A few months ago my wife and I moved our kids from California to Bali, Indonesia for the year. We've been surprised by the strength and frequency of the recent earthquakes. During one earthquake I watched waves of water slosh out of our pool. During another I paused a business call as my roof groaned and light fixtures jangled. During a few late night earthquakes I’ve been shaken awake in a daze wondering, “Why is my wife jumping on the bed so hard?” Is she doing jumping jacks?!" (She had similarly foggy and illogical thoughts about me).
We have chosen to live in the “Ring of Fire” where earthquakes and volcanic activity are a fact of life – part of what makes the landscape beautiful, lush, and dramatic. And despite the instability here, we feel very lucky – residents of the Indonesian islands Lombok and Sulawesi have lost much more than pool water.
Financial markets have their own quakes and tremors, including one this week. Intellectually, we know such events will happen. But after each many people wonder “Could this be the Big One?” As a seismologist of the human psyche, in today's newsletter I feel some responsibility to explain what we know about predicting market downturns. Today’s newsletter explores market quakes: their psychological predictors, the potential this week's tremor was a precursor to the “Big One”, and a Balinese understanding of duality.
Human trading drives market prices. Usually humans are relatively predictable, buying on good news and selling on bad. Good news has positive sentiment, and bad news promotes negative sentiment. The balance of sentiment partially explains why some stocks go up, and others go down, in trends lasting up to a year.
Empirical Research, one of the oldest and most respected equity research firms, recently tested 8 media sentiment data sources. Their researchers found that our sentiment data provides consistent predictive power for stock prices. The sentiment from one month in the past can propel stock price returns out to nine months in the future (1).
In our own research we also see sentiment as predictive of stock prices. In the following study by our Head of Research CJ Liu, we selected the top 1000 stocks by media buzz in the U.S. over the prior 12 months. We ranked them by their average past year sentiment. The quintile (200 stocks) with the highest (most positive) sentiment were bought and held for one year. The selection, ranking, and portfolio construction was then repeated each month with 1/12 of the portfolio. The model assumes no transaction costs. In the image below the blue line represents the growth of $1 invested in the most positive 200 stocks in the media since 1999. The dotted green line is the S&P 500 stock index performance.
It appears that the most positive stocks have price momentum over yearly horizons. We see similar positive sentiment outperformance for stocks in markets of Australia, Canada, China, the Eurozone, India, Japan, Germany, South Africa, Switzerland, and the UK. (We have significant additional documentation and code available to readers with a deeper interest – please reach out if so).
While this finding is true for individual stocks on average, it can’t always be true - eventually sentiment turns, and when it does, prices follow the sentiment lower.
Calling a market top is not an easy thing to do, and it’s not all about market sentiment. Sentiment often does dip before market tops, and we’ve written about the phenomenon before, most recently for Bitcoin’s top and in Chapter 17 of “Trading on Sentiment” (Wiley, 2016) called “Timing Bubble Tops.” To call a top it helps to consider the feedback among price action, fundamentals, and sentiment jointly.
One group looking at the intersection of these factors is NN Investment Partners ($300 billion AUM), the asset management spin-off from ING Bank. NN uses MarketPsych’s sentiment data as a primary feed into their risk-off signal. Their use of MarketPsych’s sentiment data signaled de-risking before a correction in 2016. A dramatization from the Dutch financial media describes their use of sentiment data to derisk at a top. NN sold eqiuities following a rise in media Stress and Gloom:
“[W]ords that are associated with stress and gloom course through digital media.... The main strategist of the Dutch asset manager [NN Investment Partners], with € 187 billion under management, taps the brakes: he reduces investments in shares and raises more cash. Later that week the stress spreads: worldwide stock markets went down sharply...”
~Lenneke Arts & Jeroen Groot. February 28, 2016. Financial Daily (Dutch).
Media sentiment can be a self-fulfilling prophecy. In charts of stock index sentiment versus price, we see sentiment momentum, often in long tends, with reversal from extremes. In this view of 2018, we often see a decline in sentiment (increased negative sentiment) preceding downturns in price, as NN described:
Quantitative research demonstrates that de-risking based on such a signal would have led to historical outperformance of a long-only strategy.
There is a Wall Street saying that captures the relationship between bad news and downturns: “The first bad news is not the last bad news.” That’s usually said about individual companies, but it can be extended to stock indexes. While the timing of this indicator isn’t perfect, it is a useful cautionary sign.
So Where are We?
Sentiment is an imprecise measure, but when combined with other indicators like calendar effects, fundamentals, and price action, its value increases substantially.
Negative fundamental factors include the recent surge in interest rates. As we’ve said previously, higher interest rates will likely mark the end of the 10-year bull market, since they render risky assets less attractive and decrease economic activity. However, interest rates have been on a slow rise since 2015, and we don’t know what level of interest rates will lead to the psychological shift away from risky assets. Globally interest rates remain near historical lows. There is still a lot of liquidity sloshing around.
Technically, we’ve had a long bull market with higher highs all year, especially for tech stocks. Mathematicians have tools to flag when prices go parabolic near market tops (2). Such indicators found topping signals in bitcoin and other cryptocurrencies in January 2018 (3). The U.S. stock market accelerated through 2017 into January (and it was parabolic), and it accelerated through the summer and was parabolic again in September to October.
The market’s future is determined by feedbacks loops between interest rates, liquidity, earnings, prices, and sentiment. Interest rates are the highest since 2011, but liquidity is still flowing. If negative sentiment persists, it could reduce confidence and the availability of credit. Prices recently hit blow-off levels. On the one hand, there are enough interesting changes afoot, and so many moving parts, that it’s difficult to call a major cyclical top at this point. On the other hand, there is no clear good news to propel stock higher, so it does appear to be at least a short term top we've reached.
That's not the definitive answer most people want to hear like "Buy!" or "Sell!", but it is a realistic assessment of where we are in the cycle.
Living Beyond Duality
In Balinese spirituality there is an explicit acknowledgment that duality - good and evil, light and dark, happiness and sadness – are interwoven. No one of these states could exist without its opposite. The faded checkered sarong on this judge outside a cemetery represents this duality between opposites.
In Bali, those individuals who quickly and casually differentiate between good and evil, right and wrong, happiness and sorrow are considered to have a spiritual level within the most primitive layer of understanding. These people are easily swayed between two opposites and cannot see the big picture.
In Balinese Hinduism, the deeper someone's spiritual understanding, the lines between "black and white" blur. The more advanced spiritual practitioners can see that sorrow is part of happiness, find enlightenment in past mistakes, understand that nothing is truly bad, and can accept both sorrow and joy as God's gift (4).
In markets, as in life, there are periods of growth and expansion followed by periods of fear and decline. Both are necessary. Our ability to understand these as part of the whole, and to learn from them, gives us a more nuanced understanding of nature, markets, and life itself. And with such an awareness, we develop wisdom.
The above research is derived using our Thomson Reuters MarketPsych Indices (TRMI). The TRMI are deployed globally to monitor the human side of financial markets. The TRMI measure and deliver real-time market psychology and macroeconomic trends from thousands of news and social media sites. If you're an academic interested in data for research, please reach out for access. If you represent an institution, please contact us. The commercial Thomson Reuters MarketPsych Indices dataset includes granluar themes and sentiments for 45 currencies, 62 sovereign bonds and stock indexes, 12,000+ companies and stocks, 36 commodities, 187 countries, and 150+ cryptocurrencies.
Richard Peterson and the MarketPsych Team
- Cahan, Rochester; Yu Bai and Sungsoo Yang. ”Big Data: Harnessing News and Social Media to Improve our Timing.” Empirical Research Partners. August 21, 2018.
- Johansen, A., Ledoit, O. and D., Sornette. 2000. “Crashes as critical points.” International Journal of Theoretical and Applied Finance, 3, no. 2: 219-255.
- Marco Bianchetti, Camilla Ricci, Marco Scaringi. 2018. ”Are cryptocurrencies real financial bubbles? Evidence from quantitative analyses.” SSRN.