Dispersed Analysts’ Forecasts of Firms’ Operational Uncertainties and Stock Returns

Abstract

A company’s operational uncertainty resulting from economic activity could be measured by its contingent liabilities. When the contingent liabilities are forecasted by the analysts, the dispersed analysts’ forecasts might constitute another source of uncertainty. In this study, the Uncertainty (contingent liabilities) of Uncertainty (dispersed analysts’ forecasts) is measured by the UOU indicator. We find that this indicator not only measures the negative correlation between the firm’s UOU and its stock returns, but also serves as a good risk and pricing indicator for the excessive stock returns. Robust tests based on industry, market conditions, and macroeconomic environments indicate our findings are still valid. We, therefore, believe our empirical findings provide a good reference for the investment decision-making process.

Keywords:

• Analysts’ forecasts
• Stock returns
• Operational uncertainties
• Risk aversion

JEL classifications:

• G14
• G24
• G32

Notes

1 Vol-of-Vol is calculated from IVs derived from option prices to measure uncertainty over risk. IV is primarily driven by expected stock price volatility. It is one of the best predictors of volatility. As a result, it can measure future stock return volatility expected by investors.

2 Hibor^®, Tushare^®, and Wind^® has 3.5 million, 2.7 million, and 5 million users, respectively. Their data commonly used by stock analysts, and the data has high credibility.

3 The reason why using three days as the selection range is that analyst reports are not published simultaneously. We adopted the window of 3 days because 3 days are generally the heights of the public’ attention. The at least 10 analysts following the stock is required since it 3 reports are the minimum numbers of the occurrence for calculating the standard deviation with acceptable accuracy.

4 The relationship between UOU(2) and these risk factors is consistent with UOU, except the relationship between Size and UOU(2) is more significant than UOU, which was −1.06% with t-statistics of −3.34.

5 For each regression of quarterly frequency, the risk free rate was taken as the close price of 3M treasure bonds at each calibration date.

6 In the “high-low” portfolio, the difference between the excess return of the high and low UOU(2) portfolios was –1.09% per quarter, with a highly significant t-statistic of –2.97.

7 The average excess returns of the UOU(2) high-low portfolio ranged from −1.72% to 1.34% for each quarter, with corresponding t-statistics of –1.03 and 3.83. And the trend of Returns of Portfolios Sorted both by Return Drivers and UOU(2) is consistent with UOU.

Additional information

Funding

This study is funded by Social Science Foundation in Sichuan, grant no: SC20TJ022.