Thursday, March 6, 2014

Employment-Population ratio


Torsten Slok keeps making interesting graphs, which make a blogger's job easy.


Lately, there has been a pretty remarkable consensus among macroeconomists that the labor market really is not doing well, despite lower unemployment rate. About 10 million people lost their jobs in the great recession,  and new employment has just about matched new people since then. The employment-population ratio -- red line -- hasn't budged. The 10 million aren't actively looking for work, so they don't count as "unemployed." Whether "discouraged" by persistent "lack of demand" or discouraged by high marginal taxes and social program disincentives, or bad match of skills and opportunities, take your pick, the consensus view on all sides has been pretty dim on the labor market.  I've seen about the same slide deck from Ed Lazear (Bush CEA chair) and Larry Summers (Obama adviser). Usually, employment and unemployment mirror each other, so it doesn't matter which measure you use.

In Torsten's view, there is nothing the Fed can do about this. I agree. The Fed seems to secretly agree too. They talk about the employment-population ratio, but if they thought there were effectively 10 million unemployed and they could do something about it, they would not be even talking about tapering, they'd be talking about buying another $2 trillion of bonds and promising zero rates into the 7th year of the Hilary Clinton administration.

But to the point of the graph: apparently, the share of 25-54 year olds mirrors this long-run trend in employment-population ratio. In Torsten's view, the 55-65 year olds made a lot of money in their 401(k)s and are retired. (The range of the vertical axes is the same, so though not quite counting people, it's close. Beware correlations among series with different vertical axes!)

I'm not quite so optimistic. The average $650,000 net worth Torsten cites is nowhere near enough to live on for 30-35 years, much of this is in housing, and half the households have less. As a member of that demographic, I think there is a lot of useful work to be had out of 55-65 year olds. The same numbers can be read dreadfully as a generation whose location, skills, health insurance arrangements, marginal tax rates (social security disability, etc.) and now long-term unemployment history leave them behind, facing a long painful old age, and the economy without their contributions.

Still, I hadn't really been thinking about demographics in the context of the employment-population ratio, and it's important. Long-run growth of the economy and tax revenue needs overall employment to rise. If the 55+ are forever out of the labor force, we'd better let some young smart taxpayers in pretty fast!

16 comments:

  1. As you say, the vertical scales are different. In addition, I expect that the split of people between the under 25s and over 54s has shifted.

    The Fed might well believe that there are ten million unemployed but feel that they cannot or should not do anything about it. It seems clear that Congress has deliberately chosen contractionary, deflationary, policies - and the main dispute between the Democrats and Republicans is really over how quickly to shrink the economy.

    The Fed is still pumping an awful lot of stimulus into the economy every month and we see it in the housing market and stock market - we are not really seeing it in labor markets.

    ReplyDelete
  2. The New York Fed has done some good research on this. They created a "demographically adjusted" employment/population ratio. Here's an explanation:

    http://libertystreeteconomics.newyorkfed.org/2014/02/a-mis-leading-labor-market-indicator.html

    The upshot: The actual E/P ratio suggests that the labor market has made relatively no progress since the end of the recession in recovering from the 4.1 percentage point decline in this measure. In contrast, the gap between the demographically adjusted E/P ratio using our normalization and the actual E/P ratio is a much smaller 0.7 percentage points.

    ReplyDelete
  3. I wonder if there hadn't been a housing bubble if that "hump" on the red line from 2003-2007 wouldn't have been a straight downtrend. It looks a lot like what you'd expect with an artificial boom followed by a nasty reality-reverting bust.

    And you don't look like you're in the 55+ demographic.

    ReplyDelete
  4. This wouldn't be a problem, but the US has a lot of debt and a shrinking tax base

    ReplyDelete
  5. Unlike many others, I don’t like Slok’s graphs, because they offer a perspective that is too partial. Take for example the graphs that you have just shown. The age class 25-54 is obviously the most important: it contains 30 years out of an average working life of less than 40 years and, above all, it is the class in which participation rates are by far the highest. Because this is the main component of the employment ratio, when you picture them, you see that the employment ratio looks like just a sort of random deviation from the share of the age class 25-54 over total population. And because the dynamics of the latter is entirely driven by demography, then the Fed looks like Don Quixote fighting the windmills.

    But this is a misrepresentation of reality: the figure de-emphasize the collapse of employment (for all age classes) that happened in the Great Recession. Take the figure on page 3 in the recent speech by Larry Summers at the NBER (sorry if I don't copy it, but I don't know how to do it). It is almost the same thing, but the picture is completely different: his red line rightly emphasizes the big collapse of employment that you had since 2008. That is what the Fed is fighting.

    Now you can argue about the effectiveness of the low policy rates (and the paper and blog by Steve Williamson do it brilliantly, as well as some of your own blog posts), you can argue about the credit easing, the quantitative easing and the forward guidance, and the negative effects of all these measures on financial stability, but you cannot give the impression that the Fed is fighting a windmill. Unemployment is a tough and sad reality. And the Fed is no Don Quixote.

    ReplyDelete
  6. This graph does not take account of the fact that the composition of the workforce. Many more women work now than did in 1980, then it was still possible to have a middle class income with just one of the spouses working, that's now simply impossible. If you want a proper comparison, you should multiply the blue line by the fraction that actually work (so when only men work, multiply by 0.5, when both men and women work equally, multiply by 1.0). This would reduce the blue line on the left, relative to the right, and the relative shift of the two lines would be much different. It would then be clear that there is still a large gap.

    ReplyDelete
  7. Just a gosh-dern minute here. If you graph the E/P ratio of just the 25-54 year olds, you get basically the same curve (see: http://research.stlouisfed.org/fred2/graph/?g=sOU). That graph is simply abstracting one level and then ignoring what's underneath.

    ReplyDelete
    Replies
    1. Which graph is that graph?

      Delete
    2. If I understand your question correctly, by "that graph" in my last sentence I meant to refer to Slok's graph, which I think obscures important information by only looking at total E/P.

      Delete
  8. The Fed? You are tslking about an FOMC that used the word more than 500 times in a meeting...in 2008. A half-day meeting. Perhaps "monomaniacal" describes the Fed. The PCE deflator at half of target and still they fret about the microscopic rates of inflation. The Fed would like to absolve itself of any responsibility for economic growth and have a single mandate. But modern economies do not fare well at zero inflation which seems to come hand in hsnd with ZLB and perma-recession. See Japan.
    I can remember the GOP screaming at the Fed to loosen up...in 1984 when inflation was running near 5 percent...

    ReplyDelete
  9. After reading sherouse's post I am greatly disturbed...the conclusion drawn is that you cannot use data to predict itself...this is thoroughly confused as mathematically it implies there is no such thing as a trend, or any structural meaning to data...

    ReplyDelete
    Replies
    1. Thanks for reading, but I think you mistook my point. I argued in that post that if you use data to predict itself, you can't draw conclusions from the fact that your predictions are close to observed values, because how could it be otherwise? That's what it strikes me that Kapon and Tracy were doing. Obviously, trends are hugely important, which is why I used a trend drawn from a pre-recession sample to make an alternate estimate.

      Delete
    2. I see, that makes more sense, thanks!

      Delete
    3. He cherry-picked the years, otherwise the result is a lot less clear. From 1948 through about 1970, the age 25-54 share of the population fell, while the employment-population ratio rose. There's a story to be told there, but it's not quite so simple. (Unless one thinks that the 1948-1970 period is irrelevant.)

      Delete
  10. "The average $650,000 net worth Torsten cites is nowhere near enough to live on for 30-35 years, much of this is in housing, and half the households have less."

    Average =/= Median I'd bet more than half have less.

    ReplyDelete
  11. "As a member of that demographic, I think there is a lot of useful work to be had out of 55-65 year olds."

    THIS IS NOT THE AGE COHORT THAT IS STRUGGLING!!!

    http://www.zerohedge.com/sites/default/files/images/user5/imageroot/2012/12/Total%20jobs%20since%20Obama%20by%20age%20-%20all%20cohorts.jpg

    ReplyDelete

Comments are welcome. Keep it short, polite, and on topic.

Thanks to a few abusers I am now moderating comments. I welcome thoughtful disagreement. I will block comments with insulting or abusive language. I'm also blocking totally inane comments. Try to make some sense. I am much more likely to allow critical comments if you have the honesty and courage to use your real name.