, I think, wrong. It really reflects psychology. The aggregate stock market reflects psychology more than fundamentals." - Yale University Economist, 2013 Nobel Prize Winner in Economic Sciences, Robert Shiller
Herd mentality is a human attribute that has been prevalent since people began forming tribes and certainly since the early days of commodity trading and the stock markets. It describes the tendency for individuals to adopt the behaviors or to follow the trends (rational or irrational) of a larger group. Individually, however, most people would not necessarily make the same choice. Herd behavior exists for a couple of main reasons, the first is due to the social pressure of conformity and the need for acceptance. The other is the common rationale that it's unlikely that a large group could be wrong. Everyone else must know something I don't!
As it pertains to investing, human emotions are strongly tied to economic and corporate news. When the news is positive optimism follows and when it's negative anxiety sets in. The stock market is reflective of these emotions, and as Benjamin Graham once said is a voting machine in the short term. Our research supports the fact that human behavior and patterns repeat over the long term as well.
The S&P 500 is down 3% over the past two trading days. The reaction appears to be due to an Emerging Markets sell-off, concerns re-surfacing regarding the Fed's bond tapering and an economic slowdown in China. "Sudden fears about emerging markets and also potential capital shortfalls for some European banks are rattling investors. People have been a bit complacent lately, so it's quite logical to get a correction," said David Thebault, head of quantitative sales trading, at Global Equities (Reuters this morning). It's hard to argue against the fact that anxiety and even some fear came to the forefront in the U.S. markets, especially in Friday's session.
Human emotion is always driving the stock market, but at times like this when it is playing an even more significant role in price movements, having a factor for investor behavior in the investment mosaic is critical. Knowing that investors react to the same trends consistently is only half the battle. Extracting that accurately and rapidly to understand how they are likely to react going forward is what we're all after.
We started out this week intending to answer some simple questions for our readers. When have we seen the S&P drop like this over two days and then what typically happens next? When have we seen the S&P start the year like this, and then what happens next? As I was running these scenarios through our software, it made me realize what makes our technology, content based search BY EXAMPLE, so powerful and precise. First, let me answer these two burning questions:
- When has the S&P dropped like this over a two day period, and what typically happens next?
To no surprise, it's happened a lot actually. Per the charts below, you'll see 8 of the top match results in the S&P 500 over the past 35 years. Notice that this is not a straight line down over two days that we're looking for, but rather one day that is down quite a bit, and the second day that has an accelerated sell off. We found 278 instances of this two day pattern with near perfect similarity (correlation score of .998 and higher) and in the next 10 days historically, the S&P is up 62% of the time and the average return of all 278 historical instances is .79% for the two weeks.
(Source: EidoSearch Data 1/27/14)
2. When has the S&P started the year like this before, and what typically happens in the next month?
Not once in the last 35 years actually, at least with high similarity, has the S&P looked like this at the start of a year. So, we decided to look for the three week start to this year (Jan 2nd to Jan 24th), and look at all three week periods in the S&P 500 since 1980 to see if we've seen this trading pattern before. We have. Per the chart below, 14 times, and in the next 1 month historically, the S&P is up 13 out of the 14 times and the average return of all 14 instances is 2.58%.
(Source: EidoSearch Data 1/27/14)
The two examples above not only highlight the value in understanding expected investor behavior, but also the power of content based search and specifically search by example. The S&P was down 3% the last two days of last week, but it wasn't a straight line. Investors traded the market down on Thursday, and then fear set in on Friday and the sell-off accelerated. This is very different than a straight line down for two days. Formulas to answer these questions, or technologies like NLP (Natural Language Processing) do not answer these questions as precisely because they are not as effective at capturing the criteria as specifically. As displayed above in the charts, EidoSearch uses search by example, the results of which are statistically driven with tremendous accuracy within the concrete parameters of what you're searching for.
The thing is, we as financial software guys don't know exactly know the right questions to ask. For each investor, whether they are looking to profit or hedge risk in the short term or long term, they are going to be interested/concerned with different questions and outcomes. Content based search, by example, accommodates this.