| This dissertation consists of four independent papers related to adaptive learning, multiple equilibria, or both. Chapter I is an introduction.; Existing RBC models with multiple equilibria suffer from two major weaknesses: they often rely on unsatisfactory assumptions, and they rarely are stable under adaptive learning. Chapter II addresses these weaknesses by making two modifications to a standard RBC model. First, newly employed labor is less productive than experienced labor. Second, short-sighted firm managers use a different effective discount rate than households. Three results arise: (1) The model is frequently indeterminate; (2) Indeterminate solutions are stable under adaptive learning; and (3) An increase in job turnover will simultaneously decrease the volatility of aggregate output while increasing that of individual households' income.; Chapter III develops a simple two-country, two-good model of international trade and borrowing that suppresses all previous sources of current account dynamics. Under rational expectations, international debt follows a random walk. Under adaptive learning, however, the model's unit root is eliminated and international debt follows either a stationary or an explosive process. Whether debt converges or diverges depends on the specific learning algorithm that agents employ. When debt diverges, a financial crisis eventually occurs to ensure that the model's transversality condition holds.; It is well known that for some policy rules, the New Keynesian model is indeterminate. Thomas Lubik and Frank Schorfheide (2004) develop a Bayesian estimation approach that suggests indeterminacy prior to 1979 but determinacy after 1982. Athanasios Orphanides (2004) uses real-time data instead of revised data and finds determinacy in both periods. Chapter IV incorporates systematic mismeasurement of the output gap into the New Keynesian model and uses the Lubik/Schorfheide approach to find indeterminacy in both periods.; Chapter V uses the technique of Lubik and Schorfheide to estimate the New Keynesian model where the Fed targets both the expected output gap and expected inflation. I then test for both indeterminacy and stability under learning. For this policy rule, I find that the economy was indeterminate and unstable under learning prior to 1979, but has been determinate and stable under learning since 1982. Finally, Chapter VI concludes. |
