Job Growth Slumps. Employment Surges.

More people have gone to work during the past three months than at any time in decades.

Those same three months have seen fewer jobs created than during any similar stretch in more than a year.

Both of those statements are true, or at least can be supported by Labor Department surveys. It is just that the surveys disagree.

According to the household survey, in January 1,717,000 more people were working than were working in October, adjusted for a small change due to changes in population estimates that are made each January. That is an average monthly gain of 572,000.

The Bureau of Labor Statistics has a series that adjusts for such changes going back to 1990. During that period, the largest three-month change was a gain of 1,449,000 in the three months through October 1994.

Meanwhile, the establishment survey reported that total employment rose by 462,000 jobs from October 2013 to last month, an average of 154,000. That was the smallest three-month increase since the summer of 2012.

The discrepancy can be traced to the fact that the two surveys are different in concept, and often different in result.

The household survey is based on calls to households, asking who in the household was working during a particular week in the month. It is much more volatile than the other survey, and is (almost) never revised.

The establishment survey is based on information from employers, and is heavily revised. The first revisions come the next month, as more responses arrive from surveyed companies. Then, in January of each year, the figures for the 12 months through the previous March are revised again, based on an entirely different source of numbers â€" unemployment insurance premiums. When we look at history, we see the figures from the revisions, not the ones from the initial surveys. It is possible that a year from now the figures for the past three months will look very different.

Over time, of course, the two surveys will tell similar stories. But they can diverge substantially over short periods of time.

The figures should not be identical, because they include different things, but they should be close. (The household survey, counting people, includes the self-employed and counts someone with two or three jobs only once. The establishment survey ignores the self-employed, and, since it counts jobs, can double count those with multiple jobs.)

The household survey is used for the unemployment rate, and for such statistics as the employment-to-population ratio. Until recently, that survey was lagging the other one in job creation, and reported a steady fall in the employment ratio. That has led to a surge in analyses asserting that the unemployment rate is down only, or at least primarily, because discouraged workers are dropping out of the labor force.

Michael Shaoul of Marketfield Asset Management has been arguing for some time that the household survey was missing jobs, and that as a result much of that negative comment was in error. In his view, “January was a well-overdue catch-up which addressed a good portion of the shortfall” but there is still more to come. In any case, the employment-to-population ratio rose during January and is now at its highest level in more than a year.

In other words, those who are depressed by the latest job figures may be overreacting. And perhaps some of the hand-wringing about people dropping out of the labor force has also been overstated.