III. Seven Regulatory Principles

The principles outlined in this section should be observed by countries in their domestic law and should also be considered as principles of multilevel governance in this area.

A. Principle 1: The supervision of banking, securities and insurance should be further integrated

The key argument in favour of separate supervisors for several problems is one of specialization and division of labour. The main source of the gains from specialization is that knowledge costs are fixed costs, and thus the average cost decreases with utilization. Single focused agencies are likely to utilize their knowledge more intensively and thus will be able to acquire deeper expertise in their domain.[1] However, coordination costs increase with the proliferation of specialized agencies. Specifically, in the context of financial supervision and regulation, there are two main issues to consider. First, the proliferation of specialized agencies may increase the likelihood of a particular firm having multiple supervisors. This may leave unregulated gaps.[2] Moreover, as financial institutions continue to diversify into a broader range of activities, a single regulator will be more efficient at monitoring these activities (e.g. by operating a single database for licensing firms).[3]

This type of coordination cost is non-trivial as exemplified by the AIG supervision system. In the words of Fed Chairman Bernanke ‘AIG built up its concentrated exposure to the subprime mortgage market largely out of the sight of its functional regulators’,[4] The AIG-Financial Products exposure was mainly handled from London, hidden from its insurance regulators and the Fed, and essentially only supervised by the OTS. Thus the OTS, a small regulator in charge of the Savings and Loan industry was tasked with supervising what was a key cog in the global financial system.[5] The ability of AIG to conduct a large-scale scam makes most clear the risks of multiple competing regulators with overlapping responsibilities: a sophisticated financial institution may engage in a particularly insidious type of regulatory arbitrage, whereby it ‘chooses’ its own regulator, one that is unlikely to have the relevant knowledge and expertise.[6]

Thus the coordination costs derived from having multiple authorities are likely to be large. How large are the gains from specialization? These gains are unlikely to be large in the financial knowledge domain, as the substantive valuation, risk analysis, liquidity and solvency issues are the same in insurance, securities and banking. Thus, on grounds of division of labour, the balance is strongly in favour of an integrated authority.

A second possible argument for diversity is to encourage innovation. Having multiple regulators is like having several independent screens, where behaviours are accepted as long as they are accepted by at least one screen. In contrast, a centralized structure is like one with successive (not alternative screens) where only projects or ideas accepted by those successive screens are accepted.[7] If innovation matters, a decentralized structure will be preferred.[8] Conversely, the centralized system generates too little innovation, and leads to fossilization.[9] In our view, recent events show that financial innovation is of limited value relative to the risk engendered. A more centralized and hierarchical system is needed.

A final argument for multiple agencies is adaptation to change. If there are multiple regulators, some of them may prove better adapted to a change in the environment. On the other hand, if a coordinated radical change through the system is needed, a single regulator will be better able to implement it.[10]

The empirical evidence on the question is small. Barth, Dopico, Nolle and Wilcox33 find that countries with multiple supervisors tend to have lower capital- adequacy ratios and hence higher insolvency risk (they take this as evidence of the ‘competition in laxity’).

Our discussion here suggests that there are some limited pluses to a system with multiple agencies, but these pluses are in this case clearly overwhelmed by the coordination costs that we have identified. Both the USA and the EU area have a large multiplicity of actors already, particularly given the ‘federal’ structure of both areas. Given that many supervised financial institutions operate in all of those areas, the argument suggests that insurance, securities and banking supervisors should be further integrated.34

  • [1] The existing literature (Clive B. Briault, ‘The Rationale for a Single National Financial ServicesRegulator’, UK FSA Occasional Paper 2, 1999; David Llewellyn, ‘The Economic Rationale forFinancial Regulation’, UK FSA Occasional Paper 1, 1999; Richard K. Abrams and MichaelW. Taylor, ‘Assessing the Case for Unified Sector Supervision’, London School ofEconomics FinancialMarkets Group Special Papers 134, 2001) tends to emphasize the economies of scale advantage fromthe perspective of IT, support, etc., of a single supervisory agency. While there are undoubtedly someutilization gains of that kind, those are likely to be small compared with the knowledge gains of havinga group of subject specialists dealing repeatedly with matters within their domain.
  • [2] See Llewellyn, above n 23; Charles Goodhart ‘The Organisational Structure of Banking Supervision’, Financial Stability Institute Occasional Papers No 1, Bank for International Settlements, 2000.
  • [3] See Briault, above n 23, and see Llewellyn, above n 23.
  • [4] ‘Second, the AIG situation highlights the need for strong, effective, consolidated supervision ofall systemically important financial firms. AIG built up its concentrated exposure to the subprimemortgage market largely out of the sight of its functional regulators. More effective supervision mighthave identified and blocked the extraordinarily reckless risk-taking at AIG-FP.’ Ben Bernanke,Testimony to the Committee on Financial Services, US House of Representatives, 24 March 2009.
  • [5] AIG had bought a savings and loan—this was the reason the OTS was the regulator.
  • [6] The literature argues that a single, large supervisory authority is better able to attract, develop andmaintain professional staff expertise. See Briault, Llewellyn, and Abrams and Taylor, all above n 23.This has not been found to be the case in other domains, where specialized agencies can offer acongenial environment to the experts in that field irrespective of size (consider the CIA and NSA inintelligence, with clearly differentiated domains and different structures), and we do not expect it to bethe case here.
  • [7] Raaj K. Sah and Joseph E. Stiglitz, ‘The Architecture of Economic Systems: Hierarchies andPolyarchies’, 76(4) American Economic Review 716 (1986).
  • [8] See Edward J. Kane, ‘Regulatory Structure in Futures Markets: Jurisdictional Competitionbetween the SEC, the CFTC and Other Agencies’, 5 Journal of Futures Markets 367 (1984); RobertaRomano, ‘The Political Dynamics of Derivative Securities Regulation’, 14 Yale Journal on Regulation279 (1997); Roberta Romano, ‘The Need for Competition in International Securities Regulation’,Mimeo, Yale International Center for Finance, 30 June 2001, (visited 10 May 2010); Paul H. Kupiec and Patricia White, ‘RegulatoryCompetition and the Efficiency of Alternative Derivative Product Margining Systems’, 16 Journal ofFutures Markets 943 (1996).
  • [9] A variant of this argument would consider learning—multiple supervisory authorities may adoptdifferent approaches to supervision which can yield valuable information that would not be generatedby a single supervisor. See Llewellyn, above n 23.
  • [10] The existing literature has also argued that a single regulator may be preferred in other areas suchas conflict resolution (a single regulator is better able to resolve conflicts that emerge between different
< Prev   CONTENTS   Source   Next >