Bond yields and debt supply: new evidence through the lens of a preferred-habitat model

Abstract

This paper examines the responsiveness of bond yields to changes in debt supply. The preferred-habitat theory predicts a positive relation between the term spread and relative supply of longer term debt, and that this relation is stronger when risk aversion is high. To capture this effect, a time-varying coefficient model is introduced and applied to German bond data. The results support the theoretical predictions and indicate substantial time variation: under high risk aversion, yield spreads react about three times more strongly than when risk aversion is low. The accumulated response of term spreads to a one standard deviation change in debt supply ranges between 4 and 46 basis points.

Keywords:

• Preferred-habitat
• Time-varying risk aversion
• Yield spreads
• Bond supply

AMS Subject Classifications:

• E43
• C22

Acknowledgements

I am grateful for comments and suggestions received from two anonymous referees, Uwe Hassler, Helmut Lütkepohl, Dieter Nautz, Sven Schreiber, Lars Winkelmann, Enzo Weber, Jürgen Wolters as well as participants of the IWH Workshop on Central Bank Communication and Decision Making, Halle, Germany, the 6th IFABS Conference on Alternative Futures for Global Banking: Competition, Regulation and Reform, Lisbon, Portugal and the 14th INFINITI Conference on International Finance, Trinity College Dublin, Ireland. Financial support from the Deutsche Forschungsgemeinschaft (DFG) through CRC 649 “Economic Risk” is gratefully acknowledged.

Notes

No potential conflict of interest was reported by the author.

1 More recently, Fukunaga et al. (2015) provide evidence for significant preferred-habitat effects in Japan.

2 Investors with a preference for shorter maturities are typically associated with banks who prefer to stay liquid whereas demand at longer maturities is often ascribed to insurance companies or pension funds.

3 The measuring of is discussed in detail in the next section.

4 Appendix 1 discusses the borderline case of extreme persistence.

5 In section 5.3 we also consider alternative measures of risk aversion.

6 We choose arbitrarily as a representative example. Since we later analyse several different yield spreads, it is noted that any of them generates almost the same scatter plots as in figure 2.

7 In section 5.3 we examine how our results change when using alternative proxies for risk aversion.

8 In principle, this model can be generalized such that the coefficients of lagged values of in the polynomial are also allowed to vary over time. The specification in (6a(6a) )–(6c(6c) ), however, already implies the long-run effect, given by , to be time-varying. Moreover, in the empirical application below, lagged values of are found insignificant.

9 In the full sample, the short rate is measure by the one-year rate. In the shorter sample we use the 6-month rate as short rate proxy.

10 Non-traded debt includes Federal Treasury financing paper and Federal savings notes of type A and B. These bonds have maturities similar to listed bonds.

11 The choice to include two lags is based on residual autocorrelation tests which showed that two lags are sufficient to remove all autocorrelation from the residuals. Lags of the independent variable were also considered but found insignificant. Since the residuals are found to be white noise, the inference in the dynamic model should be sound. Autocorrelation and heteroskedasticity specification tests as well as additional estimation results can be found in appendix 2.