Date of Award

Fall 2011

Project Type


Program or Major


Degree Name

Doctor of Philosophy

First Advisor

Howard Mayne


The main purpose of this dissertation is to determine which class of models -- bubble or Imperfect Knowledge Economics (IKE) -- provides the better account of short-term stock price fluctuations -- and thus long-swings -- on the basis of empirical evidence. However, it is not clear how to test the bubble models' implication that pure psychological and technical momentum-related factors are the primary driver of stock price movements. Moreover, IKE models' implication that fundamentals are the primary drivers of stock price movements -- but that changes in this relation are non-routine -- is also problematic.

This thesis addresses these difficulties in two main ways. One is to construct a novel dataset based on Bloomberg News' end-of-the-day equity market wrap stories. The textual data provides unambiguous support for IKE models over the bubble models. They indicate that fundamental factors are the primary driver of price fluctuations and that this relation changes at times and in ways that would be difficult to adequately capture with any overarching rule. Psychological considerations are also found to be quite important, but their impact is almost always tethered to a fundamental factor. The bubble models' implication that pure psychological and technical momentum-related considerations are the main drivers of stock prices receives little support.

The thesis also relies on formal econometric analysis to reexamine the connection between stock prices and fundamental factors. It employs recursive structural change tests and cointegration and out-of-sample fit analyses. The results support those obtained with the Bloomberg data: short-term stock price fluctuations are related to fundamentals but the relationship between prices and fundamentals is temporally unstable at times and in ways that cannot be fully foreseen.

Beyond shedding new light on the empirical validity of bubble and IKE models, the thesis examines the question of what circumstances cause market participants to pay attention to certain fundamentals over others when forecasting market outcomes. Analyses combining both the Bloomberg data and formal econometrics suggest that the frequency with which certain fundamentals merit the attention of market participants is a function of the recent variation of such factors as well as deviations of fundamentals away from estimates of common benchmark levels.