The monthly employment report from the Bureau of Labor Statistics is a highly watched set of economic data both politically and economically. With at least 10 million people unemployed, and many more either having given up or stuck with short hours, the report is a major indicator of how the economy is recovering.
A bad report reverberates through Congress, political talk shows, and the stock market – and brings fear to the dinner tables of those struggling.
That’s why the last two job reports from the BLS were as chilling as the winter ice and snow.
Growth of 200,000 jobs per month, the pace hit this summer, is considered the minimum for families to keep their heads above water as the population grows. The December gain of merely 74,000 jobs and addition of 113,000 jobs in January is considered far below expectations – and cause for something like panic.
“The American people continue to ask ‘where are the jobs?,’ and the President clearly has no answers. Republicans do,” according to Speaker Boehner.
But a closer look at the data shows that job growth may not be slowing, and in fact may be as healthy today as it has been for the last year, if not stronger. The answer lies in how these reports are calculated and other factors buried deep in the numbers.
“These data are collected via the Current Employment Statistics survey, also known as the Establishment survey,” according to Terrence McMenamin of the BLS.
“The CES is a monthly survey of approximately 140,000 businesses and government agencies. Keep in mind however, that the CES data are revised twice after original publication, once during each subsequent month. These revisions are at least partly due to establishments that are not able to provide their payroll information on time.”
Those revisions are important. The survey method of collecting the data is tedious and often incomplete from one month to the next. Revisions in methods are made at the end of the year, and all of 2013 was revised upwards.
“The over-the-year change in total nonfarm employment for 2013 was revised from 2,186,000 to 2,322,000 seasonally adjusted,” according to the BLS report. That’s an average pace of 192,000 jobs gained per month throughout the whole year.
This was not the only change made at the end of 2013, however. Ian Shepherdson, chief economist at Pantheon Macroeconomics, a private financial advisor, said in an emailed newsletter that “The BLS imposed a much more severe seasonal adjustment on January payrolls this year compared to last, or indeed compared to any year since 2009. Had the seasonal factor for 2013 been re-used, the increase in January private payrolls would have been 265K rather than 142K.”
After the 28,000 jobs lost by government are factored in that would have given us a job gain of 237,000 in January.
The importance of the “seasonal adjustment” cannot be overstated.
January is a terrible month for jobs as additional Christmas retail workers are laid off and construction workers typically have to sit out the bad weather. If you take away all seasonal adjustment to the data there is a mini-recession every January where people are laid off in droves. In 2014 the net actual, non-adjusted job loss was a staggering 2,8770,000 people – very similar to the 2,864,000 who lost their work at least temporarily in 2013.
The terrible loss of jobs from 2008 is shown clearly, but the harsh unemployment every January, and to a lesser extent in July, shows up almost as strongly.
Just as the nasty freeze of winter sets in, millions of families are out of work every year. That effect is almost as deep as the recession itself, too. The seasonal adjustments to the data are designed to make these data into a smoother line that works as a gauge for policy makers and investors, but masks the harsh reality that many people have to live through.
This particular winter has been harder than most, too. Michelle Meyer and Lisa Berlin, economists at Bank of America Merrill Lynch, performed a regression on the effect of snow and cold on the two terrible jobs reports against all prior years.
“Focusing on the winter months of December and January, we find that weather can explain about a third of the deviation from the underlying trend of nonfarm payrolls,” Meyer and Berlin reported.
The effect is not small. “Payrolls should have been biased lower by 41,000 in December and 56,000 in January,” according to their analysis. “We should expect noisy data in the coming months with a general skew toward the downside in January and perhaps February if the forecasts for cold weather prove correct. This would imply a healthy bounce back in the spring.”
Given all of this fudge, can we really put any faith in these data at all?
There is another source for job information from Automated Data Processing, a payroll firm for employers large and small that cuts about 10 percent of the payroll checks in the United States. They simply extrapolate from the real data they have as to who was paid by whom.
The data are very real-time and generally smoother. ADP, however, does not include government jobs, so there is always a note of caution at the end of the reports. They reported a net gain of 238,000 in December and 175,000 in January.
“Since January 2006 ADP has maintained a strong correlation with BLS numbers over the long haul,” said Ahu Yildirmaz, senior director of the ADP Research Institute. “Our report is designed to align with the BLS’ final, revised numbers and not those that are initially reported. This provides real-time data.”
Yildirmaz said she does not see any significant change in job growth occurring in the last two months. “What we are seeing is very much in line with what we have been seeing on a monthly basis. It’s in the ballpark, part of the noise inherent in the data. We have a strong momentum of 175-180k in job growth each month.”
In the end, both the BLS and ADP reports of job growth come from subtracting two very large numbers of the total employment, around 136 million, and find a difference of around 200 thousand – just over 0.1 percent. That comes after a large adjustment for normal cycles that occur from one month to the next. These reports are only useful for long term trends and inherently mask the real struggles of families from one month to the next.
What we can say is that at a rate of 180,000 jobs gained per month the unemployment situation is improving slowly. It will take several years before labor markets are considered “tight” and there is upward pressure on wages. But there is no reason to believe that job growth has slowed at the start of 2014, although we cannot know for sure until spring finally arrives.