| Abstract : | The genesis of academic interest in asset pricing dates back to the beginning of the previous century and coincides with the dispute of the ‘casino’ view of financial markets most economists held by that time. According to this view capital gains and losses were seen as outcomes of repeating games of chance and market participants acted as gamblers placing bets based on expectations, counter-expectations, instincts and ‘animal spirits’ (Keynes 1936). Absent of economic content, bet pricing was left within the jurisdiction of applied statistics and/or human psychology. Indeed, financial economics started to flourish soon after such a content was established for asset prices under the term ‘fundamental’ or ‘intrinsic’ asset value (Williams 1938). It represented the value of expected cash flow streams payable to asset holders by the asset issuer, which could be directly linked to the expected economic conditions faced by the issuer. After almost a century of research, fundamental analysis remains the cornerstone of the ‘normative’ tension underlying asset pricing literature (Cochrane 2005) describing how asset prices should behave under a plethora of alternative assumptions on how financial markets discount such expected cash flows. This approach hasbeen expanding towards more realistic assumptions realizing the complexity of financial markets (Johnson, Jefferies, and Hui 2003). The largest chapter of this thesis (Chapter 2) is devoted to the empirical rehabilitation of theoretical assumptions on investors behavior and beliefs. It develops a present-value asset pricing model with an econometrically useful representation that accommodates a plethora of assumptions about beliefs and an empirically plausible specification of dividend dynamics. Using 20th century S&P500 data the model is used to compare the empirical _t of belief assumptions associated with rational expectations, asymmetric information, learning, behavioral effects and evolution and to explore the empirical usefulness of some agnosticism in how beliefs are formed. Among these, asymmetric information is particularly useful both in terms of statistical criteria and in terms of ability to explain the equity premium, excess volatility and predictability of returns. The analysis delineates and confronts some of the obstacles involved in moving beyond present value models that ‘explain stylized facts’ towards models with fully edged econometric representations that can bridge theoretical and empirical work. The remainder of this thesis pertains to selected topics on the microstructure of financial markets, a term invented by Garman (1976) to define the analysis of ‘moment-to-moment’ aggregate exchange behavior.1 Time series processes observed at such high frequencies are often found subject to long memory or long horizon autocorrelation, a fact that implies potentially exploitable arbitrage opportunities that challenge market efficiency. Chapter 3 addresses this issue for the fundamental process of the direction of trades and presents empirical evidence that it is instead subject to periodic daily structural breaks. It is suggested that breaks arise as a consequence of daily variation in order flow direction independently of intra-day events and as a consequence of a natural and widespread daily periodicity in the timing of investment decisions. Therefore, a short memory autoregressive model with daily level shifts is introduced and implemented so as to capture the striking long horizon predictability of trade direction. The adopted approach is found to perform better than the standard long memory (ARFIMA) alternative in terms of statistical metrics and computational efficiency. The last topic concerns price discovery, a key function of financial markets. In the market microstructure literature it has been variously interpreted as \the search for an equilibrium price" (Schreiber and Schwartz 1986), \gathering and interpreting news" (Baillie, Booth, Tse, and Zabotina 2002) or \the incorporation of the information implicit in investor trading into market prices" (Lehmann 2002). It is accepted that the information content of market prices may vary across markets since markets process new information at different rates (Baillie, Booth, Tse, and Zabotina 2002). Furthermore, when an asset is traded in multiple markets, persistent discrepancies in price discovery across markets reflect systematic routing of informed order flow towards a particular market. Chapter 4 examines the current case of inter-market competition in Europe, where national (home) exchanges, former monopolies, compete with emerging new markets, known as alternative trading venues, to attract order flows and trading. The analysis investigates the status of price discovery in major European home exchanges compared to a fairly representative and rapidly expanding alternative trading venue, called Chi-X. My approach considers a broader notion of price discovery across markets than previous studies by analyzing (i) sub-second price changes and (ii) changes in the state of the limit order book (LOB) ran by either market. Both expansions are found useful in revealing that home markets are still more informative than Chi-X in general, conforming with the additional finding that they still attract more informed trading. Nevertheless, auxiliary panel data regressions that control for home markets’ superiority in terms of liquidity, volatility and market activity indicate significant variations in the degree of home markets’ price leadership across countries and highlight the importance of implementable mechanisms for market and execution quality harmonization across Europe.
|
|---|
Copyright © 2013 Library AUEB. All rights reserved.