In the past year there has been much discussion on (naked) short selling, frequently associated with the question of “naked CDS” (the widely-used acronym for credit default swaps). Yet naked CDS are of a different nature and deserve to be treated so. For starters, to build a naked position in CDS actually requires buying the derivative, not selling it as in the conventional naked short selling of securities.
What the academics are saying
According to the recent academic literature, it is reasonable to assume that a fraction of the ongoing demand for CDS on sovereigns seeks to hedge credit risks different from the one related to public debt, yet positively correlated with it (e.g. risks associated with public and private companies’ debt). This motive can be especially relevant in cases in which there is no other means to protect against such non-sovereign risk or, if it is available, the cost involved is prohibitive. In this scenario, the total demand for CDS on sovereigns necessarily becomes more sensitive to changes in the investors’ perceptions of the countries’ underlying fiscal fundamentals.
Some argue that regulation to restrict speculation in CDS markets could have the unintended consequences of reducing market liquidity. This in turn would raise trading execution costs for investors who are not speculating and lower the quality of information provided by CDS rates on the credit quality of bond issuers (for more, see Duffie 2010). Such regulation could, as a result, increase sovereign borrowing costs. In any case, speculation could continue via short selling of the underlying sovereign bonds, to the extent that the bond market is liquid.
On the other hand, there are those who support a ban on naked CDS point to the risk of serious destabilisation and market failure. They point to the unreliability of CDS spreads as an indicator of default probability (e.g. when corporate prices are less than those of their country of residence – see additional arguments in Portes 2010). Together with the risk of spiralling, there is an incentive for coordinated manipulation.
One of the main reasons for failures in the CDS market is the lack of post-trade transparency. The lack of information may limit the ability of participants to analyse the market, which will artificially increase the uncertainty that naturally accompanies a crisis like the current one. Thus, an alternative solution to a ban of naked CDS could be to set an obligation to provide reliable updated information on a daily basis about which players/firms are offering hedges against credit risk and in what conditions (volumes, prices, type of counterparty and contract, collateral required, etc.). Also, higher capital and collateral requirements may make manipulation strategies more difficult and risky to achieve.
Finally, it should be noted that any restrictions to “naked” CDS trading could only be effective if they are properly monitored, which is not the case presently, given the current lack of reporting on these securities to most regulators, particularly in Europe.
CDS and sovereign debt: What the regulators are saying
There is an on-going discussion on the potential for manipulative practices through naked CDS, particularly in relation to sovereign bonds. It has, for instance, been argued that the widening of bond spreads for some European countries has been led and precipitated deliberately through the CDS market. The fact that this market has a limited number of participants and the lack of post-trade transparency (but not pre-trade transparency) would favour manipulative practices on the underlying bonds.
Beyond (or before) an econometric analysis, there are two crucial questions that need to be answered. The first one relates to transmission channels from one market (CDS) to the other (bonds) – also knowing that the size of the first is quite small when compared to the second, which is highly liquid. This is certainly a topic deserving further academic research, but I take the risk of suggesting some possibilities – there might be still others.
Let us first look at some visual evidence through a series of charts for several European countries where we put the CDS prices against the bond yield spreads corrected for the risk free rate (defined as the German bonds yield minus the CDS spread for Germany).
Figure 1. CDS spreads and bond spreads
(GR = Germany, IE = Ireland, FR = France, BE = Belgium, NL = Netherlands, ES = Spain, IT = Italy).
Note that a common mistake in most analyses was to consider the Bund yield as the risk free rate, ignoring that the price of German CDS is neither nil nor constant, particularly over the past year or so. The charts do not provide by themselves any clear evidence as to the influence of CDS spreads on the bond yields. In many cases they even seem to point to the opposite. The Committee for European Securities Regulators is currently working on this issue in a more scientific way. I can mention some preliminary results on the relation between CDS spreads and bond spreads for sovereigns (that should be taken as indicative for the moment):
- the correlation between CDS spreads and bond spreads is relatively high and increasing over time for the bulk of the countries analysed; this can be interpreted as a movement towards faster price adjustments between both markets during the sovereign debt crisis;
- in the period after the European rescue package of 8 May 2010, that link seems to have intensified for almost all countries in comparison to the period September 2009 to May 2010;
- before the date of the announcement of the rescue package, there is, for 5-year and 10-year maturities, for almost all of countries analysed a bi-directional relationship between the sovereign CDS market and the sovereign bond market (i.e. past values of CDS spreads contribute explaining the bond yields and the other way around);
- these bi-directional relations break down in a large number of cases in the period after the announcement of the rescue package on 10 May 2010 both for the 5-year and the 10-year maturities.
All the results, but particularly those on the impact of the rescue package, need to be considered with due caution as the data series for the period after 8 May 2010 are relatively short and were established in an uncommon market environment.
One standard explanation for a possible link (in both directions) between the two markets is based on arbitrage considerations. If the difference between the CDS spread and the risk-adjusted bond yield is positive (positive basis), then an investor would profit (basically without any risk involved) from buying a riskless bond, shorting the risky sovereign bond and selling the sovereign CDS. If the basis is negative, profits could be generated by buying the sovereign CDS and the sovereign bond, and selling the riskless bond.
Particularly interesting for the current discussion are explanations which would favour an influence in one direction rather than the other.
In this respect, one possibility of a link from the CDS to the bond market is based on the fact that most credit institutions and asset managers adopt rules concerning their credit exposure and investment strategies which are based on CDS spread thresholds. As a result, the rise in an issuer’s CDS spread above a certain threshold might lead them to need to reduce their exposure, possibly triggering a large selling movement, thus depressing bond prices. Of course, there is nothing wrong with this if the CDS market is functioning properly and correctly reflects the issuer’s default risk.
The last explanation I would like to emphasise is also straightforward. That is, the bond market participants simply would take the CDS spreads as reliable indicators of the issuer’s risk and would integrate them into their behaviour accordingly.
Another crucial question that needs to be addressed is the fact that the data that is being used in empirical studies on CDS markets relates to pre-trade prices – so far we do not have information on effective trade prices. Although it is not excluded that some of the mechanisms described above also work on the basis of pre-trade prices, information on trade prices seems essential both to obtain a better insight in the functioning of that market and to explore further possible manipulative strategies.
Finally, measures must be adopted in order to improve the functioning of CDS (and of course other derivatives) markets, not only in terms of transparency, but also aimed at enlarging the number of participants and consequently ensuring increased liquidity and openness. This is why the reform of over-the-counter bond and derivatives markets are, in my view, a precondition, first, for a well-founded analysis of the issues at stake and, second, for identifying any additional regulatory measures that might be necessary.
Duffie, Darrell (2010), “Credit Default Swaps on Government Debt: Potential Implications of the Greek Debt Crisis”, Prepared testimony to the US Subcommittee on Capital Markets, Insurance, and Government Sponsored Enterprises Hearin, 29 April.
Portes, Richard (2010), “Ban naked CDS”, eurointelligence.com, 18 March.