Chalk up one more victim of the US government shutdown: the latest guesstimate on Q4 2018 GDP from the Atlanta Fed’s GDPNow model. The problem sits with the US Census Bureau, which today missed its release of residential construction data that feeds the algorithm here. By comparison, the Atlanta Fed is privately owned (as are all Fed branches), and the US Federal Reserve is not subject to the congressional appropriations process so it stays open as well.
In the lead up to quarterly GDP reports like January 30th’s first look at Q4 2018, we always highlight both GDPNow and the New York Fed’s Nowcasting report. Both take a purely model-based approach to estimating GDP, rather than the traditional “Blue Chip” economist surveys that dominate the market narrative. As for these models’ accuracy versus the humans, they are no better or worse on average. Rather, we like the algorithmic approach because it gives a purely numbers-based estimate with preset inputs rather than being subject to judgment biases.
So how do these models see Q4 GDP shaping up? Here’s what they have to say:
- The Atlanta Fed’s GDPNow algo is looking for 2.8% growth. This estimate has been stable since the start of 2019, so today’s missing residential construction data is not likely meaningful to the calculation. In fact, that 2.8% forecast sits in the middle of all the estimates the model has made since it started tracking Q4 economic activity in November.
- The New York Fed Nowcasting model sees 2.5% GDP growth for Q4. As with the Atlanta model, it has been predicting this outcome (within 0.1 percentage points) since November. For what it’s worth, the model also has a preliminary estimate for Q1 2019 of 2.1%.
Two points worth considering from this data:
#1: Both the 2.5% (NY) and 2.8% (Atlanta) estimates are broadly in line with consensus economist forecasts of 2.6% (the range is 2.2% to 3.1%).This does mark a slowdown from Q3’s 3.4%, but that’s no surprise given ongoing trade uncertainties and the ebbing of tax-cut related growth earlier last year.
#2: Neither the Atlanta nor New York Fed’s models showed any slowing of trend growth as Q4 unfolded, which stands in stark contrast to the US equity market tumble in the quarter. Click through to the data (link below) if you like, but both models reflect remarkably constant intra-quarter estimates of economic growth. Atlanta’s GDPNow estimates do bang around a little more than New York’s, but even then December had some of the highest forecasts (touching 3%) throughout all of Q4.
The big takeaway from all this: US economic growth is certainly slowing from its peaks last year, but the data from these regional Fed models betrays no sudden decline in activity. That’s not to say further slowing is out of the question. Trade uncertainty or further capital markets volatility (see our Beige Book note yesterday) could still dampen 2019’s outlook. But right now, the US economy is still on reasonably good footing.
Atlanta Fed: https://www.frbatlanta.org/cqer/research/gdpnow.aspx