An Econometric Analysis of Linear Factor Models using Large Dimensional Datasets of Individual Assets
I propose a research project to develop an econometric framework and methodology for the analysis of large-scale datasets of individual assets. I address unresolved issues in factor analysis that are relevant for empirical research in Finance. I work in a general class of linear models with observable as well as latent systematic risk factors. In particular, I focus on two open issues: testing for misspecification of approximate factor models, and the identification of systematic risk factors of individual assets. I develop estimation and inference methodologies that are feasible for datasets with large cross-sectional and time dimensions. I consider a large number of individual stocks as base assets, and I account for the unbalanced characteristic of these panels of data. I develop two empirical applications based on returns for about ten thousands US stocks from July 1964 to December 2012.