NEWS / Foreign Exchange

An Autopsy On The March 2007 Employment Situation Report

To say that the headlines associated with the March 2007 Employment Situation report (see Asha??™s always excellent write-up of the details below) were a surprise to me would be an understatement. Of course, I was forewarned on Thursday by another trusted colleague, Amanda Mason, to be prepared for an upside surprise today because Canada also reported a considerablystronger-than-expected March employment report. Firstly, I will briefly discuss why the ???apparent??? strength in the March 2007 Employment Situation report was such a surprise. Secondly, I will perform an ???autopsy??? on the report to help better explain where the surprises were and whether we should expect more future strong reports.

Why was I surprised? For starters, the strength exhibited in the March report does not comport with our forecast of 2% real GDP growth in the first quarter. But, of course, our forecasts have been known to be errant on occasion. What is less errant, however, are facts. And two facts were that the employment indexes in the March 2007 ISM manufacturing and non-manufacturing reports declined, as shown in Chart 1. March factory payroll employment was reported down by the BLS. However, March construction and service sector payrolls not only increased, but increased significantly more than they did in February. This clearly was at odds with the ISM non-manufacturing report.

Yes, initial jobless claims did drop 20,000 in March, but continuing unemployment insurance claims were trending higher (see Chart 2). The claims data did suggest to me that there would be an improvement in labor market conditions in March, but nowhere as much an improvement as was reported in the March 2007 Employment Situation data.

Lastly, the Manpower Employment Quarterly Outlook Survey for the first two quarters of 2007 pointed to somewhat softer labor conditions.

All right, so much for expectations. Let??™s do an autopsy on the March 2007 Employment Situation report. I will start with the Establishment Survey (nonfarm payrolls). The biggest surprises in nonfarm job growth occurred in residential specialty trade contracting (up 11,000 in March vs. down 20,400 in February) and general merchandising retailing (up 35,800 in March vs. up only 6,900 in February). Dwellings under construction are in a definite downward trend, as shown in Chart 4, so it seems peculiar that builders would be increasing the number of carpenters, plumbers and electricians on their payrolls.

With regard to the 35,800 person increase in general merchandising retail, it seems odd that this accounted for all but 100 positions in the net monthly increase in total retail payrolls. Something very volatile appears to be going on in general merchandising hiring. General merchandise employment in relation to total retail employment has gone from 18.81% in November 2006 to 19.25% in March 2007, a very sharp reversal as shown in Chart 5. I wonder if there are not some seasonal adjustment issues in play here. What ever the case, if a lot of our job growth is occurring in retailing in general, then this is unlikely to result in strong consumer spending from income growth inasmuch as the average hourly wage in this sector is only $12.74, with only leisure and hospitality paying less ($10.19 per hour). If Circuit City??™s hourly pay cut plan is successful, other retailers might opt for a variation of it, which will lower the wages in this hotbed of employment growth even more.

Now, let??™s turn to the March 2007 Household Survey, which also provided a surprise in the form of a 0.1 point decline in the unemployment rate to 4.4% -- matching a cycle low. In terms of age groups, the largest decline in the unemployment rate occurred among teenagers, where the rate fell to 14.5% in March from 14.9% in February. But the ???adult??? unemployment rate also fell a tick to 3.9%, matching its cycle low. Did teenage employment increase in March? No, it declined by 59,000. What was driving force behind the sharp decline in the teenage unemployment rate? A 0.6 point decline in the teenage participation rate to a cycle low 41.6%. In fact, as shown in Chart 6, the March 2007 teenage participation rate of 41.6% is a post-WWII low. No wonder I can??™t get a neighbor kid to shovel the snow from my driveway or mow my lawn!

In the Household Survey, total employment increased by 335,000 in March. In what industry did most of the employment increase? Retail trade accounted for 73% of the increase in people working, adding 244,000 more folks at the cash registers. At least this tended to be consistent with the Establishment Survey. By the way, a lot of those people ringing up your sales had some college. Employment among people 25 years old or older increased by 319,000. In this ???senior??? set, only those with some college or an associate??™s degree found employment (510,000). This means that all others ??“ high school diploma or less, bachelor??™s degree or more ??“ saw declining employment. With so many people being employed in retailing at relatively low hourly earnings, I guess it is not surprising that there has been a sharp increase in folks working multiple part-time jobs now that their adjustable rate mortgages are resetting. Chart 7 shows that in March 2007 vs. year-ago, there was an 11.15% increase in people working multiple part-time jobs.

With a cycle low unemployment rate, one might think that there is an excess demand for labor. But if that were the case, average hourly earnings would be expected to be soaring. But they are not. As shown in Chart 8, both on a year-over-year basis and on a three-month growth basis, growth in average hourly earnings has decelerated. Again, part-time retailing jobs do not pay much.

One area where there was extreme inconsistency between the two surveys was in construction employment. The Household Survey reported an 89,000 person decline in people working construction while the Establishment Survey reported a 56,000 increase in construction jobs. If I were a betting person, in this case, I would place my money on the Household Survey of giving a more accurate snapshot of construction employment conditions than the Establishment Survey.

In sum, an autopsy of the March 2007 Employment Situation report suggests that labor market conditions are not nearly as robust as the headlines that accompanied the report.

Federal Funds Rate Steady, For Now

Civilian Unemployment Rate: 4.4% in March, matching cycle low reading of October 2006. Payroll Employment: +180,000 in March, net additions to January and February payrolls is +32,000

Hourly earnings: +0.3% from prior month, +4.0% year-to-year gain is part of a decelerating trend

Household Survey ??“ The civilian unemployment rate fell to 4.4% in March vs. a 4.5% reading in February. The unemployment rate has held between 4.4% and 4.6% for seven straight months. The unemployment rate is a lagging indicator of economic activity. Our research indicates that the jobless rate hovers around the cycle low for about 13 months before a steady upward trend becomes visible. It should not be surprising to see these low readings of the unemployment rate for a few more months. The 2001 recession is a case in point. The jobless rate moved in the narrow window of 3.8% and 4.1% between October 1999 and December 2000. It was 3.9% in November and December 2000, close to the cycle low of 3.8%, just before the Fed lowered the federal funds rate on January 3, 2001 by 50 bps. The jobless rate rose eventually to a peak of 6.3% only by June 2003.

Establishment Survey ??“ Nonfarm payrolls increased 180,000 in March following a gain of 113,000 in February. After revisions, 32,000 new jobs were reported for January and February. The bullish headline number for payrolls fails to translate into strong year-to-year gain compared with the trend in recent months. As shown in chart 10, the year-to-year increase in nonfarm payrolls was 1.46% in March, after a peak reading of a 2.14% gain in March 2006. Therefore, the underlying pace of job creation is much slower than what is inferred from one month??™s estimate of payroll employment.

Other highlights of the employment report: Construction employment rose 56,000 in March after a decline of 61,000 in February. The rebound appears to be weather-related because the increase in construction employment in the residential sector runs counter to the reports from a housing sector suffering from a severe setback in sales. Factory employment fell 16,000 in March, retail hiring increased 36,000, and government jobs rose 23,000.

The hourly earnings and employment data point to a moderate increase in personal income during March. The 0.2% increase in the factory man-hours index suggests a small increase in factory production. Aggregate man-hours rose at an annual rate of 1.5% in the first quarter. This combined with our GDP forecast for the first quarter (2.0%) implies a soft reading for productivity growth in the first quarter.

Conclusion ??“ The FOMC is on hold, for now. The bullish labor report contains several inconsistencies mentioned in this section and Paul Kasriel??™s comments. The Fed acted in the 2001 cycle by lowering the funds when the jobless rate was close to the cycle low and the year-to-year change in nonfarm payrolls was positive but showing a decelerating trend, a situation similar to the present time. However, with the FOMC still more worried about ???uncomfortably high??? core inflation, the surface strength of the March Employment Situation report will temporarily extinguish any notions of a near-term federal funds rate cut. Nevertheless, we continue to expect considerably below-potential growth this year and a gradual moderation in core inflation as rent increases slow. So, the FOMC is on hold for now, but we expect it will be motivated to begin dropping the federal funds rate in the second half of this year.