Interpreting The Employment Data

by: Jeff Miller

Today's payroll employment report showed solid job gains as well as upward revisions to prior months.

Econocator has a very nice summary of viewpoints, well worth the look. It does not include the (predictable) Doug Kass viewpoint, described on CNBC's Kudlow show, that these were all government, health and restaurant jobs, actually calling into question cyclical economic strength.

Before the report was released, Barry Ritholtz did a preview (both on RealMoney and The Big Picture) where he raised some interesting questions about how to reconcile the reported job gains, low unemployment, and little wage pressure.

Take a look at Barry's good questions.

Here is the response that I posted on RealMoney (subscription required, and worth it for active traders):

As usual, Barry raises some good questions related to this month's non-farm payroll report. He is especially interested in the apparent discrepancies between the payroll report and the unemployment rate. We'll see a parade of (non-vacationing) economists today along with weekend commentary parsing the data. Much of the interest is in what the implications are for future Fed policy.

When I want to predict institutional policy, I always look for information about its members. How FOMC members look at employment is more interesting to me than how outside economists or pundits see it.

With that in mind, I see one factor that might help answer Barry's question. Barry "sets the bar" for required job growth at a monthly gain of 150K jobs -- needed just to keep up with population growth and inflation. That 150K growth rate has been a traditional rule of thumb.

The Fed members no longer share that view. Because boomers are taking early retirement, the labor force growth dynamics have changed. The estimates of necessary growth include 110K (Fed staff via Bies in November), 130K (Bernanke one year ago), and 100K (Moskow last August). This suggests that this year's job gains, averaging over 145K for the first half, indicate solid growth.

The decision to take early retirement is also not a binary one. I'm at an age where I have many friends who have retired from one job, yet still do some work. They often do not show up in a payroll job, but would say they are employed if asked in a survey. If they do not work, they drop out of the labor force. If compensation gets high enough, they re-enter. The same is true of young people choosing between education and work. This kind of flexibility in labor force participation relieves some of the wage pressure.

Fed members, both individually and collectively, seem pretty satisfied with how their policy has worked so far. There is no reason to expect an imminent change.

Later in the day, Greg Ip at the Wall Street Journal had a nice analysis, citing several sources on labor force participation. This is quite consistent with our argument.


There is a huge error band in payroll employment numbers. Barry's questions are good ones. The birth/death model adds jobs based upon the experience from the last few years. This could change. We will not see the "benchmark revisions" until October.

Having acknowledged this, the other economic indicators that relate to payrolls, described in our own preview, suggest that the economy and employment growth are still quite strong.

Anyone trying to find weakness in this report should be addressing the strength in the ISM manufacturing and service indicators, the yield curve, the Baltic Dry Freight indicator, Dr. Copper, the weekly jobless claims, and other economic measures formerly cited when they supported their viewpoints.

As usual, at "A Dash" we try to look for the best evidence. In particular, we try to understand how the Fed Open Market Committee looks at the data. There is a very good reason for this. They set the policy!