Interpreting Statistics

by Gene Callahan

There is always a temptation to think of a statistic as a simple, straight-forward fact: Numbers don’t lie. Consider this statement from Newsweek from an article on structural unemployment in the US:

‘Another theory is that Americans are less willing to move to take jobs. The McKinsey study reports that, in the 1950s, one in five Americans moved every year; now it’s one in 10. “Work is more mobile than workers,” says Camden.’

What could be clearer? Workers are less willing to move. Just look at the numbers!

But let me suggest that things are not so simple as they appear. (To confirm my suspicion here would take a detailed look at the raw data involved, so what I am suggesting here is right now a mere suspicion, but I think it is an illustrative one.) My guess is that this statistic is comparing apples and oranges. In the 1950s and 60s, or so it is my impression, workers did not so much move for jobs as they were moved around by their companies. One found employment with a large corporation intending to work for it for forty years, and then to retire. When the company said, “You’re moving to the Rochester office,” you moved to Rochester, or you were marked as someone who wasn’t serious about their career. From the employee’s point of view, one was not so much “willing to move” as one passively went along with someone else moving one.

With the rise of two-career families and a new attention to at least appearing family friendly on the part of corporations, there is, I think, much less of this today. Employees are now offered a position in another office, and declining is no longer seen as a black mark on one’s record. To the extent the above points are true, then it is not so much that employees are less willing to move, but that they face less pressure to move. (If we want to be formal about this, we might say that w = 1 / min(p), where w is willingness to move and min(p) is the lowest external pressure encouraging a move that will initiate one. Then we might say that, contra the article quoted, it is not that w has gone down, but that the average p has gone down, so that fewer and fewer people’s min(p) is being met.)

Now, I don’t pretend to know what percentage of the 50% decline in annual moving is accounted for by the above change. The point of this post is not to make any claim in that regard, but to illustrate that, in the social sciences, every statistic requires intelligent interpretation to even decide if we have on our plates something meaningful, and, if so, what its meaning might be.

12 thoughts on “Interpreting Statistics

  1. It could also be the case that there has, by and large, been an increase of work-at-home made possible by the advent of personal computers. No longer do so many people have to commute for work that can be done at home on the PC, in a home office.

    This probably another factor that you wouldn’t find looking at “the facts.”

  2. Also have to wonder how much of that earlier movement was local: Moving from one apartment or house to another in the same metro area. If homeownership is significantly higher than it was in 1950, and renters move more often than owners, then there is going to be a lot less moving around that is completely unrelated to jobs.

    There are plenty of stories you can tell from that statistic, it’s absolutely true that social science statistics need a whole lot of explanation.

  3. Statistics abuse abounds! Thanks for another example. Other explanations could be that people didn’t change jobs frequently in the 1950’s because organizations were more hierarchical and companies rewarded loyalty with promotions. And companies promoted from within.

    Beginning in the 1990’s companies flattened their structure and left few opportunities for promotion, so people started jumping ship more frequently. Today people think something is wrong with you if you stay at a company more than four years. The only way to advance is to hop to a higher position at another company.

    In addition, companies quit promoting from within and started hiring outside for top positions. One hops their way to the top today.

  4. “Other explanations could be that people didn’t change jobs frequently in the 1950’s…”

    But the statistic cited wasn’t about how frequently people changed jobs!

  5. As a general rule a company will pay for a manager, officer or a professional to move to another location. But they hardly ever pay for an hourly wage employee to move to a different location. For example, when GM closes an auto plant in one location an hourly wage employee has a preference at another auto plant. But GM will not pay for the hourly wage employee to move.

    But hourly wage or production workers account for some 80% of employment and that share has remained amazingly consistent over the decades. If you actually look at data on hourly wage or production workers moving from state to state you will find that it always been dominated by young, single people that hat move to an area of weak employment to an area of strong employment on their own dime. The difference this cycle is that there is no or little area of strong employment that attracts young mobile workers . For example about the only state now experiencing an influx of out of state workers is North Dakota. But who wants to move to North Dakota?

  6. One factor cited reason for the stubbornly high unemployment rate is the housing crisis. Workers cannot sell their homes in depressed areas to move to jobs in growing markets.

    America is becoming like scelrotic Europe.

  7. It MAY be (this wouldn’t take much inquiry to ascertain for someone with access to JSTOR and the like) that unemployment benefits (from the government) are higher and continue for longer than in the past.

    If an unemployed worker is being sustained well and securely on the public purse, he’ll hold out for more “convenient” employment where he already lives, rather than undergo the disruption and expense of a move.

  8. I was reading Estey’s book on business cycles from 1950 and he wrote that depressions that happen during a downturn in real estate tend to last longer and be more severe, while those that happen in an upswing in real estate tend to be mild and short. That’s because the industry employs a huge portion of the labor market.

  9. I’m a little surprised that no one mentioned the growth of the two-income family in combination with the death of job security. Why move when one spouse will have to quit his or her job and find a new one, while the spouse who is offered the job elsewhere has little or no assurance that he or she won’t be laid off sometime soon at the new location? Having been offered a job at a large corporation and then finding myself laid off with my entire division 3 months later, Ithink this factor definitely needs to be considered. 🙂

  10. Jerry,

    Does that explanation really hold water? I mean, if you can’t find a job locally then you are probably going to have to sell your house for less than what you bought it for — you’re cutting your losses, after all (unless, of course, someone else — like government — is subsidizing your costs). The housing crisis might cause some rigidity — for example, dissuading someone from leaving one job for a slightly better paying job –, but I don’t think it’s impact on unemployment is as big as some might suggest.

    By the way, speaking of statistics, does anybody know how many unemployed laborers are also homeowners?

  11. Jonathan,

    If a homeowner is “under water” (He owes more on his mortgage than the value of the home.), then selling the home requires coming up with cash. You are correct that subsidies, like prolonged unemployment insurance benefits, adds to the inertia.

    In the 1980s, German trains travelling North to South were packed with people commuting from cities like Hamburg to the affluent South (e.g., Baden-Wurtenberg). They couldn’t afford to sell homes in the North. German trains were subsized. So workers commuted long distances for the week. Subsidies favored mobility in that case.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s