Dec 28, 2009
The Importance of the Regulatory Agency Structure
It is inevitable that those working in the FIs area, or academics studying them, think that the institutional structure of the regulatory agency is very important. Abrams and Taylor (2002) put things in perspective, we think, remarking that the structure tasks are second order issues. They come after various conditions are in their right place, and are done with the right degree of intent and effort.
If the institutional structure is important (and it is, just we should remember that it is something to work on after other things are to our satisfaction) is, mainly, because the wrong organization can ruin the best team we can gather, or the best set of conditions chosen, or the best laws written.
We are told than when structure is weak, many undesirable effects can appear: monitorization is less effective or comprehensive than possible; regulatory arbitrage is more frequent; promised economies of scale are not realized; deployment of staff is not optimal; too many disputes among regulators (or among units of a unified regulator) happen; costs for the taxpayer are too large. To solve those effects (and others) perceived as caused by structure weakness, various models are proposed for regulators and supervisors, depending of particular conditions. Before any change is made, in one direction of the other, some considerations on the prerequisites for an effective supervisory structure must be added. Abrams and Taylor (2002) very cautiously think that this list, which doesn't aim to be exhaustive, "attempts to provide an indicative set of key features that constitute an effective supervisory structure":
We are told than when structure is weak, many undesirable effects can appear: monitorization is less effective or comprehensive than possible; regulatory arbitrage is more frequent; promised economies of scale are not realized; deployment of staff is not optimal; too many disputes among regulators (or among units of a unified regulator) happen; costs for the taxpayer are too large.
To solve those effects (and others) perceived as caused by structure weakness, various models are proposed for regulators and supervisors, depending of particular conditions. Before any change is made, in one direction of the other, some considerations on the prerequisites for an effective supervisory structure must be added. Abrams and Taylor (2002) very cautiously think that this list, which doesn't aim to be exhaustive, "attempts to provide an indicative set of key features that constitute an effective supervisory structure":
objectives should be clear: if possible mandated by statute to prevent the supervisory process from costing too much to the financial system or the taxpayer - which is the unavoidable path to follow, since the regulator is risk-averse and is not commended by the lack of crises but is pilloried when some occurs;
independent and accountable regulator: again with the support of a statute that protects managers from being deposed by politicians easily and with some mechanism to make the agency accountable to both the taxpayer and to industry (this seems very difficult to execute to us);
regulation must be cost effective;
resources must be adequate: ability to "recruit, train and retain a cadre" of skilled personnel;
effective & flexible enforcement powers: ability to require information from regulated firms, to assess the competence and probity of senior management and owners, and to take appropriate graduated sanctions;
regulation must be comprehensive: regulators must also be in a position to respond quickly to market innovations;
the Effectiveness Criteria. we should consider also whether other solutions are more effective or less risky, and that perhaps changes may be inappropriate due to current events.
Models of regulatory structure
The WB (2005) noted: "the institutional structure of regulatory agencies is an issue of some significance. However, the importance should not be exaggerated." We cannot be too critical with any system (specific cases of jobs badly botched can and must be criticized). And we do not think that the US or the UK systems are clearly superior to other systems, or as bad as critics say. In corporations the structure is being changed constantly, according to the needs and what is learnt. This includes the partition of the company in several ones, and after a time a reorganization that again concentrates some pillars into one (in some cases leaving just one company).
According to Rustomjee (2009), the new preference among countries for unified approaches to supervision and regulation "suggest[s]" that some governments have "identified useful advantages in the unified models of regulation." We think that it is better this nuanced formulation than a previous one in the same text ("in the past decade a number of countries have shifted towards more unified regulatory arrangements, suggesting that there may be important advantages in unified models of regulation"). We don't know for sure what is best, and if corporate life is any indication, the best organization is an elusive goal. And if one reaches that blissful state of best organization, it lasts a very short time period. Maybe it was just herd behaviour? Or subtle pressures from the EU guys that have power on this area?
A great variety of regulatory structures can be observed today. We can try to adscribe this wide range of approaches to these three categories:
Due to history, political structure (e.g., a federation) or sheer size (e.g. the US), the regulator can be highly fragmented, paying attention to just a specific financial sector activity (e.g., independent regulators for banks, for insurance companies and for securities firms).
The case par excellence is the US. A great critic of this model is Brown (2007). Spain is a similar case, although not that greatly fragmented. The Bank of Spain works with banks and S&L Associations; a Treasury directorate general with insurance companies; and an independent commision with the stock market and securities firms.
In the WB study (2005) this model represented 42% of all countries in the study.
In this case, a regulator supervises two of the three major financial activities. Examples in which a single agency works on banks and securities firms are Finland, Luxembourg or Switzerland. This comprised 6% of the countries in the WB study (2005).
Examples in which a single agency works on banks and insurance companies are Belgium and Canada (comprising 12% of countries). Examples in which a single regulator supervises securities firms and insurers are Chile, Slovakia or South Africa (11% of countries).
This case includes some varieties: if all existing regulators are merged into a single institution we are talking about the fully unified one. A close case (although preserving the central bank), is the UK's FSA.
Sometimes, instead of working on the institutions being regulated, the supervisors are split on two or more: an institution for prudential supervision of all entities, a separate institution for supervision of market conduct. If split into two, this is known as the ‘twin peaks’ model. The two cases are Australia and the Netherlands.
Some small population countries (and countries that recently entered into the EU) follow this model; notable cases are Austria, Denmark, Norway, and Sweden. More notable because of their population size and large economy are Germany, Japan, and the UK. Of the countries studied, 29% had a unified regulator system.
Some advantages of unification with disadvantages added
Paraphrasing the World Bank (2005), there is merit in the arguments for a unified model, but also several good reservations may be stated. And Abrams and Taylor (2002) offer good comments about, 1 the risks and unpredictability of the very change itself (you can end up with a system weaker than the original one due to bargaining and political capture), and 2 the risks of qualified, experienced staff turnover in their Box 6.1.
Some advantages of unified systems listed by the WB (2005) and Abrams and Taylor (2002), or weaknesses of fragmented systems follow, along with comments adding disadvantages (or strengths of fragmented systems) in italics. We simplify the very appropriate caveats, buts and ifs:
A regulator that mirrors the conglomerates' structure should monitor more effectively the full range of activities of such complex organizations. Why is this assumed to refer to unification? First of all, corporations frequently divide themselves in fully independent national groups, comprised in turn of several companies. In several countries you can find IBM itself (which has several divisions), IBM GBS, IBM Global Finance, IBM Data and a couple others. Of course, auditing units and risk management divisions are multiplied almost accordingly. Never it was intended to have all audit sections fully unified. Maybe lawmakers and supervisors need to ask themselves why corporate life is so different. Or can we be sure that private firms are so wrong on so many areas? And second, although firms have diversified, their core business keeps being dominant in the overwhelming majority of cases. The risks nature is different enough to warrant different prudential supervisory regimes, and there would be few (if any) efficiencies in bringing their supervision together.
The regulator can be understood and recognized as such by regulated firms and individuals. Moral hazard can increase: many may understand that all creditors protected by a supervisor will receive equal protection.
The regulator should "avoid problems" of duplication, gaps, inconsistencies, and competitive inequality that can arise with a regime that is based on several institutions. Maybe that "avoid" is too optimistic? Could be better to say "reduce"? And second, such regulator that could "avoid" inconsistencies, overlaps and duplication could become excessively bureaucratic and slow to confront changes in the financial world.
The single agency should minimize regulatory arbitrage (e.g., the placement of a financial product in that part of a conglomerate where the supervisory oversight is lowest). This arbitrage also can induce “competition in laxity,” or rush to the bottom. This last is not necessarily bad if the regulatory regime is not a corrupt one. It is not clear that contention of state growth is counter to the people's interests, and some argue otherwise.
Economies of scale can be gained (particularly with respect to skill requirements and recruitment of staff members). A single regulator might be more efficient because of shared resources (like shared IT systems). This might apply particularly to the “small-country” case (and small financial sector case). So maybe Australia, the Scandinavian countries and other small population countries can get the much heralded economies of scale, but these are almost ruled out by the WB in cases like the US and the UK. Besides, economies of scale, as in corporate life, may not happen, and X-inefficiencies may arise.
Deployment of staff members within a unified agency should be optimal, compared to the specialist and fragmented regulator. Experience in the corporate world says this same and the opposite.Rustomjee (2009, p. 34), shows this was not done well at the UK's FSA. Besides, that advantage can quickly disappear past some size limit.
If expertise in regulation is in short supply, expertise might be used more effectively if concentrated within a single agency, which also might offer better career prospects. Wrongly done, it could reduce prestige of the better workers, who may leave for the private sector.
Accountability of regulation also might be more certain with a simple structure if for no other reason than that it would be more difficult for different agencies to “pass the buck.” This can also happen among units of any organization. Besides, accountability might be more difficult if clear objectives are not defined, and definition of those in a single regulator can be more difficult to attain.
The costs imposed on regulated firms might be reduced if firms need to deal with just one agency; economies and greater effectiveness can be gained when all information about financial firms is within a single supervisor. Also there are opportunities for increased corruption, and more homogeneity of methods, computer models and even of vision, which runs counter to Alexander-Dhumale-Eatwell's description of increased homogeneity's dangers.
More complete coverage can be obtained, with less organizations slipping through the supervisory net because of confusion about which regulator is responsible. There should be less damaging disputes between agencies in a multiple-agency structure. Those happen frequently among any organization's units, regardless of size or structure.
A danger of a fragmented supervisor is that similar products are regulated differently because they are supplied by different types of financial firms, diminishing competitive neutrality. On the other hand, a single regulator could loss valuable information and a degree of competition and diversity in regulation – the case for not having a single regulator is alike that against any monopoly.
The WB authors recognize that the arguments for and against single prudential supervisors are "finely balanced, and the optimal structure is likely to vary between countries," depending on the country size, political considerations, past practices, and the structure of their financial system (i.e., whether it is comprised of conglomerate or specialized institutions).
The US: fragmented regulator
The US system is a cobweb of regulators created upon a confederation of states, each with its own rights of supervision in its own territory. This complex system has a good record of financial innovation are quite a big legacy of financially-driven prosperity than many other countries did not enjoy. Geithner (2008) argued that the regulatory system in the US has become
"a confusing mix of diffused accountability, regulatory competition, an enormously complex web of rules that create perverse incentives and leave huge opportunities for arbitrage and evasion, and creates the risk of large gaps in our knowledge and authority."
Among the many criticisms voiced against this system we can mention:
There is not enough emphasis in systemic risk management.
Not enough attention was given to conduct-of-business regulation. We think that the states do quite a decent job protecting the consumer.
FIs avoided supervision thru the establishment of SVIs and conduits to keep operations off-balance-sheet. Here we disagree a bit: everybody knew this was happening, just it was thought of as a clever way to deal with risks. But it was not done with the regulator unaware of this.
Little coordination resulted from the excessively fragmented structure, in which there was no lead agency. This had some effect in the crisis' genesis, but would like to add that the fact of not having such a leading supervisor is was a result sought by the lawmaker both in the federation and the states, afraid of concentration of power. This fear is an important consideration in the US.
The system didn't keep pace with market and technology innovation. Again, the lawmaker and the supervisors were aware of these developments, and thought them beneficial. And countries with a large, modern, sophisticated financial system with all technological means like Japan did experience a much softer banking crisis (although GDP took a great hit later).
The US regulator was surpassed by institutional innovation, like globalized, despecialized conglomerates, which exercised great regulatory arbitrage, avoiding capital adequacy and other limitations. But the BCBS taught that all single-family mortgages were much less risky than any commercial loan (regardless of the collateral provided, Martin S Feldstein, 1993)). And the lawmaker, supervisors and market operators thought that the CRAs computed the right ratings for those MBSs. That's why they bought them: they made money while reducing compliance costs and were keeping risks, everybody thought, manageable (Stiglitz, Orszag & Orszag, 2002).
The system is too costly. Brown (2007, pp.60-61) says total US regulatory costs are 16 and 117 times, respectively, than those in UK and Germany. Few defended that the FSA was innefficient, but following Brown's reasoning, US costs = 16*UK, US costs = 117*Germany, dividing we get UK costs = 7 times Germany costs. Who complains of this? Obviously there are other criteria to consider. Later she computes more realistic factors (down to 4 and 14 in p. 63). We'd like to add this thought: if in Spain there are four and in the UK there are three institutions working on banking, insurance and securities regulation and supervision, I think we can put 3 as a good average number for the 25 members in the EU. That is 76 institutions for 25 countries and EU. It is not so crazy to have 115 in the US fifty states, other territories and the federation.
Many proposals to rationalize the US system and its 115 regulators were made in the last decades (US Treasury Department 2008, Appendix B). We find merit in thinking that incremental change, as proposed in the US Treasury blueprint (Chapter IV, Short-Term Recommendations, and Chapter V, Intermediate-Term Recommendations), is likely to result in a better regulatory system than too pasionate proposals like Brown's (2007).
But at the same time we see:
An almost complete lack of criticism with the very perverse incentives that the lawmaker created in the US for inter alia the GSEs and the FHA (and those other quasi-governmental enterprises, the CRAs) to securitize sub-prime and non-prime loans as low-risk ones (Friedman, 2009, pp130-132, pp143-146).
That the US Treasury proposal (2008, figure in p.144) maybe is a very reductionist one: making approving reference to the Australian case, the great heterogeneity of FIs in the US and the enormous differences in FIs and among a 300-million population country and tens-of-million populations in Australia, Nerderlands or Scandinavia are arguably not adequately accounted for. We think that is apparent the risk that the amount of knowledge that can be transferred from those unified regulator experiences to the US is small.
That the prudential supervision system built in the last decade worked apparently well in a very diverse group of countries in all continents, that is, many fragmented supervision systems did well in 2007-2009. This, incidentally, again stresses the first point, that is, wrong incentives by the lawmaker that were not present in other countries.
- Add to this that low-taxation, lighly regulated jurisdictions, like the OFCs, are doing well too.
In corporate life there are many subdivisions and duplicative structures, and they are purposefully so.
The UK: unified model
The Northern Rock and UK interbank lending crises, a period in which the BoE, the FSA and the Treasury didn't coordinate effectively and gave a sorry spectacle of decisions and counterdecisions (like accepting MBSs as collateral to provide LOLR support, and changing the rules of the deposit insurance system, making it wider and deeper), is an example of real operation of the fully unified regulatory model in a large, sophisticated system like the UK's when times are difficult. Where we should have seen efficiencies we got to varying degrees high staff turnover, lack of specialists, lack of training, and regulatory paralysis.
Indeed, we got not even accountability: we are not aware of many staff ousted in any of the three agencies. On the contrary, FSA's Turner directs almost no criticism to the regulators in his speeches, just to the easy targets, and FSA's Hector Sants (2009) spells the end of the light-touch regulator with no reference to Parliament – he also speaks of contracting hundreds of supervisors for his agency. In the end, the UK will get micromanagement and a large bureaucracy with a lot of personnel and new in-house complex models and supporting computer systems, ending the savings promised.
The WB (2005) report comments on the case of Ireland and Finland, where economies of scale in infrastructure, information technology, and services were achieved by locating the agencies at the same place and by sharing resources while "nevertheless, maintaining strict separation of regulatory and supervisory" policy and practice. This is not much discussed, but maybe some variation of these intermediate solutions could be considered.
Views from corporate life
Big global firms, like Siemens or IBM, with more than 350 000 employees/contractors, have organizational problems according to their size and worldwide presence. As example: in corporate life there is a maxim about how individual divisions, project portfolios, businesses lines, products, activities, and people should be shed, and shed annually. Why is this constant pruning a rule? A reason is that there are no perfect institutional arrangements, activities, goals or teams. In the central banks and other supervisory agencies they talk (thru their officials' speeches) as if it were possible to find the perfect organization, and make it last indefinetely.
In those private, global, very dynamic organizations they live a constant review processes to make things better (although there is no much hope of ever arriving at the "this is the best organization" state). There are audits in several areas: "money" fraud (external fraud, internal controls, etc.), technological security (network security, computer security, physical access to buildings, laboratories, etc.), and quality assurance/risk management ("quality inspection" of software, of physical components, and of processes). There are several "internal affairs divisions," so to speak. That's why we cannot be too critical with the US system. Supervisory system, that is, because the lawmaker's wrong incentives are arguably worse than the too complex supervision.
This discovery program in corporations is arguably better fine tuned than what the lawmaker can design. We see no discussion about large companies "detectives" or "inspectors," or about structure reform programs, in the papers or speeches by central bankers. We see in central bankers, which work with really big companies (Sumitomo, Deutsche Bank), lack of awareness of the optimizing work permanently being done in the corporate world. This is in full display in the WB/IMF (2005) report. An exception seems to be the Walker Review (2009), that published a chapter on 'Governance of risk' on November which seems more acquainted with real corporate practice than other studies.
So yes, structure is important, but: 1 other things are first (the prerequisites of Abrams & Taylor supra), 2 those supervision tasks are also done in the supervised companies and in much bigger organizations and are seldom discussed and it seems few things are to be learnt from them, and 3 why do we think that legislative action can do a better, faster job of reforming structures than the reform processes in firms if those reforms are very, very difficult and need a constant re-work? We cannot say which of three seems more important to us.
Then, after so much conflicting data, so much questions, so few support from those in the "real world," so to speak, of corporations, why is there so much fuss about unification? Here we cannot avoid to be cynical. It is true than recent developments (financial services innovation and globalization, the emergence of international financial conglomerates, the lessons from repeated episodes of financial crisis in some countries, lessons of best (and worst) practice from other countries (WB, 2005), make an influence in the lawmaker and its counsels, but to us this very much seems a pendular experiment.
We wonder whether all this emphasis on unification could be one of those efforts by the lawmaker to appear as active (as it is demanded by the people), doing something after what happened in the 1987-2000 crisis, even if we need still more time to learn more about organization and structure.
If the lawmaker ends thinking that crises were not prevented by unified regulators like FSA, as also happened with the non-unified ones, soul-searching will start in earnest and the interest in unified regulators will fall out of favor.
Abrams, R and M Taylor (2002) ‘Issues in the Unification of Financial Sector Supervision’, Chapter 6 in Charles Enoch, David Marston and Michael Taylor (eds) Building Strong Banks Through Surveillance and Resolution, Washington, DC: International Monetary Fund.
[A previous version is downloadable in PDF: Abrams, Richard K. & Michael Taylor 2000 'Issues in the Unification of Financial Sector Supervision'. IMF Working Paper No. 00/213. Washington, DC: International Monetary Fund. http://www.imf.org/external/pubs/cat/longres.cfm?sk=3939.0]
Alexander K, R Dhumale and J Eatwell (2006) 'Global Governance of Financial Systems: The International Regulation of Systemic Risk'. Oxford: Oxford University Press.
Brown, Elizabeth (2007) ‘E Pluribus Unum – Out of Many: Why the United States Needs a Single Financial Services Agency’, University of Miami Business Law Review, Fall/Winter. Ask for a PDF.
Feldstein, M (1993), Comment to Boyd and Gertler (1993), p 375, in John Boyd & Mark Gertler, US Commercial Banking - Trends, Cycles and Policy, in NBER Macroeconomics Annual 1993, Volume 8, Olivier Blanchard and Stanley Fischer (eds), MIT Press. http://www.nber.org/chapters/c11003
Financial Stability Institute (2007) ‘Institutional Arrangements for Financial Sector Supervision’, Occasional Paper No. 7, Basel, Switzerland: Financial Stability Institute, BIS. http://www.bis.org/fsi/fsipapers07.htm
Friedman J (2009) ''A Crisis of Politics not Economics – Complexity, Ignorance, and Policy Failure'. October. Critical Review 21(2–3): 127–183. Oct 2009. DOI: 10.1080/08913810903030980
Geithner, T (2008) ‘Reducing Systemic Risk in a Dynamic Financial System’, Remarks at The Economic Club of New York, 12 June 2008, Federal Reserve Bank of New York. http://www.newyorkfed.org/newsevents/speeches/2008/tfg080609.html
Goodhart, C (2000) ‘The Organisational Structure of Banking Supervision’, FSI Occasional Paper No. 1, November, Basel, Switzerland: Financial Stability Institute, BIS. http://www.bis.org/fsi/fsipapers01.pdf
Rustomjee, C (2009) 'Bank Regulation and the Resolution of Banking Crises, Unit 6'. London: School of Oriental and Management Studies-CeFiMS.
Sants, H (2009) 'Speech at Bloomberg'. UK FSA, Nov 09, 2009, http://www.fsa.gov.uk/pages/Library/Communication/Speeches/2009/1109_hs.shtml
In the past, the FSA was primarily reactive, only making interventions on readily observable facts and adhering to the view that it should leave management to make its own decisions.
Intensive supervision, in contrast, focuses on the risks inherent in a firm’s business model and enables us to be proactive and not reactive to the management of these risks.
Our outcomes-focused philosophy requires supervisors to judge firms on the likely consequences of their decisions.
This means the proportion of our time spent looking at systems and controls will diminish relative to our focus on assessing the outcomes of a firm’s actions. This will necessarily be controversial at times, as our view and the firm’s view will not always coincide.
This divergence of judgement can normally be resolved, but the FSA recognises that this new approach may create tensions and will certainly no longer be seen as light touch!
To enable us to deliver on this approach we have equipped ourselves both to forecast and test outcomes. This capacity is needed to enable us to effectively make judgements on the judgements firms are making.
The forecasting element requires in-house modelling capability, such as that we now have for capital and liquidity. This does of course require more data to be collected from firms. And it will involve us looking more closely at the risks, both prudential and conduct, inherent in the product; from development to expiry.
The testing element requires supervisory resource to be devoted to an inspection-based approach to individual transactions.
Sinclair, Peter JN (2000) ‘Central Banks and Financial Stability’, Bank of England Quarterly Bulletin, November: 377–89. http://www.bankofengland.co.uk/publications/quarterlybulletin/qb000403.pdf
Stiglitz J, Orszag J & Orszag P (2002) 'Implications of the New Fannie Mae and Freddie Mac Risk-Based Capital Standard'. Fannie Mae Papers, Volume I, Issue 2, March. Ask for a PDF.
US Treasury Department (2008) ‘The Department of the Treasury Blueprint for a Modernised Regulatory Financial Structure’ (April), Washington, DC: US Treasury Department.
The Walker Review (2009) 'A review of corporate governance in UK banks and other financial industry entities: Final recommendations'. November 26, 2009. http://www.hm-treasury.gov.uk/d/walker_review_261109.pdf (PDF 823KB)
All documents of this review: http://www.hm-treasury.gov.uk/walker_review_information.htm
WB (The International Bank for Reconstruction and Development/The World Bank/The International Monetary Fund) (2005) 'Financial Sector Assessment: A Handbook', Appendix F: ’Institutional Structure of Financial Regulation and Supervision’. Washington, DC: IBRD/WB/IMF.
BCBS: Basel Committee on Banking Supervision
BIS: Bank of International Settlements
BoE: Bank of England
ECB: European Central Bank
CRA: credit rating agency
FHA: Federal Housing Administration
FI: financial institution
FSA: Financial Services Authority
IBRD: International Bank for Reconstruction and Development
IMF: International Monetary Fund
GBS: Global Business Services
GSE: government-sponsored enterprise, government-sponsored entity
MBS: Mortgage-backed security
OFC: offshore financial center
S&L Associations: Savings and Loans Associations
SVI: structured investment vehicleWB: World Bank