October jobs: better than expected but a long way to go

November 6th, 2020 at 11:29 am

Payrolls grew by 638,000 last month, and the unemployment rate fell sharply, by a full percentage point, to 6.9 percent. The report reveals a job market that’s healing, but at a slower pace than earlier in the year and with a long way to go to get back to full employment. Even with the large drop, the October jobless rate remains twice that of the pre-crisis rate.

Government jobs fell (by 268,000) last month, but because this is partially related the cutting back of decennial Census jobs, a clearer signal of underlying labor demand comes from the private sector, which added 906,000 jobs. That’s a decent clip, but in May and June, private sector gains rebounded at a pace of 4 million per month; since then, the pace has slowed to 1 million. At this rate, the private sector will be back to its pre-crises peak around late summer of 2021.

The figure below shows total and private payrolls. The remaining gap is clear–10 million overall and 9 million for private–as is the slower pace of gains. 

Source: BLS

This next figure shows the share of losses left to make back for many groups. The last two bars show that jobs are about halfway back. The first bar shows that because unemployment rate grew over 11 points from its low point to its high point–3.5 to 14.7 percent–and has since clawed back almost 7 points, it has made up 70 percent of its lost ground.

One important take from the figure is the slower progress by Black people. For both unemployment and employment rates, they lag other groups, having made back just half of their losses. This is an often-seen pattern for persons of color in weak labor markets, as they face labor market barriers, including racial discrimination, whose impact is accentuated in periods of weaker demand. 

Another worrisome sign in today’s report comes from the extended duration of unemployment for some jobless workers. About a third of the unemployed are now long-term unemployed (searching for work for at least six months), compared to 4 percent in April. This reflects that when this initial shutdown occurred, many workers went on temporary layoff, and, in fact, the bulk of the decline in unemployment last month came from these temporary layoffs getting back to work. But this means that the folks still left on the jobless rolls are in for potentially long spells of unemployment. 

This growth in people stuck in long-term unemployment is particularly problematic because extended unemployment insurance benefits are scheduled to expire at the end of this year. Re-extending these benefits and enhancing the dollar value of the claims should be policy makers’ first order of business in coming weeks. 

Turning to sectoral analysis, almost 70 percent of private industries added jobs last month, a relatively high share given the spiking coronavirus. Notable job gains in this regard can be seen in leisure/hospitality, which includes restaurants and bars, up about 200,000 last month and 400,000 over the past two months.

Of course, it is very possible this gains attenuate in coming months given the sharply rising virus caseloads. While the relationship between the virus and commerce is far from one-to-one, meaning we don’t see March/April-type shutdowns in places where caseloads are climbing, we should be on the lookout for negative effects on jobs in sensitive, face-to-face services.

Finally, especially given lower-for-longer interest rates from the Federal Reserve, Congress needs to get back to the business of providing relief to the millions of workers who still face considerable hardship. State and local budgets are still much reduced, forcing layoffs, and even while the the labor market is improving, it remains far from providing the income vulnerable families need to pay rents, mortgages, child care, and to meet their nutritional needs. 




September 2020 jobs report: Slowing jobs gains and a huge spike in long-term unemployment

October 2nd, 2020 at 11:09 am
Payrolls grew by 661,000 last month, well below expectations, and the jobless rate ticked down to 7.9 percent, driven not by job gains, but by people leaving the labor force. Long-term unemployment spiked sharply–in fact, its largest one-month spike on record–and shifts continue from temporary to permanent job losses. In other words, though the labor market continues to improve, it is doing so at a slower pace, and the risk of increasing numbers of job seekers stuck in long-term joblessness is rising.

Payrolls continue to climb back as commerce gradually recovers, but the pace of gains has slowed, as shown in the figure below. Private sector gains last month were stronger, at 877,000, as local education jobs fell sharply, by 231,000. At least part of that loss is due to the impact of Covid on decline job opportunities for bus drivers, school cafeteria workers, and janitors due to much less in-person teaching.

Importantly, this deceleration in job gains (see figure) is consistent with the absence of two absolutely essential policies from the federal government: virus control and fiscal stimulus. The former was never taken seriously by the Trump administration, at tremendous cost to the lives, health, and living standards of millions of Americans. The latter–stimulus–was initially implemented with real urgency, but the fact that such urgency has demonstrably faded is evident in today’s report.

Source: BLS

No labor market indicators have regained their pre-crisis peak, as shown in the figure below. It shows the percent of losses regained since payrolls started growing and unemployment started falling. For example, the unemployment rate initially rose from 3.5 percent in February to 14.7 percent in April, an increase of 11.2 percentage points. Since then, it has declined to 7.9 percent in September, meaning it filled up 6.8 points of the 11.2 point hole, or 61 percent.

The first point from the table is that all bars are well below 100 percent. Sizable holes still persist and while it has clearly improved, the job market is still operating with recessionary levels of slack. Blacks in particular reveal lagging progress. The recovery of both their unemployment and employment-to-population rates (EPOPs) are lagging behind the others. This is a serious concern which appears to confirm the strong cyclical component of Black labor market outcomes, meaning they benefit disproportionately from strong labor demand and, in the current situation, take more of the brunt of weaker labor demand.

Other signs of decelerating job opportunities include the increase in permanent, versus temporary, job losers (meaning laid-off workers who should not expect to be called back to work as commerce reopens) and the related spike in long-term unemployment, i.e., those seeking work for at least six months. Back in May, when more jobless workers still thought they were temporarily furloughed, permanent layoffs were only 14 percent of the unemployment. In September, that share was 36 percent.

The 781,000 spike in long-term unemployment in September was historically notable: it is the largest one-month increase in this indicator on record, with data starting in 1948 (note: this record holds as a share of unemployment as well, which spike up from 12 to 19 percent). The risk here is what economists call “hysteresis:” the phenomenon of a group of potential workers relegated to the labor market’s sidelines for long periods. Should their skills or even just their basic labor-market attachment fade due to long periods of joblessness, they risk a lasting disconnection from work and wages at great cost to themselves, their families, and the greater economy.

Though it is too early to tell whether these early indicators will morph into a longer-term trend, my work in this area suggests that Black, immigrant, and low-income workers are particularly vulnerable to this outcome. Certainly, the bars for Blacks in the figure above are suggestive of that result, but it is one I will be carefully tracking in coming weeks and months.

Unemployment down but so is pace of job gains

September 4th, 2020 at 11:11 am
Payrolls were up 1.4 million in August and the unemployment rate fell sharply from 10.2 to 8.4 percent according to this morning’s labor market update from the Bureau of Labor Statistics. Payrolls remain 11.5 million below their pre-pandemic peak in February and the jobless rate is still more than twice its February rate of 3.5 percent.

In years that ends in a zero, the federal government temporarily hires many workers to field the decennial Census. It is thus important to look at private sector payrolls to get a more accurate read on underlying labor demand. Private payrolls rose 1 million in August, as temporary Census hires rose by about 240,000.

Turning back to the bigger picture, the figure below shows average monthly job gains or losses in recent months. The massive job losses that occurred in March and April came to 11 million per month. The initial bounce-back when the economy began to reopen–too soon, from the perspective of public health–in May and June added 3.8 million jobs per month. Then, as the virus re-surged, hiring downshifted in the last two months, adding 1.6 million per month. The slower the pace of job growth, the longer it will take to regain the pre-pandemic jobs’ peak. This, however, is a low bar, as the working-age population has grown over this period, implying the need to more employment than existed pre-pandemic.

A broader look at the payroll jobs situation is shown in the next table. The first line shows how many jobs were lost due to the shutdown, both for total payrolls and leaving off the government sector. These massive losses amounted to 22 and 21 million, respectively. Since then, through August, 10.6 million jobs have been regained in the overall job market, 10.5 million in the private sector. This leaves holes of about 11 million in both sectors, or gaps of slightly over 50 percent. Simply put, half the job hole still remains empty.

Source: BLS

Even with the welcomed decline in the jobless rate, this is still very clearly a job market that’s recovering, and doing so more slowly, not one that has recovered. Any claims that the U.S. job market is strong, healthy, or providing ample employment opportunities for job seekers is factually indefensible.

Racial disparities

The table below shows that Black and Brown persons have been hit harder and are recovering more slowly than whites by the downturn, a familiar pattern from past cycles. Black unemployment was 13 percent last month compared to 7 percent for whites. Employment rates for Blacks and Hispanics (working age) are down 7 percentage points compared to 5 for whites.

While the jobless rate fell, we are seeing a concerning development in unemployment: a shift within the unemployment away from temporary toward permanent job losers. Here’s how the BLS put this:

“As in the prior 3 months, the decrease in unemployment in August was driven by a decline among people on temporary layoff (-3.1 million). The decline was partially offset by an increase in the number of permanent job losers, which rose by 534,000 to 3.4 million.”

In fact, the share of the unemployed who are permanent job losers, versus those on temporary furloughs, almost tripled in recent months, from 11 to 30 percent. A related, worrisome development is the shift to longer-term unemployment: the share of job losers unemployed for for at least 15 weeks has gone from 8 percent in April to 60 percent in August.

K-shaped recovery

Another important pattern developing in the current recovery is the disparity between the experience of higher and lower income households. The stock market is up, but so are reports of hunger and evictions.

Economist Jesse Rothstein did some recent analysis of the K-shaped recovery in labor market earnings. In periods like this, wage snapshots of the type we get in today’s report are inaccurate, as disproportionate job losses for low-wage workers mean that high-wage workers are a larger share of those who still have jobs, creating an upward bias in snapshot wage trends. The figure below thus tracks pay changes for the same workers across time, assigning zero earnings to job losers and comparing post-pandemic trends to pre-pandemic trends.

Because job losses have been concentrated among low earners, their earnings declines have been much steeper than those of higher earners. The figure shows the decline in labor earnings for both groups since the pandemic compared to pre-pandemic trends (these data go through July). Real earnings fell as much as 20 percent for those in the bottom half compared to 4 percent 5 percent for top-half earners.

Source: Jesse Rothstein and my analysis of CPS data.

In sum, the failure of the Trump administration to control the virus has led to a slower pace of job gains and, while the jobless rate fell significantly last month, it is still in recessionary territory and more job seekers are at risk of longer-term unemployment. Importantly, note that this shift is occurring as Congress, particularly Senate Republicans, has dropped the ball on further fiscal relief, including enhanced UI benefits. Finally, there is evidence of a K-shaped recovery underway, as both the stock market and hunger are rising, and as earnings losses are concentrated among those in the bottom half.

July jobs: Labor market keeps ticking, but virus surge is slowing pace of gains

August 7th, 2020 at 10:39 am
The labor market kept ticking in July, but the re-surging pandemic led to slower hiring across most industries. Payrolls rose 1.8 million last month, compared to average monthly gains of 3.8 million in May and June, and payrolls remain 12.9 million jobs down from their February peak (see figure). The jobless rate ticked down to 10.2 percent, but remains highly elevated–that’s still higher than the peak of the last recession–and the share of the prime-age (25-54) population working remains almost 7 percentage points below Feb’s level.

Because of an unusual interaction between pandemic-induced layoffs in local schools and the BLS seasonal adjustments (explained below), the overall job gain is biased up in July. Looking at private sector employment avoids this bias, and payrolls there rose 1.5 million in July compared to an average of 4 million per month in May/June. As usual, we must adhere to the the caveat that one month does not a new trend make. But if this slower trend persists–which could happen, given the impact of surging cases on commerce–it will take until next spring to regain the pre-crisis peak for private sector jobs.

The table below shows that the overall unemployment rate remains highly elevated relative to its February trough of 3.5 percent, and, as is always the case, the increase in joblessness and the decrease in employment rates (the share of prime-wage workers with jobs) are significantly worse for persons of color (note: because of ongoing classification issues, the BLS reports that the actual jobless rate is probably about a point higher than the published rate, but the trend is the same). The Black decline in jobless rates of 9 percentage points is just a few points shy of their total gains in this metric over the past decade, meaning they’ve gained back very little of those losses.

Other signs of a viral-surge-induced slowdown in the job market include:

–Significant weakness in the July manufacturing numbers: After adding an average of about 300,000 jobs per month in May/June, the factory sector added only 26,000 in July. The share of manufacturing industries adding jobs fell from 77 to 43 percent, and the sector remains 740,000 jobs, or 6 percent, below its pre-crisis peak.

–A shift toward longer-term unemployment: In June, 81 percent of the unemployed had been so for less than 14 weeks. In July, that share fell to about half, and those jobless for 15-26 weeks spiked from 11 to 40 percent.

Whether these one-month changes persist into deteriorating trends bears watching.

As noted, July’s topline job-gain number of 1.8 million is biased up by an unusual technical problem having to do with the interaction between the impact of the virus on education jobs and the BLS seasonal adjustments for that sector. Typically, in July, many who work in the education sector leave for the summer, consistently over 1 million, or almost 20 percent of employment in the sector (not just teachers, but also janitors and others who stop working at the school in the summer). To account for this seasonal effect, the BLS adds back about this same amount—around 1 million—to the seasonally adjusted July data for local education jobs.

But this year, because of the pandemic shut school buildings down well before the summer, these seasonal layoffs occurred earlier in the year, around April. However, the seasonal adjustment is based on the normal, historical pattern, and it thus added 1.2 million jobs to local education in July. That is, a relatively large seasonal factor was plugged in to offset a large number of expected layoffs that didn’t occur this July. This likely biased up the topline number by over 500,000 jobs.

It is therefore much more instructive in this case to look at year-over-year changes in the state and local sectors. Using that metric, jobs are down by 1 million, about 6 percent, as should be expected given the deep budgetary shortfalls faced by states and towns across the land.

This month’s snapshot of the job market is suggestive of a slower pace of recovery than we saw in May and June, surely driven by commerce pulling back as the virus has re-surged in various parts of the country. The nation is still climbing out of the huge hole resulting from the shutdown, we have a long ways to go, and we’re probably climbing out more slowly. There is certainly nothing in today’s report that should in the slightest dampen the urgency to help economically vulnerable Americans get through to the other side of this crisis. Such urgency needs to reach Congressional negotiators, who must immediately finish their negotiations and pass the next relief package.

A strong jobs report but big holes remain and we’re not outta the viral woods.

July 2nd, 2020 at 10:38 am
Payrolls popped up by 4.8 million in June, as commerce continued to gradually reopen across the country. Most industries (75 percent) added jobs, and millions of furloughed workers were called back, taking the unemployment down to 11.1 percent from 13.3 percent in May.

The strong report begs the question: are we out of the virus-infected woods? Has the pandemic-induced recession ended as we enter a strong bounce-back to a solid expansion?

The answer is as best uncertain and, based on recent state-level spikes in the virus, likely “no.” That is, absent a second wave of the virus, the economy has probably bottomed out, and yes, more labor market reports like June’s would restore a job market that would start to reliably repair the deeply damaged fortunes of working families.

But the hole in the economy and in family earnings is still large, unemployment remains at recessionary levels, racial gaps are beginning to predictably grow, continuing unemployment claims, also out this morning, are still high and stagnant (i.e., stuck at very high levels). Most importantly, the failure of national leadership to control the virus means forthcoming reports may not be this strong.

The trend is your friend, the level bedevils

The figure below shows just how far payroll jobs fell off a cliff as the virus took hold, and how, even with two strong months of job gains—4.8 million in June and 2.7 million in May—a huge hole of almost 15 million jobs remains.

Source: BLS

Similarly, the table below shows highly elevated unemployment rates relative to February, with the rates for Blacks and Hispanics up about three percentage points more than for whites. Note that in the last recession, unemployment peaked at 10 percent, below today’s 11.1 percent.

Source: BLS

As noted, most sectors added jobs in June. Restaurant and bar re-openings alone added 1.5 million jobs, accounting for almost a third of the total gains. Factory employment was up 356,000 in June, following on May’s gain of 250,000. Still, manufacturing has recovered less than half of the 1.4 million jobs lost since February.

State government—down 25,000—was an important exception to the overall trend, and states have consistently lost jobs over the crisis. This is a germane observation for Congress as they contemplate the next round of fiscal relief, in which aid to states should be an essential component.

Furloughed workers returning to work

One important and unusual characteristic of the pandemic recession has been the highly elevated share of the unemployed on temporary furlough. A key question of the strength of the nascent expansion is whether these jobless persons would move back into the employment column as they are rehired at reopening businesses or move over the more permanent jobless column.

The BLS pointed out that “June’s unemployment decline occurred primarily among people on temporary layoff. There were 10.6 million people on temporary layoff in June, down by 4.8 million.” This is a positive sign, one that bears watching in coming months as temporary layoffs as a share of the unemployed, at 60 percent, is still highly elevated. To the extent workers exit temporary layoffs to return to work, the expansion should experience a faster liftoff.

What’s the wage story?

There’s one part of the jobs story that’s been particularly hard to discern over the crisis: what’s happening with wages? What makes the wage story so tricky is the “composition effect” impacting wage results. That is, because low-wage workers were initially hardest hit by the downturn, wages rose as they left the sample, leaving a higher share of higher-paid workers still drawing paychecks. This bias reverses as they return to the job market, as was the case in June, when hourly pay fell 1.2 percent, a huge monthly decline.

One way to partially control for this bias and get a feel for what’s happening with paychecks is to look at the aggregate wage bill: basically, weekly earnings summed over all workers. This measure includes the hourly wage, but it also factors in job losses. In real terms, since February aggregate weekly pay for private-sector workers is down an historically large 7 percent and down 8 percent for middle and low-wage workers. (Note: this calculation includes my forecast for June inflation.)

In other words, the earnings of working families, while improving as per the gradual reopening, are still way down. Fiscal relief—Unemployment Insurance and checks to households—has surely made up some of the difference for them, but until their paychecks come back, they’re far from out of the woods.

In sum…

An important fact about today’s jobs report is that it is a lagging indicator. Its data were collected in the first half of June, mostly before some states began to see new spikes in cases that led to reversals of their economies’ reopenings. Subsequent reports may show less favorable trends, prolonging the closing of the persistent huge job gaps we’ve faced since April.

In this regard, there is no plausible or credible excuse for the extent to which national leadership has failed the nation’s people and our economy on controlling the virus. Simply mandating, or even modelling, wearing face masks is predicted to have a powerful impact on the spread of the virus and thereby the economy. As a recent, careful study found, “A face mask mandate could potentially substitute for lockdowns that would otherwise subtract nearly 5% from GDP.”

Finally, the jobs, earnings, and unemployment gaps, along with the continuing shedding of state workers, require another robust fiscal package focused on extended UI and state fiscal relief. Yes, we now have enough data to establish a positive trend associated with reopenings, but absent effective virus control, no one can be confident that this trend will persist.