Working Paper: CEPR ID: DP1964
Authors: Andreas Fischer; Mathias Zurlinden
Abstract: It is often argued that interventions provide news on changes in risk premia or in future monetary policy. If so, the timing of interventions offers important information for central bank watchers. One method to study the reaction function of a central bank is to consider whether the duration intervals of past interventions matter for future interventions. The time deformation of interventions is modelled as an autoregressive process following the class of ACD models first proposed by Engle and Russell (1994). The analysis considers the persistence of interventions by the Federal Reserve, the Bundesbank and the Swiss National Bank.
Keywords: time deformation; persistence; exchange rate intervention; duration
JEL Codes: E52; E58
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| timing of interventions (C41) | market perceptions of central bank behavior (E52) |
| previous interventions (Y80) | likelihood of subsequent interventions (I12) |
| duration intervals of interventions (C41) | likelihood of subsequent interventions (I12) |
| longer durations of inactivity (C41) | likelihood of subsequent interventions (I12) |
| intervention intervals (C41) | clustering of interventions (C38) |
| self-exciting behavior (C92) | persistence of interventions (C41) |