This dissertation consists of two independent chapters on economic and financial forecasting. The first chapter introduces a nonlinear forecasting framework that combines forecasts of the sign and absolute value of a time series into conditional mean forecasts. In contrast to linear models, the proposed framework allows different predictors to separately impact the sign and absolute value of the target series. An empirical application using the FRED-MD dataset shows that forecasts from the proposed model substantially outperform linear forecasts for series that exhibit persistent volatility dynamics, such as output and interest rates. The second chapter, coauthored with Christian Brownlees, provides an extensive comparison of methods to forecast downside risks to GDP growth for a panel of 24 OECD economies. We consider forecasts constructed from standard quantile regressions as well as from conditional volatility models. Our evidence suggests that standard volatility models such as the GARCH(1,1) are at least as accurate as quantile regressions.
© 2001-2024 Fundación Dialnet · Todos los derechos reservados