Yesterday's very disappointing initial claims number was quickly forgotten as algorithms latched on to any positively sounding headline out of Europe in order to push the Dow over the mythical 12,000. Alas, this is very shortsighted, because as Bloomberg economist Joseph Brusuelas is quick to point out, based on historically close Claims-NFP correlations, the June NFP number will be a big miss to expectations, and print in the 75-125K range. This ugly number which will merely further cement the case for further monetary or fiscal stimulus (and forget the latter), will come just in time for the Manufacturing ISM to print sub 50, and send the confirmation that we have all been waiting for that the US economy is now officially contracting.
Brusuelas explains why next month's jobs number could be just as ugly as the massively disappointing May NFP:
While U.S. initial jobless claims seem to be resuming their gradual trend downwards, the current level suggests another weak month of labor demand in June. That will probably be on display two weeks from today, when the monthly non-farm payroll report is published.
The four-week moving average remained at 426,000 for the week ending June 18. This is the week that the Bureau of Labor Statistics surveys payrolls, and the figure is essentially the same as it was during the May payroll survey week.
A weak payrolls number for May is supported by a Bloomberg Economic Brief model that examines the relationship between jobless claims and nonfarm payrolls — specifically, what level claims must decline to for payrolls to show sustained growth. The model is based on a regression of the monthly change in total employment against the four-week moving average of initial claims.
The model indicates that the level of initial claims necessary to support sustained labor demand sits at roughly 415,000, below the current level of claims.
Technical, seasonal and holiday issues may result in volatile claims data for the next few weeks. A Labor Department official said after the June 18 report that claims in six states all had to be re-estimated due to technical problems.
The July 4 holiday will likely also distort claims, since this year it falls on Monday, the day claims are usually reported.
Seasonal adjustments to the data may have the biggest effect. Auto shutdowns were pulled forward earlier in the year than usual as a way to handle supply shortages following the Japanese earthquake. The normal downwards seasonal adjustments to claims to remove distortions from the shutdowns may result in a sharp decline in reported claims in July, when auto production is expected to increase at 20 percent or more.
And here is the correlation between the two graphically.
Incidentally, the topic of weak claims forced Goldman Sachs to also release a note to clients, seemingly to calm them down about the 11th consecutive weekly print of 400K+ in claims.
Initial Claims are an important signal at times of cyclical uncertainty
Regular followers of our global macro and markets research will know that we attach a large degree of importance to the data on initial claims – as a timely indicator of the underlying momentum of the US economy, and as a signal for turning points in markets. We first explored this idea rigourously in January 2009 – in “When Markets Turn”, Global Economics Weekly, 09/02 – and found that around key cyclical turning points, risky assets like equities or our WF Growth basket tended to respond to the early signs of shifts in the growth momentum as captured by the data on initial claims for unemployment insurance.
It is instructive too that in this current episode of growth worries, initial UI claims hit a low-point in late-February and early-March and started to climb (i.e. started signalling a slowdown) ahead of other cyclical indicators and coincident with downdraft in the Wavefront US Growth basket, with the overall index itself lagging. So in this case too, claims have acted as the ‘canary in the coal mine’, and so the retracement in this data even as some of the other top-tier cyclical indicators are still worsening is intriguing.
In past research we have also noted that the strength of the signal for markets from shifts in data like claims is most clear around key cyclical turning points. We are probably not at a key cyclical turning point at the current juncture, although it certainly feels like we have been through more than our fair share of cyclical twists and turns two and a half years into the global recovery – a notion that Dominic Wilson explored in the GEW (“Are Cycles Getting Shorter?”, 11/23) this week. However, the unusual abruptness of the slowing in the macro data globally, and specifically in the US, means that there is a high premium to gauging the true underlying momentum of the US economy and the degree to which it is likely to recover from transient influences such as Japanese supply and weather disruptions and a fading oil shock. And while we focus on a host of macro information – hard data, surveys, and our own proprietary and co-incident and leading indicators – to arrive at that judgement, most of these are typically available on a monthly frequency, whereas initial claims have the great virtue of being available at a weekly frequency.
This higher frequency of the claims data is all the more useful currently given that we know that some of the monthly data – cyclical surveys for example – are likely to still be contaminated by the supply chain disruptions emanating out of Japan. And so for top-tier cyclical surveys like the ISM survey, it may not be until the July survey results, out in early August, that we potentially see any signs of stabilization, which are tentatively evident in production plans in Japan and in the weekly production statistics of Japanese manufacturers of transportation equipment in the US.
Pinning it all on Claims – a 30% to 50% retracement of the data weakness
Consider the following question. If we only focused on the weekly initial claims numbers, what would that tell us about the current momentum in the US economy, and the extent to which it is likely to recover from its abrupt slowing over the past three months or so?
Starting with the raw claims numbers, these were at their lowest point in late-February at 375k. From that point on they increased to a maximum of 478k in late-April – mostly on the back of transitory factors to do with auto disruptions and seasonal adjustment issues as described by our US economists (US Daily: Rise in Jobless Claims Mostly Temporary, 16 May 2011). In subsequent weeks UI claims have printed lower with the latest number yesterday coming in at 429k. This suggests a roughly 50% retracement relative to the roughly 100k deterioration from trough (375k in late-Feb) to peak (478k in late-April). Given the noise in the week-to-week fluctuations, we often look at the four-week moving average of this data. On this basis, claims bottomed at 389k in mid-March, before increasing to 440k in mid May, and the latest observation here is 426k. This implies a roughly 30% retracement relative to the trough to peak deterioration. So with all the caveats that come with extrapolating from such a narrow slice of data, based only on this one indicator, one would conjecture that the economy should regain 30% to 50% of the momentum it lost.
Our US Current Activity Indicator (CAI) suggests that the growth rate of economic activity, broadly defined, in the US slowed from close to 4.1% qoq annl. in March to 1.7% qoq annl. in May, a fall of about 2.5 percentage points. Assuming that this is the low point – and Philly Fed and Empire surveys suggest that the June outturn of our CAI may be somewhat lower still – then the 30%-50% retracement in the claims data suggests that if broad activity growth improves to the same degree, it may recover by between 0.8 and 1.2 percentage points to somewhere in the ballpark between 2.5% and 3% qoq annl. This would leave activity growth in the vicinity of, or just a little below, our 3.25% qoq annl. forecast for GDP growth in the second half of 2011.
Of course there are several reasons to skeptical about any of the precise numbers from such an exercise, and arguably we are taking the claims data “too seriously”. There is no reason why the initial low level of claims is a ‘fair’ or appropriate benchmark, given that it may also have been distorted by special factors. And while the relationship between activity growth and initial UI claims is tight, it is surely not perfect, with probably at least a short lag. The aim of this calculation therefore is to simply illustrate the extent of recovery that would be likely in activity growth if it was commensurate with the retracement in the initial claims data in a simple way.
Bottom line, no matter how you spin it, a very ugly NFP number is coming. The implications of this drop in jobs growth on Treasury tax revenues is the next topic that will be investigated.