US GDP in Q1: better than it looks

The growth capacity of the world's leading economy is once again under suspicion. This time doubts are due to the decline in GDP in Q1, which reached 0.7% quarter-on-quarter in annualised terms.1 But are such fears justified?

Part of this weakness can be attributed to an exceptionally harsh winter, as witnessed by the snowstorms that hit Boston. The bad weather has substracted about 1.0 pps from output growth. The dockworkers' strike that paralysed activity on the west coast towards the end of Q1 is another temporary factor that also reduced GDP growth in Q1.

However, these calculations would place GDP growth at a still very modest 0.6%. So why the mismatch? A recent study by the Fed has revealed that the Bureau of Economic Analysis (BEA), the US statistics office that prepares the country's accounts, tends to considerably underestimate GDP growth in the first quarter of the year and overestimate it the rest of the year when adjusting the series for seasonal effects.

Specifically, since 1985 the BEA has calculated Q1 growth in GDP at 1.8% on average, compared with 3.0% for the other quarters. This pattern, which has become more accentuated recently, should not be seen as the seasonal adjustment of the series is meant, precisely, to eliminate fluctuations occurring systematically in some quarters.
The reason for this error is that, when the BEA produces the GDP series, it seasonally adjusts it at a highly disaggregated level and the seasonality that seems to be relatively insignificant in an individual series, once the different series are aggregated it becomes obvious that it is not. In fact, if we use the same programme as the one employed by the BEA to seasonally adjust the final GDP series, a significant seasonal pattern can be observed in the series, as already suggested by the average quarterly flash figure. Once the residual seasonality has been duly corrected, GDP growth in Q1 2105 becomes 1.1%. Moreover, if we take into account the effect of the bad weather and the strike, then output would have grown by 2.35%, a figure more in line with business indicators. As shown by the third graph, the level reached by the ISM manufacturing index corresponds historically with GDP growth rate of around 3%.

In short, when all the different pieces are taken into consideration, first quarter GDP looks more positive than suggested by the figure published by the BEA. In fact, having analysed the factors slowing down GDP in Q1, the conclusion is that growth is likely to have speeded up considerably in Q2.

1. Corresponds to the second estimate for the GDP figure presented by the BEA on 29th May 2015. On 24th June the third estimate was published, with a 0.2% annualised drop, quarter-on-quarter.