Dynamic Factor Analysis For Panel Data: A Generalized Model

I develop a generalized dynamic factor model for panel data with the goal of estimating an unobserved index. While similar models have been developed in the literature of dynamic factor analysis, my contribution is threefold. First, contrary to simple dynamic factor analysis where multiple attributes of the same subject are measured at each time period, my model also accounts for multiple subjects. It is therefore applicable to a panel data framework (i.e. multiple attributes for multiple subjects observed over time). Second, it estimates an unobserved index for every subject for every time period, as opposed to previous work where a single unobserved index was estimated for all subjects for every time period. Third, I address the complexity of the model by developing a novel iterative estimation process which we call the Two-Cycle Conditional Expectation-Maximization (2CCEM) algorithm. The 2CCEM algorithm is flexible enough to handle a variety of different types of datasets. The model is applied on a panel measuring attributes related to the operation of water and sanitation utilities. The goal is to estimate a dynamic benchmarking index that will capture the financial and operational performance of these utilities