Forecasting Bond Returns Using Jumps in Intraday Prices

19 Pages Posted: 1 Dec 2010 Last revised: 5 Apr 2013

See all articles by Johan G. Duyvesteyn

Johan G. Duyvesteyn

Robeco Asset Management

Martin Martens

Erasmus University Rotterdam (EUR); Robeco Asset Management

Siawash Safavi Nic

Robeco Asset Management

Date Written: November 28, 2010

Abstract

We build on the work of Wright and Zhou (2009) who show that the average jump mean in bond prices can predict excess bond returns, capturing the countercyclical behaviour of risk premia. We show that these jumps often take place at 8:30 and 10:00 directly linking them to specific macroeconomic news announcements. Mean-reversion, which looks at the total return over the past period rather than just the part related to jumps, has no predictive ability. Hence it is important to consider excess returns that are related to macroeconomic announcements that matter to market participants, and jumps are a good market proxy for what investors believe is important news. Our improved jump measure produces a Sharpe ratio of 0.52 in an out-of-sample market-neutral investment strategy.

Keywords: Bond return predictability, mean-reversion, realized jumps

JEL Classification: E37, G14

Suggested Citation

Duyvesteyn, Johan G. and Martens, Martin P.E. and Safavi Nic, Siawash, Forecasting Bond Returns Using Jumps in Intraday Prices (November 28, 2010). Journal of Fixed Income, Vol. 20, No. 4, 2011. Available at SSRN: https://ssrn.com/abstract=1716759 or http://dx.doi.org/10.2139/ssrn.1716759

Johan G. Duyvesteyn

Robeco Asset Management ( email )

Rotterdam, 3011 AG
Netherlands

Martin P.E. Martens (Contact Author)

Erasmus University Rotterdam (EUR) ( email )

P.O. Box 1738
3000 DR Rotterdam
Netherlands
+31 10 408 1253 (Phone)
+31 10 408 9162 (Fax)

Robeco Asset Management ( email )

Rotterdam, 3011 AG
Netherlands

Siawash Safavi Nic

Robeco Asset Management ( email )

Rotterdam, 3011 AG
Netherlands

Register to save articles to
your library

Register

Paper statistics

Downloads
385
rank
72,607
Abstract Views
1,740
PlumX Metrics