Copy the page URI to the clipboard
Grayson, Paul
(2017).
DOI: https://doi.org/10.21954/ou.ro.0000ccf7
Abstract
Millions of people across the globe invest in financial markets, hoping to increase their wealth or income. Investing is risky of course, but some investors seem to battle not only the market but also themselves. In recent decades behavioural economics has uncovered many patterns of trading behaviour which appear to work against the investors who display them, resulting in lower investment returns. For this reason, these patterns of behaviour are regarded as trading biases.
This thesis investigates two main themes that follow from the observation of trading biases in financial markets, both ultimately related to helping investors make better investment decisions. First, to what extent do individual investors reliably demonstrate trading biases? Second, why do these biases occur and can their impact be mitigated?
With respect to the first theme, is there evidence that some biases have trait-like characteristics, and predict investors’ behaviour across a range of situations? This theme is related to the “state vs trait” argument from personality psychology. Applied here, the question is whether variation in trading behaviour is better explained due to environmental differences when decisions are being made (state), or because people differ in their tendencies to behave in certain ways (trait-like). If the latter, investors who tend to express biases would benefit from understanding their own decision-making patterns.
Note that the claim here is not that the biases are literally traits themselves, but that they have trait-like characteristics such as intra-individual stability and construct validity. The evidence presented in this thesis suggests that the biases examined possess these characteristics because their expression depends on some underlying behavioural tendencies; these underlying tendencies may be traits, though they are not directly investigated here.
With respect to the second theme, a suspected cause of many trading biases is the influence of irrelevant emotions on decision making. This thesis investigates evidence for the role of emotions in trading biases. It also investigates the use of cognitive reappraisal as a de-biasing technique (i.e. a method to reduce the level of bias displayed in trading decisions), which works by reducing emotions during trading.
The main bias which this thesis examines is the disposition effect. This occurs when investors are more eager to sell gains than losses, or stated another way, when they hold losses longer than gains. So, it is a bias that affects decisions about selling, based on the profit or loss made on each stock. Although the disposition effect is widely studied in behavioural finance, it has not been demonstrated to be a persistent pattern of an investor’s trading behaviour. Experimental studies measure a mathematically defined disposition effect in the lab, but assume that this research can provide insight into the disposition effect observed in financial markets.
Meanwhile cognitive reappraisal, a form of emotion regulation, has been shown to reduce the disposition effect, but only in students and only using lab conditions that lack ecological validity. Furthermore, the mechanism for reappraisal’s effect has not been investigated.
As well as investigating whether the disposition effect has trait-like characteristics, it also examines the “constituent biases” of the disposition effect, cutting gains and holding losses. Though the disposition effect is defined as a difference between selling gains and losses, the two are treated as two sides of the same behaviour rather than two separate trading patterns. This thesis argues that these constituent biases should be treated as separate biases, rather than two aspects of a unitary disposition effect. Building on the first theme, the constituent biases are also assessed for trait-like characteristics; building on the second theme, the effect of cognitive reappraisal on each constituent bias is examined.
This thesis contains three studies, two involving retail investors and one with an adult novice sample. All three involve measuring trading biases with an ecologically realistic trading simulation, played multiple times. The two studies with retail investors also include a disposition effect scale. Cognitive reappraisal is tested in one of the two studies with retail investors and in the study with novices.
I contribute to the literature in establishing that the disposition effect has trait-like characteristics aspects, by showing that it can be reliably measured in both an ecologically valid trading simulation and a self-report scale, and that these link to real-world trading behaviour. Furthermore, I show that the disposition effect is not a unitary bias but that its two constituent biases, cutting gains and holding losses, are independent of each other. They too have trait-like characteristics, and can also be reliably measured using the same ecologically valid trading simulation. However, levels of each bias in investors are not associated with each other. This contrasts with the conventional approach to the disposition effect that treats them as opposite sides of a unitary bias. I argue they are independent biases which are measured together in the disposition effect. Furthermore, as independent biases they are likely to be underpinned by different underlying traits.
I build on the existing literature by showing that the disposition effect occurs in retail investors using this trading simulation. I find that cognitive reappraisal reduces the disposition effect, while I make improvements in the external validity of this test in both the measurement of the disposition effect and participants used. I extend knowledge of emotion regulation and trading biases by showing that cognitive reappraisal reduces the disposition effect by decreasing the tendency to hold losses. Lastly, I show that cognitive reappraisal is not effective under the same conditions in novices (as opposed to retail investors). This reminds us of the merit of testing de-biasing techniques with greater ecological validity, and suggests that implementing de-biasing techniques in real world decision making may be more difficult than it first appears.
CORE (COnnecting REpositories)
CORE (COnnecting REpositories)
