re: Trump has fired the head of the Bureau of Labor Statistics

It's likely her subordinates as well. It's not like one person is cranking out the numbers.

quote:

---

It's likely her subordinates as well. It's not like one person is cranking out the numbers.

---

I think it is probably more of a system failure than any one person. That said, there needed to be some accountability. This is just terrible timing.

ETA: I'm not even arguing the accuracy
of the numbers one way or the other. It is just that when shite like this moves markets, the projections should be more in line with reality.

quote:

---

Wanting to have lower interest rates and getting mad about low job reports is an odd combo as low job reports are a primer for lowering interest rates

---

This might seem like a ridiculous comment- but maybe he wants the TRUE NUMBERS REPORTED regardless of agenda?

Any quality CEO demands that.

quote:

---

it is easy to paint him as a petty man making a petty decision

---

You would have more of a point if the press had reported on him honestly at any point in the past 10 years.

If they’re going to run with their narrative no matter what, you can’t spend every moment analyzing how they’re going to portray you.

Were the job numbers wrong?

If all we have is "Yes" men around Trump. Then we are all fricked and the government can't be trusted. Telling the president the truth should be the norm, not the exception.

As I said yesterday in the thread about how everything is fake, I think it's a little scary that Trump can essentially do what he wants now and dismiss any criticism by just saying "fake news." That said, certain sections of the government (especially the intel agencies) and the media created this situation -- Russiagate is the best example -- to where a lot of people don't trust them.

Jake Tapper was whining after the backlash on Natasha Bertrand when she had the report that the Iran bombing wasn't that successful. He pretty much said, we are the only check on whether this guy is full of shite and you should trust us. He's right in some respects, but again, they earned the criticism and skepticism -- Bertrand has proven to be a plant who says whatever the intel agencies want to undermine Trump.

quote:

---

The lag between getting all the data compiled and getting it out at useful point in time is too great.

---

The report comes out monthly. Bad data is worse than no data at all. Why not delay the release of the report until you have enough data to make a reasonable assessment? It can still come out monthly, just on a different schedule.

How can a report that comes out at +147,000, and then gets revised to +14,000 be considered useful? To be useful, it needs to be reasonably accurate.

If companies are actually reacting to the initial report, they'd be wasting a lot of effort. Are they really ramping up production based on 147K added jobs, and then cutting it back a short while later? Like a yo-yo? Makes no sense.

I think only the market is reacting, and it's just a knee-jerk reaction (fear of being left behind). Then the market reacts again to the revision. The only thing a delayed, but more accurate report would result in is less knee-jerk volatility in the market. And that would be a good thing.

quote:

---

It’s because you don’t understand the nature of data collection and statistics.

---

I know enough about statistics to know that if I'm revising the numbers this much, that I should not have released the initial report. I didn't have enough data.

If you are working with very stable numbers, you can make the assessment with less data. But when you look at something that varies every month, and this is a report you produce every month, you have a pretty good idea that you can't make an accurate assessment without a fairly complete data.

Just like you can't have any confidence in calling an election until you have a good read of the swing states.

If they are obligated to publish the report on that date, regardless of how much data is left to collect, they should include a margin of error. They should have a good idea of how much data is yet to be collected Of course, that would look something like " We estimate a + 147,000 jobs added, with a margin of error of 150,000 jobs. We will update that when our data is more complete".

They should figure out how to get the data on time, or request a change to the reporting time release.

I assure you, if one of my engineers were giving me reports with that level of uncertainty, we'd be having a talk and figuring something out. Termination would be one pathway.

You might as well put your person into run these bullshite numbers.