Goldman has kept mostly mum about tomorrow's NFP number for one simple reason: the BLS apparently has not leaked it to anyone, which explains why a rumor leaked by Larry Kudlow of all people has the power to move the EURCHF by 100 pips. Should tomorrow's BLS surprise there is just one conclusion that can be derived: any concerns of data leak from the BLS can be discarded. Which probably means that the NFP will be inline. Or not. As Goldman's Andrew Tilton explains there are just as many potential upside catalysts as downside, namely (to the upside) i) Jobless claims have moved lower since early July, ii) The Institute for Supply Management (ISM) nonmanufacturing employment index has held up, iii) The ADP employment report was reasonably healthy, while on the converse side we have i) Companies have begun to pull job advertising, ii) Manufacturing employment looks shaky, iii) Layoff announcements picked up in July. Taken together one can see why the Fed can push the data in either way, and the conclusion is that if the Fed wishes to pre-announce QE3, the NFP will be a major disappointment to where not even the robots can levitate the market higher, while on the other hand if Bernanke and kleptocratic company wish to extend and pretend for another two weeks in hopes that Mars, Alpha Centauri or Kang and Kodos will descend with an endless Jack in the Box full of Bernankebux, then it will print well over 100k. Alas, since every piece of news lately has been sold off we believe even a sizable beat will result in only a modest upside following by a massive fade, as the market refuses to rally without some evidence of QE3. Either way, for those who care about Goldman's thought, here they are, 5 short hours before the real deal.
Financial markets were in full retreat today, with the S&P500 index down 5%, Brent crude oil prices down 5%, US Treasuries continuing their massive rally (with ten-year yields falling to 2.4%, roughly equal to the low in late 2010), and the US dollar rallying.
Intensifying concerns about sovereign debt sustainability in Europe and a deterioration in the US and global growth outlook have spurred a spectacular selloff over the past week.
In this environment, the monthly US labor market report will receive even more focus than usual. The unemployment rate has risen for three consecutive months, to 9.2% in June. Historically, anything more than a very modest increase in the unemployment rate–35 basis points on the three-month moving average–has been associated with recession within six months. (See "A Recap of Our Unemployment Rate Recession Rule of Thumb", US Daily, August 1, for more details.) An increase tomorrow would put us very close to that threshold.
The labor market data have presented a very mixed picture over the past month. For those who prefer their glass half full:
1. Jobless claims have moved lower since early July. New claims averaged just under 430,000 per week in May and June, with the four-week average drifting down to 422,000 in the week of the July employment survey (July 10-16) and now at 408,000. Although the latest figures fall after the survey period, they seem inconsistent with a sharp breakdown in the labor market.
2. The Institute for Supply Management (ISM) nonmanufacturing employment index has held up. At 52.5, July's reading was equal to the average of the fourth quarter of 2010, when private services payroll growth averaged 142,000 per month. Looking at the nonmanufacturing employment index over its entire 15-year history suggests a similar pace of growth.
3. The ADP employment report was reasonably healthy. Market participants can be forgiven for skepticism regarding this report, which has had some notable misfires. Still, on average the first estimate from the Labor Department for private payroll growth has been slightly stronger than the first estimate from ADP.
For the "glass half empty" folk, some other indicators have been less encouraging:
1. Companies have begun to pull job advertising. A report from the Conference Board showed a drop in new job advertising; the total number of online job ads has fallen for two months for the first time since early 2009.
2. Manufacturing employment looks shaky. The employment index of the ISM manufacturing survey fell by more than 6 points to 53.5. Although this implies that a greater number of firms see increasing employment than declining employment, this level of the manufacturing index has historically been associated with outright declines in the Labor Department's measure of manufacturing employment. (One potential subtlety here is that survey responses generally come in the second half of the month, i.e. after the July employment survey was conducted.)
3. Layoff announcements picked up in July. The Challenger, Gray, and Christmas monthly report on layoff announcements tallied 66,000 job cuts in July, up from about 40,000 per month in the first half of the year. Although this is probably more relevant for future employment reports, it may indicate that companies had already become even more cautious on hiring.
Taken together, our models suggest a pace of payroll growth marginally better than the past two months, so we forecast +50,000 for tomorrow's report. This would also be consistent with the broad US data picture overall, at least through early July: our Current Activity Index was 1.3% as of the end of June, similar to the prior two months, and real GDP growth for Q3 appears to be tracking at a similar pace given the monthly pattern in Q2 and the very limited data in hand for July so far.
As if the employment indicators were not confusing enough, one additional wildcard is the possibility of a large swing in educational employment. Typically, employment in local government education falls by more than one million in July as the school year ends, so even small differences in the timing of layoffs could have a big effect on the seasonally-adjusted job total from this sector. The pace of job losses from the state and local sector has been fairly steady at about 20,000 per month since early 2010, but given the tough budget situation facing many local governments as property taxes erode, risks would seem to be on the side of a bigger-than-usual decline. Given that any major surprise here is unlikely to be repeated in August, market participants should probably discount any large moves from this area. Our expectations are for a private-sector job gain of roughly 100,000.
The unemployment rate will also weigh heavily in our assessment of the report, particularly given its signaling power for recessions.
It has risen from a recent low of 8.8% in March to 9.2% in the June report; the three-month average has increased 17 basis points from its low, about halfway to the 35 basis points that has historically been a consistent warning of recession.
While the trend in the unemployment rate in recent months has been worrisome, it has been unusual in at least one important respect. In contrast to the surge in unemployment during the financial crisis and the upturn last autumn, which were dominated by layoffs, this increase in unemployment is mostly due to workers voluntarily re-entering the labor market but having difficulty finding a job. The table below uses data from the Labor Department's household survey to show the composition of the rise in unemployment during four periods of rising unemployment: (1) the official recession period, (2) the further increase following the recession until the peak in late 2009, (3) the renewed uptick in the growth scare of summer and fall of 2010, and (4) the current episode.
Recent rise in unemployment due primarily to re-entrants
This observation is consistent with other known facts about the labor market. As noted above, the number of new jobless claims has been relatively modest over the past few months and has actually improved in the past few weeks. Other data sources have suggested that a larger-than-usual number of young workers left the labor force in the past couple of years to get additional training (see "Why Does the Household Employment Survey Look So Weak?", US Daily, January 11, 2010, for some discussion on this topic). Their return would therefore boost the unemployment rate (at least temporarily) given the weak hiring environment. If this is the primary reason for the increase, the rise in the unemployment rate may halt–at least temporarily–in July because the bulk of the typical springtime influx has ended.
Overall, the labor force has shrunk gradually so far in 2011, so the bar for stabilization in the unemployment rate looks fairly low. As we believe at least a few jobs were created in July, we expect the unemployment rate to remain steady at 9.2%.