## Unemployment and GSP in Minnesota

I got to thinking about this when Craig Westover dropped me a note on a recent Republican press release on the low May reading on state unemployment. He asked whether the full-employment unemployment rate was 4% like he remembered from college. As I teach students, you begin your answer with "It depends!" and then sketch out what your argument says it depends on. For Minnesota, the demographics are such that we'll have a lower full-employment unemployment rate than the nation as a whole.

But that wasn't what caught me. What did was when I also saw a link posted by a commenter at Residual Forces, which shows a preliminary estimate of gross state product (GSP -- like GDP except at the state level) with Minnesota growing 1.3% in 2005. That would be well below the national rate of 3.5% and ranks 44th out of the 50 states. And frankly that number presents a conundrum to me. Here's why:

A basic element of "growth accounting" is that we can measure the growth of an economy's output by a weighted average of the two inputs -- capital and labor -- plus changes in what we call "multifactor productivity". I can write this as an equation:

growth in output = growth in MFP + a*growth in capital + (1-a)*growth in labor

We can't measure multifactor productivity (MFP) independently. What we do is get information on capital, labor and output, make a guess at the value of the parameter 'a' (which most economists agree for the US to be 0.3), and get that number as a result. For the US as a whole, the Bureau of Labor Statistics calculates MFP to have grown 2.9% in 2003 and 2.8% in 2004.

The unemployment rate that I started this with is calculated as the number of people estimated to be unemployed -- we don't count them, we use a survey -- and divide it by the number of people in the labor force. Unemployment rates can fall because the number of people unemployed went down, either because they found jobs or they are no longer in the labor force. The Minnesota data shows that the number of people unemployed fell 10,243 workers between May 2005 and May 2006, but that the labor force fell by 16,152 workers. That means that while the unemployment rate went down, there are fewer jobs in Minnesota now than a year ago.

This might help explain that 1.3% GSP figure, but only in small part. The annual average for employment in Minnesota was 2,828,547 in 2005 and 2,807,428 in 2004, a growth rate a touch over 1%. Using the a = 0.3 estimate, that means 0.7% of the 1.3% change in GSP was due to an increase in Minnesota employment. We don't have a figure for MFP for 2005 yet, but it's unlikely much lower than 2%. If it was that, then capital would have to have fallen 4.67% in 2005 in Minnesota. (It grew 6.1% for the nation.) That would mean almost no investment going on in the state economy. That would be bad news, and not something you want to crow about in a Pawlenty press release. One might want to argue that taxes are driving capital out of the state; some thought this going into the 2006 Legislative calendar.

Or maybe MFP growth might be very different than 2%. It may be closer to zero -- if it was, then capital would have grown about 2% rather than falling near 5%. But why would some states have wildly different growth in productivity than others. For comparison, look at New Hampshire, a state that had only 0.7% labor growth in 2005 but 4.4% output growth. Why would productivity in NH grow so much more than in MN? Wouldn't technology and good management practices spill over state boundaries? Does NH just make better use of its resources? I doubt that, but it could be possible. And those MFP figures could be different due to differences in government treatment of profits, for example.

This may be a one-off phenomenon. Employment in Minnesota grew about 0.7% in 2003 and 0.8% in 2004, when real GSP grew 3.7% and 4.6%. And we should be careful to not say that just because 2005 marked the change in the Pawlenty administration from no-new-taxes to just-a-fee-please that this is what caused this. But I think the combination of a declining labor force generally and a slowing of GSP should at least give one pause before celebrating the decline in the unemployment.