Hyungsik Roger Moon - USC
Forecasting with Dynamic Panel Models
Date: 09/08/2016 (Thu)
Time: 3:30pm- 5:00pm
Location: Seminar will be held on-site: Social Sciences room 113
Organizer: Federico Bugni
Meeting Schedule: Login or email the organizer to schedule a meeting.
All meetings will be held in the same location as the seminar unless otherwise noted.
9:00am - Pickup from hotel and travel to Duke
9:30am - Matt Masten @ 202
10:00am - Luis Candelaria @ 223
10:30am - Empty
11:00am - Andrew Patton @ 228F
11:30am - Adam Rosen @ 221B
12:00pm - Lunch with Federico, Adam, and Takuya @ Faculty Commons
1:30pm - Federico @ 240
2:00pm - Arnaud Maurel @ 225
2:30pm - Fu Ouyang @ 223
3:00pm - Seminar preparation time @ 223
3:30pm - Seminar Presentation (3:30pm to 5:00pm)
5:30pm - Traver to dinner
6:00pm - Dinner with Federico and Matt @ Mateos
Additional Comments: Abstract: This paper considers the problem of forecasting panel data with a large cross-sectional and a small time-series dimension. We consider a linear correlated random effects specification and construct a predictor using Tweedie's formula for the posterior mean of the heterogeneous coefficients. This formula utilizes cross-sectional information to transform the unit-specifi c (quasi) maximum likelihood estimator into an approximation of the posterior mean under a prior distribution that equals the population distribution of the random coefficients. We show that the risk of a predictor based on a non-parametric estimate of the Tweedie correction is asymptotically equivalent to the risk of a predictor that treats the correlated-random-effects distribution as known (ratio-optimality). Our empirical Bayes predictor performs well compared to various competitors in a Monte Carlo study. In an empirical application we use the predictor to forecast revenues for a large panel of bank holding companies and compare forecasts that condition on actual and severely adverse macroeconomic conditions.