A number of key economic data were released on Friday. At 8:30AM, the monthly employment numbers came out. The Establishment Survey data seemed pretty good: Nonfarm payrolls increased by 157K, right around what was expected, and the prior two months were revised upward by a total of 127K jobs – a real positive. On the other hand, the Household Survey showed the unemployment rate moving higher, from 7.8% to 7.9%. The yield on 10-year bonds fell precipitously on this data, from 2.02% to 1.97%, and then lost another 5bps to 1.92% over the next hour:
The market clearly focused more heavily on the unemployment rate. As Bill Gross and others pointed out, if the unemployment rate remains elevated, the Fed will continue with their QE3 asset purchases for a longer period of time, meaning downward pressure on yields will continue longer than had been thought before the release. Later in the day, yields rebounded on a number of positive economic data releases, but rather than focus on that, I’m going to look at the jobs data in more detail.
Although the new jobs number and the unemployment rate come from two different surveys, when they point in opposite directions like this, it can be helpful to take a deeper look at the data to see where the difference came from and whether it can tell us anything about future reports.
The table below is from the Bureau of Labor Statstics database. The first line shows the Establishment Survey’s monthly job increases, including any revisions to prior months. Over the last three months, a total of 473,000 additional jobs were created. The next line in the table shows the number of job increases from the Household survey over the last three months, which in total was a loss of 6,000 jobs. That’s a difference of almost half a million jobs.
However, if you look at the final column of both lines, the total number of additional jobs in the last twelve months was just about equal. What this means is that the Household Survey had many more new jobs earlier in 2012, and the Establishment survey caught up in recent months.
|Nov 2012||Dec 2012||Jan 2013||Total 3 Months||Total Feb 2012-Jan 2013|
|Establishment Job Increases||247,000||196,000||157,000||473,000||1,717,000|
|Household Job Increases||(51,000)||28,000||17,000||(6,000)||1,714,000|
Here’s a chart showing the number of new jobs each month from both surveys over the last two years. Over this time period, the Household Survey (red) has been very volatile and has tended to lag the Establishment Survey (blue).
This next chart shows the cumulative number of jobs added according to each survey. Even though in any given month you might see a discrepancy between the two surveys, over time you would expect them to produce roughly the same results, because at the end of the day, a job is a job however you count it. Since the Household survey has lagged the Establishment survey over the last three months, there could be a larger bounce in the Household jobs in the next few months.
If we’d like to figure out how the unemployment rate is going to decrease based on the level of new job creation, we need to look at both the numerator and denominator. The next table shows more details on the unemployment rate. Since last January, the economy added 1,714,000 new jobs (143,000/month on average), while the labor force increased by 1,298,000, resulting in a drop in the unemployment rate from 8.26% to 7.92%. On average over the last year, the labor force increased by about 108,000 people per month. If that holds true going forward, and the economy were to continue to add 143,000 new jobs per month, the unemployment rate would drop by about 0.03% each month, or about 0.3% every ten months. By the end of 2013, if these increases were steady, unemployment would be about 7.6%.
|Jan 2012||Nov 2012||Dec 2012||Jan 2013||12-month change|
|Household Number Employed||141,608,000||143,277,000||143,305,000||143,322,000||1,714,000|
|Household Labor Force||154,356,000||155,319,000||155,511,000||155,654,000||1,298,000|
More optimistically, if we project out using the Establishment Survey average of 224,000 new jobs per month since November, the unemployment rate would fall by 0.08% each month or 0.8% in ten months. Unemployment would be at 7.1% by the end of the year.
Take these projections with a few grains of salt. For one thing, if the employment picture improves dramatically, then the labor force number might increase faster as more people not currently looking for a job will start looking for one. This will increase the unemployment rate.
What does this all mean for yields? If the Household Survey is going to catch up with the Establishment Survey, it means more jobs and lower unemployment. This should be positive for yields heading higher and is supportive of my short treasury investment thesis.