Essays in quantitative analysis of the effects of market imperfections on asset returns

Financial assets prices are not always in perfect equilibrium and deviate from their fundamental values. The dramatic rise and fall of the stock market raises concern about the rationality of sudden changes in stock valuations. The mispricing of assets contributes to financial crises, which can damage the overall economy. This dissertation analyses the effect of market imperfections at different time horizons. Starting at a macroeconomic level with a change of currency, the first essay analyses the impact of the introduction of Euro on interest rate sensitivity of European firms. We found that the connection between bond issuance and reductions in interest rate sensitivity is most significant among financially constrained firms, which suggests that financially constrained firms are the main beneficiaries of the relaxed public borrowing constraint in Europe after the introduction of euro. Releases of macroeconomic news announcements cause sudden price discontinuities, or jumps and co-jumps, on financial markets. The second essay attempts to explain the effect of US macroeconomic announcements on European equity, interest rate and foreign exchange markets at a high-frequency level. While European equity markets are more sensitive to US fundamentals, US macroeconomic announcements cause significant jumps and cojumps on all European asset classes. We found that European markets are highly co-integrated and observed a strong correlation between the type of news and the direction of the jumps. Motivated by the phenomenal success of some quantitative trading funds, the third essay describes a new pairs trading strategy, where the spread between two co-integrated portfolios is modelled stochastically. Taking into account transaction costs, the algorithm generates a systematic positive excess return based on a pure statistical arbitrage strategy. While very convenient, traditional asset pricing relies on two restrictive assumptions. First, asset returns are conventionally modelled with Gaussian-based distributions even though actual financial time series exhibit volatility clustering. The second assumption mainly affects market microstructure studies. Time series are sampled at regular interval of time and asset return distribution is used as the unique driver to model the price fluctuation of an asset over time. In reality, the time between two transactions, often called waiting-time, is stochastic and conveys important information about price formation. The two last essays relax the assumptions of log-normally of asset prices and model asset prices with a continuous-time random walk. The fourth article compares the Markovian and non-Markovian forms of the continuous-time random walk process and shows the relevance of the waiting-time distribution on price formation. The last article applies the framework to statistical arbitrage