"All Yours Janet": JPM Warns Fed About Using "Arguably Flawed Statistics"

Tyler Durden's picture

The December jobs report was an ugly mix of slowing employment growth and disappointing labor supply. JPMorgan's Mike Feroli doesn't mix his words in his brief report on today's ugly jobs data. While other 'economists' proclaimed we should "ignore it" or blamed it on the weather, Feroli notes for example, "It's hard to see how the weather -- or anything else -- was to blame for the 25,000 decrease in employment of accountants." But it his comments for the Fed that are most concerning as he worries, "the forward guidance decision could be even more difficult than the tapering call... lowering the unemployment threshold further would be doubling down on predicating policy on an arguably flawed statistic."

Via JPMorgan's Mike Feroli,

All Yours Janet

The step-down in nonfarm job growth from an upward-revised 241,000 in November to just 74,000 last month can partly be blamed on weather and, perhaps just as important, on the month-to-month noise in the series. Just as the hype over upside risks after strong November data was probably overdone, so too should the string of disappointing December data be taken in stride. Given the economic fundamentals we'd guess the underlying trend in job growth hasn't materially shifted. More troubling though is not what we are learning about business' labor demand, but what is happening in households' labor supply: the unemployment rate plunged 0.3%-point to 6.7% as the labor force participation rate fell another 0.2%-point to 62.8%.

So far, the fall in unemployment is not being accompanied by even the slightest hint of wage acceleration -- average hourly earnings were up just 0.1% last month -- but it does raise the risk that the economy may bump up against capacity constraints sooner than hoped. Bernanke's last meeting as Chairman will be a tricky one. We believe the Fed can convince itself there were enough special factors distorting this number that another $10 billion taper will be appropriate at the January FOMC meeting, though there may be some advocating a pause.

The forward guidance decision could be even more difficult than the tapering call. Rather than lower the unemployment threshold further -- which would be doubling down on predicating policy on an arguably flawed statistic -- we think the Committee will continue along the path of downplaying the significance of the unemployment rate in the setting of interest rate policy.

The weakness in the establishment survey was not just in the headline jobs number, but also in the ticking down in the average workweek to 34.4 hours, and the sluggish 0.1% gain in average hourly earnings, the lowest increase since the summer.

The big question is how much of the disappointment was weather distortion. The 16,000 decline in construction payrolls is an obvious candidate as a casualty of cold weather in the survey week. Another clue comes from the 273,000 who reported themselves as employed but not at work due to bad weather, about 100,000 more than an average December. Caution should be taken in simply adding this 100,000 to the nonfarm payroll number, as the nonfarm number counts people as employed so long as they were paid, whether or not they were at work. Our educated guess is weather may have taken 50,000 off payrolls. It's hard to see how the weather -- or anything else -- was to blame for the 25,000 decrease in employment of accountants. Another outlier was health care employment, down 6,000 and the first monthly decline in over a decade, undoubtedly a data point that will enter the civic discussion on health care reform. Going in the opposite direction, retail employment had a strong month, up 55,000, which tends to occur in Decembers following a late Thanksgiving (even after seasonal adjustment). Temp help was also strong, up 40,000. The weakness in average hourly earnings took year-ago wage gains down to 1.8%, partly a base effect as earnings late last year got a boost, perhaps ahead of scheduled tax increases.

Job gains as measured in the household survey came in a somewhat more respectable 174,000 and unemployment plunged 490,000, the most in three years. The decline in the participation rate undid the November rebound that occurred in the wake of the government shutdown. Most demographic groups saw participation decline, though it was more concentrated in persons with lower levels of educational attainment. Similarly the decline in unemployment was fairly broad-based. This decline occurred even before extended unemployment benefits began expiring at the end of December, and in fact was heavily concentrated in those unemployed less than five weeks, i.e. those who will not be affected by extended benefits. The unemployment rate of those unemployed six months or less, which some researchers have flagged as a more reliable measure of how labor market slack affects inflation, fell to 4.2%, a fair bit below the 4.9% long run average.