CEU Electronic Theses and Dissertations, 2015
| Author | Valaitis, Vytautas |
|---|---|
| Title | Adaptive Learning Explanation to non-Normal Output Growth and Output Gap |
| Summary | Empirical studies suggest that the evolution of output across time is described by a distribution with the tails fatter than normal. At the same time standard macroeconomics models are linear and can only replicate this exogenously. I show that replacing the rational expectations with ones formed under statistical learning can deliver the fat tails even with using normally distributed exogenous shocks. Monte Carlo experiments using a small-scale New Keynesian model show that propagation is more reasonable when rational expectations are replaced on the model`s final equations rather than when adaptive learning is micro founded. In both cases a model exhibits large and rare exits that become more pronounced under micro founded rule. An even less rational heuristic learning propagates the normal shocks as well, but only together with a very high variance. |
| Supervisor | Istvan Konya |
| Department | Economics MA |
| Full text | https://www.etd.ceu.edu/2015/valaitis_vytautas.pdf |
Visit the CEU Library.
© 2007-2021, Central European University