How low are those expectations? Bob Pisani at CNBC retells a joke going around the trading desks: "The nonfarm payroll report will be so bad that it can [be] given out by name, not by number: 'Mr. Smith got a job.'" The reports for May and June were disastrous, but not that bad.
Ahead of Friday's report, ADP today announced the findings of its monthly survey of the private job market. It was a bit better than expected, with a gain of 114,000 jobs. But ADP's report is not usually a good match for the government numbers. For June, it reported 100,000 more private-sector jobs were created than the Department of Labor's Bureau of Labor Statistics did. ADP also doesn't cover government jobs, which have been taking major hits for the past two years, especially in the past 12 months.
In addition to the less-than-sterling gains reported by ADP for July, a report from Challenger, Gray & Christmas said private-sector layoffs rose to a 16-month high.
None of that nor reports this week from the Institute for Supply Management on the manufacturing and non-manufacturing sectors of the economy have given any kind of boost to the optimists regarding the BLS report that will be released Friday morning.
Are there any optimists? But let's not go there. This isn't a diary about policy, the need for a WPA and other directly government-funded jobs programs and an industrial plan, and tax revisions that help real job creators. It's about process.
If anybody made money in the past couple of years predicting what would be announced each month by the BLS, it would be news to me. The consensus of experts interviewed by Bloomberg, CNBC, The Wall Street Journal and other business media almost never get it right and are sometimes stunningly off the mark. This month is unlikely to prove an exception. The consensus of experts Bloomberg has surveyed clocks in at a net of 75,000 new jobs. But some predict as many as 125,000.
Given how deep the unemployment hole is, even that top number would be paltry. Hitting the consensus would be four times better than the June report and 50 percent better than May's. If you're looking for odds on whether we'll come close to the consensus figure, you've come to wrong place. Based on previous predictions, it could less than half that, or more than half again as many. Or, for the first time since September 2010, it could even wind up in negative territory.
The problem with whatever numbers will be reported in the headlines—and that most expert analysts and pundits alike will respond to—is that they are not real.
Whoa! Hold on. I am not saying the numbers are fraudulent. It is just that, over time, the statisticians at the Bureau of Labor Statistics have developed formulae with which to adjust the numbers seasonally so there aren't huge swings based on things, for example, like hiring extra help for the retail surge of the holiday season at the end of each year and laying them off in January and February. In a controversial methodology, they also adjust the numbers based on how many jobs are likely to be gained or lost by new businesses that are born and die each month. And they do other tweaking as well.
So do the headline numbers distort reality? Yes. But even if the numbers are not quite as they first appear, do they at least represent a trend that provides useful information? Yes.
The BLS itself includes non-seasonally adjusted numbers as well, but those never make the headlines except at specialist blogs.
Also impinging on how the numbers get interpreted is the fact that the monthly report combines the results of two different surveys, one of tens of thousands of individuals and one of business establishments. From the former we get the official jobless rate; from the latter we get the number of new jobs generated for the month (modified by those adjustments already discussed). This fact almost always creates some confusion, especially when there seem to be lots of new jobs created and yet the official unemployment rate goes up.
That official rate, now at 9.2 percent, has its own very big problems. When people drop out of the work force because they've given up looking for a job—as so very many have done in the Great Recession—the numbers, adjusted or not, can appear better than they actually are.
The math is simple. Let's suppose 1000 people are in the work force and 100 of them without paychecks are looking for a job. Unemployment rate = 10 percent. Now, let's say 20 of those 100 people still want jobs, but they stop looking out of despair. Common sense says the unemployment rate is actually the same; 100 people out of 1000 want to work and haven't found any. But the official work force in this case has now declined to 980, and only 80 out-of-work people are listed as seeking a job. Unemployment rate = 8.1 percent.
As bloggers have brought to everyone's (including the reluctant traditional media's) attention in the past two-and-a-half years, the BLS does include some of those out-of-work people in its statistics, if they've looked for a job in the past year. These come under its alternative measure, labeled U6, now at 16.2 percent. This takes into account people who have no full-time jobs but are trying to get one and those who have looked for a job in the past year, but not in the past four weeks. The current count on all those people is about 25 million. But this does not include people who want a job but gave up looking a year or more ago. Should they be counted, too?
In a way, they are, and that produces one of the most telling numbers that comes out of the jobs report each month, the employment-to-population ratio. That's what the chart way up at the top of this diary shows. It's a devastating percentage.
If you dig into the interstices of the BLS report, you can find other data of interest. A beginning-of-the-year upgrade has made the site much more user-friendly. For instance, there's a calculation on how many people in the work force have been unemployed for more than six months. And tables showing specifics of various occupational fields.
The bottom line in all this? It's not a conspiracy. But when you hear those "headline" numbers Friday, you shouldn't take them literally. If you hear that 50,000 new jobs were created, which would be more than May and June combined, and that the unemployment rate has remained steady at 9.2 percent, take a deep breath and apply some caveats.