Just a quick note on this since both Mankiw and Barro brought it up. Its true that investment, particularly business investment is extremely volatile. However, that does not imply that it is the primary contributor to decreased output.
So for example, compare the change in absolute numbers of Consumer Durable Expenditure (Red) and Investment in Equipment and Software (blue) over the last recession.
You can see that E&S dipped lower and recovered higher, but not by much. That’s because even though its more volatile it is smaller in size.
Both however, are swamped by residential construction.
Now lets add residential investment (Green) and government consumption (orange).
The total carve out in residential investment is much larger than either durables or equipment and software and it has yet to recover.
Also you can see that while government whether the recession pretty well it is strongly offsetting growth in durables and equipment and software.
Another indicator - one that I prefer in some ways because its real market transaction without a lot of imputation – is to look at retail sales vs. new orders for capital goods.
These are both monthly series and are the actual amounts that companies versus consumers are plunking down cash for.
Retail and food service is just a much bigger driver in terms of the change in spending patterns. This is simply because it is far larger. We can look at the two in levels together.
Retail and Food is just much bigger. We can also mix the scales so we can see the relative performance.
You can see new orders is more volatile as you would expect. Its just so much smaller. That’s why it contributes less.