Employment: Glimmers of Hope

The latest employment report, released Friday by the U.S. Department of Labor, disappointed on many fronts. However, if we extrapolate current short-term trends -- particularly with the nonfarm payroll data -- there are some potential bright spots.

I'll discuss areas of concern first and then the reasons for optimism.

Many Numbers Remain Bad

There's a long list of bad numbers in Friday's report:

• The unemployment rate rose from 9.8% to 10.2%.
• The loss of 190,000 nonfarm payroll jobs was larger than consensus estimates of a loss of 175,000.
• The total number of job losses accelerated over the past two months.
• The civilian labor force continued to shrink (by 31,000 in October and 903,000 in the past year).
• Average weekly hours worked remained at the historic low of 33.0.
• The number of people working part-time involuntarily increased again, by 105,000 in October. It's up 637,000 since July, 2.5 million in the past year, and 5 million in the past two years).

In addition to the 8.4 million people who lost jobs in the past two years, 5 million more are forced to work part-time because full-time employment is not available. Presumably, many of these part-timers will return to full-time work before new employees are hired. This can skew extrapolations from short-term data. But with that caveat in mind, it's still worth taking a look at what recent employment numbers might imply for the coming months.

Reasons for Hope

The first glimmer comes from the 12-month change in nonfarm payrolls, shown in the first graph. The rate of change has turned up, a positive second-derivative effect. It is also positive that the monthly curve has crossed above the four-month moving average.

The trend line continues to point down, but trend lines do not signal turning points. The data line and the four-month moving average above the trend line are early signals of possible future trend reversal. The quadratic trend line is slightly concave upward (compare with the linear trend line). This is also a possible early indicator of a less negative (or more positive) trend.

The total number of unemployed is determined from the household survey by the Department of Labor. This number is much noisier than the nonfarm payroll number. The household survey has sampling error uncertainty on the order of 300,000 per month, while the establishment survey -- which yields the nonfarm payroll number -- has uncertainty of only about 100,000 per month. The total employment change year over year is shown in the following graph.

These data are behaving differently than the nonfarm payroll data. There is a similarity in that both the data line and the four-month moving average have moved above the trend line. All other comparisons show the total employment numbers failing to indicate the same potential for optimism as the nonfarm payrolls.

Specific observations:

• The data line is below the four-month moving average and moving downward nearly in parallel.
• Both the data line and the moving average are moving downward nearly in parallel with the trend lines.
• The quadratic trend line is clearly concave downward with respect to the linear trend line, indicating a possible acceleration downward.

Clearly, the year-over-year data are giving mixed signals. Total employment is still trending downward, while the nonfarm payroll data show promise of improvement. Even though the monthly data are much noisier than the year-over-year data, we can get information from them that we do not get from the 12-month changes. The two employment measurements we have been discussing are both shown in the following graph, which includes four-month moving averages.