Friday, July 8, 2011

June jobs report so bad it's difficult to believe

According to the June Establishment Survey, only 18K jobs were added, far below expectations of 105K. Private sector jobs rose only 57K, against an expected 132K, and were very far below the 157K that was predicted by the ADP survey yesterday. Worse still, according to the June Household Survey, jobs fell by more than 500K! A weak employment report might be understandable given that the economy was still suffering from its "soft patch" in June, but bad news of this magnitude—and especially considering the huge variance between the household, establishment, and ADP surveys—must be considered suspect. Sometimes these surveys just go buggy. When they all line up in the same direction, then one can be reasonably confident that the data is somewhat sound. But when they vary by almost 600K, then it's time to take a deep breath and avoid jumping to extreme conclusions.

This first chart helps put things in perspective. Note first that the household survey can be extremely volatile at times, and June's big employment decline has plenty of precedents. More often than not, a big move up or down in one month is followed by a reversal in the next month or two. Second, both surveys still appear to be in an uptrend, and both are saying that about 2 million private sector jobs have been added since early 2010. Taking an average of the six-month growth rate of the two surveys (next chart) yields an annualized growth rate of 1.3%. So it's reasonable to think that the economy is still growing, and that private sector jobs are probably increasing on average at a 1.3% annual rate, or about 120K per month.  That's a very weak number, to be sure, but it is consistent with the fact that the unemployment rate has risen in the past two months and weekly claims for unemployment have been hovering at just over 400K per week (there has been no indication in the claims data to suggest a sudden and massive loss of 500K jobs in June). Conclusion: the economy is still growing, but slowly.

For the past year or two, one consistent theme in the jobs data has been a steady—and perhaps even an accelerating—decline in the number of public sector jobs, and June was no exception. As the chart below shows, public sector jobs have declined by a little more than 500K on net since their peak in 2009 (abstracting from the temporary census jobs). This is welcome news, since bloated public sector spending and public sector payrolls (and their attendant and generous benefits) have been a drag on growth for some time.

It's also of note that the decline in public sector payrolls to date was matched only by the 500K decline (bigger in percentage terms) in public sector payrolls that occurred from 1980 through 1983. This is a nasty shakeout for the public sector, but not without precedent. Happily, the economy enjoyed very strong growth from 1983 through 1989 after suffering through wrenching adjustments. I would like to think that all the bad news we are seeing these days (e.g., a painfully slow recovery, very high rates of unemployment) is setting us up for a shift to more growth-friendly fiscal policies, and it may come sooner rather than later given the looming issue of the debt ceiling.

UPDATE: reader "brodero" suggests that seasonal adjustment factors could have played a significant role in the June jobs weakness, just as seasonal factors have sent false signals with the weekly unemployment claims data. This would be one more justification for my decision to focus on the six-month trends in the data. Note this quote from High Frequency Economics:

In trying to explain the very weak June jobs report, Ian Shepherdson of High Frequency Economics examines the June seasonal factors. Because of summer jobs, the unadjusted number of June payrolls is always higher than the adjusted number. Shepherdson says the SA-NSA spread in private payrolls was -920K in June 2010; this year it was -1084K. The swing in the seasonal factor, -164K, was the biggest ever recorded (the data go back to 1947). He concludes: "Had last June's seasonal been used again, private payrolls would have been +221K."


Kurt said...

I have heard one wag describe government jobs as "anti-jobs" since it requires several private sector workers to support each government worker. That would imply that in a proper accounting, one should actually subtract some multiple of the number of government workers from the number of private sector workers to arrive at a true employment figure.

That is perhaps a hyperbolic bit of humor, but it would be fun to analyze the behavior of such an index compared to other economic indices.

John said...

"public sector payrolls (and their attendant and generous benefits) have been a drag on growth for some time."

But remember, everybody is somebody's customer (or used to be).

brodero said...

The 6 month rate of change in private sector employment is still me when it is near
flat ( a specific %)

TradingStrategyLetter - Weekly Summary said...

'It's difficult to believe....?'

I think you hit the nail on the head again!

10 year treasury yield rises to 3.5% and all of a sudden the economy is commotose.

Benjamin Cole said...

The Fed needs to get aggressive--on the expansionary side.

Karen Carpenter never quite got fit--doesn't work for the economy either.

We are doing a Japan. Sad to see.

TradingStrategyLetter - Weekly Summary said...

Time for real leadership and making tough decisions!

Benjamin Cole said...


They may slice some jobs off at the state/local level, but check out these stats:

Employment By Department

Department of Defense 3,000,000
Veterans Affairs 275,000
Homeland Security 250,000
Treasury 115,000
Justice 112,000
Energy 109,000
USDA 109,000
Interior 71,000

Labor 17,000
HUD 10,000
Education 4,487

The chances for truly restricting federal employment strike me as very small. Every one of these federal employees can vote, and they have families that vote too. Add to that, there are 2 million employed through Defense contracting.

The private-sector bears a heavy burden to finance all those federal salaries and pensions.

John said...

"Congress has now shifted its efforts 180ยบ: no longer is the debate about how much to spend on stimulus, but how much spending to cut...In short, out with the Keynesians, and in with the Supply-Siders." CBP 5/27/2011

How long do the Supply-Siders get to fix this mess?

NormanB said...

1) Mike Martz, the NFL coach, says 'Believe What You See."

2) Anonymous says, "The Facts Are Your Friends."

3) Obama says, "Its the uncertainty about the debt ceiling."

brodero said...


June Payrolls - Further Analysis An unusually aggressive seasonal may
explain at least some of the payroll weakness. The June seasonal is
always negative (that is, the level of adjusted payrolls is less than
the unadjusted level) to offset hiring for summer jobs. Last June, the
SA-NSA spread in private payrolls was -920K; this year it was -1084K.
The swing in the seasonal factor, -164K, is the biggest ever recorded
(the data go back to 1947). Had last June's seasonal been used again,
private payrolls would have been +221K. The seasonals for May (-89K
compared to a year ago) and April (-93K) were also quite harsh. But
why? Perhaps the late Easter, but Easter was last so late in 1943, so
there's no contemporary history to go on. Maybe the BLS has been

Scott Grannis said...

John: supply-side policies can produce results very quickly. Bush's tax cuts in June '03 were followed in very short order by a surge in growth and a sustained decline in the unemployment rate. Once Reagan got his tax cuts fully implemented in '83, growth surged and unemployment collapsed.

Tax cuts are not essential this time around, since a significant cut in government spending would effectively lower future tax burdens by a significant amount. The burden of government is best measured by spending, in any event, not by taxes.

Scott Grannis said...

brodero: excellent point. Seasonal factors can wreak havoc with the data at times. When the data look crazy like they do now, then it's time to suspect that something fishy is going on.

John said...

Scott: Thanks for tolerating my contrary remarks. You're a gentleman and I appreciate your blog.

But about the Bush record:

"During his eight years in office, President Bush oversaw a large increase in government spending. In fact, President Bush increased government spending more than any of the six presidents preceding him, including LBJ. In his last term in office, President Bush increased discretionary outlays by an estimated 48.6 percent.

During his eight years in office, President Bush spent almost twice as much as his predecessor, President Clinton. Adjusted for inflation, in eight years, President Clinton increased the federal budget by 11 percent. In eight years, President Bush increased it by a whopping 104 percent."

Veronique de Rugy- George Mason University Mar 2009

Mark Holder said...

Whats crazy is that so few economists/analysts actually address the fact that the data appears very buggy. Maybe its b/c the majority of them expect/want a weak economy to back up there bearish claims for months and years now. Otherwise, its difficult to look at all the data of the last few weeks and believe that the June jobs report was really that weak. Or it backs up the theory that productivity is soaring and companies are growing without adding jobs. Very bullish for profits!

Scott Grannis said...

John: Re Bush's spending. I would be among the first to criticize Bush for spending too much and expanding the role of government, but I think that he gets a bad rap, to some extent, on this issue.

Consider that federal spending as a % of GDP was 17.9% when he took office in 2001. It rose a bit in the next few years, which is typical of recessions and the aftermath of recessions. It averaged 19.1% from 2001 through June '08, when it was 19.9%. From 1968 through mid 2008 federal spending averaged 20% of GDP. With the panic spending that ensued in the last six months of Bush's presidency, spending surged to 22.2% of GDP by Dec. '08. That spending was clearly only a one-time deal. Abstracting from that, Bush's spending was actually less, relative to GDP, than the average of all the presidents since 1968.

I would note that a good deal of the decline in spending relative to GDP during Clinton's term was due to cutbacks in military spending.

John said...

How did the Bush "surge" after the '03 tax cuts compare to growth after Clinton raised taxes in '93?

Scott Grannis said...

John: In the six months following the Bush tax cuts of '03, GDP grew at an annualized rate of 5.25%. In the six months following Clinton's '93 tax hikes, GDP grew at an annualized rate of 4.7%. Not a significant difference.

Out of curiosity, I compared the performance of the stock market over those same two 6-mo. periods. After Clinton's tax hike, the S&P 500 fell 3%; after Bush's tax cut, stocks rose 14%.

Regardless, I'll admit it is difficult to attribute changes in growth to changes in policy variables, mainly because so many variables enter into the equation. My big-picture take is that Clinton inspired more confidence, and on balance economic policies during the Clinton years were better than they were in the Bush years. Monetary policy was on average very tight and the dollar very strong during the Clinton years, but just the opposite during the Bush years. Not surprisingly, then, growth was much stronger in the Clinton years than in the Bush years.

Benjamin Cole said...

Interesting commentary by Brodero and Grannis. Let's hope for the best.

McKibbinUSA said...

The US employment to population ratio (unadjusted data0 continues to indicate stagnation in employment growth -- more at:

At some point, the US will be complelled to acknowledge the Main Street Depression that is raging across America...

db said...

Just want to say thanks for a great site to get a well informed opinion on the economy. Please keep up the good work.

Rick said...

The "data" is not actually data. It is a series of statistical extrapolations from the real data; i.e., the information provided in the surveys to BLS. This means that we are trying to interpret the results of something akin to an opinion poll.

Scott - in prior posts, you accepted the validity of the good news provided by the same statistical extrapolations. Now, when the data does not fit your thesis, you question the data instead of your thesis.

The table A-8 estimates simply reflect a downward revision of the number of new private sector jobs created over the past 12 months. The seasonally adjusted and non-seasonally adjusted stats are basically in balance and tell the same story. In the private sector, there are about 1.020 million more people employed in June 2011 than in June 2010.

In April, the economy hit a brick wall. Whatever the causes, this is what happened. Private employers had increased their payrolls in anticipation of continued greater demand. Instead, demand slackened. Employers laid off many of the newly hired.

We can argue the exact numbers. However, we cannot argue that a lot of people lost their jobs in May and June.

Scott Grannis said...

db: thanks for the encouragement!

brodero said...

Rick..."In the private sector, there are about 1.020 million more people employed in June 2011 than in June 2010."

I don't see the numbers...

Y-O-Y Change In NSA private sector

2011 06 1859.
2011 05 1792.
2011 04 1771.
2011 03 1661.
2011 02 1534.
2011 01 1146.
2010 12 1126.

Ali said...

Scott - come on now... I'll let Philippa Dunne from the Liscio Report spell it out (h/t JMauldin):

“After the release, some bulls turned to that old reliable excuse – bad seasonals. According to one analysis making the rounds, had the BLS used last year's factor – computed, of course, using exactly the same concurrent technique as this year's factor – the gain would have been 221,000! (Whoever did this made a mistake by comparing the NSA and SA levels for the two months – you have to compare the over-the-month changes.) Still, if you're going to play this game, you should be consistent, and apply last year's seasonals to several months, not just one. If you do that, May's gain of 25,000 would turn into a loss of 19,000, and June's gain would be a mere 73,000, all total payrolls. In any case, why should you do that? The seasonals are recomputed every month based on recent experience and calendar quirks, and should be more aggressive in a recovery. (Hope we won't be using the trend set in the depth of the recession as the bar going forward.) Also, there is no adjustment to the headline number – the sectors are adjusted separately (96 different industries at the 3-digit NAICS level, to be precise) and the total is the sum of those components. The whole argument is bogus.”

septizoniom said...

this is one of your most embarrassing "shill" posts ever, and as you can see from the comments, others are disappointed with your blinkered and petulant insistence that you are always right. grow up. use your real talents and not your pom poms.

Bill said...

Ok, eliminate the seasonal factors argument and explain to me the huge difference between ADP and BLS, decent ISM numbers, Monster Index growth, American Staffing Index growth, etc? Can we say perhaps that June was a bad month for jobs based on the BLS but there are other indicators that suggest the world isn't coming to an end?

Scott Grannis said...

The jobs numbers are notorious for their susceptibility to huge revisions after the fact. To begin with it is impossible for the government to get an accurate count of the number of jobs. They use statistical sampling which always has a margin of error. They use seasonal adjustment factors that can be and often are wrong. The June number could be revised up or down by hundreds of thousands a year or two from now.

The point of my post is that when the jobs numbers themselves are all over the map, and don't square with other indicators, then there is a reason for being skeptical.

brodero said...

Ali....Mauldin is hardly an unbiased source,,,and barely
knowledgable about statistical sampling...

John said...

"Happily, the economy enjoyed very strong growth from 1983 through 1989 after suffering through wrenching adjustments."

Two things to consider that helped then but hurt now: oil prices and demographics. For Reagan, oil went down to $10/barrel and the boomer generation was in its peak consuming years. Different story now, sorry to say.

Rick said...

Scott - if I use Table A-8, as you posted about last month, by subtracting Line 7 from Line 5, I get 116,717,000 seasonally adjusted private sector employed in June 2011 versus 115,699,000 in June 2010. That's a gain of 1,018,000 year over year versus a higher year over year gain from May 2010 to May 2011. Why the different stats in your reply?

Scott Grannis said...

Rick: you're mixing apples and oranges. The figures you see on that table for Jun '10 are non-seasonally adjusted, whereas the figures for Jun '11 in the right-hand column are seasonally adjusted.

Rick said...

Scott - I don't agree. I'm using Column 5 and Column 10. Columns 2 and 4 are not seasonally adjusted. At least, that's how I read the BLS spreadsheet.