Use this URL to cite or link to this record in EThOS:
Title: Measuring and modelling patterns of behaviour in datasets of individual investor trading records using flexible methods
Author: Burgess, Matthew
ISNI:       0000 0004 7227 7199
Awarding Body: University of Warwick
Current Institution: University of Warwick
Date of Award: 2017
Availability of Full Text:
Access from EThOS:
Access from Institution:
Datasets of individual investor trading records have been an important source of empirical evidence in the field of behavioural finance. This thesis contributes to two topics within this empirical literature using a dataset of trading records from a discount brokerage. The first topic is the disposition effect (DE), the tendency for investors to sell winning positions at a faster rate than losing positions. A version of the aggregate DE score introduced by Odean (1998) is analysed, as a time series and at the level of individual stocks. The influence of each stock on the aggregate DE score is calculated, and the characteristics of high and low influence stocks compared. A formal relationship is derived between this DE score and the hazard ratio estimated in a proportional hazards (PH) model. PH models have been used in the literature to measure the effect of covariates on the DE at the investor level. Past approaches have used a marginal model to address the problem of correlation between positions at the investor-level, which involves computing robust standard errors after estimation of the model. A shared frailty model is tested as a more flexible alternative, where unobserved heterogeneity is modelled through the use of latent variables. It provides a significantly improved fit relative to the corresponding marginal model, and adheres more closely to the PH assumption. The second topic is the preference of investors for lottery stocks. These are stocks that are low in price and high in volatility and skewness, a scheme of stock categorisation suggested by Kumar (2009a). The theme of using more flexible models to accommodate investor-level correlation is continued, with a mixed-effects logistic regression being used to study the factors affecting the decision to purchase a lottery stock. This allows the comparison of both time-varying and static factors.
Supervisor: Not available Sponsor: Engineering and Physical Sciences Research Council
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available
Keywords: QA Mathematics