Pages

Wednesday, December 31, 2014

Robots Writing Weekly Fiction Disguised As News Articles -- December 31, 2014 -- Jobless Claims Surge; No One Knows Why -- Bloomberg

This is so incredibly cool, when the pieces to the puzzle start fitting together -- I owe it to a reader for sending me a story from Platts this morning. I will get to it later, in another stand-alone post, but it's a big story.

******************************
The Weekly Jobs Report

But first, from the Department of Weekly Fiction. This is the headline, LOL, I can't make this stuff up:

U.S. jobless claims rise, but labor market still strengthening

Say what?

I don't know if readers in general know this, but the major news services do use robots to write their re-occurring news articles. From Slate:
One of the world’s largest news organizations is letting the robots take over the task [of writing business stories], at least for certain types of articles.
The Associated Press announced last month that “the majority of U.S. corporate earnings stories for our business news report will eventually be produced using automation technology.”
In the coming weeks, its first machine-written articles—written with software supplied by the Durham, North Carolina­–based startup Automated Insights and corporate earnings data from Zacks Investment Research—will go live on the AP’s global news wires.
I can't make that up either: the tag line/teaser for that article --  “Robots” are surprisingly good at writing news stories, but humans still have one big edge. LOL. I can't wait to see what that "one big edge" is. I assume the one big edge is slanted reporting, to push an agenda; robots simply report the facts, the who, what, where, but not the "why." [One major exception: I do believe the New York Times uses robots to write their editorials; they are all the same; it is a fact that all global warming stories are written by robots.]

Now back to the story written by robots (there's another gem hidden in this story -- another piece of the puzzle falls into place). This is the jobs data as reported by the robots:
Initial claims for state unemployment benefits increased by 17,000 to a seasonally adjusted 298,000 for the week ended Dec. 27, the Labor Department said on Wednesday. That followed four straight weeks of declines.
So, that's the robotic data.

How does the US State Department and mainstream media slant it?  The number was way higher than expected, way, way higher, but the trend, yes, the trend, that four-week average, the trend remained consistent with sustained strength, yes, sustained strength, in the labor market. Ignore the number, trust us; the labor market is strengthening.

So, how about that "trend." It went UP.  LOL. But only a little.  But with the huge number this week, the four-week average will trend up over the next few weeks. 
The four-week moving average of claims, considered a better measure of labor market trends as it irons out week-to-week volatility, rose only 250 to 290,750 last week. It has remained below the 300,000 mark for 16 straight weeks.
Back to the headline. It all makes sense. When I saw the headline that jobs claims rose way more than expected but and yet the headline said the "labor market was still strengthening," I experienced what is called cognitive dissonance. Many people had never heard of cognitive dissonance until Barack Obama became president. In elementary school, most of us learned that lack of common sense was synonymous with cognitive dissonance.

But then I realized: from the robot that wrote the story and the headline, it was entirely accurate. Robots continue to replace tasks that Americans don't want to do, and thus for robots, the "robot labor market is strengthening."

It finally makes sense.

So, for jobs Americans don't want:
  • do those tasks in China;
  • for tasks that can't be shipped to China, shift to undocumented immigrants; and,
  • for tasks that even undocumented immigrants don't want, hire robots.
 And let the robots write the weekly jobs article AND the headlines.

*****************************
Bloomberg's Spin

Bloomberg's spin: the numbers don't make sense this time of year.
More Americans filed applications for unemployment benefits for the first time in five weeks, displaying the typical year-end holiday swings that make the data difficult to interpret.
Jobless claims rose by 17,000 to 298,000 in the week ended December 27, 2014, from a revised 281,000 in the prior period, a Labor Department report showed today in Washington. The median forecast of 22 economists surveyed by Bloomberg called for 290,000. No states estimated data and there was nothing unusual in the report, a spokesman said as the figures were released.

The number of applications can fluctuate during this time of year as the holidays make it tough to adjust the data for seasonal variations. Employers are dismissing fewer workers and adding staff as household purchases pick up, driven by a drop in gasoline costs that will keep boosting the economic expansion.
By the way, what was the revision. Bloomberg didn't say. The original report was 289,000. The revision was 281,000. I have to assume those numbers were correct. But that means there was a decrease in the number of applications between the original report and the revised report. One has to ask how 8,000 applications disappeared between the original report and the revised report.

This is the answer. Remember: this is not raw data. The raw data is massaged to give us the "seasonally adjusted numbers." 

Disclaimer: I am not denigrating robots. It's just the way it is. Some of my best friends are robots. With regard to everything else in the post above, don't make any investment, financial, or relationship decisions based on that data. I do not have any formal training in reporting labor statistics and I assume my interpretation(s) above are really, really wrong.

By the way, I'm not going to go through the explanation, but the analysts' forecasts were not all that wrong, despite the "surprise." It has to do with the government's revision of the previous weeks' number. Seriously.

No comments:

Post a Comment

Note: Only a member of this blog may post a comment.