By Tony Hollaar
The first research question that this report theoretically investigates is: Why would investors react to changes in sovereign credit ratings? In order to answer this question the distinction is made between rational and irrational investors. If investors are rational they would normally1 only use credit ratings as the sole information source for estimating the default risk of a country. These ratings are public information and therefore can be retrieved with relatively little transaction costs whereas estimating the default risk oneself cost more money, time and effort because there are a lot of economic, political and social factors one must consider. But investors can also behave irrational as a result of behavioral biases2, which can cause an over- or underestimation of the default risk of a sovereign country. In such a case investors do not fully rely on credit ratings and consider other factors as well, for example the business cycle a country is in. The second research question that is empirically examined is: Do investors react to changes in sovereign credit ratings? To test if this is the case, a statistical analysis is performed using OLS. Hereby rating changes for European Union countries are considered for the period: 1990-2011. The rating changes of the three major credit rating agencies: Standard&Poor’s, Moody’s and Fitch-Ratings are used, whereby the rating scales are linearly transformed (see Appendix, Table A.1). The reaction of investors is measured by analyzing the changes in bond- and CDS spreads during credit rating changes, either being an upgrade or downgrade. The timeframe chosen to measure the actual changes in spreads due to changes in sovereign credit ratings is a two-day window surrounding a rating change, from one day before the rating event up until one day after the rating event. Also a three-day window, namely from one day before the rating event up until two days after the event, is used to examine if the sovereign bond market and the CDS market respond differently because previous research suggested that the latter market is more liquid and thus reacts faster to new information. The strong assumption made here is that changes in spreads (within the timeframes discussed) are only determined by changes in default risk which are entirely reflected by the rating changes. Under this assumption a regression analysis is executed and the results are shown in Table 1. 1
Under the assumption that credit rating agencies calculate default risk fairly well and there is no close substitute. This
assumption implies that changes in sovereign credit ratings add new information and thus value to market participants. However an exception to this assumption is public information, which has come available prior to rating changes, which informs investors about a change in the default risk of a sovereign. Such an example is Ecuador who announced at the 15th of December in 2008 a default, a day before the rating adjustment had taken place. In such a case the downgrade did not add new information to market participants. 2
These biases could be caused by herding, home bias, availability bias or risk aversion but are too extensive to cover in this summary.
Table 1 shows that investors do react to sovereign rating changes. Namely a significant effect is found in both the CDS and the bond market within two and three day windows surrounding a credit rating event. The participants in the CDS market appear to react faster and stronger to a rating change than in the bond market when comparing spreads within the two- and three-day window surrounding the rating event. Furthermore in the bond and CDS markets the response to a rating downgrade is much larger than to a rating upgrade, for which the latter market has a larger change in spread in absolute values in both instances. Table 1: Effect of changes in...