Uncertainty plays a key role in economic dynamics, in particular when agents face irreversible choices. This dissertation devotes to address the effects of uncertainty on the growth rate of an economy, on the decisions and frequency of adoption of new technology and aims at shedding some light on current financial markets' new features. In the first chapter, a continuous time stochastic model is developed to study a small open economy experiencing real and monetary shocks. Martingale representation and convex analysis techniques in the presence of constraints are used to study the representative agent's optimal consumption and investment strategies. We provide a closed form solution for a Cobb-Douglas economy. The consumption growth rate and risk free interest rate exhibit a U shape relationship; a low real interest rate and positive consumption growth rate can coexist. Uncertainty negatively affects expected consumption growth rate, and some empirical evidence supports this result. The second chapter studies optimal scrapping and adoption of technology under technological and market uncertainty. The optimal decision rule is a (s, S ) style policy where the trigger and target technology levels are negatively (positively) correlated with recession (boom) persistence and technological uncertainty. The average time between two adoptions is governed by several factors. Technological uncertainty reinforces depreciation and thus hastens replacement. Moreover, both types of uncertainty have an impact on the scrapping and upgrading levels. Overall, adoption is more frequent for economies spending a large fraction of time in booms. The end of a recession can trigger updating; this result implies that investment spikes are procyclical. The third chapter investigates the evolution through time of the price-earnings ratio (PER) and, in particular, its interaction with the business cycle. Using a simplified version of the technology adoption model developed in chapter two, we find that PER levels are high when the firm is close to implementing a new technology. Moreover, earnings are much more responsive than the value of the firm to recessions and expansions. As a consequence, the PER can be higher in a recession than in a boom.
