B1224
Title: An unrestricted MIDAS Poisson regression model
Authors: Talha Omer - Jonkoping University, Sweden (Sweden) [presenting]
Par Sjolander - Jonkoping University, Jonkoping International Business School (Sweden)
Kristofer Mansson - JU JIBS (Sweden)
BM Golam Kibria - Department of Mathematics and Statistics Florida International University USA (United States)
Abstract: An unrestricted mixed data sampling (U-MIDAS) model estimated using maximum likelihood (ML) for the Poisson regression model is proposed. An issue when using the standard U-MIDAS model is the overfitting due to a potentially large number of lags of the high-frequency variable used to predict the low-frequency regressand. Therefore, as a remedy, we suggest a regularized ridge approach. In terms of mean square error (MSE), we analytically prove the superiority of the ridge approach over the ML approach. Moreover, in a simulation study, we demonstrate that the ridge approach is superior in finite samples.
