Economic news affects asset prices, including bond yields, stock prices, and exchange rates. Basic economic thinking would lead one to expect that shocks to some indicators, such as unemployment or GDP growth, should cause interest rates to rise and stocks to fall, whereas shifts in other indicators are largely shrugged off by the markets. But the size and direction of asset price responses to unexpected news announcements can be surprisingly difficult to pin down.
This article compares two different approaches for estimating the effects of economic news on financial assets. The standard approach uses the predictions of an empirical forecasting model, which makes the measured impact of news volatile and dependent on the specific model used to generate the forecasts. In contrast, the Rigobon–Sack approach estimates the effects of “true news” (an indicator as reported minus its expected value at survey time) by using a statistical filter to eliminate measurement errors and informa- tional noise that accumulates between surveys and data releases.
We find that for most key economic indicators, the estimated effect of true news is very similar in sign and magnitude to the measured effect. However, the difference in size is greater for some measures than others. The strongest effects of economic news are seen in bond yields, while the weakest are in stock prices. Moreover, the immediate impact of news on asset prices is easier to assess than their full-day impacts because the accumulation of other information in the market through the business day makes measuring the per- sistent effects of new economic news more difficult.