Luca Sala, Professor of Economics at Bocconi University, presented a series of lectures on factor models in March at the Economic Modeling Development Center, NAC Analytica. Factor models are widely used in finance, macroeconomics and big data to explain the behavior of a large number of random variables in terms of a smaller number of common factors. More specifically, factor models can be applied to extract business cycles, real interest rates, potential output and other latent variables important for forecasting and modeling the economy of a country.

Lectures also covered Principal Component Analysis, its comparison with factor models and the conditions under which they are equal. Furthermore, the EM-algorithm (expectation-maximization) for model estimation, application of Kalman filter/smoother and Vector Autoregression (VAR) were considered for using in nowcasting.

Finally, the applications of factor models were studied by considering the seminal works by Stock and Watson (1989) who applied factor models to extract business cycle indicators, Forni and Reichlin (2001) who determined commonalities among European regions, and Giannone, Reichlin and Small (2008) who used factor analysis in big data for nowcasting GDP growth.