Credit default swaps and the UK 2008–09 short sales ban

ABSTRACT

Most studies of the short sales ban of UK financial stocks from September 2008 to January 2009 fail to control for the UK’s worst ever banking crisis and the underlying increase in risk. This paper studies the ban’s impact on the 13 large financials with credit default swaps (CDS) and 20 smaller stocks without CDS. The results reveal that returns of banned stocks Granger cause CDS returns in the pre- and post-ban periods, but causality runs from CDS to stock returns during the ban period. Underlying risk proxied by the CDS probability of default increased during the ban and the immediate pre- and post-ban periods which highlights an endogeneity problem ignored in some studies. This increased risk provides a plausible rationale for why CDS and related equity bid-ask spreads - which increased during the ban period – failed to fall significantly in the post-ban period. Panel regression results indicate that probability of default was an important economic determinant of stock bid-ask spreads during the ban period. Finally, our results suggest that the ban offered direct price support for the smaller non-CDS stocks during the ban period and indirect support for CDS stocks from their pre-ban to their post-ban levels.

KEYWORDS:

• Bans
• short sales
• limits to arbitrage
• CDS spreads

JEL CLASSIFICATIONS CODE:

• G01
• G12
• G14
• G18

Acknowledgement

We are grateful to four anonymous referees for their helpful comments. These have helped us to clarify a number of issues and improve the overall exposition of the paper.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1 See Shin (2009) for a fascinating account of the collapse of Northern Rock.

2 The ban included options, futures, depository receipts, contracts for differences, spread bets and dual line stocks.

3 An investor could take a bearish position in a banned stock during the ban period by buying its CDS. If its default risk increases, then the investor could close her position at a profit by selling the CDS.

4 See columns (5) and (6) in Table II, p.360.

5 Sifat and Mohamad (2015) reach similar conclusions on the market microstructure impact.

6 These included Alliance & Leicester plc and Bradford & Bingley plc (both taken over by Banco Santander), HBOS plc (taken over by Lloyds TSB), Lloyds TSB Group plc (became Lloyds Banking Group plc), London Scottish Bank plc (went into administration on November 20, 2008) and Highway Insurance group (takeover by Liverpool Victoria announced October 17, 2008).

7 In response to the perception of some market participants (particularly in Europe) that the modified restructuring had been too severe in its limitation of deliverable obligations, a further modification of the modified restructuring clause was introduced in 2003. Under the modified-modified restructuring, the remaining maturity of deliverable assets must be shorter than 60 months for restructured obligations and 30 months for all other obligations (Packer and Zhu 2005).

8 As demonstrated by Hilscher, Pollet, and Wilson (2015), the credit protection return is equal to the percentage change in CDS prices adjusted by the ratio of two annuity factors. As the adjusted ratio will always be close to one, the credit protection return is thus well approximated by the percentage change in CDS prices. In this respect, we follow a number of recent studies in leading international journals that investigate the effects of CDS markets on information transmission by using the first log-difference of CDS mid prices as the proxy for CDS returns (e.g. Batta, Qiu, and Yu 2016; Friewald, Wagner, and Zechner 2014; Oehmke and Zawadowski 2017; Qiu and Yu 2012).

9 The GC results from estimating a base-line model without exogenous variables are qualitatively similar.

10 Some studies (e.g. Pan and Singleton 2008) have identified an idiosyncratic component in the short-term CDS maturity, possibly related with liquidity issues and short-term distress. We therefore extend our analysis to the 1-year CDS maturity whose results are very similar to those of the 5-year CDS maturity in Table 3. The estimation results of the 1-year CDS maturity are available on request.

11 This reflects the exclusion of Alliance and Leicester whose board accepted a takeover by Santander in mid-July 2008.

12 We note that risk proxies derived from derivatives – e.g. implied volatility or risk neutral skewness from options – are generally derived under the risk neutral measure or assumption of risk neutrality. However, our PD measure disregards both risk premia arising due to agents being risk-averse and illiquidity premia that both are known to be typically non-trivial and to rise during crises (e.g. Amato 2005; Corò, Dufour, and Varotto 2013). Ignoring this may overstate default probability estimates, particularly during the crisis period.

13 Note that their US post-ban results cover a period of only 14 days.

14 The focus is on the medians as these are robust to outliers.

15 Lloyds opened on the 19 January at 52.43, falling to a low of 29.76 before closing at 32.24. The following day saw a low of 16.76 (Bloomberg).

16 ‘Fears spark uneasy start to Lloyds Banking Group’, Financial Times (20/1/09)

17 ‘Paulson reaps £270m “shorting” RBS’, Financial Times (27/1/09).

18 We are grateful to two anonymous referees for these suggestions.

Additional information

Funding

Coakley gratefully acknowledges support from grant number ES/L011859/1, from the Business and Local Government Data Research Centre, funded by the Economic and Social Research Council.