Diffusion indices are best at indicating direction

2 Jun 2010 by Jim Fickett.

The PMI is a diffusion index, and diffusion indices are designed to measure the prevalence of change in a particular direction. The PMI dropped pretty consistently as the boom times before the current bust wore on. Although output continued to increase, the weakening of consensus, when added to other indications, might have served as a useful warning.

Following on to a recent post on the PMI, I've added a Reference page on the PMI and related indices. The PMI is a diffusion index and, since I was unable to find a concise overview of diffusion indices anywhere else, I've added one.

Diffusion indices are based on questions about direction, for example, “In your business, did new orders decrease, increase, or stay about the same?” Because the questions are about direction, and not amount (i.e. not, “How many more dollars did you make from sales this month?”), diffusion indices are best at providing information on whether there is a consensus about direction, rather than providing information on the amount of activity. Imagine, for example, that a few firms greatly reduce output, while many firms slightly increase output. Then output might rise only slightly while the index would still give a high reading. An article on a diffusion index for GDP, from the Canadian Economic Observer, sums it up:

Developed by the National Bureau of Economic Research (NBER), diffusion indices look only at the direction, not the rate of change. …

Like all summary measures of the economy, diffusion indices also hide important information. One cannot tell from their aggregate values the industries that are driving the overall change, their relative importance, their recent trend, or how fast they are changing.

Earlier we highlighted the negative side of this – the earlier graph showed a period during which output growth hardly changed, while the PMI varied considerably. In this post I'd like to highlight the positive side.

The graph below is the same as yesterday, with the only difference being the region highlighted. Note during this period (the boom preceding the current bust) that the PMI went down pretty consistently even as output continued to grow. By the end of the period highlighted (late 2007), output was still growing but the consensus supporting it had evaporated. For those who understood what a diffusion index really tells you, this, along with other, more fundamental, indications, could have been an important warning of the crash to come.