A Better Gauge for Job Growth

Bob Johnson: This week, we look at three different charts and three different ways to look at monthly employment growth.

First, we're going to start with nonseasonally adjusted data. This is the data as it literally comes in. You can look at the monthly data and see there is huge volatility ranging from something close to 3 million jobs lost in a typical January to something as high as more than a million jobs added in April. The volatility is caused by various known seasonal factors. In January, it's retailers coming back from the holidays and shutting down a little bit from their holiday employment. You see another big spike in jobs in the other direction in April and May when we add almost a million jobs. That happens as the normal college graduation class comes out and as normal seasonal hiring for the summer begins. And then we have a fall-off in July--people going back to school as well as some of the things going on in the auto industry. Then, the other big spike you see there on the upside is in October, and that's largely related to the retailers beginning to hire on the other end of the cycle for the holiday season.

So, this presents an extremely volatile look that's confusing and hard to interpret. So, what's another way to look at the data? Well, the government tries to use seasonal-adjustment factors, which is shown in the second chart here. This takes the number of jobs and applies a seasonal factor. Here, you can see a much smoother pattern, and you can actually see a little bit more of what's going on. But even still, there is volatility here, too, where we've had anywhere from roughly 150,000 jobs in any one month to 250,000 jobs in another month. So, again, there is still some volatility. You'd expect these bars to be roughly the same height and slowly sloping upward if the seasonal-adjustment factors were perfect.

Instead, I suggest that maybe we look at the data a third way, and that is looking at the percentage growth on a year-over-year basis and averaging that over three months. This chart looks almost flat. There's very, very little variation in the data. We've been in the range of 2.1% to 2.3% almost all the time over the last four years. So, this method shows the incredible stability of employment, and I think it better reflects what's going on in the underlying economy than the other two graphs, which both have their own set of issues.

By the way, the graph right now suggests that employment is still healthy, which is great news for the economy because we've never had a recession while employment was still in a healthy growth mode.