Predicting the Health of the Economy - The Surprise Index

Updated March 11, 2014

There are thousands of statistical methods that are used to measure the health of the economy but one that I find eye-catching is the interestingly named Citi Economic Surprise Index (CESI).  This metric is calculated daily on a three month rolling basis and measures a combination of both qualitative and quantitative macroeconomic indicators of economic news using weighted historical standard deviations of economic data surprises (i.e. when economists project that a metric will change by a certain amount and the metric changes by far more in a positive or negative direction with respect to the anticipated consensus level).  A positive Surprise Index suggests that economic data releases have been generally better than the consensus of economists and a negative Surprise Index suggests the opposite, that economic data releases have been generally worse than consensus.  As many of us observe, markets frequently move most substantially when macroeconomic data surprises to the upside or downside.  One recent example was the Philadelphia Fed's manufacturing survey for February; economists polled expected a 7.3 reading however, the actual reading was negative 6.3 showing that conditions were far worse than expected.  This would result in a negative Surprise Index number.  Please keep in mind that a positive or negative index value has nothing to do with the actual health of the economy, rather it reflects the relationship between expectations and reality.

Citi is not the only promoter of this metric.  A 2012 paper by Chiara Scotti at the Federal Reserve Board examines the construction and use of a surprise index which is calculated using this equation:

The indicators used by the author to construct the Surprise Index for each nation in the study are as follows:

1.) First GDP release for each quarter.

2.) Monthly industrial production.

3.) Number of employees on non-agricultural payrolls or unemployment.

4.) Retail sales.

5.) Measure of the manufacturing sector (i.e. ISM (U.S.), Composite PMI (Europe), PMI (United Kingdom and Canada) and the Tankan Survey (Japan).  

6.) BEA personal income for the United States.

Announcement surprises are calculated as the difference between the actual announcement data and the Bloomberg expectations divided by their sample standard deviation.

Now, let's look at the Surprise Index calculated by the author for all five economies individually plus an aggregate of all five economies (solid lines), comparing it to the CESI (dashed lines) for the time period between 2003 and 2012:

Notice that in every case but Japan, the Surprise Index was strongly negative by the middle of the Great Recession as a result of the fact that economists were far less pessimistic about the outcome of the crisis than the actual data showed.  In the case of Japan, you'll even notice that the index dropped sharply in April 27, 2011 as the actual number for industrial production was far lower than expected following the March 2011 earthquake.  That shows us the power of this index.  Generally, economists lag the market.  As the economy enters the growth portion of a cycle, economists tend to be overly pessimistic from their recent experience, expecting that many measures of the economy will not perform as well as they do in reality.  This results in a positive Surprise Index reading.  On the other end of the cycle, as the economy enters the contraction portion of a cycle, economists maintain the optimism that was built up during the "good times" and the real world decline in economic metrics results in a negative Surprise Index reading.  That is why the Surprise metric can be an interesting and powerful tool for predicting where the economy is headed.   

Now, let's look at Citi's recent Surprise Index data starting with the United States:

Notice that the Surprise Index has fallen sharply over the past few weeks.  This shows us that economists are tending to be overly optimistic when compared to economic metrics that have not met consensus.  This could be giving us a hint that an economic slowdown is in the works.  Interestingly, Citi economists have ramped down their economic growth prospects from 1.5 to 2 percent to 1 percent.

Here's Citi's recent Surprise Index for Europe:

The Surprise Index for the period throughout the month of November was frequently negative, showing that economic data releases were worse than consensus.  The Surprise Index rose during December and January with signs of an economy that was improving more rapidly than economists expected, however, the index has fallen considerably from highs seen in early 2014 as was the case for the United States.

Lastly, here's Citi's Surprise Index for China:

China's index has been very volatile over the past three months, dropping sharply in November as economic data was worse than economists projected.  During December and January, the Surprise Index was in positive territory as the economy outperformed expectations, however it fell sharply in mid-January and then rose very sharply in February as data showed stronger than expected growth in exports and a rising trade surplus with the United States.  As March has rolled on, the Surprise Index for China has fallen as the Chinese economy, once again, is under performing expectations.

It certainly appears that the economy could be in for a bit of a rough ride over the coming months if we include the indices for Europe and Japan.  While the market sages like to finger the recent winter weather for the recent poor economic showing, the problems may be more structural than anticipated at this point.

The aptly named Surprise Index is a fascinating measure of the economy and by using trends established over the past decade, can provide us with some sense of where the economy may be headed.  In any case, it's a measure of the economy that bears watching.