Knock me over with a feather — and probably most investors on Wall Street, too. Despite continuing weekly jobless claims in the millions and an ADP report showing nearly three million jobs lost last month, the Bureau of Labor Statistics reports that the US economy added 2.5 million jobs in May. The unemployment rate dropped to 13.3%, down from 14.7% last month, as America apparently reopened much faster than first thought:

Total nonfarm payroll employment rose by 2.5 million in May, and the unemployment rate declined to 13.3 percent, the U.S. Bureau of Labor Statistics reported today. These improvements in the labor market reflected a limited resumption of economic activity that had been curtailed in March and April due to the coronavirus (COVID-19) pandemic and efforts to contain it. In May, employment rose sharply in leisure and hospitality, construction, education and health services, and retail trade. By contrast, employment in government continued to decline sharply. …

The unemployment rate declined by 1.4 percentage points to 13.3 percent in May, and the number of unemployed persons fell by 2.1 million to 21.0 million. Reflecting the effects of the coronavirus pandemic and efforts to contain it, the unemployment rate and the number of unemployed persons are up by 9.8 percentage points and 15.2 million, respectively, since February.

We used to skewer Reuters when it used the word “unexpected” to describe lackluster jobs reports during the Obama administration. In this case, “unexpected” is an understatement. The debate this week was whether job losses would be as limited as in the ADP report, or whether it would stretch out to economists’ expectations in the range of -8 million. No one had predicted a rebound to hit in May; most eyes were on June, at best.

The report captures the movement of people from furloughed status, mainly, but clearly job losses were low enough for this impact to push the job market back into the black. As the report notes, permanent job losses are still on the rise, but at a much lower level than one might have imagined:

The number of unemployed persons who were on temporary layoff decreased by 2.7 million in May to 15.3 million, following a sharp increase of 16.2 million in April. Among those not on temporary layoff, the number of permanent job losers continued to rise, increasing by 295,000 in May to 2.3 million. (See table A-11.)

In May, the number of unemployed persons who were jobless less than 5 weeks decreased by 10.4 million to 3.9 million. These individuals made up 18.5 percent of the unemployed. The number of unemployed persons who were jobless 5 to 14 weeks rose by 7.8 million to 14.8 million, accounting for about 70.8 percent of the unemployed. The number of long-term unemployed (those jobless for 27 weeks or more), at 1.2 million, increased by 225,000 over the month and represented 5.6 percent of the unemployed.

This looks like the reopening mainly benefited the furloughed, not those whose jobs were lost. In the latter category, the more recent the job losses, the better odds of getting a new job, too. That’s not unusual, but the continuing unemployment numbers suggest that we might not see a rapid full recovery for a while, a point to which we’ll return in a bit.

While jobs got restored, wages fell, which indicates the kinds of jobs lost and restored:

In May, average hourly earnings for all employees on private nonfarm payrolls fell by 29 cents to $29.75, following a gain of $1.35 in April. Average hourly earnings of private-sector production and nonsupervisory employees decreased by 14 cents to $25.00 in May. The decreases in average hourly earnings largely reflect job gains among lower-paid workers; this change put downward pressure on the average hourly earnings estimates.

The reason for the “gain” in April was that job losses were primarily in lower-wage positions. The wage “loss” in May reflects the reopening of some of those same jobs. It’s an artifact of the force majeur of COVID-19, not an actual reflection of wage trends, but it does indicate that a lot of lower-wage workers are still out of jobs. The unemployment benefits might disincentivize their return, or employers might have to offer higher wages if they need those positions filled. In fact, the smaller wage drop might indicate employers already are offering higher wages for earlier returns to work.

Still, this is great news, and hopefully shows some momentum. Guess who’s happiest about it at the moment?

This is a rare case of understatement from Trump, too. CNBC’s Jeff Cox notes that this is the biggest jobs report in history:

Employment stunningly rose by 2.5 million in May and the jobless rate declined to 13.3% according to data Friday from the Labor Department that was far better than economists had been expecting and indicated that an economic turnaround could be close at hand.

Economists surveyed by Dow Jones had been expecting payrolls to drop by 8.333 million and the unemployment rate to rise to 19.5% from April’s 14.7%.

Stock market futures burst higher following the report and indicated an open of nearly 600 points higher for the Dow Jones Industrial Average.

The May gain was by far the biggest one-month jobs gain in U.S. history since at least 1939.

“Barring a second surge of Covid-19, the overall U.S. economy may have turned a corner, as evidenced by the surprise job gains today, even though it still remains to be seen exactly what the new normal will look like,” said Tony Bedikian, head of global markets at Citizens Bank.

The Wall Street Journal was a bit more cautious, considering the events of the past week or so:

Friday’s jobs report offers a labor-market snapshot from mid-May, when the government conducted its monthly survey of households and businesses. Fresher data suggest the labor market has since stabilized, though it likely suffered another setback from riots and looting after George Floyd was killed in police custody May 25 in Minneapolis.

Many businesses in big cities had reopened or were set to reopen only to be looted or forced to board up during the protests. That could delay their reopening by days or weeks and cause another round of job losses.

It might take months to see a full restoration of commerce in the urban areas hit by the riots, but the rest of the country could take up much of that slack. It also might send some city residents out to the suburbs more permanently, a trend that COVID-19 might have already started before the riots.

Still, all this is curious while jobless claims remain stuck in the 1.8 million per week range. Did BLS get good data out of these surveys, or are they missing something in their responses? Or have states screwed up the jobless claims systems due to overwhelming initial demand?