A wavelet transform analysis of the relationship between unexpected macroeconomic news and foreign exchange rates

Abstract

This study applies the wavelet transform technique to analyse the relationship between exchange rates of major currencies and the unexpected macroeconomic news. Hourly closing exchange rates of major currencies from 2000 to 2009 are utilized for wavelet analysis and those of 2010 are used for ex ante test. It is found that JPY and CHF are the most affected exchange rates, which are followed by EUR and GBP. The least affected exchange rates are AUD and CAD. In general, the most influential news is NFP, which is followed by ISM_M. These two factors have an impact on at least four currency exchange rates in the short, intermediate and long terms. UR is a short-term factor and the effect of NFP lasts for all the three time horizons considered. Ex ante test is carried out and the accuracy rates for EUR, AUD, GBP, CAD, CHF and JPY are 0.75, 0.5, 0.67, 0.83, 0.83 and 0.58, respectively.

Keywords:

• unexpected macroeconomic news
• wavelet transform
• currency rates
• time horizon

JEL Classification:

• F31
• G15

Acknowledgements

We thank an anonymous referee for useful comments. Any remaining errors are ours. Both Chens also thank the National Science Council for financial support through grant NSC 100-2410-H-030-032.

Notes

¹ According to Rebitzky (2010), news is defined as , which is the difference between the actual realization and its prior expectation .

² We use Daubechies biorthogonal 5/3 filter to perform the discrete wavelet transform.

³ In practice, EUR, AUD and GBP are quoted directly in US dollars and are indicated as EUR/USD, AUD/USD and GBP/USD, respectively.

⁴ Results on USD/AUD, USD/GBP, USD/CAD, USD/CHF and USD/JPY can be provided upon request.

⁵ Factors that have unexpected negative signs more than twice for a single currency pair will be explained later.

⁶ Ex ante results on USD/AUD, USD/GBP, USD/CAD, USD/CHF and USD/JPY can be provided upon request.