Rating agencies are information providers that deliver services aimed at reducing information asymmetry on the financial market (Partnoy, 1999), by expressing judgement summarizing the available public and private information (Cowan, 1991). The information value of the service offered depends, in part, on the agency's degree of independence on the rated entity, given that, based on the agencies' pricing policies, the rating services are charged directly to the rated entity (Kuhner, 2001)2: by virtue of this independence they are acknowledged as External Credit Assessment Institutions (Honohan, 2001).
Investigations into the economic and management balance of the agencies have highlighted limited competition on the market (Partnoy, 2002), developing theoretical models for defining the fee levels to apply for maximizing profit in the period (White, 2002). In the wake of the recent bankruptcy of large companies rated by the agencies, the studies produced on the subject of the independence of rating agencies tend to focus primarily on issues of governance and internal controls (Coskun, 2008), or on the actual exploitation of situations of conflict of interest (Covitz, Harrison 2003), while there are currently no papers investigating exposure to conflicts of interest, as a result of the excessive economic dependence of the agencies from their customers.
The customer lifetime value (hereinafer CLV) is a consolidated model in the field of marketing for discriminating customers on the basis of their economic relevance. The literature highlights the preference for this type of approach, compared to the available alternatives (Kumar, 2008). By applying this model to the customers of rating agencies as well, it is possible to assess the latters' exposure to conflicts of interest (Coffee,1984), based on the favor accorded to issuer-customers, to the detriment of investor-customers (Walter, 2004). The scarce transparency of certain significant information for assessing relations with the single customers, typical of several rating agencies (Committee of European Securities Regulators, 2009), however, requires a degree of simplification/approximation when implementing the analysis model.
This article focuses on the aspects of the agencies' economic independence (section 2.1), the pricing policies adopted by the three top international rating agencies (Fitch, Moody's and S&P) (section 2.2) and the critical aspects related to the application of the CLV model to the agencies' customers (section 2.3). The empirical analysis investigates the customers of the three agencies, considering all the ratings issued in the observed time horizon and taking into account the fees and costs related to the service offered, defines the value of each business relationship in the 1999-2008 time horizon. The description of the sample (section 3.1) is followed by a explanation of the methodology to measure the agency's customer portfolio and to estimate measures useful to identify any potential conflicts of interest (section 3.2), with a discussion of the results for each of the three agencies considered (section 3.3). The last section contains brief conclusions.
Generally speaking, results highlight a low level of exposure to collusion between agencies and each customer, as a function of the revenues generated exclusively by the rating activities, identifying several recurrent features of the agencies' best customers. By differentiating the pricing policy among customers, results change significantly: if higher fees are applied to more relevant customers, the exposure to collusion increase significantly.
The Relevance Of Economic Independence For Rating Agencies
Rating agencies are firms that sell information services about investment risks (Sinclair, 1994), taking on the role of certifiers of the value of the rated assets (Millon, Thakor 1985): the profitability of these enterprises is primarily based on their income, which is a result of the fees paid by the customers requesting the rating services, and the costs incurred in delivering the service (Mattarocci, 2005). The fees are paid to the rating agencies by the issuers of the debt securities to be rated (hereinafter the issue ratings) (Cantor, Packer, 1994): remuneration by the customers, however, exposes the agencies to potential conflicts of interest (Coffee, 1984), as a result of the favour accorded to the issuer-customers, to the detriment of the investor-customers (Walter, 2004), the solution of which depends on the agencies' capacity to retain their reputation towards the investors (Brookfield, Ormrod 2000).
Exposure to conflicts of interest, and the effectiveness of the prevention measures adopted, depend on the importance of the personal advantages accruing to the party concerned (Demsky, 2003); as a rule, the impact of the conflicts of interest is expected to increase, the larger the size of the rated entity compared to the certifier (DeAngelo, 1981) and the longer the duration of their business relationship (Barber et al., 1987): the two key factors are relevant to assess the degree of financial dependence of the certifier from its customers (Reynolds, Francis, 2000). The agencies analysed here are the three top international rating agencies (Fight, 2001), which rate--both on a solicited and unsolicited basis--all the security issues negotiated on the capital markets supervised by the Securities and Exchange Commission (Basel Committee on Banking Supervision, 2006), applying fees (3)--in the case of solicited ratings--which do not appear to be consistent with a profit maximization approach in light of the competitive conduct that features the market (White, 2002). The agencies' market coverage guarantees the possibility of diversifying their portfolios of operations, albeit limitedly to the composition and volume of the capital markets: the market supply, in fact, is rather concentrated, compared to the demand (Jewell, Livingston 2000), thus leading to the repetition, over time, of relations with the agencies (Cantor, Packer 1995). Rating agencies are exposed to the risk of economically depending, in a systemic manner, on the capital market (Partnoy, 2006): economic dependence on their customers is an issue confirmed by the development of activities aimed at diversifying the geographical areas, as well as the customers and products (S&P, 2009). The pursuit of these objectives has determined the introduction, by the agencies, of concentration limits to the turnover with respect to single issuers (Securities and Exchange Commission, 2003). Generally speaking, smaller agencies feature a higher capacity to introduce innovative products, due to their greater operational flexibility, despite the fact that their size inevitably exposes them to a greater risk of economic dependence on the issuers (Fight, 2001).
The issue of economic dependence on their customers is addressed in the fundamental principles of the code of conduct of the rating agencies laid down by the International Organization Of Securities Commissions (IOSCO, 2004). In the section on the policies and procedures to ensure their independence and prevent any potential conflicts of interest, the principles provide that, as a rule, agencies should: abstain from any rating decisions when under the economic influence of external subjects, including the issuers; guarantee their formal and substantial independence and objectivity; assure that the rating assigned to the issuer is not influenced by any business relationships. Among the practices and policies adopted by the agency for the purpose of achieving the above mentioned objectives, the principles (as reviewed in 2008, IOSCO) provide that the agency must disclose to the market those issuers from which it receives more than 10% of its annual rating revenues. Based on the principles published by the IOSCO, the agencies have introduced codes of conduct that take into account the recommendations also concerning their independence from the issuers (Fitch Ratings, 2009; Moody's 2008; Standard and Poor's, 2007). The progress made in the implementation of the fundamental principles of the code of conduct recommended by the IOSCO has recently been investigated by the Committee of European Securities Regulators: in the case of Fitch it was found that there were areas of improvement, with regard to disclosure on the fees applied to issuers (Committee of European Securities Regulators, 2009).
As regards the relations between rating agencies and issuers, regulatory recognition plays a key role, firstly in the United States, following the introduction of the register of Nationally Recognized Statistical Ratings Organizations, followed by the New Basel Accord (Basel Committee on Banking Supervision, 2008): by laying down a set of minimum requirements for agency authorization, under the regulations, demand by the issuers has gradually shifted towards the recognized agencies, determining a further concentration of the market (Partnoy,1999). A primary role in the SEC's focus on the financial independence of the Nationally Recognized Statistical Ratings Organizations is played by structured finance: in this case, the issuers and/or arrangers of the operations, which are generally very concentrated, frequently request ratings thus generating a considerable flow of revenue for the agencies, with respect to their activities (Securities and Exchange Commission, 2008). Within the framework of the translation of the New Basel Accord into the EU Directives 2006/48/EC and 2006/49/EC (Official Journal of the European Union, 2006), economic independence is one of the factors that the supervisory authorities must take into account when deciding on agency recognition: exposure to conflicts of interest may be particularly serious when an agency's revenue-producing capacity depends on its principal customers, which might affect risk assessment (Committee of...