Unemployment and Government Jobs

Often it conflicts heavily with other private or more reliable sources of data, which have been collectively telling a grimmer story than today’s headline news release of an additional loss of 217,000 jobs for August 2009.

Regardless, here’s the news, with my analysis:

Unemployment rate jumps to 26-year high of 9.7%

By Rex Nutting

WASHINGTON (MarketWatch) – The U.S. unemployment rate jumped to a 26-year high of 9.7% in August as nonfarm payrolls fell by 216,000, the 20th consecutive monthly decline, the Labor Department estimated Friday.

Of course, 9.7% would be dream rate to teenagers who are now suffering their own quiet depression with a 25% unemployment rate, and for blacks and Hispanics, who are pulling up an uncomfortable second and third behind the teens.

U.S. payrolls have dropped by 6.9 million to 131.2 million since the recession began in December 2007, the government data showed. Unemployment has increased by 7.4 million during the recession to 14.9 million.

As you probably know, unemployment is measured in two ways:  the percentage rate is derived from the household data, while the actual number of jobs gained or lost is sampled from ‘establishments,’ meaning businesses.  In the paragraph above, we might note that there is now a 500,000 job difference (7.4 – 6.9) between the number of jobs reported lost by households vs. establishments.

The 216,000 decline in payrolls was close to market expectations of a 233,000 drop, but the unemployment rate rose higher than the 9.5% level expected. The unemployment rate was 9.4% in July.

Payroll losses have moderated in most industries in the past two months. Payrolls declined an upwardly revised 276,000 in July. In June and July, payroll losses were revised up by 49,000.

Oops.  A 9.5% rate expected, but 9.7% hit.  That’s a miss.  Also note the large downward revisions, totaling nearly 50k jobs for June and July.   Downward revisions are a sign of weakness.

Government Jobs

By way of commentary, we might also note in this report that government jobs are now more numerous than the entire goods-producing sector (construction + manufacturing combined), larger than the retail trade, larger than business & professional, larger than education and health services and even leisure and hospitality.

In shorthand, government jobs are now more numerous than every major subcategory and are only outnumbered by "Service-providing" as a super-category. 

There are now only 4.8 private sector workers for every government worker:

And this ratio may be even more unfavorable, because the number of private contractors working on government projects, which has grown magnificently over the past decades, is not included in the "government employee" category.

Sooner or later, we are going to have to come to grips with this enormous imbalance between the producers and the distributors of wealth.

How have we managed to even support this ratio of government-to-private sector workers?

Easy.  We’ve borrowed their salaries from the future:

When you don’t really have to pay for employees, but can effortlessly borrow their salaries with the aid of your friendly thin-air banker, then it’s quite easy to add them.  If government debt had been held constant (taxes would have to have been raised to cover the costs), you can bet your bottom dollar that the number of government employees would be a lot lower than it currently is.

As always, I am not making any value judgments about these jobs, only pointing out that I think there is a serious misalignment between how many we can afford and how many we’d like to be able to afford.

My concern here is that we’re piling up debts that could ruin the country without having a decent conversation over whether 4.8 to 1 is a sustainable ratio or not.  Perhaps we should have that dialogue.

The Birth-Death Model

Of course, as you should know by now, the 217,000 lost jobs were mitigated by the addition of 118,000 jobs by the "birth-death" model.  [Note: They cannot be simply summed, because the 217k is seasonally adjusted, while the B-D numbers are tossed into the "not-seasonally-adjusted" bucket, which is then adjusted…but the direction and magnitude of the impact is clear enough].

My intellectual concerns about this model stem from the fact that it is supposed to account for both business births and deaths (that are largely otherwise invisible to the job counters), but the model only ever seems to be able to conclude that business births are outnumbering deaths.

This is not really surprising, since the model is drawing from an insurance database that seriously lags the current realities and therefore covers a period of growth, not contraction.  That an extrapolation from old data continues to show a persistent upside bias is not really all that surprising to me.  What is surprising is that extrapolations from this data set are being used today, apparently without any modifications to correct for what is happening these days.

But back to the data, the birth-death addition of 118,000 jobs is a 28% increase over the number of modeled jobs added last August.

If anyone can provide evidence that would support the economy being 28% more supportive to new business creation this year than last year, I would love to see the data.

However, as it stands, the credit and retail environment right now are about as hostile to new businesses as any in decades, and it seems a stretch to imagine that business births continue to outpace business deaths by a healthy margin.

As always, collect multiple views drawn from numerous data sources, and then trust yourself.