Ambiguity, also known as Knightian uncertainty, is rooted in nearly every real-life decision process. It refers to situations in which both the outcome and the probabilities governing the set of possible outcomes are unknown. Risk, on the other hand, refers to situations where the future outcome is unknown, but the set of possible outcomes is known, with certain probabilities attached to them. To illustrate, if I toss a coin, I know that it will land with heads or tails, but I do not know which side it will land on until after the toss. If I know the probabilities are 0.5 and 0.5, then I face risk. If I do not know what the probability of heads or tails is (the coin may be fair or unfair), then I face ambiguity. Thus, the key distinction between risk and ambiguity is whether the probability distributions associated with the possible outcomes are known or unknown.

Frank Knight is credited as the first to formalize the notion of ambiguity, also known as Knightian uncertainty claiming that to capture the complete implications of uncertainty, one must consider both risk and ambiguity. Since the introduction of the Ellsberg Paradox in 1961, there is a wealth of evidence demonstrating that not only are people averse to both these aspects of uncertainty, but they also behave differently when confronted with ambiguity compared to risk. However, the literature on voluntary disclosure has yet to take ambiguity into consideration. This somewhat simplistic view of firms’ disclosure practices overly relies on the influence of risk by either assuming away ambiguity or implicitly assuming that individuals are neutral to it. Thus, ignoring ambiguity likely results in an incomplete understanding of firms’ disclosure decisions under uncertainty.

To understand the possible difference in managers’ voluntary disclosure decisions, one must understand general investor behavior under these two dimensions of uncertainty. Theoretical models of the relationship between risk and voluntary disclosure have posited that by providing risk-averse investors with more information, managers increase the rate at which investors learn about the firms’ economic fundamentals, causing a reduction in risk and, consequently, stock price volatility. Conversely, an increase in ambiguity, in the presence of both risk- and ambiguity-averse investors, may cause the provision of more information to lead to severe investor reactions. This happens because a higher degree of ambiguity results in lower perceived probabilities of good states and higher perceived probabilities of bad states. Models of information processing that include ambiguity-averse investors have shown that such pessimistic behavior may lead to sharp price fluctuations as a reaction to forecast news when ambiguity increases, resulting in firms’ reduced willingness to provide new information compared to situations of heightened risk. For example, one model shows that small cash-flow news may cause a sudden increase in ambiguity-averse investors’ perceived probabilities of bad states, amplifying the effects of negative news shock on stock prices, consequently leading to excess volatility and an increased probability of a stock price crash. As such, increases in ambiguity may cause shifts in perceptions, causing ambiguity-averse investors to overweight the perceived probabilities of future bad states, leading to higher price volatility when news is announced

This paper extends the literature by providing new insights into the distinct effect of ambiguity on firms’ voluntary disclosure decisions. Using a novel, firm-level ambiguity measure, I find strong evidence that firms’ forecast decisions are influenced by both risk and ambiguity. Specifically, managers issue less earnings guidance during periods of heightened ambiguity, which contrasts with their behavior during periods of heightened risk. The results indicate the average effect of ambiguity on voluntary disclosure to be opposite to that of risk, showing a negative relationship between increases in ambiguity and voluntary management forecasts. This finding alone stresses the importance of including both ambiguity and risk when analyzing uncertainty and voluntary disclosure decisions.

In the second stage, I provide novel empirical evidence showing the adverse capital market consequences that guiding under ambiguity might have. When ambiguity is high, bad news forecasts are associated with increased share price volatility and significant negative abnormal returns. I do not observe the same reaction when ambiguity is low. These finding underscore managers’ awareness of the severe repercussions of disclosing information under ambiguity, which may explain their inclination to avoid it.

The paper contributes to the literature by studying the most prominent voluntary disclosure practice, earnings forecasts, under a larger construct of uncertainty that includes both risk and ambiguity. The evidence I provide lends support to the recent interest in ambiguity, showing that when assessing uncertainty, one must consider ambiguity.

Chart: Cumulative Abnormal Return Around Earnings Announcements Based on Ambiguity and Bundling of Earnings Forecasts with Earnings Announcements

Note: The chart shows the mean cumulative abnormal return (CAR) around the five-day window of a quarterly earnings announcement. This is based on whether firms issue earnings forecasts with the quarterly earnings announcement (bundle), the news of the forecast, and the ambiguity quartile prior to the earnings announcement. Bundle is an indicator variable if the firm provides a voluntary management earnings forecast. BundleGoodForecast is an indicator variable if the forecast is above the median analyst forecast, and BundleBadForecast is an indicator variable if the forecast is below the median analyst forecast. Low ambiguity refers to firms in the lowest ambiguity quartile, while high ambiguity refers to firms in the highest ambiguity quartile prior to the earnings announcement.