Goldman Raises December Rate Hike Odds To 80%
Having pointed out a glaring error in today’s payrolls report, which indicated that there was at least one math error in calculating the average hourly earnings number, and as a result casts doubt on every other piece of data released by the BLS, we urge algos and the handful of carbon-based traders, to take anything released by the BLS with a boulder of salt, especially data on wage inflation, until the BLS provides an explanation for what is going on.
Until then, here is Goldman methodically going “by the numbers”, and validating the market’s reaction that sent December rate hike odds to the highest in one year, as moments ago Goldman chief economist, Jan Hatzius, revised his odds of a December hike from 75% to 80%.
Nonfarm payrolls fell 33k in September—considerably below expectations—however, we believe temporary hurricane effects likely explain all or most of the weakness. In fact, the employment report appears strong on net after taking into account hurricane effects, given the drop in the unemployment rate to a new cycle low and the upward revisions to average hourly earnings. We increased our Fed probabilities, with subjective odds of a December hike at 80% (vs. 75% previously).
And the breakdown:
- Nonfarm payrolls fell by 33k in September—113k below expectations—and growth in prior months was revised down by 38k on net. However, we believe temporary hurricane effects likely explain all or most of the weakness. We had previously estimated a drag of 125k from hurricane effects, and it appears to have been even larger: the BLS commissioner reported “a sharp employment decline in food services and drinking places and below-trend growth in some other industries likely reflected the impact of Hurricanes Irma and Harvey” (food service employment alone fell by 105k). Gauging the magnitude of the impact is difficult, but given the commissioner’s statement and given the sharp rise in the household “not at work due to weather” series, our working assumption is that all or most of the weakness was hurricane-related—and likely transitory. The state-level payrolls data released October 20th will provide substantial clarity on the magnitude of the impact. By industry, goods-producing industries added 9k jobs in September reflecting a rise in construction employment (+8k). Private service-providing employment fell 49k, as a 111k drop in leisure and hospitality payrolls was partially offset by growth in education and health (+27k) and trade, transportation, and utilities (+26k) jobs. Government payrolls rose 7k. The breadth of job gains weakened likely due to hurricane effects, with the payrolls diffusion index – the net share of industries adding jobs during the month – falling to 55.7% from 60.2%
- The household measure of employment was very strong, rising 906k in September following the 74k decline in September. The 939k gap between employment growth in the household and payroll reports was the largest ever excluding months with level adjustments to population controls. Household employment on a population- and establishment survey-adjusted basis rose only 7k, as the numbers of workers on unpaid leave from their jobs—which are included in household employment but not in establishment employment—rose by 908k (SA), presumably largely driven by the hurricanes. The unemployment rate fell in September to 4.22% from 4.44%, as the surge in household jobs more than offset the increase in the participation rate to 63.1% (from 62.9%). While the unemployment rate may in principle have been pushed down by hurricane effects (for instance if the response rate drops more among unemployed), we think this is unlikely because the BLS noted that “there was no discernible effect on the national unemployment rate” and because historically natural disasters have led to moderate increases in the unemployment rate. Another reason to doubt that the decline in the unemployment rate was due to the hurricanes is that household employment was strong (as opposed to the labor labor force being weak). The broader U6 underemployment rate fell 3 tenths to 8.3%, as the shares of marginally attached and the U3 rate fell.
- Average hourly earnings increased by a larger-than-expected 0.45% in September (mom), and a significant upward revision to July growth (+0.2pp to +0.5%) resulted in the year-over-year rate increased to +2.9% from a previously reported pace of +2.5% in August. While the composition effect due to the decline in the share of low-wage leisure and hospitality workers may have boosted September earnings, the upward revisions in prior months were large. Average weekly hours held steady at 34.4.
- Our preliminary wage tracker—which distills signals from several wage measures—shows 2.4% for Q3, up from +2.2% in Q2.
- We believe the headline payrolls miss is considerably less important than usual for the monetary policy outlook, because hurricanes clearly affected the data, other US growth data has been firm, and there are two more employment reports between now and the December meeting to make up for the weakness. We actually think the most important takeaway from the report was the upward revision to average hourly earnings, with wage growth now reported at just below 3%. Given this and given the drop in the unemployment rate to a new cycle low, we increased our Fed probabilities, with subjective odds of a December hike at 80% (vs. 75% previously).
Of course, if the BLS’ wage growth calculation is wrong, everything changes. Until then, here’s a look at where the market-implied odds of a December hike are: not surprisingly… right at 80%.