Effects of Japanese intervention on yen/dollar exchange rate volatility: a conditional jump dynamics approach

Abstract

This article investigates the effects of foreign exchange interventions by the Japanese authorities on the level as well as the volatility of yen/dollar exchange rate. The empirical results show interventions in the last decade were effective not only in altering the exchange rate level, but also in volatility reduction. Jump events that tended to drive yen's appreciations and volatility increases have been effectively reduced by the ‘large-in-size’ interventions during the last decade. Our findings are in compliance with the coordination channel that explains the effectiveness in intervention proposed by Taylor (2004) and Reitz and Taylor (2006).

Notes

¹  Taylor (1995) gives a comprehensive review of literatures for nonspecialists. Advanced discusses on the exchange rate economics are provided by Sarno and Taylor (2002).

² Ito (2002) uses GARCH Model to analyse the volatility of yen exchange rates. Watanabe and Harada (2006) uses the component GARCH Model with long-run and short-run volatility components to examine the effects of central bank interventions.

³ The coordination channel is proposed as a relatively new route of effectiveness of CBIs in addition to the traditional portfolio balance and signaling channels. The coordination channel states that the CBIs may act as a coordinating signal, increasing fundamentalist traders' confidence in the validity of PPP and encouraging stabilizing speculation to re-enter the market. One can refer to the above mentioned literatures for more details on the coordination channel and refer to Sarno and Taylor (2001) for the mechanics of the portfolio balance and signalling channels.

⁴ According to Taylor (2004) and Reitz and Taylor (2006), the amount and direction of intervention are important when we evaluate the effectiveness of CBIs. Intervention with right direction and large enough amount is supposed to be credible and effective through the signaling and coordination channels. Therefore, we adopt the actual amounts of intervention instead of the 0-1 dummies in this study.

⁵ The Federal Reserve Bank of St Louis's Archival Federal Reserve Economics Data.

⁶ Watanabe and Harada (2006) also use the same date to divide their sample into two sub-periods.

⁷ Please see Maheu and McCurdy (2004).

⁸ The estimate of is marginally significant in the first half. The impact of jumps on the conditional mean of returns tends to centered around zero in this period, which means the yen is not expected to appreciate nor depreciate in response to a jump event in the early 1990s.