Ellis W. Tallman

Overview

Ellis W. Tallman is a senior economist in the macropolicy section at the Federal Reserve Bank of Atlanta. His research examines VAR-based forecasting for policy analysis, with a focus on prior specification, conditional projections, and real-time forecasting challenges including staggered data release and temporal disaggregation of quarterly GDP to monthly frequency.

Key Contributions

• BVAR forecasting guide (Robertson-Tallman 1999): joint work with John Robertson documenting the Atlanta Fed's 6-variable monthly BVAR model; comparative RMSE evaluation 1986–1997 across six prior specifications; key finding that $\lambda_5$ (sum-of-coefficients) and $\lambda_6$ (cointegration dummy) explain most of the Sims-Zha prior's forecast advantage over the base Minnesota prior.
• Conditional forecasting (Robertson-Tallman 1999): clear exposition of the Waggoner-Zha (1999) minimum-MSE procedure for conditioning VAR forecasts on known data values at staggered monthly release dates.

Related

• John C. Robertson
• Minnesota Prior
• Conditional Forecasting
• Vector Autoregression
• Sims-Zha Prior
• Daniel Waggoner
• Tao Zha