This week, the 2015 first quarter GDP growth came in at just a 0.2% annualized rate came in well below the expected 1.0% consensus forecast among Wall Street analysts. However, an innovative forecasting tool pioneered by economists at the Atlanta Fed was nearly spot on, forecasting 0.1% growth. GDPNow, is a real-time GDP forecasting methodology that sources data from previous months which has already been released, and attempts to predict using this data what GDP number the already available data implies.
This methodology, commonly referred to as ‘nowcasting’, takes advantage of the fact that GDP is released at nearly a month lag from quarter-end, weeks after the final jobs report for the last month of the quarter is released. It’s become an increasingly popular trend to ‘nowcast’ key economic indicators like GDP, though many nowcasts differ from their underlying data to their statistical method. Macroeconomic Advisers, which puts together one of the industry’s most popular GDP nowcasts, predicted a 1.4% growth rate for the first quarter, well above the actual number.
According to The Wall Street Journal, a number of Wall Street economists have increasingly started tracking the Atlanta Fed GDPNow methodology, citing economists like Eric Green from TD Securities, who follow the index closely.
Federal Reserve Bank of Atlanta President Dennis Lockhart (Photo Credit: Scott Eells/Getty Images)
The GDPNow methodology and small forecast error relative to official GDP figures from the Bureau of Economic Analysis
The GDPNow model forecast is updated on average five or six times a month on various weekdays, following nearly every major economic data release, including the Manufacturing ISM Report on Business, the U.S. International Trade in Goods and Services (FT900) Report, the Monthly Retail Trade Report, the New Residential Construction Report, the Advance Report on Durable Goods Manufacturers, and the Personal Income and Outlays Report.
The method works well, boasting mean squared errors relative to the official GDP figures that are smaller than most other competing 'nowcasting' methdologies. The GDPNow forecast becomes incredibly accurate within one month of the release of the official GDP figures from the Bureau of Economic Analysis, once the jobs report for the final month of the quarter is released, supplying the last piece of the puzzle for the GDPNow forecasting methodology.
Source: Federal Reserve Bank of Atlanta
Other popular ‘nowcasting’ forecasts are produced by other Fed Banks, including the Minneapolis Mixed Frequency Vector Autoregression (MF-VAR) model, the Philadelphia Research Intertemporal Stochastic Model (PRISM), and the Federal Reserve Bank of Cleveland’s prediction model for GDP growth based on the slope of the yield curve. In addition, Moody’s Analytics, Now-Casting.com, and Macroeconomic Advisors all produce proprietary model short-run GDP forecasts.
Atlanta Fed GDPNow predicts a 2Q real GDP number of 0.9% (as of April 30, 2015)
The Atlanta Fed's GDPNow is already predicting a GDP shortfall in the second quarter of 2015. According to the Atlanta Fed, "The initial GDPNow model forecast for real GDP growth (seasonally adjusted annual rate) in the second quarter of 2015 was 0.9 percent on April 30." This is somewhat hirer than the first quarter forecast of 0.1%, but would still be a suprising indication of continued weakness, should the official estimate from the Bureau of Economic Analysis fall in that neighborhood, well below the historical real GDP growth rate of 3% for the greater part of the 20th century postwar period.
Source: Business Insider
Perhaps, the first quarter’s weak GDP growth number of 0.2% should be met with less surprise given inclement weather (which has exerted a toll on the first quarter GDP numbers in the past like in 2014), the low price of oil negatively affecting employment in energy-focused states, and new regulations coinciding with one of the sharpest declines in the growth of nonresidential fixed investment since 2008. That being said, the Atlanta Fed's new GDPNow index, along with other rigorous 'nowcasting' methodologies, should give economists, policymakers and investors a more objective and reliable way to forecast GDP growth.
