Actual employment in the United States is highly seasonal, so one thing the Bureau of Labor Statistics does is apply a statistical formula to adjust employment data to smooth out seasonal trends. Over long terms, the seasonally adjusted data is clearly the superior indicator. Compare the nice smooth curves of the past 20 years’ worth of seasonally adjusted jobs data to the spiky unadjusted data and you’ll see why. But if you’re interested in short-term issues, it may for some purposes be better to pay attention to what’s actually happening rather than to the statistically formulae.
As you can see above, the strong January jobs numbers didn’t mean that there was lots of net hiring in January. January is, instead, always the worst month of the year for employment. What happened in January 2012 is that the rate of job losses, though extremely high, was unusually small for a January. Conversely, the “terrible” May data we just got actually involved a large net increase in employment. It was a bad month relative to the historical average May just as our January was a good month relative to the average January. But the shift in the state of the economy between January and May was very small compared with the underlying seasonal drivers of employment.