No abstract

We develop a common factor approach to reconstruct new business cycle indices for Argentina, Brazil, Chile, and Mexico ("LAC-4") from a new dataset spanning 135 years. We establish the robustness of our indices through extensive testing and use them to explore business cycle properties in LAC-4 across outward-and inward-looking policy regimes. We find that output persistence in LAC-4 has been consistently high across regimes, whereas volatility has been markedly time-varying but without displaying a clear-cut relationship with openness. We also find a sizeable common regional factor driven by output and interest rates in advanced countries, including during inward-looking regimes.

This article studies two issues in forecast combination, first considering ways to combine forecasts from surveys and time series models. Second, it considers the possibility, advanced by Hendry and Clements (2004), that model instability can help explain the gains in forecasting performance resulting from combination. The article is organized as follows. Section 2 discusses the design of the universe of forecasting models used in combining forecasts from time series models and subjective survey forecasts. Section 3 undertakes an empirical analysis using forecasts from univariate and multivariate linear models, nonlinear models, and survey forecasts. Section 4 provides analytical results that shed light on the performance of forecast combinations under model instability. Section 5 presents empirical results on forecast combinations under breaks. Section 6 concludes.

Recent financial research has provided evidence on the predictability of asset returns. In this paper we consider the results contained in Pesaran and Timmerman (1995), which provided evidence on predictability of excess returns in the US stock market over the sample . We show that the extension of the sample to the nineties weakens considerably the statistical and economic significance of the predictability of stock returns based on earlier data. We propose an extension of their framework, based on the explicit consideration of model uncertainty under rich parameterizations for the predictive models. We propose a novel methodology to deal with model uncertainty based on 'thick' modelling, i.e. on considering a multiplicity of predictive models rather than a single predictive model. We show that portfolio allocations based on a thick modelling strategy systematically outperform thin modelling.

No abstract