An empirical investigation of the nexus between sovereign bond yields and stock market returns – a multiscale approach

Silvo Dajcman


This paper studies comovement between changes in sovereign bond yields and stock market returns for ten Eurozone countries (Austria, Belgium, Finland, France, Ireland, Italy, Germany, Netherlands, Portugal, and Spain) in the period from January 2000 – end of August 2011, applying the maximal overlap discrete wavelet transform (MODWT) variance and correlation tools.

We found that the for all but one country (namely Portugal) the stock return volatility is higher than the volatility of sovereign bond yield changes across all wavelet scales. The highest volatility in stock indices’ returns and bond yield changes is captured by lower level MODWT wavelet coefficients (i.e., level 1 ( ) MODWT wavelet coefficients, associated with dynamics over 2 to 4 days and level 2 ( ) MODWT wavelet coefficients associated with dynamics over 4 to 8 days). A practical implication stemming from this finding is that investors with short investment horizons that want to efficiently manage the risk of their portfolio investments have to respond to every fluctuation in realized returns or bond yield changes, while for an investor with a much longer horizon, the need to do this is reduces, as long-run risk is significantly smaller as indicated by the variance.

The results of the paper show also that the estimated wavelet correlation changes with the time scale and is mostly positive for all countries but Portugal. The statistical evidence against hypothesis of no multiscale dependence yet is weak. Thus we cannot statistically claim that the wavelet coefficients at higher scales are significantly different (either higher or lower) that those at lower scales.

For the financial markets of Portugal and Germany we proved that the dependence between stock and sovereign bond market dynamics may not be just a multiscale phenomenon, but may also exhibits time dynamics across scales. The rolling-window wavelet correlation estimates show that at all scales correlation turned negative for Portugal at the start of 2010 – this is when the Eurozone debt crisis started which also hit hard the Portugal sovereign bond market – thus probably causing the ˝average˝ wavelet correlation for Portugal for the whole observed period to be negative.



Eurozone; stock markets;sovereign bonds; wavelet analysis; Eurozone crisis

Full Text: PDF

Print ISSN: 1392-2785
Online ISSN: 2029-5839