Model-Based Measures of GDP: 2005–2014

Like most macroeconomic variables, real gross domestic product is subject to measurement error. Because the U.S. Bureau of Economic Analysis lacks complete information at the time it publishes its initial GDP estimates, revisions are often substantial.2 Analysts concerned about the accuracy of these early estimates for expenditure GDP could focus instead on gross domestic income, the BEA’s measure of U.S. output on the income side of the national accounts.3 Conceptually, GDP on the expenditure side should equal GDP on the income side, and there should be no choice to make between the two series. As a practical matter, however, the two measures can differ by a significant amount because each measure is constructed using “largely independent ” source data, which themselves are “less than perfect ” [BEA (2014)]. Since November 2013, the Federal Reserve Bank of Philadelphia’s Real-Time Data Research Center has been publishing GDPplus, a new measure of U.S. real GDP growth that combines the BEA’s official estimates of expenditure GDP and gross domestic income. Based on the work of Aruoba, Diebold, Nalewaik, Schorfheide, and Song (2013, henceforth ADNSS), GDPplus represents an appealing complement to the BEA’s official estimates because it combines information in those estimates in a statistically optimal manner. GDPplus combines the information in expenditure and income GDP using a signal extraction method based on the Kalman filter. In the ADNSS framework, true GDP is treated as an unobserved variable. At th