Sovereign debt management and monetary conditions
(Extract from page 67 of BIS Quarterly Review, September 2011)
The relative supply of government bonds can affect interest rates if arbitrage is imperfect. Such imperfections, or "preferred habitat" effects, can arise from investors looking to match the duration or other risk characteristics of their liabilities.1 These effects are likely to be especially relevant under the current strained circumstances, with financial weakness and uncertainty, including about interest rates themselves, limiting market participants' willingness or ability to take risk and to arbitrage. Evidence suggests that such effects are generally small, but significant.2
Central banks using large-scale government bond purchases to lower long-term interest rates may thus need to take account of increased government debt issuance, and of debt management operations shifting the relative supply of securities. In recent years, central bank asset purchases and increased government debt issuance have been of roughly the same magnitude. However, in practice, unconventional monetary policy seems to have achieved its objective of easing monetary conditions, without being materially impeded by any yield effects of government issuance (see, for example, Gagnon et al (2010)).3
This probably reflects two factors. First, (non-sterilised) central bank asset purchases increase the monetary base, whereas government debt issues usually fund spending or the maturation of existing debt, leaving the monetary base unchanged overall. Second, the agencies' communications, bolstered by clear institutional separation, strongly signalled their distinct policy intentions and objective functions. The monetary authorities emphasised price or macroeconomic stability, and the debt managers focused on steady and predictable issuance.
When central banks come to sell the government debt they hold, they will operate on the same side of the market as debt managers. This could amplify the impact on yields, although the gradual return of normal arbitrage and risk appetite may reduce this effect somewhat. The respective agencies will again need to communicate their objectives and ensure their respective operational plans are clearly understood.
1 See the discussion in, for example, McCauley and Ueda (2009) and Turner (2011).
2 See eg Swanson (2011).
3 See Borio and Disyatat (2010) for a discussion of the different channels by which unconventional monetary policy can act on monetary conditions.