Aggregate Investor Confidence In The Stock Market
December 15, 2015
Overconfidence is one of the most robust findings in the field of Behavioral Finance, and is associated with excessive trading and risk taking among market participants. Assessment of the level of confidence in their abilities and skills is well-documented for individuals. However, the literature lacks a measure of aggregate investor confidence, in order to test its implications on a macro-level. This paper introduces a simple measure of aggregate investor confidence by adapting a formal model of overconfidence. Applications of the measure suggest that, in aggregate, trading activity is high when investor confidence is high, particularly pronounced for smaller stocks. The effect partially reverses thereafter, suggesting partial correction of initial overreaction through overconfidence. The newly introduced investor confidence index possesses better ability to predict trading activity than past returns, as used in prior studies. Additionally, investors tend to increase risk appetite when confident, represented by increased investment in stocks associated with higher risk.
Aggregate Investor Confidence In The Stock Market – Introduction
In this study I investigate the time-series relationship between aggregate investor confidence and trading behavior, using an adaptation of a simple theoretical model of the formation of confidence about the precision of information. The aim is to explore the effect of aggregate investor confidence on trading activity and risk appetite.
A basic assumption in finance theory is that agents in stock markets behave rationally. That is, investors determine security prices by appropriately discounting future cash flows based on their respective level of systematic risk. Even if temporary mispricing occurs due to irrational beliefs or incomplete information of some agents, arbitrageurs swiftly restore equilibria. In other words, mispricing due to non-rational beliefs will never prevail, as it only requires few informed and rational market participants to detect mispricing.
In contrast, the history of stock markets yields rich evidence of events that are difficult to align with this assumption. The Wall Street Crash of 1929, the Black Monday in 1987, the dot-com bubble in 2000, or the 2010 Flash Crash. Baker and Wurgler (2007) add the ‘Tronics Boom and the Go-Go Years of the 1960s, and the Nifty Fifty bubble of the 1970s to the list. As a consequence, a multitude of systematic behavioral biases of stock market participants gained popularity in the field, as their application allowed a better understand such phenomena.
One behavioural perspective is the role of aggregate investor confidence. That is, investor optimism and pessimism generally tends to affect market outcomes. Formal investor confidence models propose that investors are overly confident about the accuracy of their private information (Daniel, Hirshleifer, and Subrahmanyam, 1998), which should be higher (lower) subsequent market-wide gains (losses) (Gervais and Odean, 2001). Statman, Thorley, and Vorkink (2006), Griffin, Nardari, and Stulz (2007), and Glaser and Weber (2007) find that past returns are associated with higher trading
The novelty of this paper is the application of a direct measure of aggregate investor confidence. Starting with a simple theoretical model on the formation of confidence, a measure of aggregate investor confidence in their own abilities to pick stocks is developed, based on feedback information available to all market participants. The measure incorporates the suggestions of Griffin and Tversky (1992), who identify two primary drivers of confidence: Strength and weight. Strength is the magnitude of arriving feedback, whereas weight is the reliability thereof. In this study, strength is the impulse of recent trading performance compared to a ‘typical’ trading performance in a given point in time, whereas weight is the reliability of those impulses. As investors tend to attribute market gains to their own trading abilities, even if they are shared by the entire market (Gervais and Odean, 2001; Odean, 1998), aggregate confidence should be higher after market gains, especially if the source of success is difficult to attribute (Griffin and Tversky, 1992). In other words, aggregate investor confidence is high (low) when most recent performance is higher than a ‘typical’ return given a certain market state, and when the reliability thereof is low.
The empirical results show that aggregate investor self-confidence is related to trading activity and risk appetite. In other words, high aggregate investor confidence is associated with high trading activity in the subsequent months, which partially reverses for those stocks where the effect is initially strongest. A series of tests suggest that the investor confidence index introduced in this study is a better predictor of trading activity than past return, which is used as a proxy for investor overconfidence in prior studies1. Furthermore, confident investors tend to have increased levels of aggregate risk appetite, and increase the proportion of small stocks subsequent to periods of high investor confidence.
The remainder of this paper is organized as follows. Section two summarizes related literature, and section three outlines and motivates the empirical approach. Section four summarizes the data and intuitionally inspects the newly introduced investor confidence index in its ability to reflect major historical events. Section five validates the measure, and tests its applicability. Section 6 concludes.
See full PDF below.