Working Paper: CEPR ID: DP4359
Authors: Andreas M. Fischer
Abstract: Reuters news reports have become an accepted tool for empirical studies analysing informational asymmetries in FX markets. This Paper tests the accuracy of the Reuters announcements against Swiss interventions. The evidence finds that the Reuters announcements are not tightly clustered just after the first intervention transaction as is commonly assumed in empirical studies. The variance of the prediction error is measured in hours and not in minutes. A further assumption that the Reuters announcement always follows after the first intervention transaction is found to be violated. These and other empirical results question the accuracy of Reuters reports for Swiss interventions.
Keywords: Central Bank Interventions; Intradaily Data; Reuters News Reports
JEL Codes: F31; F33
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Timing of Reuters reports (G14) | Timing of actual SNB interventions (F31) |
| Frequency of Reuters reports (G14) | Number of SNB transactions (E42) |
| Exchange rate volatility (F31) | Actual transaction data (Y10) |
| Timing of actual SNB interventions (F31) | Timing of Reuters reports (G14) |
| Frequency of Reuters reports (G14) | Exchange rate volatility (F31) |