May Payrolls Preview: The Tiebreaker

After a poor March jobs report, followed by an April scorcher, the May payrolls report due at 8:30am on Friday will be the tiebreaker, not only for the current state of the economy where both soft and hard data have been deteriorating in recent weeks, but perhaps also for the June rate hike decision, which as the Fed noted in its May FOMC minutes, may not take place without "evidence" that the recent "transitory weakness" in the economy is over. Here are the consensus expectations for tomorrow's report:

  • May Nonfarm Payrolls Exp. 185K (Range 140K to 235K) vs April 211K
  • Unemployment Rate Exp. 4.4% (Range 4.30%-4.60%) vs April 4.4%
  • Average Hourly Earnings M/M Exp. 0.20%, vs April 0.30%; Y/Y Exp. 2.60%, vs April 2.50%

Payrolls Expectation

In terms of overall expectations, the consensus is looking for 185k nonfarm payrolls to be added to the US economy in May - the same as the April consensus - compared to 211k actual jobs added in April. That according to RanSquawk would be in line with the 185k/month pace seen in 2017 thus far. On one hand, there is potential for upside surprise, as per today's stellar ADP report which came in at 253K, far above the 185K expected. On the other, Goldman believes a favorable swing in the weather between the March and April survey periods boosted last month's hiring pace, and suggests the 211k pace of April job growth "likely overstates the near-term underlying trend", as such there will be payback in the May report. Also, Goldman cautions that the ADP measure has been running above official private payroll growth so far this year, by 60k per month on average, so take it with a grain of salt.

Unemployment rate

The unemployment rate is forecast to hold steady at 4.40%, matching the lowest reading recorded since 2001, and beneath the FOMC’s NAIRU projection between 4.70% and 5.00% (made in its March forecasts). A 4.4% print would be stronger than the Fed’s own year end forecast of 4.50%. If May unemployment stays at or near that level, it would be further evidence the economy has reached full employment and is at full capacity, meaning virtually everyone seeking work has found a job, even if that doesn't explain why wage growth remains anemics. If the rate dips lower, that could put upward pressure on wages and inflation, or alternatively it will prompt questions about the quality of jobs added.


As a result, most of the attention is likely to fall on the earnings data for signs of inflationary pressures. Average hourly earnings (AHE) are seen rising by 0.20% M/M, easing a touch from the +0.30% pace seen in April. On an annualised basis, the pace of AHE growth is seen rising by 0.10 ppts to 2.60%. In its latest Beige Book, the Fed stated that “most firms across the districts noted little change to the recent trend of modest to moderate wage growth,” though many firms reported offering higher wages to attract workers “where shortages were most severe.”  According to RanSquawk, HSBC notes that though wage growth has picked up, as of late, it remains sluggish when compared to previous cycles. Looking at the May wage number in particular, Goldman warns there may be a negative surprise pointing out that the May payroll period ended on the 13th, which is associated with meaningfully below-average wage growth.

Goldman's summary:

We estimate nonfarm payrolls increased 170k in May, a moderate slowdown from April’s +211k pace and modestly below the three-month moving average of +174k. While labor market fundamentals remained broadly stable – featuring a further decline in continuing jobless claims – recent deterioration in service sector employment surveys suggests hiring may be slowing at the margin. We also believe a favorable swing in the weather between the March and April survey periods suggests the 211k pace of April job growth likely overstates the near-term underlying trend, which we believe is closer to 175k (and should slow further as the economy moves beyond full employment). Relatedly, May is also an important hiring month, and labor supply constraints in some geographies and industries suggest some additional downside risk. On the positive side, both jobless claims and the ADP report suggest more favorable labor market fundamentals, and the end of the federal hiring freeze suggests scope for above-trend growth in federal employment.

On wages, Goldman warns there may be disappointment:

We estimate average hourly earnings increased 0.2% month over month and 2.5% year over year in May, reflecting the interaction of firming wage growth with negative calendar effects. The May payroll period ended on the 13th, which in our model is associated with meaningfully below-average wage growth. However, we are more constructive on wage growth generally, exemplified by the acceleration in the  employment cost index to a cycle-high pace in Q1.

Factors arguing for a stronger report:

  • Jobless claims. Initial claims for unemployment insurance benefits declined, averaging 241k during the four weeks between the April and May payroll survey  periods, a new cycle low. Additionally, continuing claims dropped by an encouraging 63k from survey week to survey week, roughly the same pace as in the prior month.
  • ADP. The payroll processing firm ADP reported a 253k increase in private payroll employment in April – above consensus expectations – suggesting a solid underlying pace of job growth. The ADP measure has been running above official private payroll growth so far this year (by 60k per month on average), and we believe the May ADP reading received a boost from the net strength in the financial and economic indicators also used in their model. These considerations make the task of teasing out the underlying signal from the report more difficult.
  • End of federal hiring freeze. The administration’s hiring freeze n for federal workers (excluding defense and public safety) went into effect on January 23 and concluded on April 11 – the Tuesday of the April survey week. Its impact on overall payrolls appears fairly limited, with average monthly payroll growth in these categories slowing from +3k in 2016 to -4k during the three months of the freeze. The impact also seems minor when compared to federal job growth during the 1981 federal hiring freeze at the start of the Reagan administration (see Exhibit 2). Assuming the 2016 trend in labor demand growth continued this year, the cumulative impact of the 2017 freeze was approximately -20k (on the level of federal payrolls). Accordingly, we see some scope for an above-trend reading in tomorrow’s report,  reflecting pent-up labor demand (we assume +10k for total government payrolls).

Arguing for a weaker report:

  • Service sector surveys. Service-sector employment surveys n have deteriorated somewhat in recent months, with the ISM non-manufacturing survey falling to 51.4 in April (from its recent high of 55.2 in February) and available May surveys weakening on net. Our overall non-manufacturing employment tracker fell to 53.4 in May from 54.4 in April, with declines in the Philly Fed and Richmond Fed employment subindices but improvement in the New York Fed and Dallas Fed measures. More encouragingly, the key labor market subcomponent of the Consumer Confidence report remained strong, rebounding 0.8pt to 11.7, not far from its cycle-high reading. Service sector payroll employment grew 173k in April and has increased 129k on average over the last six months.
  • Labor supply constraints. We view the labor market as close to full employment, with the unemployment rate roughly 0.3pp below its structural rate and yesterday’s Beige Book referencing increased reports of labor supply constraints. As slack diminishes further, this should exert both upward pressure on wages and downward pressure on job growth. From a hiring perspective, May is a particularly important month, with non-seasonally adjusted payroll growth averaging 838k over the last five May reports. As shown in Exhibit 3, we find that payroll growth tends to slow during late spring in years with relatively tight labor markets, as defined by an above-median Q1 employment gap (i.e. 2017).1. Labor constraints appear particularly binding in May (and August) in these years. One potential explanation is that the May payroll  period occurs after much of the start-of-year seasonal slack has been wound down (earlier in the Spring hiring season) but before the entry of students and recent graduates into the labor force (in late May and June).

  • Continued retail weakness. Retail employment growth has fallen n from its historical trend of 15-20k per month to -2k on average over the past six months. We believe the structural shift of retail sales from brick and mortar stores toward less labor-intensive e-commerce firms will continue to weigh on payrolls growth in that industry, with the impact on the order of 10k per month relative to its previous trend. This drag on retail employment has appeared particularly pronounced recently – with a 50k cumulative drop in retail payrolls over the last three months – and we note the possibility that weak brick and mortar sales trends in Q1 may be accelerating the pace of this structural shift. Similarly, we note the possibility that the weakness in April home sales and housing construction may have weighed on hiring in that industry.
  • Seasonals. Since 2010, May payroll growth has surprised negatively relative to consensus in four of the seven instances. While this is only slightly more than half the time, the average surprise has been fairly sizeable at -50k over this period. This may suggest downside risk to the extent the BLS seasonal factors have not fully evolved to reflect this tendency.
  • Job cuts. Announced layoffs reported by Challenger, Gray & Christmas after our seasonal adjustment rose sharply (+28k to 59k, a one-year high). Over half of the increase reflects a 20k layoff announcement at Ford Motor that was announced after the May payroll survey period. After taking into this account, the increase in May job cuts was more modest

Neutral Factors:

  • Return to Normal Weather. We believe the early-March winter storms likely exerted a meaningful drag on March payroll growth and provided a boost to April. Winter Storm Stella hit the Midwest and East Coast at the beginning of the March survey week, with the level of population-weighted snowfall during a March survey week at its highest since at least 2005. This suggests the April employment report may have benefitted from workers in the establishment survey returning to their jobs in some industries. Our preferred aggregate of weather-sensitive industries (construction, retail, and leisure and hospitality) also rebounded, to +66k from -17k in March. Accordingly, we believe headline job growth in April likely overstates the near-term trend, suggesting scope for moderation in May.
  • Manufacturing sector surveys. Employment components of manufacturing n sector surveys were mixed in May, with improvement in the ISM manufacturing employment component (+1.5 to 53.5), but deterioration in several regional surveys, including the Philly Fed, New York Fed, and Dallas Fed employment components as well as the Markit PMI subindex. Our overall manufacturing employment tracker pulled back to 0.6pt to 55.7, still a healthy level. Manufacturing payroll employment rose 6k in April, its fifth consecutive increase, and has increased 12k on average over the last six months.
  • Job availability. The Conference Board’s Help Wanted Online (HWOL) report showed a rebound in May online job postings (+4%) following April’s 1% pullback. However, we continue to place limited weight on this indicator at the moment, in light of research by Fed economists that suggests the HWOL ad count has been depressed by higher prices for online job ads.

* * *

Other observations:

Impact on Fed policy:

  • With the implied probability of a June hike at around 96%, it would likely take a horrific report to stop the Fed from lifting rates by 25bps. With the rate of joblessness below the Fed’s estimate of NAIRU, as well as its end-2017 target, it would likely look through a big headline miss, so long as wages don’t collapse.
  • Many Fed speakers have been paying particularly close attention to wages, observing that they have been a notable weakness.
  • Fed’s Kashkari (voter, dove) last month said there may be more slack in the labour market, arguing that stronger wage growth may pull more people back into the labour market, helping participation to rise.
  • Fed’s Evans (voter, dove) points out that across the board wage growth has not proceeded as quickly as the Fed would have thought. A sentiment that has also been touched on by the Fed’s Kaplan (voter, slightly hawkish) too.
  • Fed’s Williams (non-voter, centrist) went further, and described wages as “stubbornly soft.”
  • In terms of Fed hikes, even if wages missed, it may still not be enough to derail the Fed’s hike plans. Pantheon Macroeconomics has argued that in the previous cycle, the Fed lifted rates when AHE were running at 2.60% Y/Y, and it then accelerated sharply to 4.00% within five quarters.
  • Given rate changes operate with a four/five quarter lag, Pantheon says the Fed will be aware of the dangers of leaving it too late to raise rates.

Possible market reaction

  • The market is pricing in just one full hike in 2017, with the implied probability of two hikes slightly better than a coin flip.
  • An upside surprise in the Employment Situation Report may contribute to a repricing where the market converges towards the Fed’s forecasts, though with clear doubts about whether inflation can sustainably pick-up towards the Fed’s inflation goal (PCE has been easing as of late), it is unlikely the market and Fed’s view will converge.
  • Given past market reactions, a likely expression to an upside surprise may be a flattening of 2s10s, a sell-off in the long-end, which could help to lift the dollar.
  • Stocks are almost guaranteed to go up no matter the actual data.