JOB MARKET PAPER

 

“Localized Level Crossing Random Walk Test Robust to the Presence of Structural Breaks”, advisor Dr. Alex Maynard.

 

The random walk hypothesis has been important in numerous disciplines including economics, finance, and international finance, in which random walk models have been used to model variables such as consumption, stock prices, and exchange rates. Likewise, tests of the random walk hypothesis are frequently employed to test deeper theoretical models, such as the permanent income hypothesis and weak form market efficiency. It is typical in such tests to allow for a linear trend. This is often necessary on economic grounds as well. For example, economic growth and inflation, give rise to an upward long-run trend in stock prices. On the other hand, few of the tests allow for changes or breaks to occur in the trend term. This is arguably a somewhat restrictive assumption, which may result in unreliable inference. For example, changes to the trend growth rates or long-term average inflation rates would imply a break in trend for nominal stock prices. This paper proposes a modified version of the nonparametric level crossing random walk test, in which the crossing level is determined locally. This modification results in a test that is robust to unknown multiple structural breaks in the level and slope of the trend function under both the null and alternative hypothesis. No knowledge regarding the number or timing of the breaks is required. A data driven method is suggested to select the extent of the localization in order to maximize power in a proximate model. We propose bootstrap critical values that control size well in simulations with moderate sample sizes. We apply our testing methodology to a real time series and compare the results with those obtained by means of existing random walk and unit root tests. We conclude that the test proposed in this paper can be considered a good complement to existing level crossing literature.   

 

RESEARCH PAPERS

 

“Testing Weak Form Efficiency on the Toronto Stock Exchange”, with Dr. Francis Tapon.

 

We believe that in order to test for weak form efficiency in the market a vast pool of individual stocks must be analyzed rather than a stock market index. In this paper we use a model-based bootstrap to generate a series of simulated trials and apply a modified chart pattern recognition algorithm to all stocks listed on the Toronto Stock Exchange (TSX), Canada’s largest stock market. We compare the number of patterns detected in the original price series with the number of patterns found in simulated series. By simulating the price path we eliminate specific time dependencies present in real data making price changes purely random. Patterns, if consistently identified, carry additional incremental information which adds value to the investment process. This informativeness does not guarantee profitability. However, conclusion on relative efficiency of particular sectors of the economy can be drawn. Although, we fail to reject the null hypothesis of weak form efficiency on TSX, some sectors of Canadian economy appear to be less efficient than others. In addition, we find negative dependency of pattern frequencies on the two moments of return distribution, namely variance and kurtosis.

 

RESEARCH IN PROGRESS

 

“Discrete Choice Modeling with Nonstationary Panels Applied to Exchange Rate Regime Choice”, with Dr. Yugio Sun.

 

Most of the empirical literature on exchange rate regimes uses the IMF de jure classification based on the self-reported or announced regimes, despite the existing inconsistency from actual policies in place. To address this problem, we analyze the differences between the official bilateral exchange rates reported by central banks and implied exchange rates from a triangular relation among three major world currencies. We attempt to recover the true foreign exchange policy of some targeted counties. Specifically, we employ Dynamic Conditional Correlation Multivariate GARCH (DCC-MGARCH) models to estimate the degree of time varying correlation in bilateral and implied exchange rates for the countries under consideration. Using the estimated parameters as factors we classify countries based on the degree of dependence. We find that countries with more flexible exchange rate regimes tend to have higher degree of dependence between bilateral and implied exchange rates.