Page 285 - Special Topic Session (STS) - Volume 3
P. 285
STS544 Baoline C. et al.
Figure 2: Revision comparison from estimated bridge equation
model for selected PCE services
5. Conclusion
In this study we have demonstrated that bridge equation framework and
bridging with factors model are potentially useful methods for compiling
advance estimates of detailed PCE services, because these methods allow all
available information on the dynamics of the quarterly target variables and the
monthly indicators to be incorporated in the estimation. The one-step-ahead
out-of-sample predictions from these two models resulted in reductions in the
RMSR. However, we do not see reductions in revisions for every component,
nor in every period. To explore further improvements in accuracy, we plan to
experiment with various forecast combination and model averaging
techniques.
References
1. Bańbura M., D. Giannone, M. Modugno and L. Reichlin, (2013) “Now-
casting and the real-time data flow, European Central Bank, Working
Paper Series: no. 1564, pp. 1-54, European Central Bank.
2. Giannone, D., L. Reichlin, and D. Small (2008): “Nowcasting: The real-time
informational content of macroeconomic data,” Journal of Monetary
Economics, 55(4), 665–676.
3. Klein, L.R. and E. Sojo, (1989) “Combinations of High and Low Frequency
Data in Macroeconometric Models,” pp. 1-13 in Economics in Theory and
Practice: An Eclectic Approach, L.R. Klein and Marquez (eds.), Kluwer
Academic Publisher.
4. Kitchen, J. and R. Monaco, (2003) “Real-Time Forecasting in Practice: The
U.S. Treasury Staff’s Real-time GDP Forecast System,” MPRA Paper: no.
21068, U.S. Department of Treasury.
5. Higgins, P., (2014) “GDP Now: A Model for GDP ‘Now-casting,’” Working
paper: 2014-7, Federal Reserve Bank of Atlanta.
274 | I S I W S C 2 0 1 9