Median Real Weekly Earnings for Workers back to 2016Q4 Levels

Figure 1: Employed full time: Median usual weekly real earnings for Wage and salary workers, 16 years and over, 1982-84 CPI Adjusted Dollars, seasonally adjusted (blue), and Employed full time: Median usual weekly nominal earnings (second quartile)f or Wage and salary workers, High School graduates, no college, 25 years and over, seasonally adjusted, deflated by CPI, not seasonally adjusted (red). NBER defined recession dates shaded gray. Orange shading denotes Trump administration. Source: BLS via FRED, and author’s calculations. (updated 8:30PM Pacific)

36 thoughts on “Median Real Weekly Earnings for Workers back to 2016Q4 Levels

  1. 2slugbaits

    I thought Larry Summers had an interesting take on the recent bonuses that some companies are handing out. Some sectors are seeing tightening labor markets, which would normally signal higher wages. But we’re not actually seeing higher wages. Instead, companies are compensating employees with temporary bonuses rather than permanent wage hikes. That says a lot about how long businesses expect economic conditions to continue. If the economy goes south, then employers just won’t give bonuses. This has some interesting macroeconomic consequences because it’s one way of reducing nominal wage rigidity. We might see lower unemployment rates during recessions, but lower wages during recoveries. And will workers view bonuses as temporary windfalls, or will they adjust permanent income expectations?

  2. Bruce Hall

    With the unemployment rate at rock bottom levels, it would seem that:
    1. marginally added employees are not in “core” high paying jobs, but perhaps lower paying service sector jobs… or
    2. companies were holding the cost line until the tax laws were changed and may open up the coffers if they attempt expansion with a tight labor market… or
    3. the companies are miraculously able to get more employees in a tight job market by paying them less.

    I suppose there are other explanations, but they haven’t occurred to me yet.

  3. PeakTrader

    The employment-population ratio increased from 59.8 in December 2016 to 60.1 in December 2017. Perhaps, older workers, who earned more were replaced by younger workers earning less. Over time, real wages will rise for younger workers. Nonetheless, the minimum wage hasn’t kept up with productivity gains and should be raised.

    Also, I suspect, out of control health care costs depressed wages, since those costs were absorbed in compensation. Moreover, it’s possible, less health care was used, because of higher deductibles, which would slow health care inflation. Furthermore, it’s possible, the booming stock market facilitated early retirements.

  4. Moses Herzog

    For whatever it’s worth coming from a non-expert who, in an amateur fashion, reads tons of Economics and finance materials, I wouldn’t be surprised to see that adjusted median earnings go back down to around the 315 level it had in the early ’90s. In fact, with the condition knotted to it that the VSG lasts past 2020, I would predict that EXACTLY THAT will happen. Guys like the VSG and his political sponsors weigh much of life success in how low they can drive down wages for the working man/woman. I doubt sitting at the White House has changed the VSG’s agenda on the wages issue. The VSG and his wealthy pals often sell or market this agenda through coded lexicon, like “low-inflation”, “added burden on the employer”, but those reading good sources and the events unfolding can see through the “low-inflation goals”, “burden on the job-creators” smoke screen. THE REAL GOAL FOR REPUBLICANS IS DRIVING “LABOR COSTS”/WAGES DOWN

    1. pgl

      “I wouldn’t be surprised to see that adjusted median earnings go back down to around the 315 level it had in the early ’90s.”

      What is amazing is how this is not much higher than when Jimmy Carter was President.

      1. Moses Herzog

        Let’s say for the moment you are correct. At least towards the last 2 years of Carter’s presidency, the oil embargo might have played a “small role” in that. That would have affected ANY presidency, as it did Nixon’s in ’73. It’s too bad Carter wasn’t willing to use debt to invigorate the economy like Reagan loved to use exponentially increased government debt to fund fiscal policy. Then maybe the toothless illiterate rednecks of the Tea Party would be singing Carter’s praises to this day. Carter and Mondale were “fools” (politically) enough to believe Americans would want revenues (taxes) to be MATCHED with government expenditures, when MOST Americans wanted a “free ride” on their children’s back. Pity for those 2 guys. Mondale and Carter just didn’t understand how guys like Newt Gingrich and Michael Deaver could lead dumb*ss Americans around by their nose.

  5. SecondLook

    I’ve always thought that just focusing on wages is a bit misleading without taking into account the total cost of labor, i.e wages plus non-cash compensation.

    To provide some context:

    Employer costs for employee compensation averaged $35.64 per hour worked in September 2017,
    the U.S. Bureau of Labor Statistics reported today. Wages and salaries averaged $24.33 per
    hour worked and accounted for 68.3 percent of these costs, while benefit costs averaged
    $11.31 and accounted for the remaining 31.7 percent. Total employer compensation costs for
    private industry workers averaged $33.55 per hour worked in September 2017.

    It’s worth noting that the percentage ratio between wages and non-wage costs in 1986 was 73% to 27% respectively.
    It should be self-evident what has been the single biggest driver for non-wage compensation.

  6. Ed Hanson


    You have written extensively, usually in connection with your Wisconsin posts, that the household survey for employment is subject to wide correction and you prefer, properly, the establishment survey statistics. Why is it different with this post? The data is derived from the household survey.

    The closest to the category you charted from the establishment survey I could find was production and nonsupervisory employees. It is seasonably adjusted and also adjusted for the cpi-u. The BLS writes;

    “From December 2016 to December 2017, real average hourly earnings increased 0.1 percent, seasonally
    adjusted. The increase in real average hourly earnings combined with a 0.6-percent increase in the
    average workweek resulted in a 0.7-percent increase in real average weekly earnings over this period.”

    Your chart does not show such an increase. Is there a better establishment category? Does your author’s calculations change the results?


    1. Menzie Chinn Post author

      Ed Hanson: My criticism pertained to the household survey estimates at the state level, where the sample sizes are relatively small. In this post, I am citing estimates from the household survey at the national level. My use of the household survey is based upon the need (1) to report a median (as opposed to arithmetic average) figure, and (2) to disaggregate to a demographic category, namely workers with high school-no college education (red line in graph). I hope you understand why a median might be a particularly interesting statistic relative to arithmetic average given the skew in the distribution.

      1. Moses Herzog

        I’m asking a sincere question here, not my usual smart-A ones. I can kind of imagine the bell curve in my mind, and you are saying the tails of the bell curve throw the numbers off, yes?? Are you saying the HIGH end (the higher number wages) of the tail would be thicker than the low end of the bell curve?? That the arithmetic mean would make the wage number look FALSELY or MISLEADINGLY high??

        1. Menzie Chinn Post author

          Moses Herzog: Just saying median is a better representation of the ordinary person’s experience than the average because of skewed distribution with many some at very high end. Hence, the arithmetic mean is not the most informative measure of central tendency.

      2. Ed Hanson


        Sorry about the rookie mistake, I missed that the establishment was mean not median.

        I am somewhat surprised that interviews with only 15,000 nation wide is not considered considered relatively small. That plus a question usin what I can only assume a purposely vague, leaves question in my mind that this part of the household survey is not subject to volatility and difference from a corresponding establishment survey.

        What is important about the category of high school-no college education? Except maybe to emphasize that our government education system is not flexible enough to provide decent education to those who can not or do not want to go to college. But then again, don’t get me wrong, I would hate for the inefficient and political government education system to expand. A straight forward voucher system applicable to variety of choices including vocational would solve more problems than it would create, and over time would allow this group to choose better paths to maximizing their earning potential.


        1. Moses Herzog

          @Ed Hanson
          You’re not dyslexic by any chance, are you?? Sometimes that can affect reading comprehension. Because I’m assuming you read the whole post BEFORE commenting.

  7. rtd

    Let’s just hope this trend doesn’t continue to the extent witnessed during the Carter administration (of which Menzie omitted). Although, I’m not even certain why we’re denoting presidential administrations… Menzie, would you mind pointing readers to the literature focusing on the impact of the executive office on Median Real Weekly Earnings? Thanks.

    1. Menzie Chinn Post author

      rtd: I started with the beginning of the Great Moderation, but since you think it important, I extended graph back to beginning of series, 1979Q1. Does it change anything in your mind?

      1. rtd

        I think it’s important to the extent that you seem to be implying the recent trend is to be of interest. If so, it’s nice to show a similar trend. I just found it strange to omit a few years – seemed strangely selective without an explaination. As made obvious by my statement, I already knew what the time series looked like, so I certainly am not in need of you extending the graph to change my mind. Although my mind should’ve been clear in my comment “Let’s just hope this trend doesn’t continue to the extent witnessed during the Carter administration.”

        Anyway, I’m much more interested in the literature focusing on the executive office’s impact on Median Real Weekly Earnings.

        It’s worrisome that I’ve had to nearly recreate my initial post in its entirety – I wonder if you even read it. Was my comment too long for you to read? Or did it not contain your name enough times? Maybe you’re simply disinterested (altough you did reply). Boy, your resemblance to POTUS seems to grow daily.

      2. rtd

        I just think it strange to omit a few years of the time series. I also thin it’s strange to omit the part of the time series that displays a precipitious drop – which is what it seems you’re trying to point out.

        Why would you extending the graph change my mind? I already knew what was on the graph – this should be obvious by my comment.

        I’m more interested in you providing the literature focusing on the executive office’s impact on Median Real Weekly Earnings – as stated by my comment.

        I wish you would’ve posted my original comment further showing the similarities between you and POTUS. I’m too exhausted to try to recreate it. Censorship is what I would expect from your twin Donny Boy.

        1. Menzie Chinn Post author

          rtd: You know GDP data extends back to 1947 on a quarterly basis? But I haven’t always graphed GDP back to 1947, preferring to start in 1967 or 1986 (Great Moderation start). That’s so one can focus the more recent period and see the details. Industrial production goes back to 1919; guess I better start plotting the entire time series because something back in the 1920’s could look like something in the 2000’s…

          1. rtd

            No, you’re confused (or didn’t fully read/comprehend again)… please reread my comment a few more times and try again.

            You’re being silly. You omitted a less than a decade of data in the initial attempt at “analysis” – not nearly 40 – 60 years. Your arguments are ridiculously weak in saying tongue-in-cheek “guess I better start plotting the entire time series because something back in the 1920’s could look like something in the 2000’s”. Showing an additional 7? or so years isn’t the same as an additional 60 years and you know it.

            More so I find it comical that you agree to make the change and then continue to try to argue your point after. Kind of reminds me of when Trump endorsed Roy More before an election and after defeat claimed he knew Moore couldn’t win. The similarities between you and TheDonald are compounding.

            Also, you again forgot to provide resources on the how Median Real Weekly Earnings are impacted by the executive branch.

          2. Menzie Chinn Post author

            rtd: Who’s Roy More?

            In a rat-x world, expectations of tax cuts spur GDP growth and additional aggregate demand can pull up wages. Work for you?

          3. rtd

            Obviously a typo on Roy Moore, my phone’s autocorrect isn’t always correct.
            Do you have a Razr?

            I believe the central bank influences gdp growth and could (would???) offset fiscal policy that might have a tangible impact on gdp growth (that’s what works for me). But either you don’t believe the rat-x theory you’ve expressed or the data doesn’t fit the theory. Or something altogether different. Toy see, other than seemingly implying “here I’ll present the tenure of the guy i don’t like accompanied with a dataset that isn’t in his favor”, your initial post couldn’t be more vague.

        2. Menzie Chinn Post author

          rtd: A quadratic loss function which is typically assumed in formulations of the Taylor rule would imply the central bank offsets some — but not all — of the stimulus arising from a tax cut, particularly in the short run. What Fed reaction function are you assuming?

  8. joseph

    “I’ve always thought that just focusing on wages is a bit misleading without taking into account the total cost of labor, i.e wages plus non-cash compensation.”

    You don’t have to speculate. The BLS collects data on both wage and total compensation including benefits.

    In the last 10 years, real wages have been flat — almost no change. If you look at real total compensation, the increase has been a pathetic 3% over the last 10 years.

    So no, you can’t just wave you hands and say vague things about “benefits”. We have real numbers, including benefits, and they don’t look very good.

    In fact, you can go back to around 1975 and there was an abrupt change in the rate of change of total compensation. Since, it is been very mediocre except for a short jump in the late 1990s.

    1. SecondLook

      I think you misunderstood me, I wasn’t speculating, I was just pointing out that the column’s focus was on wage growth and not on the often ignored larger picture of total benefits/costs.
      It has been postulated that poor wage growth is partly due to increased total costs, most particularly health care.
      By the way, my comment was purely agenda-free, simply a note about the tendency to narrowcast some significant economic questions – mostly I suspect out of habit.

  9. Neil

    It is sort of hard to take this data point seriously given the ECI data on private wages and salaries released today. Shows a steady uptrend continuing.

    1. pgl

      Upward trend in nominal or real? Remember – inflation is not zero. I just looked at the overall ECI since 2017. Average nominal increase per year has been only 2.6% whereas consumer prices have risen by 2% per year on average during this period. So real ECI has been growing at about 0.6% per year since this series has been reported. Better than nothing but less than productivity growth.

    2. 2slugbaits

      Neil Keep in mind that Menzie is reporting median earnings rather than mean earnings. Earnings are bounded by zero at one end of the scale but unbounded at the other end. There’s a lot of skew in the earnings data…the 1% and all that.

  10. Moses Herzog

    I was thinking of scanning a small section of a book that pertains to this topic (and many others), turning it into a file, and then trying to “upload” the file inside a comment on this blog, do you know if that will work technically or not?? Is the comment section here able to take uploaded files inside of a comment??

    1. Menzie Chinn Post author

      Moses Herzog: I don’t think so. You can post the file somewhere and provide the URL, but not imbed (e.g., graphic) as far as I’ve had experience.

      1. Moses Herzog

        I am exceedingly lazy, but sometime soon I will try to put a blog post up with THAT file and then put the URL either in this comment section or in other posts future posts that seem related. Mondale’s book has never “flown off the shelves” so I’m thinking a page or two to put the situation in context is not “out of line”

        It may seem pretty far off topic, but I wanted to post a section of Mondale’s book where he discusses his infamous debate “snafu”. Where basically he tells Americans he will raise taxes. I thought it was one of the most courageous stances by a national politician in American history (telling Americans the truth) He explains his thinking at the time. This may be “a stretch” to connect Mondale’s stance on HOW fiscal spending should be done and our nation’s budget balance to the topic of Median earnings for workers. But in “the big picture” I think they are actually pretty closely related, and I also think Mondale’s stance is one of the most misunderstood, misinterpreted, and misrepresented moments in American political history.

  11. Moses Herzog

    This is the actual speech, I had incorrectly remembered it was in the Mondale/Reagan debate, but it was actually in Mondale’s nomination speech. The file that I plan to put up in the near future, explains in Mondale’s own words his thinking at the time of making the statement he makes towards THE END of this video. and Telling the TRUTH to voters was “the end” of Mondale. This is why politicians (outside of Mondale and a minute few others) continually tell LIES, because that is all the American voter wants to hear.

  12. Moses Herzog

    I guess thematically it fits better with the Kansas post, but since I rambled on so much about it, I guess I’ll put the URL here Menzie. The book pictures are awfully small, but if you click on the book pages (files) they’ll become big enough you can read them. I hope Menzie or someone appreciates the ideals and conviction in Mondale’s words, that are so missing in today’s politics.

  13. Erik Poole

    Doesn’t that graph make you just make you wish that everybody had elected Obama for a third term?

Comments are closed.