Earlier today, Goldman initiated its Q2 GDP tracking at 3.0%.
A few hours later, the Atlanta Fed, whose "preposterous" below-consensus forecast we first flagged in early March [11], and which nailed the Q1 GDP to within 0.1% of the final print, came out with its own first Q2 GDP tracking estimate [12]. The number: a recession-worthy 0.9%.
[13]
And cue panic as the entire Wall Street sell side scrambles to converge with the forecasting superstar Atlanta Fed, which unless sees a dramatic pick up in its own estimate, means that the US is looking at a 0.5% GDP for the first half, and anything less than 4% average GDP growth in the second half will lead to the weakest US growth in 2015 since 2011!
So for all those weathermen "economists" who scramble to do actual math in their GDP calculation instead of merely goalseeking meaningless numbers, we would like to remind you that we showed precisely how to recreate the Atlanta Fed number almost two months ago in "The Scariest Spreadsheet In Fed Possession Revealed [14]."
Most ignored it, and most were massively wrong.
Something tells us this time everyone will be poring for hours over the Atlanta Fed GDPNow model.
This is how the Fed bank presents its GDP predicting spreadsheet [15]:
Is GDPNow an official forecast of the Atlanta Fed or the Bank’s president?
No, it is not an official forecast of the Atlanta Fed, its president, the Federal Reserve System, or the FOMC.
Is any judgment used to adjust the forecasts?
No. Once the GDPNow model begins forecasting GDP growth for a particular quarter, the code will not be adjusted until after the “advance” estimate. If we improve the model over time, we will roll out changes right after the “advance” estimate so that forecasts for the subsequent quarter use a fixed methodology for their entire evolution.
When will forecasts of GDP growth in a particular quarter start being made?
About 90 days before the “advance” estimate for GDP growth for the quarter is released. GDPNow forecasts begin the weekday after the BEA’s “advance” estimate of GDP growth for the previous quarter is released. For example, the advance estimate of real GDP growth in the fourth quarter of 2013 was released on January 30, 2014. The GDPNow forecasts for real GDP growth in the first quarter of 2014 were tracked from January 31, 2014, until the day before the “advance” estimate released on April 30, 2014.
How frequently is the GDPNow forecast updated?
The model forecast is updated five or six times a month on weekdays, with at least one following six data releases: Manufacturing ISM Report on Business [16], U.S. International Trade in Goods and Services (FT900) [17], Monthly Retail Trade Report [18], New Residential Construction [19], Advance Report on Durable Goods Manufacturers [20], and Personal Income and Outlays [21]. Other data releases, such as Wholesale Trade [22] and Existing-Home Sales [23], are incorporated in the model as well and their impact on the model’s forecast will be shown on the weekday after one of the six data releases. The proprietary forecasts from Blue Chip Economic Indicators and Blue Chip Financial Forecasts shown in the chart are available from Aspen Publishers [24].
Where can I read about the methods and source data used in the model?
A detailed description is given in a working paper [25] describing the model. To summarize, the BEA’s NIPA Handbook [26] provides very detailed documentation on both the source data and methods used for estimating the subcomponents of GDP. The late Nobel Prize–winning economist Lawrence Klein pioneered many of the “bridge equation” methods used for making short-run forecasts of GDP growth using this source data; a 1989 paper [27] he coauthored with E. Sojo describes the approach. Ben Herzon, an economist at Macroeconomic Advisers, provides a bird’s-eye view illustrating how to use a bridge equation approach in practice to improve GDP forecasts in this 2013 presentation [28]. The econometric techniques used in our GDPNow model were heavily adapted from the GDP nowcasting models described in a 1996 Minneapolis Fed Quarterly Review article by Preston J. Miller and Daniel M. Chin [29] and a 2008 paper [30] by the Board’s David Small and economists Domenico Giannone and Lucrezia Reichlin.
Where can I find alternative forecasts of GDP growth?
For model forecasts from other Fed Banks, see the Minneapolis Mixed Frequency Vector Autoregression (MF-VAR) model [31], the Philadelphia Research Intertemporal Stochastic Model (PRISM) [32], and the Federal Reserve Bank of Cleveland’s prediction model for GDP growth based on the slope of the yield curve [33]. Moody’s Analytics [34] and Now-Casting.com [35] produce proprietary model short-run GDP forecasts. For survey-based forecasts, see the Philadelphia Fed’s quarterly Survey of Professional Forecasters, [36] which includes forecasts of real GDP and its major subcomponents. The Wall Street Journal’s Economic Forecasting Survey [37] occurs monthly but does not include forecasts of the subcomponents of GDP.
How accurate are the GDPNow forecasts? Are they more accurate than “professional” forecasts?
Since we started tracking GDP growth with versions of this model in 2011, the average absolute error of the model’s real-time forecast made just prior to the release of the “advance” (first) estimate of the annualized growth rate of real GDP is 0.68 percentage point. The root-mean-squared error [38] of the forecasts has been 0.94 percentage point. These accuracy measures cover “advance” estimates for 2011Q3–2014Q1. We have made some improvements to the model from its earlier versions and the model forecasts have become more accurate over time (the complete track record is here [39]). When back-testing with revised data, the root mean-squared error of the model’s out-of sample forecast with the same data coverage that an analyst would have just before the “advance” estimate is 1.15 percentage points for the 2000q1–2013q4 period. The figure below shows how the forecasts become more accurate as the interval between the date the forecast is made and the forthcoming GDP release date narrows.
Overall, these accuracy metrics do not give compelling evidence that the model is more accurate than professional forecasters. The model does appear to fare well compared to other conventional statistical models.
How are revisions to data not yet reflected in the latest GDP release handled?
In general, the model does not attempt to anticipate how data releases after the latest GDP report will affect the revisions made in the forthcoming GDP release. The exception is the “change in private inventories” subcomponent, where revisions to the prior quarter reading affect GDP growth in the current quarter. Users of the GDPNow forecast should generally use the forecasts of the change in “net exports” and the change in the “change in private inventories,” and not forecasts of the levels. Revisions to retail sales are used to anticipate revisions to real monthly expenditures in the “PCE control group [40]” and revisions to housing starts are used to anticipate revisions in the monthly value of private residential construction spending put in place [41].
Do you share your code?
At this point, no. However, the Excel spreadsheet [39] gives the numerical details—including the raw data and model parameters—of how the monthly data map into forecasts of the subcomponents of GDP
They may not share their code, but backing into how the Atlant Fed gets their GDP number is actually not difficult at all: the spreadsheet is even full of helpful guides so even the likes of Joe LaVorgna, Ethan Harris and Jan Hatzius can navigate it.
Just start here (link [42]).
[43]
... And explore, because unlike every economist "model" on Wall Street, this one actually does work.
And if the Atlanta Fed it is again correct to within 0.1% of the final print, forget about a rate hike in June, September or ever because the time for the Fed to unleash the next QE has officially arrived.