IMF-估计美国的每月国民帐户（英）-2025_53页_1mb

Summary
This working paper estimates monthly GDP and its eight subcomponents for the U.S. since 1950 using a Kalman filter approach. The methodology combines quarterly National Income and Product Accounts (NIPA), exact monthly data for some components (e.g., consumption), and other monthly indicators (e.g., trade, inventories). Generalized Method of Moments (GMM) is employed to estimate Kalman filter parameters, addressing parameter uncertainty and providing tight confidence intervals (less than 0.3% of GDP since the 1960s).

Key contributions include:

1. Monthly series directly comparable to quarterly NIPA data, derived from a linear stochastic state space model.
2. Confidence intervals accounting for both filter and parameter estimation uncertainties, validated using auxiliary models.
3. Evidence that monthly GDP peaks often align with NBER-recession start dates, but end dates remain less precise.
4. The baseline model shows minimal improvement from alternative specifications, with gains negligible in practical terms.

The estimated series are precise enough for use in applied research, enabling better analysis of economic activity at monthly frequencies. The methodological approach leverages available data to bridge the gap between monthly indicators and comprehensive quarterly statistics.