Development of a pricing methodology for CDS on Small and Medium Enterprises

The present thesis is aimed to investigate the possibility to expand the hedging purpose of Credit Default Swaps (CDS), to hedge credit risk on trade credits of Small and Medium (industrial) Enterprises (SME). Recent years of severe economic crisis have led to some deep changings in the practices of credit management within industrial entities. Year after year such entities have stressed their attention on credit risk management. At the end of the chain of credit risk management we find Trade credit insurance, which is a fast growing sector, with raising interest by the industrial entities. Conceptually, Trade credit insurance’s purpose is not quite different than the purpose of a Credit Default Swap (i.e. hedging the buyer from the insolvency risk of an underlying entity), so it would be its most direct competitor. The main issue in extending the pricing of CDS to SME is represented by the fact that SME, in general terms, are neither listed nor rated and do not issue debt instruments. This entails the fact that, for such entities, risk neutral Probabilities of Default, needed by standard pricing models, cannot be retrieved from market traded instruments. To overcome such issue, we constructed an “equivalent risk neutral PD”. The approach was similar to the Radon-Nikodym theorem: we defined a corrective factor to be applied to the “real world PDs”, in order to obtain an equivalent risk neutral Probability of Default. Obviously, what obtained is a proxy. Real world PDs for SME were calculated starting by the databases put into our disposal by modeFinance Srl, which is a certified credit rating agency, specialized in the creditworthiness evaluation of Small and Medium Enterprises. On the other hand, CDS risk neutral default probabilities were bootstrapped from real CDS trades, by different maturities and rating classes. The data source was in this case Bloomberg. Being such hypothetical CDS a new instrument, there is no real benchmark to state whether the obtained spreads resulted into a fair cost, or not. So we first applied the developed pricing model to a set of 1,000 Italian SME. The results obtained were quite positive, since the average SME-CDS spread obtained by rating class and maturity, other than being monotonically increasing with rating class worsening and maturity increasing, was also in average 42% higher than the (average) spreads observed for real traded CDS, as common sense would suggest it had to be. As a final step we compared the obtained cost of these new instrument with the cost of 4 standard Trade credit insurance policies (which were kindly disclosed, in anonymous form, by the insurance broker Willis Italia S.p.A.). Even if CDS and Trade Credit Insurance may have the same conceptual goal (to protect from default risk of an underlying entity), as instruments they are radically different and such cost comparison of the two is not quite straightforward. In order to make the comparison more plausible, we made two hypotheses on how to calculate the equivalent CDS cost of such trade credit insurance policies. For all the analyzed policies, the real policy cost was lying between the two CDS cost proxies. In general terms at the moment, it is hard to tell whether the CDS prices obtained with the model developed, could be defined “fair”. Certainly, the results of the benchmarking tell us that they are at least plausible.