The Surprise Index - What Does It Measure?

Here’s an interesting indicator: it’s called the Surprise Index. Calculated by the economists at MFC Global Investment Management, it quantifies in one measure the extent to which U.S. economic indicators exceed or fall short of consensus estimates.

An economic report with better-than-expected news is assigned a value of 1; a report with worse-than-expected news is assigned a value of -1; a report meeting expectations gets a 0 value. Add up the values of the reports for the week, and you have the Surprise Index’s reading for that week.

The indicator is published on the second page of MFC Global’s weekly Market Commentary. Upward momentum in the index (which would arise if economic reports increasingly exceeded on the upside) could, in theory, foreshadow a rebound in the economy; vice versa for downward momentum.

As can be seen in the Aug. 15 issue, there was a good run-up over May and June in the Surprise Index, but in July it collapsed suddenly to negative levels as deep as April’s level. The trend has edged up since, but it’s still less than zero – indicating that the latest U.S. economic indicators are, on average, disappointing projections.

Comments

[whidbey:]

What this index proves is that consensus estimates are prone to human emotions just like anything else. If we just assume all participants have a 50% chance of being wrong, we can compare the list of money scores with a "runs test". It suggests the results are not random, but a correlation , a serial correlation in which past data predicts future outcomes. Autocorrelation in other words. Probably not worth the time it takes to produce, and certainly of no value to those who feel chance is the rule.

2008 Aug 22 03:16 PM