You Can Ignore Non Farm Payroll Day

Over the past few years, we have admonished traders to ignore the gyrations of any given NFP report. A single monthly number in a series is meaningless, and the overall trend is what matters more, as do certain specific components.

We should generally ignore a single month for many reasons. First, they are the output of a model — increasingly counted less, and hypothesized more. Second, they are highly subject to revisions, as a real time snapshot of a very large and complex economy. Lastly, the monthly changes in a 135 million person labor pool are practically a rounding error. 135k change is one tenth of one percent — 0.1%. And as we have seen on some recent months, a change of ~10k or so is a 0.01% of the full pool, well within the margin of error.

That’s why we emphasize trend, or a 3 month moving average, versus getting too bent out of shape over any given month. And we have done this in good times and bad, when we were bullish or bearish.

Given the potential impact of the snowstorms across the United States during the BLS survey week, the potential range of NFP for this month is half a million people wide. Any jobs report between +100k and -400k would not surprise us.

Here’s a real world example of how easily economists can be led astray: Robert Barbera, chief economist of ITG, noted that January 1996 saw a -200k due to snow, versus the previously run rate of +200k. Economists freaked, and two forecast recession BASED ON THIS SINGLE DATA POINT. The following month? The next month’s NFP was +700k.

That’s why you never rely on a single, potentially aberrational data point oin a long series.

Employment Situation out in one hour at 8:30 am

Employment Situation Summary

Print Friendly, PDF & Email

What's been said:

Discussions found on the web:

Posted Under