Reverse 'Tweetment': AP Tweet Asserts That Jobless Claims Rise 'Signals Steady Job Gains'; That Day's Stories Say No Such Thing
On Thursday, the Department of Labor announced that initial unemployment claims during the week ended August 3 rose to a seasonally adjusted 333,000, up from a revised 328,000 the previous week.
A "breaking" tweet from the Associated Press issued just a few minutes after the report's 8:30 a.m. (5:30 PT) release read as follows: "U.S. unemployment aid applications up only 5,000 to 333,000 - a level that signals steady job gains." The folks at Twitchy.com properly wondered how rising jobless claims can lead to more jobs. The wire service abandoned the tweet's claim only 19 minutes after its release, and went as far as admitting that "hiring lags" in a longer, late afternoon item.
Martin Crutsinger's 8:55 a.m. item went to the four-week moving average statistic for comfort, and did not make the leap made in the initial tweet. Eventually, instead of pretending to have a crystal ball, the AP reporter went with a plausible rendering of what has happened this year (bolds are mine throughout this post):
MEASURE OF US JOBLESS CLAIMS FALLS TO 6-YEAR LOW
A measure of Americans who applied for unemployment benefits over the past month has fallen to its lowest level in almost six years, signaling fewer layoffs.
The Labor Department said Thursday that the average number of people who applied for benefits over the past four weeks dropped 6,250 to 335,500. That's the lowest level since November 2007, the month before the Great Recession began.
The four-week average smooths week to week fluctuations.
Weekly applications for unemployment aid increased by 5,000 last week to a seasonally adjusted 333,000. But that's up only slightly from the previous week's 5 1/2-year low.
The decrease in the four-week average points to a positive trend in recent months. Applications, which are a proxy for layoffs, have fallen more than 10 percent since the start of the year. That's helped drive net job gains this year, which are the number of people hired minus the number who lose or quit their jobs.
The afternoon report by Crutsinger and Paul Wiseman went further, directly contradicting the wire service's morning tweet. It also misstated the day's claims number:
US JOBLESS CLAIMS AT 6-YEAR LOW, BUT HIRING LAGS
Americans who have a job may take comfort in knowing that companies are laying off fewer people than at any time since before the Great Recession.
The government said Thursday that weekly applications for U.S. unemployment benefits have averaged 335,500 over the past month. That's the lowest level since November 2007, which was one month before the recession began.
But while most companies have stopped cutting jobs, many remain reluctant to hire. That's bad news for the roughly 11.5 million Americans who are unemployed and a major reason the unemployment rate is still so high four years after the recession officially ended.
"We have seen a disconnect between the level of hiring and firing," said Bricklin Dwyer, an economist at BNP Paribas.
... The number of first-time applications did rise slightly last week, to a seasonally adjusted 330,000. But that's just 5,000 higher than the 5 1/2-year low reached two weeks ago.
... The drop in layoffs helps explain why job growth has increased this year to an average of 192,000 net jobs a month, even while overall economic growth has stayed sluggish.
... The Labor Department says layoffs have averaged 1.6 million a month through June, fewer than a monthly average of nearly 1.8 million in the pre-recession year 2006.
The actual averages in the final excerpted sentence are 1.635 million and 1.765 million in 2013 (through six reported months) and 2006, respectively, meaning that average monthly layoffs have been 7.4 percent lower than 2006 so far this year instead of the 11.1 percent reduction indicated in Crutsinger's rounded figures.
AP reporters have frequently written that initial claims consistently below 350,000 per week should lead to a lower unemployment rate (a figure which somehow increased from 325,000 back in 2009). Based on the tweet, I expected the wire service's reports later in the day to repeat that assertion. But they didn't, leaving one to wonder what really prompted that early morning tweet.
Unfortunately, the tweet's impact is that many who get their news on the go and don't look past headlines and tweets will really think that Thursday's initial claims report "signals stead job gains," when the AP's own reports that day ended up denying that.
So was the tweet a snafu, or a calculated deception of casual news consumers?
I wouldn't have to ask the question if AP reporters — and tweeters — would stick to reporting facts and stop engaging in amateur "analysis."
Cross-posted at BizzyBlog.com.