Can Sentiment and Momentum Predict Prices?
The Federal Reserve Bank of San Francisco put out a really interesting paper the other day that examines recent academic research that has found that sentiment and momentum, when used together, can be used to successfully predict stock market returns. For years technicians have believed this and accepted the fact that tools like sentiment and momentum can forecast stock market returns. Academics have refuted any claims that stock market prices can be predicted. They constantly cite Eugene Fama’s efficient-market hypothesis and the random walk theory. Both of these theories suggest that prices move randomly and that it is impossible to predict fluctuations in stock prices. Technicians and others on Wall St. scoff at these theories. They argue that it is, in fact, possible to outperform the market using technical analysis tools and sentiment data. The research shared in this paper is evidence that academics are moving further away from flawed theories like the efficient-market hypothesis.
The idea that sentiment and momentum can drive stock prices is nothing new. The paper begins with a brief history of this belief. All the way back in 1936 the legendary economist John Maynard Keynes compared the stock market to a “beauty contest” where participants aren't judging the underlying concept of beauty, but rather “anticipating what average opinion expects the average opinion to be.” Another famous economist, Robert Shiller described the phenomenon as a sort of positive feedback loop where an initial price rise attracts public attention which in turn fuels enthusiasm pushing prices even higher. This loop can continue to cycle until it pushes prices up to a point of “irrational exuberance.”
This is really the primary cause of any speculative bubble. Take for example the bitcoin bubble of 2017. The price of a single bitcoin went from around $1,000 at the beginning of 2017 and surged to a high of over $19,000 by year-end. An initial rally caught the attention of the general public, which fueled the rally. Bitcoin mania continued to spread to the public until it reached an irrational point where there were no marginal buyers, and the price inevitably collapsed over 80% to where it currently sits around $4,000.
The paper details the process that researchers went through to prove that sentiment and momentum drive stock prices. In order to quantify sentiment, researchers used Google search volume for the term “Stock Market” as well as the University of Michigan’s consumer sentiment index. They explain using the chart below that S&P 500 12-month percentage change is positively correlated to the sentiment index they used.
This only established correlation but the researchers were interested in proving causation as well. They wanted to figure out if the sentiment and momentum alone can predict prices. They used scatterplots with a fitted linear regression line to understand the relationship between the two variables. In doing so, they found sentiment alone does not predict returns, however, they found that sentiment, when combined with momentum, can successfully predict the returns of the S&P 500.
It’s refreshing for technicians to see academics abandoning Efficient Market Hypothesis. Instead, they’re beginning to test and understand some of the tools we use. To read the paper and learn more about this study click here.