@futurebird fired the head of the Bureau of Labor Statistics which is supposed to be non partisan. The labor secretary serves at the pleasure of the president. She is a one-term congressional rep from the Portland suburbs who lost her reelection bid.

@michael_w_busch

I keep thinking about that thing where Trump would change how much he thought his business was worth based on "his feelings at the time" and I assume he thinks numbers aren't connected to anything and you can just say any number you want. I mean when he wants a number to be large each time he says it, it grows exponentially*.

*I have the sudden urge to do a regression on "trump numbers" and find out exactly how they tend to grow and warp over time...

@futurebird @michael_w_busch

This comment of his says everything you need to know about his relationship with how numbers actually work.

“We're going to get the drug prices down ... 1000%, 600%, 500%, 1500%, numbers that are not even thought to be achievable"

How this man ever ran a business is beyond me.

Our president just fired head of the Bureau of Labor Statistics because the most recent jobs numbers didn't match his feelings about how good the economy is. He did not explain WHY he thought the report was incorrect. Naturally many people now say we won't be able to trust future reports.

To me, faking such numbers seems difficult.

Dr. Reich, @rbreich as a former Secretary of Labor and guy who knows about The Economy can you tells us what might be tell-tale signs of a "Fake Jobs Report" ?

@futurebird Just ask the worst chatbot you can find to write a positive jobs numbers report with some convincing looking references and quotes and paste it on social

@rbreich

I think some people are also alarmed by the "revisions" to numbers in previous months. I remember this always happening in previous reports but these revisions have been very large.

What are some reasons you might need to make a big revision?

@Nazani @rbreich

The finance guys I know use those numbers in their models and they really need them to be based on reality or the models won't work.

Fake numbers will break SO MANY things.

@futurebird @Nazani @rbreich PhDs have been earned finding ways to get around countries that conceal their numbers. They'll just have to apply the methods to the US.

@futurebird

When the economy is holding steady, the jobs numbers only need small revisions.
When the economy is going up or down, the model lags, so the revisions when actual data comes in are bigger.
Larger revisions are a sign that the economy is going up or down faster.

@Kathmandu @futurebird

What is the source of the Job Report numbers? Is it the state job reports or is it some sort of cross country Federal assessment? Be interesting to find out. Based on TACO Turd's nose dive into the National Reporting I would think it would be a system he can disguise the real numbers so as to meet his political wishes.

@BagOfNails @Kathmandu @futurebird The jobs reports come from data collected in surveys sent to employers all over the country. One of the reasons revisions are regularly needed is that many companies return the surveys late, so as more data comes in for prior months they make updates. Some info on the methods they use is here:

Access Denied

(www.bls.gov)

@BagOfNails @Kathmandu @futurebird The jobs reports come from data collected in surveys sent to employers all over the country. One of the reasons revisions are regularly needed is that many companies return the surveys late, so as more data comes in for prior months they make updates. Some info on the methods they use is here:

Access Denied

(www.bls.gov)

@BagOfNails @Kathmandu @futurebird The jobs reports come from data collected in surveys sent to employers all over the country. One of the reasons revisions are regularly needed is that many companies return the surveys late, so as more data comes in for prior months they make updates. Some info on the methods they use is here:

Access Denied

(www.bls.gov)

@BagOfNails @Kathmandu @futurebird The jobs reports come from data collected in surveys sent to employers all over the country. One of the reasons revisions are regularly needed is that many companies return the surveys late, so as more data comes in for prior months they make updates. Some info on the methods they use is here:

Access Denied

(www.bls.gov)

@btrinen @BagOfNails @Kathmandu @futurebird There are two surveys, known as the "household survey" and the "establishment survey". The household survey asks people if they have a job (and if so, how many hours they are working, or if not whether they are looking for one); the establishment survey asks employers how many employees they have. These methods often disagree — and also disagree with official state-based unemployment figures.

@btrinen @BagOfNails @Kathmandu @futurebird Oh, and they're always limited to "non-farm payrolls" because it's really hard to get reliable data from farms.