Mass Manipulation, Trading, and the Technology Polarizing Our Societies

"If we understand the mechanism and motives of the group mind, is it not possible to control and regiment the masses according to our will without their knowing about it? The recent practice of propaganda has proved that it is possible, at least up to a certain point and within certain limits."
~ Bernays, Edward. "Propaganda" (1928).

A nephew of Sigmund Freud - Edward Bernays (1891−1995) - pioneered the psychological manipulation of mass behavior through propaganda from the 1920s through the 1940s.  We discuss his ideas in today's newsletter because we've been pleased by the results we've achieved in live trading based on similar assumptions (SEE DISCLAIMERS BELOW).  The trading is entirely systematic, and it appears from our results (so far) that the news must be swaying mass behavior in ways that are difficult for human intelligence to exploit, but much easier for modern machine learning algorithms to take advantage of.  Today's newsletter describes Edward Bernays' insights and their modern application in marketing and investing, the social dangers of such work, and the now-public results of our own quantitative trading account.

Profiting from Crowd Psychology in Markets

"This is an age of mass production. In the mass production of materials a broad technique has been developed and applied to their distribution. In this age, too, there must be a technique for the mass distribution of ideas." 
~ Edward Bernays, "Manipulating Public Opinion" (1928)

Since we stopped trading our hedge fund (MarketPsy Long-Short Fund LP, 2008 through 2010), we decided not to relaunch trading until we found a consistent and easy-to-understand trading strategy.  After several years of research, we think we've identified a solid strategy.  PLEASE SEE THE DISCLAIMERS AT THE END OF THIS NEWSLETTER.  We believe that this strategy takes advantage of mass behavior in response to news.

The following equity curve depicts the growth of $1 in our one active trading account.  It earned a bit more than a 28% gross return since launch on February 22, 2016.  The daily gross returns of the account are plotted in the image below, assuming a starting amount of $1.

Using our media sentiment data (the Thomson Reuters MarketPsych Indices - TRMI), machine learning algorithms search for patterns in media tones and topics that are predictive the following day's price direction.  The machine learning algorithms then make probabilistic estimates of the following day's price change in the S&P 500.  We can trade on these predictions using SPY (an ETF) or S&P 500 emini futures.

Daily trading in the SPY started on February 22, 2016 with 1x leverage which was increased to 2x leverage on March 14, 2016.  The net monthly percentage returns of this strategy since it launched, assuming that 1% management and 20% performance fees were charged, is visible in the table below.



Please contact us if you are a Qualified Investor and would like more information about managed accounts utilizing this strategy.  SEE DISCLAIMERS BELOW.

Machine intelligence seems to make better forecasts of collective human behavior - manifest in market prices - than most individual humans.  But before we jump into the machine learning behind such results, we'll examine Bernays thoughts on patterns in (and the manipulation of) mass behavior.

Engineering Consent

"We are governed, our minds are molded, our tastes formed, our ideas suggested, largely by men we have never heard of. This is a logical result of the way in which our democratic society is organized. Vast numbers of human beings must cooperate in this manner if they are to live together as a smoothly functioning society."
~ Bernays, Edward. "Propaganda" (1928).

Edward Bernays developed news campaigns to alter the collective beliefs and behaviors of Americans and, in an industrial twist, to improve product sales.  The 2002 BBC documentary, The Century of the Self, describes Bernays methods.  Bernays was named one of the 100 most influential Americans of the 20th century by Life magazine.  Bernays called the scientific technique of public opinion-molding the engineering of consent.  Bernays argued that the manipulation of public opinion was a necessary part of democracy.  Yet that same belief in the benefits of manipulation eventually spread into his work for corporations, changing consumer behavior to the benefit of his industrial clients.

In one campaign, which he later regretted, Bernays helped the smoking industry overcome the social taboo of women smoking in public. When he launched this campaign, women were only allowed to smoke in designated areas, or not at all, and were subject to arrest if violating this law.  At the 1929 Easter parade in New York City, Bernays staged models holding lit Lucky Strike cigarettes - "Torches of Freedom" - in order to associate women smoking as a socially desirable activity (and increase sales for his client Lucky Strike). News outlets picked up on the story and provided free marketing by disseminating the story widely.  Bernays portrayed this event as news, which increased its influence.  He demonstrated to his client companies that the news, not advertising, was the best medium to carry their message to an unsuspecting public.  Because the news typically reported on the behavior of celebrities and leaders, Bernays believed that "If you can influence the leaders, either with or without their conscious cooperation, you automatically influence the group which they sway." 

Since Bernays thoughts were written in 1928, advancements in computing and statistics have refined the craft of propaganda from loosely anecdotal to precision targeting of individuals.  Refinements of Bernays techniques are now applied in behavioral targeting on platforms such as Google and Facebook, where ads (and news) are presented based on users' past preferences and that of similar others'.  Such algorithms are not necessarily benign - they may also lead like-minded groups to focus their attention on the same topics and the same dangers, as evidenced in this New York Times article comparing the content of the Red versus Blue Facebook news feeds.  Bernays noted that "The conscious and intelligent manipulation of the organized habits and opinions of the masses is an important element in democratic society. Those who manipulate this unseen mechanism of society constitute an invisible government which is the true ruling power of our country."

It is possible that such behavioral algorithms, as they parry with our own click behaviors and unconscious preferences, are co-opting the power of mass manipulation from the political leaders who formerly held it (witness how the Republican party establishment lost control of their party's primary).  The algorithms, through the news they disseminate, are directing attention, provoking outrage, and accelerating the polarization of political leanings and belief.  Targeted advertising and news delivery creates an echo-chamber.  News that confirms ones assumptions - because most humans tend to seek confirmation rather than cognitive dissonance - is recycled into one's news feed.  Gradually such algorithms fracture social mores and beliefs into, for example, the distinct belief bubbles of Republicans and Democrats.  

We have a conflict of interest in all of this.  Many people are swayed by the topics and tones of the news media, and we are able to quantify those factors.  With this quantified media sentiment data, we construct predictive models.  Using such predictive models, our trading company (MarketPsy Capital) takes advantage of news-driven crowd behavior.  The strength and persistence of the behavioral patterns we've found is both profitable for our trading but also - in fact - disturbing for what it says about the frequency with which the mass of investors is predictably swayed in their over and underreaction to the ebb and flow of information.  

When Are Machines Smarter than Humans?

Most investors - especially contrarians - are aware of simple sentiment-driven rules for trading such as “Buy when investors are pessimistic, sell when they are optimistic" (courtesy of Ben Graham).  Such contrarian sayings make intuitive sense to most traders, and they are frequently re-iterated by top investors (e.g., in this New York Times article Warren Buffett recommended to buy U.S. stocks in October 2008).  However, while such rules are typically correct more than 50% of the time, timing is a challenge (e.g., Warren Buffett could have similarly written a bullish article in January 2008 after Countrywide failed).  

Sophisticated machine learning algorithms can help us improve the timing and better understand the nuances of such rules.  Modern algorithms can judge the precise probabilities of hundreds of simple rules and consolidate multiple simple predictions into a single daily forecast.  

At MarketPsych our deep understanding of the TRMI, our research into investor psychology, and our expertise in big data and machine learning helped us to produce algorithms that decipher sentimental influences on the markets.  While applying these algorithms we observed that the more emotionally charged a situation is (i.e. more attention-capturing newsflow), the better our algorithms are at prediction. The market correction in August 2015 and Brexit were both high performance periods for the algorithms.  When collective attention is focused on the news, then that information is more likely to influence behavior.  News-rich environments and price volatility (which also captures attention) thus appear more beneficial to the trading model's results.  SEE DISCLAIMERS BELOW.

We've not alone in this endeavor.  When I speak to large quant funds about our TRMI data, I hear similar comments and requests - most want to know what investors are paying attention to (reading, clicking on), and how they feel about such content (their emotional perspective).  It is becoming widely accepted that those factors are the keys to predicting human (and market) behavior.  Yet such findings also raise questions about human free will.  If investors in such a competitive arena as the financial markets are predictably influenced by the newsflow, then in an environment with lesser personal stakes such as the political arena, Bernays was likely accurate in his claims that the behavior of the crowd is easily swayed by the news media.

Housekeeping and Closing

"...In almost every act of our daily lives, whether in the sphere of politics or business, in our social conduct or our ethical thinking, we are dominated by the relatively small number of persons...who understand the mental processes and social patterns of the masses. It is they who pull the wires which control the public mind."
~ Bernays, Edward. "Propaganda" (1928).

Bernays life work was controversial.  No one wants to believe that their own behavior is manipulated by a few elite masterminds.  Furthermore, Bernays' belief that the "manipulation of the masses" is a natural and necessary feature of a democratic society was discredited by the fascist rise to power in Germany which demonstrated that propaganda could be used to subvert democracy.

Many people feel that Hillary Clinton (and Jeb Bush's failed campaign) were establishment-engineered to sway the masses in their favor.  In part, the appeal of populists including Donald Trump is that they were not designed and presented for mass consumption.  Perhaps a vote for Trump is a rebellion against the establishment-deployed formulae that seeks to manipulate behavior.  But then perhaps it's not so simple, political populist's portrayal in the media may be on the cutting edge of such algorithms - using controversy to capture attention, generate clicks, and manipulate the human mind's anxiety to his electoral advantage while persuading voters that a vote for them is a rebellious act that will secure their liberty.  The reality is that all political candidates are trying to sway voting behavior, and they are all trying to manipulate the masses, some consciously and deliberately, others intuitively.  The news media is their tool.

In a letter to President Franklin D. Roosevelt, Supreme Court Justice Felix Frankfurter described Bernays and one of his contemporaries as "professional poisoners of the public mind, exploiters of foolishness, fanaticism and self-interest".  And in his 1965 autobiography, Bernays recalls a dinner at his home in 1933 where he learned that the Nazi leader "Goebbels ... was using my book Crystallizing Public Opinion as a basis for his destructive campaign against the Jews of Germany. This shocked me. ... Obviously the attack on the Jews of Germany was no emotional outburst of the Nazis, but a deliberate, planned campaign."

As seen with the Nazis, manipulation of the public mind is a very dangerous science.  In recent years this science has been delegated to amoral algorithms.  Those who understand the nuances of such algorithms are able to find significant business (and political) opportunities.  Yet too often these opportunities exist at the expense of social (and market) stability.  We're keeping an eye on developments in this space, as the implications for our societies are profound and sometimes disturbing.

We discuss insights from sentiment analysis of the financial herd in our new book, “Trading on Sentiment: The Power of Minds Over Markets” (Wiley, 2016).  

We love to chat with our readers about their experience with psychology in the markets.  Please send us feedback on what you'd like to hear more about in this area.  

If you represent an institution, please contact us if you'd like to see into the mind of the market using our Thomson Reuters MarketPsych Indices to monitor real-time market psychology and macroeconomic trends for 30 currencies, 50 commodities, 130 countries, 50 equity sectors and indexes, and 9,000 global equities extracted in real-time from millions of social and news media articles daily.

Monitoring the world's psychological vital signs,
Richard Peterson M.D. and the MarketPsych Team


DISCLAIMERS

Dr. Richard Peterson is the sole owner and sole investment adviser representative of MarketPsy Capital, LLC.  MarketPsy Capital, LLC is registered as an investment adviser with the State of California.  No other person or entity, affiliated or otherwise, is qualified or registered to provide investment advisory services to potential clients.  

An investment using our proprietary trading strategy is speculative and is subject to a risk of loss, including a risk of loss of principal.  No assurance can be given that the strategy will achieve its objective or that an investor will receive a return of all or part of its investment.

The performance results discussed herein do not represent the performance of the strategy or of any account managed by MarketPsy Capital, LLC (the “Adviser”).  Rather, these data represent hypothetical results of the Adviser’s strategy and are included for informational purposes only.  The results shown reflect the deduction of: (i) an annual management fee of 1%, charged monthly; (ii) an annual performance allocation of 20% of net profits, subject to a “high water mark,” and (iii) estimated transaction costs.  Fees and expenses were applied retroactively and do not reflect actual fees or expenses deducted from a client’s account. 

Hypothetical and back-tested performance results have inherent limitations, some of which are described below.  Back-tested returns do not represent the results of actual trading and are calculated through the retroactive application of the Adviser’s model portfolio and strategy configurations, designed with the benefit of hindsight.  Since back-tested results do not represent actual trading, they may not reflect the impact that material economic and market factors might have had on the decision-making processes of the Adviser, if the Adviser was actually managing client assets during the back-test period.

NO REPRESENTATION IS BEING MADE THAT THE ADVISER’S STRATEGY WILL OR IS LIKELY TO ACHIEVE PROFITS OR LOSSES SIMILAR TO THOSE SHOWN.  IN FACT, THERE ARE FREQUENTLY SIGNIFICANT MATERIAL DIFFERENCES BETWEEN BACK-TESTED PERFORMANCE RESULTS AND PERFORMANCE RESULTS SUBSEQUENTLY ACHIEVED BY FOLLOWING A PARTICULAR STRATEGY.

In addition, back-tested performance does not involve financial risk, and no hypothetical trading record can completely account for the impact of financial risk in actual trading.  For example, the ability to withstand losses or adhere to a particular trading program in spite of trading losses could have material adverse effects on trading results.  There are numerous other factors related to the markets and to the implementation of specific investment programs that cannot be fully accounted for in the preparation of back-tested performance results and all of which can adversely affect actual trading results.

Results are compared to the performance of the S&P 500 Total Return Index (the “SPX”) for informational purposes only.  The Adviser’s investment strategy does not mirror the SPX and the volatility of the Adviser’s investment strategy may be materially different than that of the SPX.  The securities or other instruments included in the SPX are not necessarily included in the Adviser’s investment strategy and criteria for inclusion in the SPX are different than those for investments in the investment strategy by client. 

The performance of the SPX and the trade data used to create the back-tested returns contained herein were obtained from published sources believed to be reliable, but which are not warranted as to accuracy or completeness.  Unless noted otherwise, returns presented herein do not reflect fees or transaction costs, but do reflect net dividends, if any.

This material contains certain forward-looking statements and projections regarding the future performance and asset allocation of the investment strategy.  These projections are included for illustrative purposes only, are inherently speculative as they relate to future events, and may not be realized as described.  These forward-looking statements will not be updated in future.

The performance results shown were executed through the account of an affiliated entity of the Adviser and is not necessarily indicative of typical investment positions using the proprietary trading strategy. It should not be assumed that recommendations made in the future will be profitable or will appreciate in a manner similar to the results shown herein.

PAST PERFORMANCE, INCLUDING HYPOTHETICAL PERFORMANCE, IS NOT NECESSARILY INDICATIVE OF FUTURE RESULTS.