Why the Fed is Flummoxed by the U.S. Labor Market:
A New Analysis of Employment in the Era of Secular Stagnation and Global Oversupply
U.S. Jobs in 2014―A New Conundrum
Source: Bureau of Labor Statistics and author’s re-indexation.
Aggregate payroll growth in 2014 has been concentrated in low wage sectors and, for reasons discussed in this report, is not transmitting effectively to consumption growth or accelerating GDP growth.
At times I imagine Fed Chair Janet Yellen walking purposefully within the cloister of the Fed’s Eccles Building in Washington muttering the words of Ludwig Bemelmans’ Miss Clavel (of “Madeline” fame)….”Something is not right!”…as she moves from meeting to meeting. A few months ago Chair Yellen revealed that she had misgivings about the jobs market―citing evidence of “labor slack.” More recently, as equity markets gyrated wildly in October, she focused on global deflationary concerns. By the end of the month, however, she crafted a post-meeting statement from the Federal Open Market Committee, taking comfort in a lower unemployment rate and a number of other data points indicating economic improvement.
Yet we are left with the feeling that Yellen is profoundly uncomfortable about declaring victory over a seven year economic slump and moving on to monetary business as usual. Our Federal Reserve Board chair is, after all, one of the best labor economists in the profession and must see that present circumstances in the jobs market just don’t feel like anything approaching normal.
One elephant in the room is a set of statistics describing the participation in the labor force of those who are employable―the so-called Labor Force Participation Rate (LFPR)―which remains at low levels unseen since the early 1970’s, when two income families were not yet the norm. The LFPR, together with the ratio of employed individuals to the population of employable people (the Employment-Population Ratio), are flashing bright red warnings about substantial slack in the U.S. labor force. Such slack would seem to indicate that there is little concern to be had about sustainable inflation in wages and, moreover, that the increased wages that are necessary for the vibrant recovery and reflation of the U.S. economy are not likely to manifest themselves any time soon. Much has been written, by me and many others, about these two ratios of late. So I won’t tie up the reader here with information that has already reached the mainstream press.
The other, more brazen but less meaningful elephant that is standing in the corner of the Fed’s boardroom, is the unemployment rate itself, now down to 5.9% from a peak of 10.0% during the Great Recession. Not quite normal, but given the magnitude and speed of the decline, something on which to ruminate. I don’t dwell on this figure because the bulk of the reduction in the headline unemployment rate is due to the decline in the LFPR; it is a breathless statistical aberration, not an indication of vibrancy. The creation of around 9 million private sector jobs since the Great Recession, is of far greater meaning than the unemployment rate. But these new jobs are offset to in part by continued population increases. That wages have not risen, and in most sectors have fallen relative to inflation for a considerable period of time (even with historically low inflation), is proof that these broader measures are not accurately telling the story.
During 2013, job creation did not transmit efficiently into growth because of the quality of jobs created. Simply put, a significant majority of jobs created last year were in the low-wage sectors of Retail Trade, Administrative and Waste Services, Social Assistance, and Leisure and Hospitality (56% of jobs created, in sectors representing only 36% of total employment). In 2014, the situation improved, with the majority of jobs (61% through September) created being in the higher wage sector.
Source: Bureau of Labor Statistics
As shown in the Figure 1, 2014 has seen a more robust creation of high wage jobs (although the prior two months of September and October appear to be reversing that trend. And, for the most part as shown in Figure 2, the pace of overall job creation has been roughly consistent with that of 2103.
So why aren’t we feeling that more in the overall economy, or―for that matter―in average wages across all job sectors? The answer is twofold. First, for much of 2014, and for the first time since employment levels began recovering on a sustained basis, we saw wages increase in low wage sectors at a far more rapid and ongoing pace than those of high wage jobs (Figure 3).
Source: Bureau of Labor Statistics
Second, and more importantly, the growth in the index of aggregate payrolls for low wage jobs has substantially exceeded that of high wage jobs, which have remained nearly flat since March 2014 (Figures 4 and 5). Effectively, almost all payroll growth has been jammed into the low wage sectors. The payroll index produced by the U.S. Bureau of Labor Statistics is the product of estimates of average hourly earnings, average weekly hours, and employment (number of jobs) in each sector. Figure 5, below, is just a zoom in on Figure 4―but note in the latter figure that this is the first appearance of such sustained inflection in the period studied:
Source: Bureau of Labor Statistics
Before we go on, I should define what I mean by low and high wage jobs. While both exist in every super-sector, there is a very sharp difference between the two super-sectors―Retail Trade, and Leisure and Hospitality―and two large sub-sectors―Administrative and Waste Services (mostly temporary workers) and Social Assistance―and pretty much any other large grouping of employees. Together, these categories total over 42 million jobs, in contrast to the approximately 75 million jobs in the high wage sectors.
As of this writing, the data on the distribution of jobs and wages, between high wage (>$22/hour) and low wage (<$19/hour) sectors, is as follows (all jobs data in this report is shown in thousands):
The hourly pay gulf between low wage positions and high wage positions is not only very substantial (by a factor of 1.75x) but is exacerbated by the low number of hours worked by people who hold low wage positions―particularly in the Leisure and Hospitality and Social Assistance sectors. As a result, as shown in Table 2, the weighted average annual gross pay for these 42.3 million jobs is well under $26,000 per year.
Why is Job Creation not Transmitting to Consumption Growth?
So there we have it, 36% of the jobs in the U.S. and, more importantly, 45% of all the jobs created since January 2011, gross an average of less than $500/week. Some a bit more, but also many less. If we are right to be pay attention to both the mix of jobs created as well wage behavior for those jobs in 2014―stalling in higher wage jobs (which, after accounting for inflation, saw wages fall for much of the year) yet rising in low wage jobs (where wage growth was actually positive on a real basis in 2014)―then we need to focus on the reasons that the sustained growth in absolute number of jobs (and total payrolls) is not transmitting to consumption.
Here we move from raw data to interpretation of the above data in light of other factors. There is, unfortunately, a paucity of currently available data to validate some of these observations, but―taken together―I believe the below explanation is both robust and highly likely.
First, let’s address consumption itself. It is worth noting that after growing at an annual pace of 4.0% for the twelve months through June of 2014, the annual rate of growth of personal consumption expenditures slumped to 1.5% in the third quarter of this year (Figure 6). Is this a trend reflective of the inflection shown in Figure 5, or is it a passing phenomenon? Q4 2014 will tell the tale, but given that the pace of U.S. job growth has remained robust, the foregoing is at least notable.
The most compelling explanation for this disconnect between job numbers and consumption, to my mind, rests in the combination of pay and hours for low wage workers. These workers―a good example being the average leisure and hospitality sector worker making just over $19,000/year―are essentially the working poor who are, thankfully, the beneficiaries of federal employee subsidies including the Earned Income Tax Credit (EITC) and, if they make less than certain thresholds at the individual and family levels, the Supplemental Nutrition Assistance Program (SNAP―what we used to call food stamps). The Child Tax Credit (CTC) also helps lower income families, but because it doesn’t phase out until much higher levels of income, it is not really a factor.
These programs are all “means tested,” meaning that the less you make (to a point) the more you receive in support. The converse, of course, is also true. The more your pay goes up, the less you receive in benefits.
Consequently, many in our population of low wage workers (especially those who earn less than even the low averages set forth in Table 2) really don’t “take home” all net increases in gross pay, as they lose benefits that they were receiving as well. The burden on government is, of course, relieved―but there is no new money to flow back into the economy under such a scenario.
Given the magnitude of these supports relative to the low levels of employee compensation that we observe in the four sectors I have identified above, they would appear to be a factor in addressing the transmission issue. Depending on the number of children in a household, the credit can be quite substantial. A single parent, with only one child can qualify for a benefit of up to $3,305, which begins to phase out at $17,830 in income and doesn’t fully phase out until $38,511 in income. For individuals or couples with more children, the benefit is even more substantial―a couple with 3 or more children can qualify for $6,143 in tax credits which phase out between $23,260 and $52,427 in income. SNAP benefits for the lowest income workers―especially those with children―can amount to another few thousand dollars a year.
To appreciate the impact of these benefits on the argument I am making in this paper, if the average worker of the 42.3 million in low wage sector jobs were to get a 3% raise, the amount of that raise would be only $765/year, several hundred of which would be lost to a reduction in the EITC, increased employment taxes and other tax payments. If that worker also had two children, that raise would knock the worker out of SNAP eligibility for benefits that would previously have been worth more than the rest of the above raise.
Please note that I am not writing about social benefit programs to decry the construction of those programs or to argue that they be changed, but merely to point out the math underlying the impact on consumption when wage increases disproportionately impact lower income households (which is, after all, a good thing in many other ways).
At the higher end of the jobs spectrum, it would appear―given the flattening of hourly wage growth and aggregate payrolls (and the actual nominal declines lately in some high wage sectors)―that reentrants and new entrants into several sectors are coming in at substantially lower than prevailing pay levels for existing employees. As the U.S. in particular exhibits nominal wage rigidity (i.e. we tend not to reduce employee compensation but instead to dismiss excess labor as a way of cutting costs), it is far more likely that soft wage growth (or declines) is spurred by a bifurcation in wage rates between long term employees and those who have entered the 1.3 million high wage positions created thus far in 2014.
It is likely that a number of the foregoing reentrants are taking positions after their long term (extended) unemployment insurance benefits have expired. Since the U.S. Congress failed to extend payments to the long term unemployed in December of 2013, millions lost their insurance benefits (there were 3.1 million of non-qualifying long term unemployed as of July 2014). While some of these people have dropped out of the labor force, many have also accepted positions at lower wage rates than they had been previously receiving and, likely, were holding out for. As I forecast in January of this year, this phenomenon appears to have contributed to stalling wages during 2014.
There has been some indication of wage firmness in some sectors in recent months, and real wage declines have reversed a bit. Certain analysts are viewing this as a tightening of the labor market. I think that is premature, at best, because of (i) the enormous overhang of U.S. labor as illustrated by the LFPR, and the fact that the LFPR has seen no net improvement since the end of 2013, and (ii) the recent improvement in real wages has been to an extent driven by declines in the rate of inflation, as opposed to robust nominal wage growth.
Persistent disinflation is a major factor going forward. Declining oil prices will reflect themselves in coming months’ CPI and PCE inflation statistics. And the toxic combination of disinflation/deflation in other major economies and a strengthening U.S. dollar will see cheap imports build and reduce demand for U.S. exports (witness the growing U.S. trade deficit widening in September to $43 billion from $40 billion in the prior month, even with lower oil prices) at present price levels. It is hard to imagine U.S. labor having any increased pricing power under such circumstances.
We are still playing out the ramifications of the enormous global oversupply of labor and productive capacity relative to global aggregate demand. This phenomenon has dramatically reduced labor’s share of production (GDP) since the late 1990’s, and increased capital’s share of GDP in the U.S. even while the world experiences a capital glut. The disinflation/deflation we are seeing in the worlds’ largest economies are a direct result of that oversupply and the failure of inadequate policy “plasters” to heal underlying imbalances.
The impact on U.S. wages and payrolls in 2014 has been different from the experience of 2013, but the underlying cause is the same. And so is the result―limited increases in aggregate disposable income. Chair Yellen should heed the inner voice telling her that something is not right, because the foregoing demonstrates a still-ailing labor situation in the U.S., not the vibrancy read into headline statistics.
Comments are closed.