Escobal, J. y Castillo, M. (1993). Un indicador del nivel de actividad de la economía peruana en base a un modelo de variables latentes. Notas para el Debate, 7, 9-29.

The construction of an indicator of the level of economic activity for Peru is proposed here, based on a set of easily accessible time series (production, employment, income, etc.). The model assumes that the dynamics of these series respond to the evolution of a single common factor that is not directly observable and that underlies all of them (latent variable), in this case the level of activity. The first step is to perform statistical tests to determine whether it is possible to specify a single common latent variable in the system. Then a model is constructed in which the evolution of the chosen time series is explained by the dynamics of the latent variable, which, in turn, follows a first-order autoregressive process. Under the dimensional representation of state variables (State Space Form) it is possible to estimate the parameters of the system (the determination of the optimal values of the parameters is performed from the Expectation-Maximization/EM algorithm). Finally, using the Kalman filter algorithm, the evolution of the state variable, which represents the level of activity of the economy, is identified.