Despite ADP's blowout print this week, consensus January payrolls is 175k (somewhat below the 6- and 12-month averages), but Goldman Sachs expects a higher 200k print thanks to a combination of lower-than-usual year-end layoffs, favorable weather effects, and further improvement in labor market indicators.
Notably, Bloomberg Intelligence explains that over the past decade, consensus has tended to overestimate the January payrolls gain, but it appears to have corrected this bias over the past five years -- a period in which the average miss has been less than 5k. As such, the current consensus forecast implies a significantly weaker outcome than was suggested by the latest ADP employment survey, which showed a 246k increase in private payrolls.
The seasonal adjustment of the payroll data turns sharply negative in January, as recurring winter layoffs commence in a range of sectors. In recent years, the seasonal adjustment factor for January has averaged near -3036k. (In other words, employers typically layoff 3 million workers in January.)
Furthermore, as Goldman notes, the report will also be accompanied by the annual benchmark revision to the establishment survey as well as the annual introduction of new population controls in the household survey.
Overall Goldman forecasts that nonfarm payroll employment increased 200k in January, after an increase of 156k in December and 204k in November. Labor market indicators generally strengthened last month, with a drop in jobless claims between the survey periods to four-decade-lows and further improvement in the employment components of many service-sector and manufacturing surveys. The key labor market subcomponent of the consumer confidence report also rose to levels close to cycle highs. We also expect two temporary factors to boost January employment growth relative to surrounding months, specifically a weather-related rebound (on the order of 20-40k) and lower-than-usual year-end layoffs in the retail sector. On the negative side, we expect some pullback in transportation and warehousing payrolls following elevated growth in December, itself likely related to the secular shift toward online holiday sales.
Factors arguing for a stronger report:
- Weather rebound from December. Our analysis of NOAA (National Oceanic and Atmospheric Administration) weather station data suggests that unusually high snowfall during the December payroll survey period may have reduced job growth by roughly 20-40k. As shown in the left panel of Exhibit 1, snowfall in early January was much more in line with seasonal norms, suggesting scope for a weather-related rebound. Updating this analysis to incorporate regional granularity from the December state and local employment survey bolsters the case for January improvement, in our view. In addition to a 3k drop in the weather-sensitive construction category, the softness in overall December payrolls was particularly pronounced in the East North Central and Pacific regions, two parts of the country where snowfall was above seasonal norms during the week of December 17th. More specifically, in the right panel of Exhibit 1 we show that below-trend December payroll growth occurred in several states that also exhibited unusually high snowfall (in the weeks relevant for December payroll growth). Michigan, Illinois, Wisconsin, and Indiana stand out as four relatively large states where payrolls may have been depressed by weather in December.
Exhibit 1: Weather Likely to Boost January Payrolls Relative to December
Source: Source: National Centers for Environmental Information, National Oceanic and Atmospheric Administration, Department of Labor, Goldman Sachs Global Investment Research
- Jobless claims and retail layoffs. Initial claims for unemployment insurance benefits moved lower, averaging a four-decade-low 248k during the four weeks between the December and January payroll survey periods. While seasonally adjustment difficulties likely account for much of the drop, we believe some of the decline reflects underlying labor market improvement and relatively fewer January retail layoffs – both of which would suggest scope for a rebound in January payrolls growth.
- ADP. The payroll processing firm ADP reported a 246k rise in private payroll employment in January, up from +151k in December and well above expectations of +168k. In past research, we have found that large surprises in the ADP report tend to be predictive of the subsequent nonfarm payroll surprise. Additionally, we believe the above-trend growth in ADP construction employment (+25k) provides evidence for a rebound in weather-sensitive payrolls categories.
- Manufacturing sector surveys. The employment components of manufacturing surveys generally improved in December, with most now in expansionary territory. The ISM manufacturing employment component rose to a 30-month high (+3.3pt to 56.1), and the Philly Fed (+9.2pt to +12.8), Dallas Fed (+9.5pt to +6.1), Richmond Fed (+9pt to +8), and Empire State (+10.5pt to -1.7) employment indices also improved. On the negative side, the Kansas City Fed (-2pt to +6) employment component softened, and the Chicago PMI employment index declined into contractionary territory. Manufacturing payroll employment rose 17k in December, its first increase in five months.
- Service sector surveys. Most of the employment components of service sector surveys improved or remained at encouraging levels in January. The Philly Fed non-manufacturing employment index rose to a 1-year high (+2.8pt to +19.5) and the New York Fed index increased to an 18-month high (+4.5pt to +16.9, SA by GS). Meanwhile, the Richmond Fed (-4pt to +8.0) and Dallas Fed (-2.1pt to +4.8) measures pulled back modestly to levels still consistent with expansion. The ISM non-manufacturing survey will be released tomorrow, though the December employment index of 52.7 was consistent with moderate growth in service-sector jobs. Service sector payroll employment increased 132k in December and has increased 167k on average over the last six months.
- Job availability. The Conference Board’s Help Wanted Online (HWOL) report showed an increase in online job postings for a second consecutive month in January, though their total job ad count remains 9.5% lower on a year-over-year basis (vs. -8.8% in December and -14.9% in November). We place a limited weight on this indicator at the moment in light of research by Fed economists, which suggests the HWOL ad count has been depressed by higher prices for online job ads.
Arguing for a weaker report:
- Transportation jobs. Transportation and warehousing payrolls have seen elevated growth in December in recent years followed by softer growth or outright declines the next month, a phenomenon that may be driven by a secular shift toward online holiday sales. This winter seems to be exhibiting the same pattern, as transportation payrolls increased at an above-trend pace of +15k in December. With the holidays now behind us, we expect tomorrow’s report to show restrained growth or a modest decline in this category.
- Early-month storms in the South. Despite fairly unremarkable snowfall on a national basis during early January, one counterargument against the case for a weather-related rebound is the early-month snowstorms in the South, a region less accustomed to severe winter weather. While admittedly a risk, we would note that these storms generally occurred the Friday and Saturday before the survey week, and our analysis of snow cover – the depth of snow recorded by weather stations on a given day – indicates that most of the snow accumulation had melted by the Monday or Tuesday of the survey week (see Exhibit 2). Accordingly, we believe the negative payrolls impact of these storms is likely to be fairly modest, though storm-related absences could exert some downward pressure on hours worked (for employees with multi-week pay periods).
Exhibit 2: Most of the Snow Accumulated in the Southern Winter Storms Had Melted by the Monday and Tuesday of the Payroll Survey Week
Source: Source: National Centers for Environmental Information, National Oceanic and Atmospheric Administration, Goldman Sachs Global Investment Research
- Seasonals. Since 2010, January payroll growth has surprised negatively relative to consensus in five of the seven instances, with an average miss of -18k. While this may suggest some downside risk at the margin, we would note that severe winter weather was likely a factor in some of these observations (2010 and 2011, for example).
Neutral factors:
- Federal Hiring Freeze. The new administration’s announced hiring freeze for Federal works took place on January 23rd, over a week after the January payrolls survey period had ended. As a result, we do not expect any negative impact on the January employment report. In fact, we see some possibility of a positive impact if, for example, some federal departments anticipated the hiring freeze and front-loaded employment growth accordingly. That being said, we doubt we will observe evidence of a material impact in the aggregate data.
- Job cuts. Announced layoffs reported by Challenger, Gray & Christmas after our seasonal adjustment declined by 5k to 36k in January, though the level of announced layoffs remains moderately above the October lows.
* * *
We believe that the unemployment rate is likely to fall one-tenth to 4.6% (from 4.716% unrounded) – which would mark a return to the cycle low – driven by reduced year-end retail layoffs and our expectation of broader labor market improvement in January. We forecast average hourly earnings to rise 0.3% month over month and 2.8% year over year, reflecting firming labor markets and state-level minimum wage hikes. Our estimate incorporates a boost of 10 basis points from minimum wage hikes, which affected 19 states and increased the effective national minimum wage by about $0.25 (to $8.50 per hour).
Tomorrow’s employment report will be accompanied by the annual benchmark revision to the establishment survey as well as the annual introduction of new population controls in the household survey. In September, the BLS released a preliminary estimate of the establishment survey revision, which suggested a downward adjustment of 150k to the level of March 2016 employment (or -12.5k per month from April 2015 to March 2016). This preliminary estimate appears broadly consistent with the trends in the Quarterly Census of Employment and Wages, from which the benchmark revision is primarily derived. We would note that the composition of September’s preliminary estimates was somewhat unfavorable, with a larger downward revision expected in private payrolls (-224k) – concentrated in professional services (-133k) and retail trade (-120k) – offset by an expected upward revision to the level of government employment (+74k).
As a reminder the entire world is long stocks, short vol, and short bonds, so any surprise tomorrow that would upset The Fed's carefully choreographed 3-rate-hikes plan could spook a very one-sided ship.
But if that fails Buy The Fucking Payrolls Dip no matter what...