The BLS JOLTS (Jobs Openings and Labor Turnover) report came out today.

The BLS claims jobs openings are up. Based on an alternate reports, I suggest opening are not only down, but falling steeply.

The implications are huge, so let’s dive into the discrepancies.

BLS: Job Openings vs. Hires

Jobs Openings

That’s how the BLS sees things. Before we dive into the alternate view, let’s take a peek at mainstream media analysis.

Mainstream JOLTS Comments

Bloomberg Econoday spins it this way:

Job openings are up but hiring isn’t, in what are mixed but still favorable results from the April JOLTS report. Job openings rose to 5.788 million from a downward revised 5.670 million in March. The job openings rate also rose, up 1 tenth to 3.9 percent. The hiring rate, in contrast, fell a sharp 2 tenths to 3.5 percent in what perhaps confirms anecdotal reports that employers are having a hard time finding qualified applicants for skilled positions. In a separate indication that points to weakness in worker confidence, the quits rate fell 1 tenth to 2.0 percent suggesting that workers are not shopping their skills around to other employers. Back on the positive side, the layoff rate fell 1 tenth to a low 1.1 percent. This report is mixed and embodies what are increasingly mixed signals across employment indicators in general.

Supposedly hiring is down because employers cannot find qualified workers.

How about the fact that part-time employment for economic reasons soared by 468,000? For more grim details, please see Fed Hiking Not: Payroll Jobs +38K, Employed +26K, Labor Force -458K, Revisions -59K.

Beat the Street

CNBC reported 5.8 Million Job Openings in April vs 5.7 Million Expected.


Job Openings – Real Time Macroeconomics

Jon Hartley, Researcher and Policy Analyst for Real Time Macroeconomics see things this way (anecdotes Mish).

New Jobs Listings

Clearly someone is wrong. Who is it?

Red Flag on the US Economy

Please consider the Financial Sense report Real Time Online Jobs Data Continues to Raise a Red Flag on the US Economy.

Following the big jobs miss last Friday, we recently showed that this trend is likely to continue without a turnaround in leading employment indicators.

Another important data point that we’ve been keeping our eye on is online jobs listings, which, as we noted last month, were beginning to plunge and rollover in a manner similar to the last US recession

Our contact is Jon Hartley at Real Time Macroeconomics who previously worked at the Federal Reserve, Dallas Cowboys, and Goldman Sachs Asset Management as a quantitative analyst, economic researcher, and data scientist of sorts.

When we spoke to him a couple months ago on our podcast about his company and using real-time online data for tracking changes in the economy, he said there were some signs of leveling off but no clear red flags just yet.

Since that time, however, the data they track has moved from leveling off to outright decline and Jon just recently emailed us to say he is now “leaning toward forecasting a likely downturn/recessionary scenario” and said, “I don’t think last week’s poor jobs report was just a blip.”

Though the data only goes back around 10 years, it’s clearly rolled over in a manner similar to the last recession. As we noted last month, if the online data is providing a more accurate picture of the US economy vs. the official government surveys, then “expect to see a string of downward revisions” and continued misses moving forward. Last week that prediction was fulfilled quite strongly with only 38,000 jobs added and heavy downward revisions to the two prior months.

We recently spoke with Bob Eisenbeis, Cumberland Advisors’ Chief Monetary Economist and former Executive Vice-President at the Federal Reserve Bank of Atlanta, who made a very interesting observation on our show (airing Wednesday). Given the economic uncertainties currently present (he specifically cited the Great Depression as our closest parallel), forecast errors on government data will tend to be larger than normal.

The big worry, Eisenbeis said, is that US productivity continues to slow, which, as we recently discussed (See Chad Syverson Busts the Labor Productivity Mismeasurement Thesis) , is in all likelihood not a measurement issue, but more structural in nature. Unfortunately, there’s not much that monetary policy can do about this.

Why Believe Hartley?

I asked Cris Sheridan at Financial Sense a simple question: Why believe Hartley?

The answer is found in the May 17 FS report Plunge in Online Job Postings Raises Red Flag.

  1. The official JOLTS data is collected using a survey and has a smaller sample size.
  2. Online data may skew towards higher paying white-collar jobs (a forthcoming paper to be provided on this, Hartley noted)
  3. Real Time Macroeconomics shows only new online job openings whereas JOLTS may include unfilled job postings from prior months.

Given the above, the real-time online data is likely providing a more accurate picture of the US labor market vs. the official survey. Expect to see a string of downward revisions if that’s the case.

Downward Revisions

I have been discussing downward revisions for some time. If they come, revisions are likely to show the US is in recession already.

More and more things simply do not add up.

I have a report tomorrow listing 10 reasons the May jobs report was not an outlier. If you prefer to believe cheerleaders, I have another answer.

Jobs Report Just Noise Says J.P.Morgan

Earlier today ZeroHedge tweeted a comment from JPM. Here’s the conversation.

Whom do you believe?

Also consider Fun with the BLS Birth/Death Model.


I received this email clarification from Hartley …

Hey Mish,

On the data, a couple things which I outlined in my last interview with Cris:

  1. Our figure is a series of “new” job openings, while the BLS JOLTS number is “total” job openings.
  2. Our data is purely online, whereas BLS JOLTS is offline, which can create specific biases in either series.
  3. JOLTS uses a small sample size and hence can introduce a lot of noise, whereas our sample size is much larger.

Hope this shines some light on the differences between the series.


Mike “Mish” Shedlock