An intraday examination of the role of priors in the price discovery process

Abstract

The informational efficiency of financial markets is the most debated topic of financial economics. Asset price formation in the absence of new market relevant information is described as a puzzle unresolved by standard economic theory. This thesis asks if this puzzle is simply a case of missing information. If there is market relevant information in existence, which has been identified by market agents, but not discerned by economists. Traders/Investors have prior expectations for the outcome of future market relevant information events (subsequently termed ‘priors´). Such priors relate to the scope, scale, timing and probability of an upcoming information event. These priors play a significant role in the price formation process upon the arrival of the expected information event. Priors are updated frequently by individual investors concurrently with information flows which may change the individuals’ expectation of the upcoming high scope information event. Such changes to priors alter the fundamental valuation of a given asset and can be considered completely in line with rational expectation hypotheses. The subtle changes to the priors of a few market agents alone is not enough to result in a large-scale price formation process. However, an information event sufficiently absorbed by a large number of market agents, which results in an alteration to prior’s en masse, will significantly alter the weighted average valuation of a given asset sufficiently that a large-scale price formation process should be observable. This research identifies a database of market relevant information, in the form of market rumours, broadcast by market agents and commentators. This information by nature is not published or archived by the incumbent and regulated financial information sources such as Reuters and Bloomberg, thus potentially missed by research economists. Empirical results of this thesis show that at intraday observations of market price, large scale and persistent volatility events are observable at the time of rumour broadcast. The instantaneous increase in volatility during the first minute of rumour arrival is up to 211%, while the cumulative increase in volatility over a 60-minute window is as much as 2614%. Such large-scale volatility events had previously been attributed to ‘noise’ or private information flows. Further findings show that large excess returns in the run-up to central bank announcements can be attributed to market rumours dispersion. Such pre-announcement excess returns had been observed in the past but unexplained in the literature. Results show that trading on the pre-announcement rumour for 10 days per year can generate almost 100% greater return than holding the market portfolio on all other days of the year. The thesis also documents the existence of a new price formation process undocumented in the literature. Empirical results identify the existence of large excess returns in the day following European Central Bank (ECB) announcements but only when the day is a Friday. This is termed ‘the ECB conditional Friday effect’.