Why Yellen Might Turn Her Back on a Labor Market Barometer She Once Championed
In a 2014 speech, Federal Reserve Chair Janet Yellen drew attention to one comprehensive metric that would serve as a better pulse on the state of the jobs market than the unemployment rate alone: the 19-part labor market conditions index, or LMCI.
This composite barometer draws on a host of data, including employment, underemployment, workers' wages and average workweek length, job openings, as well as hirings and firings, to produce a summary statistic showing if conditions in the labor market improved from one month to the next.
And what it's showing looks to be cause for concern: The LMCI has declined for five consecutive months, falling to its lowest level since the aftermath of the financial crisis in May.
Yellen's June 6 remarks, however, contained no reference to the LMCI, instead indicating that she believed "we are now close to eliminating the slack that has weighed on the labor market since the recession."
This prompted some economists to wonder whether Yellen had turned her back on the labor market indicator she helped champion. However, a deeper dive into how the LMCI is compiled and what goes into the index offers some clues as to why this streak of declines might not have the chair on edge.
Not necessarily indicative of an economic slump
Analysts at Bespoke Investment Group note that in the 1985 low-oil environment as well as in 1994, there were larger drawdowns in the LMCI that did not presage a recession, but proved to be "mid-cycle pullbacks."
"We continue to see signs the labor market is not as weak as the soft payrolls print might suggest; wage growth remains solid and accelerating, while there are no signs of businesses cutting payrolls," wrote Bespoke.
A questionable barometer...
Moreover, other broad-based metrics of the labor market's health have diverged from the LMCI.
It's "hard to take [this trend in LMCI] seriously when its pace was average during the most extraordinary participation and employment-to-population ratio surge of the cycle," tweeted Ernie Tedeschi, economist at ISI Group.
While this indicator used to lead changes in the Congressional Budget Office's employment-to-population ratio gap, its predictive power waned in this regard over the past year, he observed.
..with faulty signposts
And of the various metrics that go into the LMCI, not all of them agree with their closest and more prominent cousins. Notably, the "help wanted" portion of the LMCI draws upon data from the Conference Board Help Wanted Online series rather than the Bureau of Labor Statistics' Job Openings and Labor Turnover Survey.
The former trended sideways for most of 2015 before starting to slide; the latter is near cycle highs.
Anomalies in the data might be contributing to the mixed signals these indicators are sending. After December's outsized drop in the Conference Board's index (which also affected January's reading) was attributed to a "single event," the agency advised that users "may want to consider this as a non-economic event for analysis purposes," as Bespoke Macro Strategist George Pearkes pointed out.
The use of the Conference Board metric is presumably to address continuity concerns, as a version of this index based on job postings in newspapers goes back to the 1950s.
But growth is slowing
Goldman Sachs Group Inc. Economists Daan Struyven and Zach Pandl also found that a sizeable portion of the decline in the LMCI can also be attributed to second derivative deterioration — or a slowdown in the rate of improvement.
"The LMCI inputs are detrended [that is, have cyclical elements stripped out], and the estimated trends likely 'soak up' some of the growth in labor market activity (such that only growth in excess of the trend contributes positively)," they wrote. "By summarizing the first difference of detrended indicators, the LMCI in effect reflects a combination of the rate of change in labor market conditions — the first difference —as well recent acceleration or deceleration—the second difference."
That is, inputs in the LMCI don't merely tell us how conditions in the labor market are changing; they give a read on how the conditions are changing relative to an estimated trend based on a presumed statistical cycle.
This factor has fueled some of the LMCI's recent softness. Think of it like the monthly manufacturing purchasing managers' indexes: A move from 53 to 51 suggests that the sector is expanding—but at a slower pace than the previous reading.
However, the economists acknowledge that the downturn in hiring plans from a National Federation of Independent Business survey, the decline in the average hours worked for production employees, and the drop-off in the hiring rate are suggestive of more genuine weakness in the job market.
"The U.S. labor market may be in better shape than in May 2009, but it is also not firing on all cylinders," conclude Struyven and Pandl.