2024-01-28-纽约联储-纽约联储_DSGE_模型_疫情后评估_18页_1mb

报告摘要

The New York Fed Dynamic Stochastic General Equilibrium (DSGE) model's real-time forecasting performance for output growth was comparable to that of professional forecasters (e.g., Blue Chip Economic Indicators and Survey of Economic Forecasters) before the Covid-19 pandemic, from 2011 through 2019.
After Covid, forecasting accuracy declined significantly for both output and inflation, with RMSEs being at least twice as large in the post-Covid period compared to pre-Covid.
Core PCE inflation forecasts were less accurate with the DSGE model post-Covid, partly due to its failure to fully capture inflation persistence and the impact of factors like cost-push shocks and distributive fiscal policies.
The model's accuracy deterioration was partly driven by incorrect predictions on key events, such as the nature of inflation transience, soft landing possibilities, and the pace of policy normalization post-Covid.

During Covid, the model incorporated temporary shocks and gradually adjusted for policy changes, including the shift to average inflation targeting, but still failed to accurately forecast economic resilience and inflation trends.
Notable forecast errors included overly pessimistic output projections in some contexts and over-optimistic inflation forecasts in others, often linked to financial conditions and monetary policy stance.
Post-assessment, the model is acknowledged to have limitations in handling certain economic shocks, prompting recommendations for enhancements, such as incorporating more heterogeneous agent behaviors.

Overall, the DSGE's performance is considered adequate for output forecasting but deficient for inflation, particularly in the post-Covid era.
Future work should focus on improving inflation dynamics and exploring alternative modeling approaches to better reflect real-world economic uncertainties.

展开完整摘要