GDPplus is the Federal Reserve Bank of Philadelphia's measure of the quarter-over-quarter rate of growth of real output in continuously compounded annualized percentage points. It improves on the Bureau of Economic Analysis's expenditure-side and income-side measures.
The expenditure-side measure (real GDP) is more commonly used by economic analysts, whereas the income-side measure, sometimes called real gross domestic income (real GDI), is little used, but each contains useful information. In particular, as proposed in Aruoba, Diebold, Nalewaik, Schorfheide, and Song (ADNSS) , one can view both real GDP and real GDI as noisy indicators of true, underlying, latent real gross domestic product, which can then be estimated using optimal filtering methods. Here we implement the ADNSS statistical framework. We call the optimal estimate GDPplus. (We construct GDPplus via the Kalman smoother.)
On this web page, we provide GDPplus data. We will update the page whenever we construct a new GDPplus vintage, whether due to the BEA's release of an additional observation or a revision of previously available observations for the real GDP and real GDI indicators. Note that our annualized growth rates use the formula for continuous compounding and are expressed in percentage points. We estimate the ADNSS two-equation model with the identifying restriction on the variance matrix for the shocks. We re-estimate the model for every new vintage using the method of maximum likelihood.
The GDPplus methodology is based on the measurement-error approach discussed in the following papers:
Aruoba, S.B., F.X. Diebold, J. Nalewaik, F. Schorfheide, and D. Song, "Improving GDP Measurement: A Measurement-Error Perspective ," FRB Philadelphia Working Paper 13-16 (May 2013).
Aruoba, S.B., F.X. Diebold, J. Nalewaik, F. Schorfheide, and D. Song, "Improving GDP Measurement: A Measurement-Error Perspective," Journal of Econometrics, 191 (2016), pp. 384-397.
Additional references are provided in the above papers.