30 October 2017

ECB: Dissecting long-term Bund yields in the run-up to the ECB’s public sector purchase programme

This paper intends to help understand via what channels the inception of the PSPP has led to the observed bond yield compression. Under the so-called signalling channel, bond purchases would underline the central bank’s commitment to keeping its key interest rates at low levels for an extended period of time.

This in turn stabilises market expectations of short-term rates at least for the short- to medium-term horizon. Under the portfolio rebalancing channel, central bank purchases reduce the supply of government bonds available to the private sector and make investors attempt to re-shuffle their own portfolio allocations in response.

In order to detect what channels are at work, it is useful to think about long-term bond yields as being the sum of a so-called expectations component (average expected future interest rates) and a term premium (comprising inter alia compensation for interest rate risk). The signalling channel would mainly lead to a compression of the expectations component, whereas the portfolio re-balancing channel would mainly make the term premium shrink. While yield changes are observable, individual changes in the two components are not: hence, authors deploy a dynamic term structure model in order to decompose changes of long-term Bund yields into changes in the expectations and the term premium component.

They find that the large decline in government bond yields since the summer of 2014 can be attributed almost exclusively to decreasing term premia. On the day of the PSPP announcement, it was likewise a decrease in the term premium rather than the expectations component, which moved long-term yields. Their results thus speak in favor of the portfolio re-balancing channel being at work.

Authors also show that there were episodes in the past (notably during the financial and euro area sovereign debt crisis), when the expectations component rather than the term premium was the more dominant driver of yields. Hence, the identified strong dominance of term premium changes during the PSPP period is not an artefact in the sense that the model would always tend to attribute bond yield changes mainly to premia. A series of robustness checks confirms the main results.

Full working paper

© ECB - European Central Bank