Standard-Nutzungsbedingungen:Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden.Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen.Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte.
Terms of use:
Documents in EconStor may
AbstractThis paper investigates the power of macroeconomic factors to explain euro area bond risk premia using (i) a large dataset that captures the nowadays data-rich environment (ii) the Elastic Net variable selection. We find that macroeconomic factors, in particular economic activity and sentiment indicators, explain 40% of the variability of risk premia before the crisis, and up to 55% during the financial crisis, and both for core countries (from 40% to 60%) and periphery countries (from 35% to 44%). Moreover, macroeconomic factor models clearly outperform financial indicators like the CP-factor and credit default swap (CDS) premia, even in periods of significant market turbulence.JEL codes: E43, E44, G01, G12, C52, C55 We search for quantitative evidence on the extent to which macroeconomic factors are priced in in bond premia. To determine whether, when and by how much bond premia is related to price, economic activity, business sentiment or financial factors, or a combination of those, we employ the Elastic Net estimator (EN henceforth, Zou and Hastie, 2005), 1 a variable selection approach that helps overcome some specific challenges of euro area bond market data. First, we can evaluate a large number of potential determinants: 132 monthly macroeconomic indicators from nine macroeconomic groups, and, whenever possible, we also consider country-specific in addition to EA wide data (see the online appendix for details of our data and some additional results). Second, we can select observable factors based on their explanatory power for bond premia, which provides higher transparency and interpretability than principal components or other statistical techniques that instead summarise the information content of the explanatory variables (e.g. Stock and Watson 2002). Finally, the EN is particularly suitable for small sample analysis, which fits well with the short history of the EA and our goal of investigating the financial crisis impact.We first document the strong impact of the financial crisis on bond risk premia across 11 EA markets. We show that the fairly strong commonality in bond risk premia dynamics across euro area 1 The EN estimator belongs to a broad class of Least Angle Regression estimators (LARS) that are designed to rank the importance of every explanatory variable using a response vector (in our case excess returns) and help select a parsimonious model using a regularization parameter.ECB Wo...