Can intervention indices detect central bank's actual intervention behaviour?

Abstract

This article employs the Weymark (1997) and Baig et al. (2003) models to evaluate the central bank's intervention behaviour and utilizes ex post intervention news to check the models' estimation results. Empirical results show that the Baig et al. model's evaluation performance in intervention behaviour is inferior to the Weymark model. The scale of intervention index is not an excellent indicator to detect intervention behaviour. Intervention indices and intervention news all support that the intervention frequency in expansion period is higher than in recession period and in face of fatal events the intervention exhibits cluster phenomenon.

Acknowledgements

We appreciate B. R. Brooks, M. Tatsuyoshi and Y. Yushi for helpful discussions and comments. All remaining errors of course are our own.

Notes

¹Tanner (2000) considered that under a managed floating exchange rate regime, the focus of the monetary authority concerns shifts from exchange rate and foreign reserves to the pressure of exchange rate.

²For detailed derivation, see Weymark (1997, pp. 62–64).

³For example, when the external excess demand of domestic currency is negative and central bank adopts leaning with the wind exchange rate policy to induce domestic currency depreciation, this will enlarge depreciation degree and induce negative intervention index.

⁴The 11th business cycle in Taiwan is not over yet, so we cannot extend the sample period to cover it.

⁵Although the central bank has intervened exchange market for 8 months during the recession period of the 10th business cycle, the successive intervention is to cope with the dot.com bubble.