A Quick Question

Recently jobs numbers came out … two data points are of interest and can be used, perhaps, to judge the bias of the sources. One set, points out that January job numbers are up and by one metric unemployment has dropped to 8.3%, getting firmly below 9. The other set, which is new as well, points out the divergence between two proxies which normally track but recently have diverged. Unemployment, as tracked by those applying for unemployment benefits and jobs, normally tracks well with the number of unemployed. Yet in the last 18 months this tracking has diverged. More and more people have (according to the second unemployment proxy) have given up seeking work. By that second metric, unemployment is above 10.5%.

Honest reporting would, I offer, report both points. There are many, who for political reasons, decide on or the other figure is more significant. Are there good reasons besides the nominally “bad” political partisanship ones for not noting both?

Oh, by the by, I’ve got to run early to get to a job site about an hours drive north. My links post will go out tonight.

9 Responses to A Quick Question

  1. In theory unemployment rate is defined as a fraction with the labor force in the denominator (which is those working plus those unemployed) and the numerator being the unemployed. To be unemployed you have to be willing and able to accept a job and seeking one if not already employed. Therein lies the rub. Unemployment could go down because some people give up looking entirely…or it can go down because some may get jobs. The first is ‘bad’ and the second is ‘good’.

    Honest reporting would entail not necesarily reporting both metrics but using one consistently, not switching to one that gives a higher figure to make it appear that unemployment is getting worse rather than better. The standard metric has just as much downside as any other. If people think the economy is getting better and decide maybe it’s worth trying to find a job again, that could drive the rate up rather than down even though people are entering the market because they feel it’s improving.

    The business media usually digs into the unemployment rate metric in more detail, detailing out how many raw jobs were added or workers left the labor force entirely. Rather than confusing the issue by trying to have two or more metrics measuring the same thing, I think it would be more efficient to use the labor force participation rate to pair with unemployment. It essentially measures the sum of the employed and unemployed divided by who is able to work in the economy.

    There is already, though, enough confusion about accounting for metrics built from different data sources. For example, the unemployment rate is built from a household survey. There is also a payroll survey which measures incoming payroll taxes by worker to see how many new jobs are showing up. There are also raw stats like the number collecting unemployment, the number of new unemployment claims, the number expiring etc. All of them have pros and cons so it’s not like one metric can be said to be so much better than the others that others can be dropped entirely.

  2. One important addition, it’s not at all clear those who leave the workforce should be ‘forced back in’ by being counted as unemployed. A coworker of mine has a mother who was a highly paid nurse supervisor. For a long time she has been unemployed after getting laid off. He wanted her to stop looking for work and move in with him and his wife, esp. as they have young kids and can help with childcare during the day when they both work. Let’s say she does this because she is getting older anyway and closing in on retirement. While she may be able to get a lower paid job as a regular nurse if she keeps looking, she would rather verge into retirement.

    Should she be counted as unemployed and part of the labor force? If tomorrow the nurse supervisor field started hopping again and she was getting multiple offers of $100K+, she might very well take them. But right now she isn’t looking for work and would be counted as not part of the labor force. A decline in the labor force participation rate may be due to a bad job market, but it also may be due to people deciding that work is not something they have to do given their circumstances.

    This is why they often pair reporting on the unemployment rate with the raw number of new jobs added to the economy. That metric indicates we’ve crossed the 200K jobs per month required to match the growth of the workforce population and have left over jobs to bring down the unemployed numbers.

  3. Boonton,
    You seem to be missing the point of the two news points. Point one is jobs were “up” in the last two months, pushing one metric of unemployment under 8.5%. The other piece was that of two metrics unemployment the second measures unemployment at 10.5% currently and (this is the important bit) until quite recently, i.e., the last 18 months or so, these two different ways of measuring unemployment tracked closely but recently have diverged. So when you offer,

    A coworker of mine has a mother who was a highly paid nurse supervisor. For a long time she has been unemployed after getting laid off. He wanted her to stop looking for work and move in with him and his wife, esp. as they have young kids and can help with childcare during the day when they both work. Let’s say she does this because she is getting older anyway and closing in on retirement.

    So? Why has this sort of thing gone become so popular as to create a 3% divergence? That’s the question.

  4. I’m not sure what’s “up” with your “quote” marks. Jobs are indeed up in the last few months since the raw number of new jobs are reported when the common unemployment rate is reported.

    So? Why has this sort of thing gone become so popular as to create a 3% divergence? That’s the question.

    Well the standard metric would report her as leaving the workforce thereby no longer contributing to the unemployment rate. That would have no bearing on the fact that jobs are ‘up’ since that’s a measure of actual jobs, not a rate. I presume your alternative metric would still include her on the presumption she’s really still unemployed. But the story about her illustrates a point of fuzziness. If she decides to help out with her grandkids instead of looking for a new job at what point will she stop being considered a ‘discouraged worker’?

    As for why one metric has started tracking down while the other hasn’t, that’s not all that unusual. Some measures by nature lag others. A monthly survey, for example, will capture a surge of newly hired workers within a month or two of it happening while counting them by incoming payroll tax receipts will lag by up to 3 months. Even within the same metric frequent revisions are common. It’s interesting that in the report on the standard metric, not only was last month’s new job creation high, the previous months were revised upwards as well. On top of that metrics that are computed using different methodologies will often never be the same.

    So here’s the thing, if you had two different unemployment metrics that diverged in different directions, one going down the other up, I think it’s a valid concern. Maybe one trumps the other or maybe both cancel themselves out meaning the underlying truth is no real change in unemployment. Two different metrics moving in the same direction is a good sign that the underlying truth is being reflected in the metrics. One metric moving while the other isn’t, therefore, hints that the underlying truth is starting to move…but not as strongly as if you saw both move. Here I think experience has shown that the divergence of metrics isn’t as useful a way to judge the reality of a movement as other factors like consecuative months of the same movement, revisions of previous months consistent with the current movement etc.

  5. Boonton,

    That would have no bearing on the fact that jobs are ‘up’ since that’s a measure of actual jobs, not a rate. I presume your alternative metric would still include her on the presumption she’s really still unemployed. But the story about her illustrates a point of fuzziness. If she decides to help out with her grandkids instead of looking for a new job at what point will she stop being considered a ‘discouraged worker’?

    Irrelevant. Why is this behavior accounting from a 3% discrepancy which newly appears in the last 18 months.

    if you had two different unemployment metrics that diverged in different directions, one going down the other up,

    Which is what we have.

  6. Boonton,
    Or at least have had. Here’s the reference I couldn’t find last night.

  7. Which is what we have.

    Err no you said we have one staying level while the other is falling.

    Why is this behavior accounting from a 3% discrepancy which newly appears in the last 18 months.

    Consider the nature of the story again. The person can work but they can also live comfortably without working. Who are these people? Well my friend’s mom who straddles between working and retirement. Maybe kids who are living on their parents dime. Maybe spouses who work to supplement income but the cost of child care puts a floor below which it ends up costing more for them to work than they get. Maybe spouses who work for benefits rather than cash and would leave the workforce if their spouses get benefits.

    So the answer to your question is pretty obvious, what drove them out of the economy is a bad job market and they won’t come back in unless the job market gets really good again. Over time, some of these workers are going to be out so long that they will effectively cease being in the labor force. The example of my friend’s mom fits that case. If she is out of work another 5 years raising grandchildren instead, at some point she is not going to return to work ever again and your proposed alternative metric is going to have to figure out a way to drop her out of the rolls of the unemployed and put her into the ‘not in the labor force’ block.

    The other factor here is the structurally unemployed. Those with outdated skills or low marginal productivity who were the first fired. Because the recession put so many others out of work, they will be the last to be rehired hence that rate will remain sticky until the ‘official rate’ goes down a lot more and starts to pull down the ‘unofficial rate’.

  8. BTW, you may want to follow Modeled Behavior for a less spinny analysis of economic metrics.
    http://modeledbehavior.com/2012/02/07/jolts-day-4/

  9. Boonton,
    OK. Thanks … I appreciate following econ blogs less spinny than the ones you’ve recommended in the past (de Long and Krugman).

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>