27th May 2015 by Simon Ward
Researchers at the San Francisco Fed claim that US GDP seasonal adjustments are faulty and the economy grew by 1.8% at an annualised rate in the first quarter, rather than the reported 0.2%. Since unusually bad weather probably depressed performance, this could be taken to imply that the recent US slowdown is more apparent than real. Such an interpretation, however, would be unwise, for several reasons.
First, the official 0.2% growth estimate is likely to be revised down, based on subsequently-released data. The consensus expects a fall to -0.8%, suggesting “true” growth of only 0.8%, using the San Francisco Fed seasonal adjustment. (The revised number is due for release on Friday.)
Secondly, the Fed analysis is open to question. The corrected GDP numbers are obtained by applying a second seasonal adjustment to the official data. The results, however, are sensitive to the choice of statistical filter. The researchers use the standard X12-ARIMA procedure without adjustments. If allowance is also made for Easter / leap year effects, the double correction has a much smaller impact: first-quarter growth, for example, is estimated at 0.6% – see first chart.
Thirdly, the recent economic slowdown is consistent with monetary trends. A post in September noted that US real narrow money growth had fallen sharply and was signalling weaker economic data in early 2015. This fall has since reversed, suggesting that the US economy will reaccelerate strongly in the second half – second chart.
US economic weakness, therefore, is genuine but is likely to prove temporary.
Applying the Fed analysis to UK GDP data, incidentally, does not suggest a significant seasonal adjustment problem: double-adjusted numbers are closer to the official series than in the US – third chart.