The released unemployment claims numbers are high, but they’re not right

Our economy is undergoing a Great Pause as a result of the COVID-19 public health crisis. Economists and policy makers are on a hunt for numbers that can help them measure the crisis and devise an appropriate response.  One number that everyone has been waiting for has been the number of new unemployment claims.  We have all known that it will be a number that we’ve never seen before on a national scale and have been thinking of what to make of such a large number. On March 22, I wrote that a Great Pause naturally brings unprecedented unemployment claims numbers and that we shouldn’t panic and see this a sign of an economic meltdown.

Keeping in mind that we need to look at the number of unemployment claims differently than we usually do, let’s look at the official number released by the Department of Labor this morning: 2.9 million.

You may have seen the number 3.3 million in the headlines. But that’s a seasonally adjusted number, which makes no sense. There’s nothing seasonal about what’s going on now.

The seasonal adjustment is an obvious flaw in the number being reported. But, when you take a closer look, it’s clear that there are deeper flaws. I’ve spent some time in my previous post estimating the number of new claims based on data from local news sources. (Credit goes to Jacob Robbins for starting to compile these data online).  Comparing to these estimates, there seem to be problems with the official numbers for several states. The chart below shows the difference between my estimates and the official release for states for which I had direct estimates.  For 12 states, my estimates are lower than the official release and for 12 states they are higher.  For four states, my estimates are more than double the official release numbers.  There is no way the official release numbers can be correct for these states.

Take California, the largest of these four states.  The released number is 187,000 for the week.  But, during a press conference Monday night, Governor Newsom reported that the daily average of claims over the last seven days was 106,000. These means a weekly total of 742,000, almost four times the official release.

Or take Oregon, whose official number is 23,000 for the week.  A news source last Wednesday reported that nearly 18,500 people filed for an unemployment claim on Tuesday alone.  For Connecticut, a news source reported last Friday 72,000 claims for the week, against an official release of 25,000.  For Colorado, it has been reported that there were 8,000 claims per day, but the official weekly estimate is 19,000.

It is important to note that the Department of Labor was not shy to release big numbers.  For Pennsylvania, they report 379,000 initial claims, 5 percent above my estimate.  For Ohio, they report 188,000 initial claims, 12 percent above my estimate.  For Louisiana, they report 73,000 initial claims, 73 percent above my estimate, and a whopping 39 times the number of claims in the preceding week.   There is thus no evidence that the Department of Labor is trying to systematically underreport. 

That said, it would be useful to know why the numbers for some states seem so far off.  Well-informed estimates, as were provided for Pennsylvania and Ohio, are more useful than numbers that are too low for some technical reason.

And, please, do not generate an additional 385,000 unemployed workers by seasonally adjusting, we have enough unemployed workers to take care of as is.

One reply on “The released unemployment claims numbers are high, but they’re not right”

[…] Surely, the economy of the District does not reflect the economy of the United States as a whole.  At a quick glance, though, there was nothing unique about DC prior to the Great Pause that would make it more susceptible to new unemployment claims.  If anything, the opposite is true.  DC does not have a particularly generous unemployment benefit program compared to the states.  According to the American Community Survey, its workforce is more educated than the American workforce at large: 57% of its adult residents have a bachelor’s degree compared with 31% for the US at large.  All available evidence to date points to sectors employing less educated workers (e.g., hotels and restaurants) accounting for a disproportionate share of unemployment claims. So DC is not some exaggerated outlier.  DC did respond to the crisis somewhat earlier than some of the states.  But by the second week of elevated claims, the states will have had a chance to catch up and log their own claims 25 times above levels seen before the Great Pause.  When that happens, we can expect to see weekly claims surpass 6 million.  We can only hope the Department of Labor will not create an additional 850,000 claims by seasonally adjusting the data… […]


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s