Employment Report – Nonfarm Payrolls – Easy Trends (thru January 2013)
Employment report figures are perhaps the most closely tracked economic numbers of all. I’m continuing a feature called “Easy Trends” – a place where I’ll analyze the recent trend for the nonfarm payrolls from the monthly employment report and discuss whether it is currently going up, down or neither. You can read the basics of my methodology on the FAQ page.
NOTE: You may be reading an outdated analysis. Please visit my latest Employment Report trend analysis for more info.
Quick ‘n Easy
Nonfarm payrolls, a component of the employment report called “Employment Situation Summary,” is meant to measure the number of jobs in the economy except for those in the farming industry. The large ups and downs in farming make it harder to see the true underlying trend, so jobs in that sector are excluded. What people usually focus on is the change in nonfarm payrolls from the previous month. We typically want to see something above 150,000 during average economic times. (Easy Intro to Nonfarm Payrolls)
Nonfarm payrolls from the employment report is meant to measure the number of jobs in the economy except for those in the farming industry. The large ups and downs in farming make it harder to see the true underlying trend, so jobs in that sector are excluded. What people usually focus on is the change in nonfarm payrolls from the employment report for the previous month. We typically want to see something above 150,000 during average economic times. You can read more in my Easy Intro to Nonfarm Payrolls.
I have decided to analyze the two different methods the Bureau of Labor Statistics uses in estimating the number of new nonfarm payroll jobs added per month in its employment report. The one that most articles quote is from the “Payroll Survey” (a survey of businesses). The second one is from the “Household Survey” (survey of households, duh!). Because the “Household Survey” includes a whole bunch of workers who aren’t included in the “Payroll Survey” the BLS has to adjust the data a bit to help them match up. In the end, they should track each other fairly well. Nevertheless, I felt it would be helpful to look at the trends from both of them. In fact, my personal preference is to look at the average of the two to get a better estimate of the true trend of job creation.
Below is a graph showing the trends in both the “Payroll Survey” and the “Household Survey” after adjustments were made to the latter to represent a similar population. As you can see, they track each other fairly well after adjustments (see the unadjusted data in a line that’s way above the others):
Employment Report Trends and Projections
Below, I will discuss whether the nonfarm payrolls number from the employment report is currently in a trend, when the last confirmed trend was and what that says about projecting the next data point to be released.
Employment Report Trend Analysis
Quick ‘n Easy
The trend of the average of the two surveys (my preferred metric) shows about 184,000 jobs being added per month from Oct 2010 thru Dec 2012. The Household Survey (adjusted) and the average of the two surveys were both too low to be included in the recent rising trend, so keep an eye on that for the possibility of a change in the trend. Clearly, we have job growth but not at a pace that is necessary to make a rapid difference for the millions who lost their jobs in the Great Recession. Details of the report are mildly positive.
Here is a chart of the recent trends:
|Payroll Survey||Household Survey (Adjusted to Allow Comparison)||Average of Payroll and Household|
|Current Trend: Feb 2012 – Jan 2013. During that time, there was a confirmed upward trend with payrolls rising by 154,000 per month.||Current Trend: Oct 2010 – Dec 2012. During that time, there was a confirmed upward trend with payrolls rising by 190,000 per month.||Current Trend: Oct 2010 – Dec 2012. During that time, there was a confirmed upward trend with payrolls rising by 184,000 per month.|
|Last Confirmed Trend: Dec 2011 – Feb 2012. During that time, payrolls were rising by 285,000 per month.||Last Confirmed Trend: Feb – Aug 2010. During that time, payrolls were rising by 171,000 per month.||Last Confirmed Trend: Aug – Oct 2010. During that time, payrolls were falling by 71,000 per month.|
|Projected Change Next Month: +89,000 jobs for Feb 2013 (excluding any off trend data points)||Projected Change Next Month: +896,000 jobs for Feb 2013 (excluding any off trend data points)||Projected Change Next Month: +558,000 jobs for Feb 2013 (excluding any off trend data points)|
The employment report for January 2013 had a headline number that was below expectations (+157,000 net jobs added in January versus an expected +175,000 or so). My own preference is to look at the average of the changes in the Payroll Survey (survey of businesses) and the adjusted version of the Household Data (survey of people) to see the underlying trends in the number of employed people. Looking at it that way, last month we lost 97,000 net jobs, but we’ve been adding 184,000 jobs per month from Oct 2010 thru Dec 2012. The latest report’s average was too low to include in that rising trend, potentially signaling a slowdown in the jobs trend. Since we need about 150,000 jobs per month to keep up with the growing workforce, that means for a little over two years, we’ve been adding slightly more than what’s needed. There’s no way that has (or will) put a dent in the massive numbers of people who lost their jobs during the Great Recession.
In addition to the headline change in jobs, there are several other elements of the report it’s always good to look at:
|Statistic||Basic Definition||Monthly Change||Assessment||Comments|
|Unemployment Rate (Official)||Percentage of job-seekers who have no job||(+) 0.1 pct, now at 7.9 pct||Negative||It isn’t surprising to see an increase when there are more people looking for work (see “Civilian Labor Force” below). But of course we’d like to see this number stay low, not increase.|
|Revisions to Previous Months||Data are always being revised to improve accuracy||Numerous revisions to last several years of data, mostly positive ones||Positive||Every year, the Bureau of Labor Statistics (BLS) takes a look at more comprehensive data to calibrate what the true number of jobs is compared to what its monthly surveys had been suggesting. For 2012, this resulted in 424,000 more jobs than what the BLS had been reporting on a monthly basis.|
|Civilian Labor Force||Number of Americans eligible for and seeking work||(+) 143,000||Positive||It’s always good to see more people looking for work.|
|Participation Rate||Americans who want to work as pct of all available||Unchanged at 63.6 pct||Neutral||It’s still well below the 66-67 pct of normal years in the recent past, but at least it didn’t drop last month.|
|Employment-Population Ratio||Americans who are employed as pct of all available||Unchanged at 58.6 pct||Neutral||We want this number to increase, but at least it didn’t drop.|
|Diffusion Indexes||Degree to which gains in employment were widespread||(-) for 1,3, 6 and 12-month comparisons||Negative||These numbers suggest that the gains in employment are becoming less widespread.|
|Weekly Hours Worked (Aggregate Index)||Measure of how many hours workers put in||(+) 0.1 pct vs last month||Positive||Although hours worked are already factored into “weekly payrolls” (below), it’s good to see the hours worked go up separately because companies will often increase worker hours before they are ready to hire more workers.|
|Weekly Payrolls (Aggregate Index)||Measure of how much money workers take home||(+) 0.4 pct vs last month||Positive||This is a good sign for personal income levels.|
This report contained significant revisions to past data, generally positive ones. However, the Household Survey and the average of both the Payroll and Household Surveys suggest that January may have been much weaker than the recent trend, which itself already was not strong enough to make a big difference in the labor market. In fact, the average of the two surveys suggests that January saw a drop in overall jobs. The internal details tell an overall slightly positive story, so perhaps the trend can remain positive, even if the slope of that trend ends up being a bit less steep than we’d like. We have a confirmed upward trend in nonfarm payroll jobs as assessed by both the Payroll Survey and Household Survey (adjusted), including the average of the two surveys – although the latest figure for the Household and for the average of the Payroll & Household were both too low to be included in that rising trend. It’s just that the rate of rise is not what we need it to be. The end result is still somewhat positive.
It’s extremely unlikely that we were in a recession during January 2013 with numbers like these (i.e., rising trends). Keep in mind though, jobs numbers are usually a lagging indicator. They do not typically predict what is coming.