Jobs report for March: A tale of two surveys (but not of a jobs slump)

April 7th, 2017 at 9:52 am

The monthly survey of workplaces revealed a slower pace of job growth in March, as payrolls grew by 98,000, the lowest gain since last May, and down from the pace in recent months, while downward revisions reduced payroll gains from the first two months in the quarter by 38,000. The survey of households, however, told a more robust story of the March labor market, with unemployment down to a cyclical low of 4.5%, and for the right reasons: job seekers finding work, not giving up and leaving the labor market.

That’s the lowest unemployment rate since May 2007, while the underemployment rate–a more encompassing measure of labor demand important gauge–fell from 9.2% in February to 8.9% in March, its lowest rate since December 2007.

While some news sources will be tempted by this below-trend payroll number to declare a slump in employment growth–“if it bleeds, it leads”–that would be a mistake. One month does not a new trend make. The monthly confidence interval for change in payrolls is 120,000 (meaning that there is a 90 percent chance that the true change in payroll employment for the month of March lies between about -20,000 and 220,000; that’s the statistical noise I’m always going on about), the underlying trend remains solid, weather effects may have been in play in March, and the household survey looks strong.

It is not unusual for the two surveys released on jobs day to tell somewhat different stories, and the key point to keep in mind re these monthly numbers is that they are noisy. Therefore, we want to be careful not to over-interpret one month’s results. Instead, we should look at the underlying trends.

Our monthly smoother helps amp up the payroll signal by averaging out some of the noise in the monthly data, taking averages of monthly payroll gains over 3-, 6-, and 12-month spans. Over the past three months, payrolls added 178,000 jobs on average, close to the underlying trend for the past year of 182,000. Given the size and growth of the US labor force, these averages represent a solid pace of job gains that is clearly and steadily moving the job market to full employment. It is, however, a slower pace of monthly payroll gains compared to earlier in the recovery (a year ago, the 12-month average was 229,000).

Source: BLS, Author’s calculations

Still, today’s lower-than-average (and lower-than-expected) payroll number does not alter my assessment that the job market is closing in on full employment. Of course, if future months show a clear deceleration of the ongoing trend–say, a downshift from close to 200K/month to 100K/month, this would signal a decline in hiring activity and could (I’d argue “should”) slow the Federal Reserve’s plans to raise rates.

Meanwhile, tighter job markets provide wage earners with more bargaining power. On average, as the next figure shows, average wage growth has accelerated in recent months, from around 2% to a pace north of 2.5% (March came in at 2.7%). A few caveats, however, are notable. First, inflation has also picked up in recent months, partially due to normalizing energy costs, and was growing most recently at about the pace of hourly wage growth, implying flat real hourly earnings. Also, on an annualized quarterly basis, wage growth was 2.4% in 2017q1, below its recent trend. Given data volatility, this doesn’t yet imply a slowdown, but we’ll track this going forward.

Source: BLS, Author’s calculations

A few other notable results:

–The employment rate for prime-age workers, a closely watched measure to see if the labor market recovery is pulling working-age persons into the job market, was 78.5% last month, up half-a-percentage point over the past year, and another sign of progress. These 25-54 year-old workers have clawed back 3.7 out of 5.5 percentage points, or two-thirds, of their losses since the big downturn. It is thus extremely important to heavily discount reports that such workers are out of the reach of a strong labor market. Some surely are, but many others are clearly not. Be very careful, my friends, not to conflate the cyclical with the structural!

–The decline in the underemployment rate reflects the monthly tick down of about 150,000 involuntary part-timers. Over the past year, that measure of slack is down by about half-a-million workers (6.1 million last March to 5.6 million this March). The figure shows a steady, improving trend in the number and employment share of part-timers who’d prefer full-time work, though the series is not quite back to pre-recession levels.

Source: BLS, Author’s calculations

–Fans of seasonal adjustment were concerned that the March payroll number would be biased down due to weather effects, specifically unseasonably warm weather that raised February’s job gains and, conversely, a winter storm in March that blanketed parts of the Northeast and Midwest during the week in which the job surveys are in the field. But BLS data on absences from work due to weather show a huge spike in full-timers who had to work part-time due to weather (the largest on record for March with data back to the 1970s), but not much of one for people not at work at all. Still, the part-time issue could have dampened March’s payroll count, yet another reason not to worry yet about slumping job creation.

In sum, moderate, steady GDP growth amidst low productivity continues to equal solid job creation that is squeezing slack out of the labor market. The overall pace of job gains has probably slowed a bit but that is not unusual as we move towards full employment and face utilization constraints. However, there’s still room to run on the supply side of the job market, as the prime-age employment rates and involuntary-part-time series reveal. Also, it will be very important for the Fed to carefully track the wage growth point I made above regarding the slower quarterly rate. If that sticks, they’ll want to be very careful not to shut down wage gains just as they’re catching on.

Print Friendly, PDF & Email

8 comments in reply to "Jobs report for March: A tale of two surveys (but not of a jobs slump)"

  1. Smith says:

    Two months about we were about two years away from full employment, so I’m thinking we’re now about 1 year ans 10 months away, at a minimum. How is that possibly characterized as near full employment? How is near 1 year and 10 months? Since when? Why don’t your figures for participation of prime age or under employed ever include actual numbers instead of meaningless percents? How are people who don’t have time to google for hours supposed to match the 180,000/per month new jobs (100,000 of which are needed for 1% population growth rate with a employed workforce of 120 million currently) to those who dropped out or are still stuck with part time work. Ok, I’ll google one more time. Fred says 204 million, but that’s ages 15 to 64. Let’s adjust this 50 year interval to your 30 year definition to come up with a conservative estimate of 30/50 * 204 = 120 million. We have 1.8% left to get back to pre recession numbers, (your 5.5%-3.8%) so even half is .9% which times 120 million equal over a million. It’s easy to estimate from the blue line on the graph above there is more than a million left for the underemployed to reach some ballpark prerecession level. That’s how you get 2 million, by adding 1 million prime age drop-outs and 1 million underemployed. Then 180,000/month less 100,000 for population growth leaves 80,000 per month. 2 million divided by 80,000 equals 25 months. Ok, so it’s a bit over two years. You must either demonstrate how this math is wrong or acknowledge we are not near full employment. How did I figure out how to calculate all this? I learned from you.

    • Smith says:

      Some editing needed where words got garbled in first two sentences above, left in uppercase to identify changes:
      “ago” not “about”, and “and” not “ans”
      Two months AGO, we were about two years away from full employment, so I’m thinking we’re now about 1 year AND 10 months away, at a minimum.
      Quoting word “near” and adding “supposed to mean” to make clear questioning the characterization of full employment to be “near” as opposed to what something a year and 10 months away.
      How is “NEAR” SUPPOSED TO MEAN 1 year and 10 months?
      Also, I cite the figures 5.5% – 3.8% but the actual figures in the blog post are 5.5% – 3.7%, (prime age labor force drop outs), however I nonetheless used the correct figure of 1.8% difference.

      I’m still left perplexed, and have asked this before, what does it mean when an economist ignores his own data?

    • Smith says:

      Your friend Dean Baker seems to agree with me as his latest post shows wages are basically flat or declining, another indication of a weak job market, and an economy far from being near full employment.

      • AngloSaxon says:

        The weakness in wages is over. The totals since November 2016 have showed solid acceleration from the mid-cycle slump. Baker is looking at the lag and not where things are going. What will happen when last summers and early falls weakness is moved off the yry? Weak labor market my………….why do you think inflation has been firming. Another sign, it usual front runs the overall wage acceleration a well. You can’t blame it all on oil prices either. That won’t work.

        • Smith says:

          What inflation are you talking about? His graph shows 3 month moving average, not yearly.
          Check FRED

          Employed full time: Median usual weekly real earnings: Wage and salary workers: 16 years and over (LES1252881600Q)
          “Percent change from a year ago” is trending downward the past three months.

          Consumer Price Index for All Urban Consumers: All Items (CPIAUCSL)
          “Percent change from a year ago” is flat and 1%

          Lest you get too excited about the real median wage climbing to 2.6% in 2015, that was oil related, and since 2009, real median wage climbed only .87 percent, meaning the $45,656 median wage in 2009, after falling about $500 and remaining there for 3 years, is now up $397, 7 years later, and costing the economy zero, no inflation at all. See productivity note below.

          Productivity climbed 7% in that time. 7 – .87 in real wage increase = 6.13%
          Nonfarm Business Sector: Real Output Per Hour of All Persons (OPHNFB)

          Want to guess who got the 6.13%?

  2. AngloSaxon says:

    We are at full employment. Household growth rocketed and NFP will in the 2nd quarter. Follow the past 70 years on this guys, when these series started(from 1948-93, you will have to calculate the U-6 on your own).

    Wage growth is a lagging indicator always. Dean Baker probably thought we weren’t near full employment in July 2005. Yes we were at full employment. There was just a lag.

  3. Smith says:

    What is your data? What are your sources? Why do I have to guess?
    Household growth, I assume you mean the ill-conceived median household income, was reported jumping for the year 2016, even in this blog. Trouble was liberal Democratic economists and people like you had no idea what it meant. The economists kind of wanted to pat themselves on the back and Obama a “job well done” sendoff. Unfortunately as voters knew (especially a few tens of thousands in Wisconsin, Michigan, and Pennsylvania), this was not the case.
    I posted replies on this blog which you evidently can be forgiven for missing. Pay attention, and I’ll repeat the info. Sorry for the attitude, but you are grossly wrong and it affects policy, and the professional economists are getting this wrong also. It’s very frustrating to observe. I googled this to get the information. I have links on previous posts.

    Median income rose for households because of a drop in unemployment, full time and part time, not because of wage increase. When more people in a household are able to find jobs, full time or part time, the income for the average household goes up. That is how household income rises faster than wages. There is also a second effect where if more jobs are in the upper half of income, the median rises, but neither jobs nor wages for the lower half need to change one iota.

    in the case of where we are for full employment, any trend that doesn’t include U6 is absolutely worthless in the current environment. Aside from stagnant wages, and incomes barely above pre recession levels, U6 is one of the primary measurements indicating we are not at full employment. Another important metric is the labor participation rate of prime age population (25 – 55 years old), also still depressed.

    The 2000s were marked by a jobless recovery and stagnant wages (declining even for those with lower incomes and or less education). Even at it’s peak, the occupation with greatest wage growth (STEM and IT) exhibited pathetic increases of .3% per year on average. Check State of Working America for those facts.

    We are not full employment, you have no data to assert we are, we were not in 2005, and inflation has never been 2% or below with unemployment 5% or below in a non recession year in the last 60 years except 1998, 1965, 1955, 1954.

    Study this some more before disputing my facts. I need others helping gather the truth.

    • Smith says:

      Below (scroll down and skip my rant) is a interesting quote from today’s Times about the inflation. This fact bolsters my argument that when even liberal Democratic economists agree to an inflation target of 2%, they are showing their ignorance and stupidity, of history and how the economy works. Why is the 2% figure acceptable? Why doesn’t Krugman crusade against 2% as clearly as he fought moves to austerity? With a 2% target, the stagnation won’t be secular, it will be a deliberate self inflicted wound executed to benefit bankers and corporate interests. One might excuse Yellen’s actions with the hope she is deliberately lying to everyone that 2% is acceptable, in order to prevent further harm by the rest of the Fed or a new Trump replacement. But what is the excuse for economists who should know better? You know, the Democrats who aren’t captives of corporate interests like Prof Summers, or looking for mysteries to solve like Krugman, instead seeing plainly the corruption (which he used articulate).

      “All of this is the result of two broad trends. First, inflation is lower than in the past. From 1950 through 2011, it averaged around 3.5 percent. In January 2012, the Federal Reserve committed to a target of 2 percent, and actual inflation levels have been even lower.”