30.09.2015

Market Disagreement and Stock Returns

Analysts and investors may agree to disagree on certain portfolio picks, but the nature of that disagreement and its interplay with the portfolio could, in itself, have implications for returns and asset prices.

Previous research has already shown that greater disagreement is associated with higher prices and lower subsequent returns. Professor Jialin Yu builds on that finding to define the characteristics of disagreement that are most likely to have an impact, as well as the types of stocks that are most affected.

One key concern is that there can be a big difference between the “bottom-up” measure that looks at disagreement among individual stocks forecasts, and the “top-down” one which refers to forecasts of the market index.

“Bottom-up disagreement is constructed using thousands of individual stock forecasts, while there are on average only 20 or so analysts in our sample covering the S&P500,” Professor Yu said.

“The analysts’ effort likely mirrors investors’ focus on stock pickings, but some individual stock disagreements may not be reflected in the top-down disagreement. For example, investors may agree on the future prospect of the market but disagree on which stock will lead or lag. Hence, strong disagreement may exist and can be discerned from the bottom up, which makes this a better proxy of belief dispersion in the market.”

One illustration of how this might work is the housing market, he said. A top-down measure might attempt to capture the mood of homebuyers at the national market, but could it succeed in this when homebuyers tend to focus on local housing market conditions?

Professor Yu tested the bottom-up versus top-down measure using analyst forecasts from 1981-2005 and confirmed that the bottom-up measure was indeed a better approach to measuring asset price when there is disagreement. The data showed that a one-standard deviation increase in disagreement using the bottom-up measure was associated with a drop in stock market returns after one year of 6.6 per cent.

His research also showed that stocks were slow to revert to the mean in the face of a shock disagreement. The shocks had a half-life of about one year and it took about three years to largely revert to the mean.

Prices tended to become higher when there was greater disagreement but this correlated more with discount rate news than cash flow news, such that higher disagreement led to a lower discount rate.

Growth stocks were found to be more sensitive to portfolio disagreement than value stocks, indicating value premium was affected by disagreement.

Overall, as might be expected, high disagreement stocks underperformed low disagreement stocks.

“Taken together, the evidence on equity premium, discount rate and value premium provides strong support for the hypothesis that disagreement matters for the equilibrium price and expected return in the stock market.

“It may also be promising to explore the implications of disagreement for other markets because the conditions of disagreement and short-sales constraint may hold for anything form houses and art to tulip bulbs with various shapes,” Professor Yu concluded.


YU, Jialin

Academic Director, HKUST-NYU Stern MSc in Global Finance, MSc in Investment Management (Full Time), Associate Professor
Finance