The USA Aug 2012 BLS Jobs Report

The USA Bureau of Labor and Statistics released the jobs report on Sept. 7. I've dissected it into different (sub-)sectors to see how the employment market has been changing recently.

You can get the jobs report here (pdf) or here (html). There are two components to the jobs report: the “A” tables and the “B” tables. You can get the historical data here: “A” tables and “B” tables.

Those two components measure different things. The “A” tables contain data from the household survey; they survey households and ask about the employment status of all the individuals in the household who are at least 16 years old. The “B” tables contain data from the establishment survey; they survey private non-farm business and the government and ask about how many people they are employing. (For both of these surveys, they ask about other things too, like demographic data, hours worked, wages, etc.)

The differences in the two surveys mean that they are somewhat incomparable, in a few ways: the two main reasons are that the “B” survey counts one person with two jobs as two jobs whereas the “A” survey counts that as just one job; the “A” survey includes jobs such as agricultural and self-employment; the “B” survey is not age-restricted; and the “A” survey counts unpaid leave as employed whereas the “B” survey does not (I'm paraphrasing all of this from the jobs report.)

The one-sentence summary of this jobs report is that the unemployment rate decreased from 8.3% to 8.1% (from the “A” tables) and the non-farm payroll employment increased by 96000 jobs (from the “B” tables).

I'm going to talk about the “B” tables, the establishment survey, the number of jobs^[1]. The BLS does seasonal adjustment on the data; this is so that you can meaningfully compare month-to-month differences. I've made two additional adjustments: first, there was a big spike in federal government employment due to the census; I removed that (replacing it with a straight line connecting the before-and-after values). Second, I adjusted for population growth, assuming a 1% annual growth (more precisely, I adjusted by multiplying numbers in the past by the appropriate growth factor so that total population is equivalent to current population).

Just to be clear about this: the graphs I'm going to show you have been adjusted for population growth.

The jobs are categorized hierarchically: total jobs are divided into Total Private and Government; Total Private is divided into Goods-Producing and Private Service-Providing; Goods-Producing is divided into Mining and Logging, Construction, and Manufacturing; and so on. There are a total of 150 categories and sub categories. I'm not going to analyze them all; rather, I'm going to focus on the ones where interesting things seem to be happening when you subdivide the categories. (Also, I'm rearranging the hierarchy to put Private Goods, Private Services, and Government all at the same level.) You can find the definitions of the categories here if you work hard enough.

First, total jobs:

Figure 1: Total USA Jobs

It's obvious from this graph that goods and services jobs dropped significantly, starting in early 2008 and hitting bottom in early 2010. It's also obvious that, since early 2010, services jobs have been increasing steadily and government jobs have been decreasing somewhat. Goods jobs have been declining since (at least) 2002, but (although you can't quite see it in this graph) increasing slightly since 2010.

Let's look more closely at goods jobs:

Figure 2: Goods-Producing Jobs

The steady decline in goods jobs since 2002 is due almost entirely to the decline in manufacturing jobs; all three categories of goods jobs declined significantly from 2008-2010, and since 2010, manufacturing has been increasing, mining and logging have been increasing, and construction has been nearly constant.

Now let's take a closer look at private service-providing jobs:

Figure 3: Private Service-Providing Jobs

Almost all of the sub-categories of services jobs dropped significantly from 2008 to 2010; the one exception was education and health services. Most of the jobs from 2008 to 2010 in these categories were in the Trade/Transport/Utilities and Professional/Business categories. The T/T/U category includes Wholesale Trade, Retail Trade, Transport/Warehousing, and Utilities; taking a closer look at these doesn't reveal anything interesting, other than a steady decline in utilities (by an average 1.5% per year), which are roughly 2.2% of total T/T/U.

Let's take a closer look at Professional and Business Services:

Figure 4: Professional and Business Services

and a closer look at Professional and Technical:

Figure 5: Professional and Technical Services

(“NS” is, as described by the BLS, “other industries, not shown separately”) and now the same categories but (because it's a bit easier to see what's happening) their change relative to January 2002:

Figure 6: Change in Professional and Technical Services

and a closer look at Administrative and Waste:

Figure 7: Change in Administrative and Waste

It's worth looking at Education and Health:

Figure 8: Education and Health Jobs

They've both been growing since (at least) 2002 at a steady 0.6% rate.

Now let's look at government, the final category:

Figure 9: Government Jobs

and at change in government jobs relative to mid-2007:

Figure 10: Change in Government Jobs

(I don't know why there was a spike in April-May 2009.)

Since mid-2007, a total of almost 1.5 million government jobs have been lost, after adjusting for population growth.

We can look at changes in government jobs over the last few years.

Changes in Government Jobs (Thousands)

	Jan. 2009/Jan. 2010		Jan. 2010/Jan. 2011		Jan. 2011/Jan. 2012		Jan. 2012/Aug. 2012
Federal Non-USPS	+108	+5.0%	+19	+0.8%	−38	−1.7%	−31	−1.4%
USPS	−60	−7.9%	−40	−5.7%	−32	−4.9%	−13	−2.0%
State Education	−52	−2.1%	+16	+0.7%	−21	−0.9%	−6	−0.2%
State Non-Education	−69	−2.3%	−76	−2.7%	−103	−3.7%	−35	−1.3%
Local Education	−102	−1.2%	−203	−2.5%	−171	−2.1%	−107	−1.4%
Local Non-Education	−156	−2.3%	−178	−2.7%	−116	−1.8%	−27	−0.4%

Finally, let's look at all this in a more historical context. The data below have been seasonally adjusted but, unlike the data bove, not adjusted for population change.

Figure 11: Total Non-Farm Jobs, 1939 to present (log scale)

Figure 12: Rate of Change of Jobs, from 1939 to present (% per year)

Notes
1. ^The reason I am focusing on the number of jobs, rather than the unemployment rate, is simple: it is a more direct measure^[2] of how much money is getting to the consumers. The private sector is sitting on vast piles of cash, something like $1.8 trillion (up from $800 billion a year ago), and if the consumers were buying more stuff, demand would increase, and businesses would invest this cash in either capital or labour in order to meet the demand. The main reason consumers aren't buying more stuff is because they don't have money. (The hyper-Republican talking-point explanation for the vast piles of cash is government intervention and regulation and the uncertainty over so-called “Obamacare” and global warming, and that the only way to fix it is to lower taxes and de-regulate. This trickle-down drivel ignores the basic fact that a business won't invest unless there is, or will be, demand for their products.)
2. ^It is a more direct measure because, as shown in the tables in the job report, wages and weekly hours are pretty much constant.