Economic Surprises!

Many successful investors claim to profit from the gradual unfolding of economic news in markets. Participants in financial markets observe a rich stream of economic news on a daily basis. The informational content in this news flow is used by these participants in order to update their perception of the state of the economy, and serves as an input for the financial decision-making process of investors. The economic cycle is an important driver of financial asset returns, yet in many ways the stately march of expansion and recession moves too slowly for traders and portfolio managers. After all, the swirl of perceptions about economic growth can exert a significant influence on market prices even when the economy itself is little changed. One question is how this economic news impact financial market returns. Macroeconomic news can be decomposed in an expected and unexpected component (relative to a consensus forecast), or ’surprise’. Unexpected news about the state of macroeconomic fundamentals should impact prices shortly after their release, but have no impact on markets afterwards. 

To answer this question, we should therefore study macroeconomic surprises that are contained in economic releases, and use these surprises to predict asset returns. A popular way to do this is to extract a signal from the noise of economic releases, and construct a surprise index. Popular surprise indexes that are widely used in the industry are the Bloomberg and Citigroup surprise index. The rationale behind these indexes is a compelling one. When the economy is accelerating, economic releases typically exceed forecasts (positive surprise) which should result into higher stock prices, market interest rates, and plausibly a stronger currency. A deteriorating economy produces negative surprises, with opposite impact on markets. Can these surprise indexes actually predict future returns? Let’s take a look at some simple correlations derived by a Bloomberg researcher:

First, we compare the level of the relevant surprise index with asset returns using weekly data starting in 2003. Asset returns are computed over a rolling 3-month period. For example, we calculate the for the Euro Stoxx 600 and compare that to the level of the Euro Area Surprise Index. We find that the correlation was highest with stocks, while there appears to be a weak relationship with currencies and surprises. What is interesting is that EU and UK swaps and equity markets are more correlated to the US economic surprises index than to their local indexes. This implies that global markets co-move significantly with economic releases in the U.S, i.e. there might be spill-over effects from the U.S. to other major markets. 

So far, we looked at the correlations between past price movements and past surprises. What is more interesting is whether past surprises are able to predict future returns. Let’s now compare the 13-week changes in surprise indexes with the forward-looking five-week change in various asset markets. The results are weaker here. When regressing the level of surprise indexes with future price action, the relationships in Europe and the U.K. were stronger than those in the U.S. Combined with the findings above, this might suggest that non-U.S. markets are relatively inefficient, reacting to American economic developments in real time while reflecting domestic fundamentals only with a lag. Meanwhile, 13-week changes in the various indexes offer little insight into future market returns.

To summarize, economic surprises do seem to contemporaneously explain asset market returns, although the strength of the relationship is a bit on the week side. Surprisingly, the correlations are strongest for equities, where company-specific factors should play a larger role. On the other hand, there is almost no relationship between currencies and economic surprises, while surprise analysis is most often used in these markets. The levels of an economic surprise index appear to be moderately more useful than changes in surprise, though neither is particularly successful in forecasting future asset price moves. Perhaps the most interesting finding is that developments in the U.S. can be more important than domestic economic activity. So when making investment decisions, take these economic surprises at least into consideration.


Amar Soebhag