Essays on learning and information-processing in financial markets

Risteska, Simona (2021) Essays on learning and information-processing in financial markets. PhD thesis, London School of Economics and Political Science.

Text - Submitted Version Download (11MB)

Abstract

This dissertation consists of three chapters studying how economic agents learn, form their beliefs and make economically relevant decisions. The main theme of the thesis is to infer beliefs from observable actions and test whether agents process information in a way that is consistent with various theoretical models. In the first chapter I use price data from the real-estate market to infer agent beliefs that are consistent with their pricing behaviour. In particular, I study the way agents make inference from the observable actions of others. In the housing market, where the use of comparables for pricing is most common, I show that the inability to fully grasp the structure of information flows leads agents to overweight stale news due to their repeated use by intermediate agents. The findings are inconsistent with a fully Bayesian model and might instead be reconciled with a model of na¨ıve learning. The second chapter is conjoint work with Francesco Nicolai. We use data on mutual fund portfolio holdings to extract fund managers’ stock return expectations in a fairly general model of portfolio formation. We employ panel regressions to partial out the effect of time-varying stock and manager characteristics and show that subjective expected returns are significantly affected by personal experience. In particular, we provide evidence that professional managers are more strongly influenced by recent returns and those experienced at the early stages of their holding period. The third chapter, co-authored with Francesco Nicolai and Marco Pelosi, provides evidence of the disparity in the incidence of property taxes levied at different points in time. We show that housing demand is significantly less elastic to taxes deferred to the future relative to taxes levied at the moment of purchase, even after accounting for liquidity constraints. We attribute these findings to lack of salience, implying that the burden of deferred taxes will be borne in the future when they are levied. We develop a model to show that the trade-off between lack of salience and liquidity constraints gives rise to an optimal tax mix.

Item Type:	Thesis (PhD)
Additional Information:	© 2021 Simona Risteska
Library of Congress subject classification:	H Social Sciences > HG Finance
Sets:	Departments > Finance
Supervisor:	Julliard, Christian and Verardo, Michela
URI:	http://etheses.lse.ac.uk/id/eprint/4314

Actions (login required)

Record administration - authorised staff only