From macro to micro: causal inference, firm valuation and trading conditions

The aim of the Ph.D. thesis is twofold. First, we investigate possible stock market mispricing to eventually build profitable investment strategies. Second, we analyze how the microscopic interactions among agents influence trading conditions thereby leading to market instabilities. As regards the study of possible mispricing, we identify via the vector autoregressive approach revenues as the primary driver process of firm growth. To do so, we employ the recent Independent Component Analysis (ICA) technique which allows us to identify contemporaneous causal relations among the considered variables. In particular, the first original contribution of the thesis is to extend the ICA methodology for singular and noisy structural vector autoregressive models; see Chapter 2. As a second original contribution, starting from the revenues, we propose a firm valuation framework incorporating the associated intrinsic uncertainty. We derive a probability distribution of fair values, we construct a market factor capturing misvaluation comovements and we propose two stock recommendation systems that hinge on the fair value distribution; see Chapters 3, 4 and 5. Finally, in the last contribution, we analyze asymptotically market stability as the number of assets and traders increase. Market instability is defined as a result of oscillating equilibrium strategies of optimal execution problems in market impact games, where the dynamical equilibrium between the activity of simultaneously trading agents generates the price dynamics. One of the main results is the connection of market instability to the market cross-impact structure when portfolios execution orders are considered; see Chapter 7, 8 and 9.