CMStatistics 2022: Start Registration
View Submission - CMStatistics
B0276
Title: Modal Grids in MIDAS estimation with large number of regressors Authors:  Lorenzo Frattarolo - European Commission Joint Research Centre (JRC) (Italy) [presenting]
Abstract: MIDAS estimation handles regressors with lower frequency using temporal aggregation with a parametrized weight distribution. Once the aggregation is done, estimation is equivalent to OLS. The proposed method exploits this feature and, given the weight function, computes a grid of weights such that each set of weights has its mode on a different lag. Then aggregation is performed for each set of weights and each regressor, resulting in a number of new aggregated regressors equal to the number of original regressors multiplied by the number of weight sets. The selection of aggregated regressors is then performed using the generalized least squares screening (GLSS). Values of parameters of the weight function originating the most significant aggregated regressors are stored and reused as initial values in a final maximum-likelihood estimation of the MIDAS regression. This methodology allows pre-selection among a large number of variables while maintaining contributions from a wide distribution of lags in the final estimation.