In this paper we forecast daily returns of crypto‐currencies using a wide variety of different econometric models. To capture salient features commonly observed in financial time series like rapid changes in the conditional variance, non‐normality of the measurement errors and sharply increasing trends, we develop a time‐varying parameter VAR with t‐distributed measurement errors and stochastic volatility. To control for overparametrization, we rely on the Bayesian literature on shrinkage priors, which enables us to shrink coefficients associated with irrelevant predictors and/or perform model specification in a flexible manner. Using around one year of daily data, we perform a real‐time forecasting exercise and investigate whether any of the proposed models is able to outperform the naive random walk benchmark. To assess the economic relevance of the forecasting gains produced by the proposed models we, moreover, run a simple trading exercise.

No abstract

Consumers widely use music genres (e.g., pop, rock) for finding the right products. However, they are commonly arbitrary, not-standardized, disputed, and closely related genres often overlap. In this paper, we challenge established music genres (e.g., pop, rock) by comparing them to an entirely data-driven approach.To this end, we use a unique data set of revealed user preferences to carry out a context-based artist similarity. This measure is used in turn to find high-density artist clusters. The contribution of this paper is twofold. First, we investigate the differences between established music genres and data-driven clustering. Second, we provide implications for researchers and practitioners.