Discouraged Workers And Seasonality

I want to provide a little more detail related to my earlier analysis on this morning's unemployment numbers. I provided a bit of optimism in my prior post, and as Derek noted, some seem to be scoffing. First, don't get me wrong. 9.7% is a heinous unemployment rate. That's the highest number in decades. In fact, it's the highest rate we've seen other than a short period from 1982-1983 since as long as Bureau of Labor Statistics data goes back (to 1948). I don't mean to minimize this. I'm just trying to understand the overall trend. That's where I think you need to consider how seasonality and discouraged workers are affecting the picture.

Let's begin with discouraged workers. I created this chart using BLS data, going back two years:

09-08 discouraged workers.PNG


As you can see, there was a pretty steady movement upwards in discouraged workers since the recession began. And the number has risen dramatically. That's why I find this month's decrease to be relevant. It could just be a blip, and discouraged workers could go back up next month. But if this is a new trend, it's certainly a good thing for our economy. I also like fewer discouraged workers because that helps the reported unemployment rate to more accurately depict Americans' actual employment status (since the reported rate does not take discouraged workers into account, though it probably should).

Here are two curves, neither of which is seasonally adjusted. They show unemployment rates with and without discouraged workers.

09-08 discouraged worker variance.PNG


When the recession wasn't so bad, you can see that the curves were much closer together. As things got worse, they began to stray apart a bit further, implying more discouraged workers. You'll see that the latest statistic (for August) begins to bring these curves back together. Again, I think this is positive.

Seasonality also has some wacky effects. I'm not saying seasonality is completely irrelevant, but I worry that, during a recession, it could skew things. Here's a graph of two curves. They show the seasonally adjusted and not seasonally adjusted unemployment rates (neither of which includes discouraged workers).

09-08 seasonality variance.PNG


As you can see, with seasonal adjustments, you have a much smoother rise. But without seasonality, there appears to be a more positive trend recently. Since these are actual numbers, I find this more comforting than when seasonally adjusted numbers suggest an improvement, but the actual numbers suggest otherwise (like in July).

Finally, I thought one more graph might be interesting. This one is a little strange, but still kind of useful. It shows non-seasonally adjusted unemployment percentage including discouraged workers and the widely reported seasonally adjusted unemployment percentage not including discouraged workers. Here you get wide variances, because you've got two variables that can lead the curves in different directions.

09-08 discouraged and seasonality variance.PNG


This one is difficult to explain, but I'll take a shot. First, seasonal blips are very clear in April of 2008 and 2009. In each of these months, the green line leaps downward briefly before trending back up. But what about the decrease we're seeing in the green line from July to August in 2009? In 2008, the not seasonally unemployment rate including discouraged workers was unchanged at 6.3% for July and August. Yet, for 2009, we saw 10.2% and 10.0% for July and August, respectively. It's hard to know exactly how seasonality is affecting this, but if you compare 2008 to 2009, it seems that this year's decrease might be meaningful.

It's impossible to know for sure if the potentially positive signs we're seeing for the August numbers are just blips or new trends, but I think there could be something there. Of course, in context, unemployment is still ugly. I don't mean to dispute that. I'm just trying to see if there's anything deeper to note.