X-Files, 2021 Edition

It seems every decade or so, there is a bout of paranoia about government statistics, to wit, a reader asserts the BLS is hiding data on median wages:

Scylla and Charybda. SSA doesn’t have any data since 2019. And BLS has no current data on median compensation at all, except in 1980 dollars or in indices.

I guess the American voter is just not supposed to know how the average American worker is doing in terms of wages at any given point in time. Personally, I’m glad that SSA has actually made data available that is easily comprehensible, even though it’s somewhat dated. It’s better than nothing.

Apparently only educated elites with sophisticated methods of calculating the data are supposed to know.

What’s wrong with that picture? What is the government hiding? And why?

When confronted with BLS series Employed full time: Median usual weekly real earnings: Wage and salary workers: 16 years and over (LES1252881600Q) stored as on FRED, the same reader writes:

The problem is that the BLS data series both track average wages over time. This can be verified by reflating 2019 BLS data from 1982 dollars to 2019 dollars and comparing the result to the SSA data, which is highly reliable, since it relies on IRS data.

In other words, he/she is saying what the BLS has labeled as median is actually average wages…

This episode reflects the general tendency among some commentators to reject or disparage the data (or accuse the government of malfeasance) when the data do not cooperate with a preferred narrative.

Paranoia is everywhere. For instance, regarding employment data, FoxNews is a purveyor, as are these examples: Senator BarrasoJack Welchformer Rep. Allan WestZerohedge . And who can forget longtime reader Ricardo (aka Dick/DickF/RicardoZ) writing in 2014:

…Our government is doing a serious disservice by falsifying the employment condition in our country. Policy changes that could actually help are being delayed with false information.

The belief the data is either by design or accident distorting the data is so widespread that even people who have PhD’s in finance have suspicions.

I would be the last one to assert government statistics are perfect. For instance, TIC data are known to have problems in tracking actual capital flows. GDP data are known to be revised. But the misleading conclusions usually arrive with the person with the narrative, not the persons generating the data (unless it’s ShadowStats).

I don’t usually end with a plea — but this time I do. Before one starts asserting that (1) the government statistical agency is hiding the data, (2) the government statistical agency is mislabeling the data, (3) the government statistical agency has miscalculated a growth rate, or ratio, or difference, please, please, please spend 15 seconds on FRED.

And before accusing your blogger of misattributing the data, please, please, please read the notes to graphs and at the data source portal (often FRED).







10 thoughts on “X-Files, 2021 Edition

  1. pgl

    JohnH, CoRev, and Judy Shelton are very much the same. And two of these three are also gold bugs. Well I don’t think CoRev is a gold bug.

  2. macroduck

    Menzie, you are giving your new stalker a lot of attention. For someone as damaged as Johnny, negative attention feels better than no attention. It gives him a false sense that he matters.

    Here on your blog, you are the Freudian daddy, the fairy tale king, the sun diety. Now, there may he a bit of Cain in Johnny’s behavior, trying to eliminate his rivals because you found his offering inadequate, but it’s still your attention that motivates him. And that’s especially true if his masters see that you’re paying attention to him. Double points!

    If you could lead him back to the land of honest discussion, then this would be the story of the prodigal. All signs indicate that it’s Cain and Abel. Johnny is to be marked forever as the killer of truth. That mark requires that we spare him, but also that we shun him.

  3. Not Trampis

    you could be right however in doing this Menzie ( or Chinny as we call him down) under performs a valuable educational service.

  4. J. Barkley Rosser, Jr


    Judy Shelton’s PhD was not in finance but in business administration from U. of Utah, kind of funny given that the econ dept there has probably the most radically leftist econ PhD program in the US.

    1. Menzie Chinn Post author

      J. Barkley Rosser, Jr. Well, yes, but her dissertation was a finance dissertation. Many places, the degree is Bus. Ad. while the specific subject is finance (similarly, they might have a dissertation in marketing or accounting, but the degree is Bus. Ad.)

  5. baffling

    as menzie pointed out in a previous comment, the ssa data is simply the average per worker. there does not appear to be any criteria related to whether the worker is full time or not. the bls, on the other hand, appears to be median wage for full time workers. john, are you aware of this difference in the data?

    1. pgl

      “john, are you aware of this difference in the data?”

      One would hope he is aware of the difference. Now getting to admit as much – do not hold your breath.

  6. Dr. Dysmalist

    My phone, or something, has eaten my comment three times; let’s see if a fourth time works before I try the large mallet remedy.

    In a previous life, I was collecting data for a large (for it’s day) project. This was in the Stone Age when the “Internet” seemed to comprise mostly words (email, newsgroups, chat rooms, etc.) and there wwas not much, if anything, of a World Wide Web. Data existed in large, dusty volumes housed in large, dusty libraries, at least until you copied it and entered it into a spreadsheet so that you could print it to an ASCII file that your stat software may or may not decide to read. You saved your work to 5-1/4 inch floppy disks.

    Anyway, I wanted to use a lot of data published by USDA ERS, and I had the opportunity to travel to D.C. to meet with the economists who curated a lot of the data. They were great, and showed me that, despite my days of searching and reading footnotes, endnotes, technical notes, and notes about notes, I didn’t know much. “This series sounds great but don’t use it: political influence means it doesn’t show what they say it does. Use this other one that sounds worse but has high integrity.” “That description is the best translation to English but it doesn’t really mean what you think; that government actually means X when they use those words.” “Don’t use the most recent 3 years of that series; it takes 3 years for all the revisions to be finalized.” I was blown away.

    They were highly professional, extremely friendly, courteous, and helpful, truly experts. I know that most government employees that collect, curate, and publish economic data are the same way. They’re primary motivation is, far and away, to publish the best data possible. There are very few slackers or time-servers in these jobs; the work is too difficult for that kind of person. There is a damn good reason that policymakers, private analysts, and academicians trust these people and their data: the gov’t folks have earned that trust over many years of exemplary, unbiased service.

    So when I hear or read of someone, especially someone who starts with a predetermined and biased conclusion, impugning the integrity and character of those government employees, I become very, very angry. I immediately know several things about the impugner: 1) they themselves have little if any integrity; 2) On the subject at hand, they don’t know asshole from shoe sole; 3) They’ve never met even one person who does that job yet they’re comfortable accusing them of incompetence, dishonesty, and bias; 4) Given 1, 2, and 3, a reasonable conclusion is that the accuser is a pi$$-poor excuse for a human being. Just because the accuser is lazy, ignorant, biased, dishonest, and lacks integrity, that doesn’t mean that others are the same. The vast majority of people do their jobs with honesty, integrity, and a sincere effort to do it well.

    This kind of disingenuous crap really pi$$es me off. If highly trusted data doesn’t confirm your biases, it’s your biases that are wrong. Stop bloviating and start learning. Otherwise, you’re an ignorant cretin.

Comments are closed.