Recently John Taylor, the Economics Professor at Stanford and the author of the Taylor Rule which is often used in monetary policy, caused a bit of a dust up in the Economics blogging community with the following graph showing a very tight correlation between the percentage of the economy that is going towards investment (Y axis), and the unemployment rate since 1990 (X axis).

Of course, correlation is not the same thing as causality, and it is not exactly clear which way the causality would run if it does exist. Taylor, who strongly leans conservative in his views, was suggesting that the way to bring down unemployment is to raise the share of the economy going towards investment. Thus we should do things like cut taxes on investment.

That is a plausible interpretation of the graph. When businesses invest, they also tend to hire people. Who is going to build a factory and then not hire people to work in it?

However, it could run the other way: when the unemployment rate is high, there is a lot of slack in the economy, and thus unused capacity. Why should a company invest to expand production when it has unused factories? After all, there is a very strong inverse correlation between capacity utilization and unemployment, as shown in the next graph.

As some others in the Economics blogging world pointed out, the data did not start in 1990, and some even went so far as to accuse Taylor of cherry-picking his data. The data actually starts in 1948, and if we put up a similar graph for the pre-1990 data we get a very different picture — with almost no relationship at all, a regression line through this data would be almost horizontal, and if anything would have a positive slope.

To me, the most interesting question is why should there be a strong relationship (regardless of which direction it runs) since 1990, when there was none before then? I’m not sure I have the answer, but I think perhaps I can shed a little light on it, and perhaps others can run with it and find the answer.

Fixed Private Investment is broken down into two categories, Residential and non-residential (or business) investment. The graph that Taylor produced was the sum of the two. Clearly, though, Taylor’s interpretation and implied remedy would rely on the business investment side, not the homebuilding side.

I thus decided to break down the percentage of GDP spent on housing vs. the unemployment rate, and then look at business investment as a share of GDP vs. unemployment for the two periods. In the more recent period, we can see that there is an inverse relationship between residential investment and the unemployment rate, although it is clearly weaker than for total fixed investment. Lots of homebuilding when there is low unemployment, not so much when unemployment is high.

In the earlier period, there is still a relationship, but it does not look nearly as strong, in part because the range of residential investment as a share of GDP is much smaller. It is still there, though.

On the business investment front, the relationship looks very strong in the more recent period, more so than for residential investment, but interestingly not as strong as the relationship with total fixed investment.

However, things really break down when we look at the earlier period. If anything, it looks like the relationship runs the other way — with low levels of business fixed investment associated with low unemployment rates.

I’m not really sure of what changed starting in 1990, but it seems pretty significant. The nature of recessions was a bit different prior to the 1991 downturn. Earlier recessions were often “deliberately” caused by the Fed in an effort to bring down inflation, and also had a much bigger “inventory cycle” component to them.

The more recent recessions have a bigger “balance sheet” component to them, with the wealth lost from popping bubbles (S&L’s, Tech, Housing, et. al.) causing people to cut back on spending to repair their balance sheets. I’m not sure that is the cause, but it would be a plausible interpretation. I would love to see more work done to explain why the world changed in this way in 1990.
 
Zacks Investment Research