Monday, June 6, 2011

A Very Geeky Look at Jobs

It would be great fun to comment about Anthony Weiner's Twitter account. Sarah Palin's claims about Paul Revere (followed by her claim that she didn't misspeak and her followers attempts to adjust Wikipedia to get those bells in there with Revere warning those British) are equally ripe for comment. Mitt Romney says he "Believes in America"and thinks that "the Obama experiment"didn't work, which could also benefit from some careful analysis.

I will leave the commentariat to blather about those things and try to focus on the one item Americans actually say they care about (hint: it's not Obamacare, the Deficit, Medicare, or Pensions). Consistently, polls have shown that people are concerned about jobs in America. Politicians are pointing fingers but nobody has any real proposals to make a difference.

Last Friday the media was buzzing about the employment report. After several months of significant private sector job creation, May was a relatively weak month. Jobs were created but not enough to offset the rise in population. Right away, stories showed up about how this would affect Obama's reelection chances. Someone came up with a statistic that no modern president has been reelected when the unemployment rate was over 7.2% which then got repeated as gospel. An introductory statistics class tells you that you can't predict from such a small number of cases. This is analogous to describing how Ichiro bats in late innings of games in June with a man on second.

The May report could have been affected by seasonal factors. Professor Karl Smith, filling in while Ezra Klein was on vacation, said that this was a result of gas prices being high (before their recent decline). A number of other sources claim that the auto slowdown by Japanese manufactures also played a role. Austan Goolsbee went on the Sunday morning talk shows and said that we "shouldn't make too much of one month's jobs reports". All of these economists are right, of course. We've got to move out of the instant evaluation of absolutely everything and get back to taking the long view.

Here's the geeky thing. Recently, the Washington Post ran an interactive graphic showing the changes in employment from January 2007 to the present. You can see it at

The first page presented here shows that most of the decline in cumulative non-farm employment occurred between July 2008 and July 2009. Since July 2009, employment has been relatively stable with a slight upward tick. But what's fascinating about this graphic is that it lets you break down employment by "supersectors" and "subsectors". It quickly demonstrates the intuitive notion that the jobs recovery is uneven.

There's a button on the upper right of the graph that lets you look at the supersectors. When you click on those, you learn that some sectors are doing better than employment as a whole. Government hiring is up slightly (the census), as is leisure services. Financial services runs a little behind. Mining and Logging run way ahead. So does Education and Health Services. Manufacturing lags behind, losing 15% of its 2007 level. Construction takes a huge hit, losing a whopping 25% of jobs.

When you have the graph of a "supersector", you can click on its curve and it will break out "subsectors' within that. Doing this shows that while Mining and Logging is up, Logging is way down. In the Professional and Business services category, management is up from 2007 while administrative assistants and sanitation workers took a big hit and are slowly recovering (still down 10%). The Financial Sector is off about 10% with real estate hit a little worse than financial services. People who build and finish those houses are significantly worse off than those who sold and financed them. In the area called Trade, Transportation, and Utilities, retail and other trade is down about 5% while utilities remained level. The information sector is down overall, with newspapers and media downsizing, but new web companies show a 25% gain over 2007.

I have had great fun taking apart these numbers and trying to make sense of them (told you it was geeky). Here is what I learned:

1. There are some areas that won't be recovering quickly. The loss of construction and logging is significant and it's hard to imagine what will take its place. The building boom masked a larger problem we had as semi-skilled jobs left manufacturing.

2. While it doesn't break out this way, it's pretty easy to separate upper-class jobs from working-class jobs. In the vast majority of cases, the lower the occupation on the social class hierarchy, the more jobs have been lost since 2007.

3. The areas that are down are NOT those areas most often impacted by upper class investment. These are structural issues and not simply cyclical issues. Lower taxes and decreased regulation may allow some of the flexible industries like finance to expand,  but don't affect jobs overall (just the jobs politicians are most familiar with)

4. But the free market strategies will not fix the structural problems. Those will require an infusion of funds to support the structural realignment -- the "auto bailout" for GM and Chrysler being an ideal example. Infrastructure investment would directly meet the skill sets of the manufacturing workers, construction trades, and even loggers. Such an action could provide support to folks who (as Bill Clinton liked to say) "work hard and play by the rules" while the structural changes work their way through the system.

You can dispute the value of "big government" investment and call it socialism if you want, but the numbers actually do speak for themselves.

That's all my geekiness for today. You can now go back to being outraged at Anthony Weiner or reading about Paul Revere's Midnight ride..

No comments:

Post a Comment