Once a year we get a slew of news about U.S. economic performance, when the Commerce Department’s Bureau of Economic Analysis (BEA) provides both its initial estimate for recent economic performance and detailed revisions on previous economic trajectories. The initial estimate for Q2 real GDP growth confirms what a great many other high frequency times series have made clear. The turn of the year moves to increase government spending, cut taxes and relax financial and environmental regulations have combined to catalyze a noticeable pickup for U.S. economic growth. Revisions to previous estimates of real output and income growth did little to change the output picture, but much to estimates of income. Proprietors—small businesses—did much better than initially estimated over the 2012-2017 period. And wage earners, we now know, did much better in 2017, than initial estimates suggested. Lastly, in a major change of heart, the BEA blessed the geek community with raw not seasonally adjusted data. These data allow adventurous analysts to do their own seasonal massaging, which, for the recent past, we think, allows for eliminating the crazy Q1 swoon for growth that was a reality for the previous half dozen years.
Causes of Faster Growth
The administration’s narrative for the jump to 4.1% for Q2:2018 real GDP growth is that it indicates the success of Trump policies in pushing the economy to a permanently faster growth trajectory. This view gets short shrift from mainstream economists. The supply-side argument for recently enacted tax cuts is centered on increasing the economy’s long run expansion capacity—by boosting productivity growth. Nothing that we see over a quarter or two will shed much light on this question. And such an increase in productivity growth strikes us a very implausible.
However, in their zeal to criticize voodoo supply side dogma, Keynesian economists overlook the standard textbook implications of a tax cut and spending increase: growth accelerates. Faster growth will not last for ever, but at the same time it is not simply a mirage. The reality of higher growth, supported by fiscal stimulus, can be seen in other important data, including and especially jobs growth. Non-farm payrolls, over the first half of 2018, rose by 215,000 per month, compared to 165,000 per month—the expectation for 2018 job gains before policies were changed. First half real GDP growth, 3.1%, compares with a 2.1% gain for calendar year 2017. That looks like genuine acceleration to us.
From a mainstream economic perspective, the policies enacted were perfectly perverse (see Barbera, International Economy Magazine, forthcoming). After insisting on fiscal austerity for 6 years of tepid expansion, amid the Obama Presidency, the GOP delivered big fiscal stimulus with the jobless rate at 4%. The policies also exacerbated after-tax income inequality. Lastly, our strong hunch is that the policies will do little to spur long term productivity performance.
The policies, however, have done what standard models suggest they would do—they have elevated the near-term trajectory for U.S. real output and employment growth. The multiplier is smaller than it would have been when monetary policy was stuck at zero lower bound, but it isn’t zero.
Jobs growth, at present, is running well ahead of almost anyone’s definition of a long term sustainable pace. We think that raises the probability that destabilizing excesses will flourish in the quarters ahead. And the history of unwinding destabilizing excesses is not pretty.
Wage Earners: Not As Bleak As Originally Estimated
The BEA revisions to 2012-2017 estimates, as we noted above, did little to change the picture for output growth, but much to change estimates of income. Proprietor’s income was substantially higher than previously thought throughout the period, lifting the estimate for personal income, and thereby lifting the estimate for the savings rate. More important, from our perspective, the 2017 estimate for wage income jumped. The 4.6% revised estimate for 2017 wage growth makes much more sense in a low a falling jobless rate world. The increase remains soft, given the end of year 4% jobless rate. Nonetheless, the original estimates told a story of flagging compensation amid a super low jobless rate. The new data suggest that the firming jobs market is lifting worker compensation.
|Wage Income (yoy)||Wage Income (yoy)|
Raw Data: Grist for the Mill of Enterprising Economists
Wall Street economists have noted a puzzle in GDP data in that growth has appeared systematically slow in the first quarter of the year. This led BEA to undertake an overhaul of its seasonal adjustment procedures. The final step of this overhaul was included in the 2018 benchmark revision. In this revision, the first quarter puzzle got less pronounced, but is still there. The root of the problem of residual seasonality is that BEA seasonally adjusts at a very disaggregate level and decides not to seasonally adjust some components at all. This may be the right decision for those components, but if these components have some correlated seasonality, material seasonal patterns may emerge in aggregate quarterly data.
As another part of the 2018 benchmark revision, BEA compiled not seasonally adjusted quarterly data—both real and price indexes. Before this, they unfortunately only published seasonally adjusted data (until 10 years ago they had released unadjusted data, but only nominal). Users can now take the unadjusted data and seasonally adjust it themselves. Of course, there are many modeling choices involved in seasonal adjustment, and there is alas never going to be any unique right answer. See Seasonal Adjustment of NIPA data for more discussion of the impact of the 2018 benchmark revision on seasonal adjustment, and an attempt at directly seasonally adjusting the data. The average absolute difference between currently published data and this directly seasonally adjusted data is 1.1 percentage points in quarter-over-quarter annualized growth rates. Whereas the published growth rates for the first two quarters of 2018 are 2.2 and 4.1 percentage points, respectively, with direct seasonal adjustment in this paper, the growth rates are instead 3.1 and 4.0 percentage points. More broadly, direct seasonal adjustment eliminates the first quarter growth puzzle.