McGill Law Journal ~ Revue de droit de McGill
REGULATING FINANCIAL INSTITUTIONS:
THE VALUE OF OPACITY
Anita Anand & Andrew Green*
In this article, we explore a question of in-
stitutional design: What characteristics make a
regulatory agency effective? We build on the
growing body of administrative law literature
that rigorously examines the impacts of trans-
parency, insulation, and related administrative
processes. We argue that there are certain ben-
efits associated with an opaque and insulated
structure, including the ability to regulate un-
fettered by partisan politics and majoritarian
preferences. We examine Canadas financial in-
stitution regulator, the Office of the Superin-
tendent of Financial Institutions (OSFI), whose
efficacy in part explains the resilience of Cana-
das banking sector throughout the financial
crisis of 2008. In particular, OSFI operates in a
black box, keeping information about the for-
mation of policy and its enforcement of this pol-
icy confidential. With its informational ad-
vantage, it is able to undermine the possibility
that banks will collude or rent-seek. Our con-
clusions regarding the value of opacity cut
against generally held views about the benefits
of transparency in regulatory bodies.
Dans cet article, nous explorons une ques-
tion dorganisation
institutionnelle : quelles
sont les caractristiques dune agence de rgle-
mentation efficace ? Nous nous appuyons sur
une littrature croissante, en droit administra-
tif, portant sur les impacts de la transparence,
de lisolation, et dautres processus administra-
tifs relis. Nous soutenons quil y a certains
avantages lis une structure opaque et isole,
savoir la capacit de rglementer sans
linfluence des politiques partisanes ou des pr-
frences majoritaires. Nous
examinons
lorganisme de rglementation des institutions
financires du Canada, le Bureau du surinten-
dant des institutions financires (BSIF), dont
lefficacit explique en partie la rsilience du
secteur bancaire canadien lors de la crise finan-
cire de 2008. Le BSIF travaille huis clos ; il
ne dvoile ni le processus dlaboration de ses
politiques, ni sa manire de les appliquer. Cet
avantage au niveau de linformation lui permet
de diminuer les risques de collusion et de re-
cherche de rente de la part des banques. Nos
conclusions sur la valeur de lopacit font con-
trepoids aux positions gnralement admises
concernant les avantages de la transparence au
sein des organismes de rglementation.
* Professors, Faculty of Law, University of Toronto. Thanks to the participants of the
2010 meetings of the 4-Sided Law and Economics Group (Faculty members from Toron-
to, Sienna, and Tel Aviv), the 2010 annual meeting of the Canadian Law and Econom-
ics Associations, UBCs Business Law Institutes 2011 Seminar on Banks, Markets and
Regulation, and Grant Bishop for helpful comments. Thanks to Emily Bala, Adam
Friedlan, Haakim Nainar, Chava Schwebel, and Dharshini Vigneshawaran for valua-
ble research assistance funded by the Social Sciences and Humanities Research Council
of Canada.
Anita Anand & Andrew Green 2012
Citation: (2012) 57:3 McGill LJ 399 ~ Rfrence : (2012) 57 : 3 RD McGill 399
400 (2012) 57:3 MCGILL LAW JOURNAL ~ REVUE DE DROIT DE MCGILL
Introduction
I.
Financial Institution Regulation: Rationales and
Structural Characteristics
II. Optimizing Administrative Design
A. Insulation
B. Transparency and Opacity
C. Combining Insulation and Opacity
III. Regulating Financial Institutions
IV. Policy Implications
Conclusion
401
406
408
410
413
415
418
423
426
REGULATING FINANCIAL INSTITUTIONS 401
Introduction
The global financial crisis highlighted profound difficulties in the
banking sector, which in turn cast attention on differing regulatory ap-
proaches across countries.1 Drawing on the administrative law literature,
we examine the characteristics of a regulatory agency that oversees finan-
cial institutions and ask how these characteristics influence the strength
of its oversight. In particular, we analyze the agencys insulation from po-
litical control and the opacity in its operations.2 We discuss both the value
of these characteristics in the regulation of financial institutions them-
selves and also how these characteristics interactwhen are they com-
plements and when are they substitutes?
In order to ground the theoretical discussion, we examine the regula-
tor of Canadian financial institutions, which is called the Office of the Su-
perintendent of Financial Institutions (OSFI). Canadas financial institu-
tions weathered the crisis well relative to their international peers, an
outcome that has been attributed at least in part to the presence of an ef-
fective regulator.3 The academic literature points to Canadas regulatory
structure as a factor that discouraged banks from taking excessive risks.
1 The ratio of non-performing loans to total loans in the United States rose from 0.8 in
2006 to 3 in 2008 and the return on average equity (ROE) dropped from 13 to 1.4. The
European Union-wide average ratio of non-performing loans rose from 1.9 in 2006 to 2.9
in 2008, while after-tax ROE dropped from 15.8 to -8.4 in the same period. See
International Monetary Fund, Global Financial Stability Report: Navigating the
Financial Challenges Ahead, (Washington, DC: IMF, 2009) at 188-190, online: IMF
Kunt, Enrica Detragiache & Ouarda Merrouche, Bank Capital: Lessons from the
Financial Crisis (2010) International Monetary Fund Working Paper No 286 at 32,
online:
(showing
average quarterly bank stock returns by country from quarter 1, 2006 to quarter 1,
2009 and reporting quite severe drops in valuation); Andrea Beltratti & Ren M Stulz,
Why Did Some Banks Perform Better During the Credit Crisis? A Cross-Country
Study of the Impact of Governance and Regulation (2009), online: Social Science
Research Network
& Thierry Tressel, A New Database of Financial Reforms (2008) IMF Working
Paper No 266, online: IMF
DoddFrank Acts Expansion of State Authority to Protect Consumers of Financial
Services (2011) 36:4 J Corp L 893.
IMF
2 We use the terms opacity and transparency throughout the text as two sides of the
same coin. Each term refers to certain costs and benefits related to institutional design.
3 See Canadian Securities Institute, Canadian Best Practices Take Centre Stage at Fi-
nancial Conference in China (25 February 2009), online: Focus Communications Inc
Bank of England & Treasury, Financial Stability and Depositor Protection: Further
Consultation (London: HM Treasury, 2008), online: HM Treasury
402 (2012) 57:3 MCGILL LAW JOURNAL ~ REVUE DE DROIT DE MCGILL
Ratnovski and Huang, for example, examine the performance of the sev-
enty-two largest commercial banks in OECD countries during the finan-
cial crisis, analyzing the factors behind Canadian banks relative resili-
ence at this time.4 They identify two main causes, one of which is regula-
tory factors that reduced banks incentives to take excessive risks.5 This
view of the efficacy of banking regulation permeated Canadas own gov-
ernment. For example, OSFI was not subject to regulatory reform follow-
ing the crisis, and the government in fact relied on the structure of OSFI,
and new guidelines that OSFI proposed, to stave off an international bank
tax.6
Building on the existing literature, this article assumes that OSFI
played a role in ensuring the stability of Canadian financial institutions
relative to their international peers during the financial market meltdown
of 2008.7 Of course, we do not mean to exclude factors other than Canadas
regulatory structure that likely contributed to the stability and perfor-
mance of Canadian financial institutions, such as: Canadas oligopolistic
market (consisting of five big banks);8 macroeconomic policies emanating
from the Bank of Canada; the historical structure of the banking system;
the approach to regulating the Canadian mortgage market; and, a gener-
ally conservative approach that pervades the banking sector.9 Given the
4 Lev Ratnovski & Rocco Huang, Why Are Canadian Banks More Resilient? (2009) IMF
Working Paper No 152, online: Social Science Research Network
5 Ibid. Other factors included a higher degree of retail depository funding, and to a lesser
extent, sufficient capital and liquidity.
6 The government argues that such a plan is unnecessary in Canada and has instead
proposed a solution that it calls embedded capital. This proposal would require each
countrys regulator to set rules or guidelines concerning financial institutions that issue
debt convertible to equity as a means of self-insuring against failure. The proposal is
controversial and is not supported by other G-20 countries: See e.g. Paul Vieira, G20
May Fail to Agree on Financial Reform: Flaherty, Financial Post (16 April 2010) (Fac-
tiva); Andrew Mayeda, Flaherty Calls for G20 to Refocus Financial Reforms on Bank
Capital, Canwest News Service (18 May 2010), online: Global News
7 See e.g. Paul Krugman, Good and Boring, The New York Times (31 January 2010)
online: The New York Times
8 See Franklin Allen & Douglas Gale, Competition and Financial Stability (2004) 36:3
Part 2, Journal of Money, Credit and Banking 453 at 472. Comparing the United
States, whose history is marked by greater financial instability, to the United Kingdom
and Canada, where the banking sector is dominated by a few big banks, they argue that
supervision is more effective in a concentrated banking system.
9 International Monetary Fund, Canada2008 Article IV Consultation: Preliminary
Conclusions of the IMF Mission (2007) IMF Staff Report, online: IMF
online: University of Toronto, Faculty of Law, Faculty Blog
Regulatory System: Testimony Before the Senate Committee on Banking, Housing, and
Urban Affairs, 111th Cong (2009) (Richard J Hillman, GAO Managing Director), online:
GAO
10 Canadian Banks Forum, Big Five Canadian Banks (2011), online: Canadian Banks
Forum
11 The Office of the Superintendent of Financial Institutions was established in 1987 pur-
suant to the Office of the Superintendent of Financial Institutions Act, being Part I of
the Financial Institutions and Deposit Insurance System Amendment Act, RSC 1985, c
18 (3d Supp), as amended by SC 1996, c 6 [OSFI Act].
404 (2012) 57:3 MCGILL LAW JOURNAL ~ REVUE DE DROIT DE MCGILL
promotes the adoption of policies and procedures designed to control and
manage risk. It must also monitor and evaluate system-wide or sectoral
issues that may negatively impact institutions.12 OSFI functions within a
web of other bodies discussed below.13
We argue that OSFI exhibits two traits in particular that are connect-
ed to its efficacy: it is both insulated and opaque. Insulation refers to its
separation from elected officials in key respects. This separation can be
seen, for example, in the appointment and removal requirements for OSFI
members, the lack of control by elected officials over the structure and
procedures of OSFI, and the weak powers of elected officials to review or
issue directives to the body.
OSFIs opacityor lack of transparencyexists in the processes for
establishing rules and guidelines. While the federal cabinet makes regula-
tions that OSFI must administer, OSFI itself develops and implements
other forms of regulation, such as guidelines and policy statements. There
is no prescribed process, no mandated public consultation, no necessary
stakeholder input or cost-benefit analyses required with regards to the
formulation of either policy statements or guidelines.14 The process by
which OSFIs capital adequacy guideline is developed is a particularly
striking example of opacity. The guideline seeks to provide a framework
within which the Superintendent assesses whether a bank … maintains
12 Ibid, s 4(2). As part of discharging this mandate, OSFI administers the Bank Act, which
grants authority to Cabinet for establishing regulations that define the regulatory capi-
tal of a bank and authority to OSFI for capital and liquidity requirements (SC 1991, c
46, s 485). Section 648 sets out the conditions under which supervisory intervention
may occur, one of which is contravention of OSFI guidelines for capital and liquidity
(ibid).
The OSFI also administers a host of other pieces of financial legislation that pre-
scribe constraints on where and how it can regulate: OSFI Act, supra note 11, s 6,
Schedule to Part I. Other statutes include: Trust and Loan Companies Act, SC 1991, c
45; Cooperative Credit Associations Act, SC 1991, c 48; Insurance Companies Act, SC
1991, c 47; Pension Benefits Standards Act, 1985, RSC 1985, c 32 (2d Supp).
13 For example, the Financial Consumer Agency of Canada (FCAC) has responsibility for
consumer protection and exercises authority in relation to the consumer provisions of
the Bank Act (supra note 12): Financial Consumer Agency of Canada Act, SC 2001, c 9,
ss 2-3. The Bank of Canada is responsible for the economic and financial welfare of
Canada. It regulates credit and currency in the best interests of the economic life of the
nation and controls national monetary policy: Bank of Canada Act, RSC 1985, c B-2,
Preamble.
14 There is a public process for formal regulations involving pre-publication consultations
and analysis. For federal guidance on the regulation-making process, see Canada,
Treasury Board Secretariat, Cabinet Directive on Streamlining Regulations (Ottawa:
Treasury Board Secretariat, 2007), online: Treasury Board Secretariat
REGULATING FINANCIAL INSTITUTIONS 405
adequate capital pursuant to the acts.15 In the guideline, the superinten-
dent establishes two minimum standards: assets to capital multiple and
risk-based capital ratio. The guideline indicates that OSFI may also issue
further notes to clarify expectations on compliance with the technical
provisions of the internal ratings approach set out in the guideline.16
While OSFI is actively engaged in international discussions around ap-
propriate standards for financial regulation, such as for the Basel III
standards, OSFI is not mandated to adopt these standards. For our pur-
poses, the key point is that in an area of crucial importance to the stabil-
ity of financial institutions, OSFI is able to pass and implement guidelines
without Cabinet approval or an open process. Rather, they are established
by a relatively insulated body through a partially opaque process.
Despite its insulation and opacity, however, OSFI is almost universal-
ly viewed to be an effective regulator.17 In this article, we ask: What is it
about institutional insulation and opacity that may lead to effective regu-
lation of banks? Part I examines the purposes of financial institution reg-
ulation. Part II then discusses the characteristics of administrative agen-
cies, and particularly the relationship between transparency (or opacity)
and insulation. We build on the growing body of administrative law litera-
ture that rigorously examines the impacts of transparency, insulation,
and related administrative processes.18 We argue that there are certain
benefits associated with an opaque and insulated structure, including the
ability to regulate unfettered by partisan politics and majoritarian prefer-
ences. Our conclusions regarding the value of opacity cut against general-
ly held views about the benefits of transparency in regulatory bodies.
The article then uses OSFI to illustrate the theoretic discussion of fi-
nancial regulation and administrative structures. Part III outlines the
administrative structure, accountability, and functioning of OSFIto
whom it is accountable and how it works paying particular attention to
its level of transparency and insulation. Part IV explains why insulation
15 Canada, Office of the Superintendent of Financial Institutions, Capital Adequacy Re-
quirement (CAR) Simpler Approaches, No A (Guideline), (Ottawa: Office of the Super-
intendent of Financial Institutions, 2007) at i.
16 Ibid.
17 The Strategic Counsel, Qualitative Research: Deposit-Taking Institutions Sector Consul-
tation, (Toronto: 2010) at 2-6, online: OSFI
18 See e.g. Matthew C Stephenson, Optimal Political Control of the Bureaucracy (2008)
107:1 Mich L Rev 53; Adrian Vermeule, Mechanisms of Democracy: Institutional Design
Writ Small (New York: Oxford University Press, 2007) [Vermeule, Mechanisms of De-
mocracy]; Mark Fenster, The Opacity of Transparency (2006) 91:3 Iowa L Rev 885;
Rachel E Barkow, Insulating Agencies: Avoiding Capture Through Institutional De-
sign (2010) 89:1 Tex L Rev 15.
406 (2012) 57:3 MCGILL LAW JOURNAL ~ REVUE DE DROIT DE MCGILL
and opacity may be positive attributes that enable OSFI to function effec-
tively. We argue that OSFI operates in a black box of sorts relative to
both the government and financial institutions. It encourages co-operation
between itself and financial institutions by protecting information about
its policy agenda and formation. We argue that OSFI profits from its in-
formational advantage by weakening the ability of regulated entities to
collude amongst themselves or rent-seek. We conclude by drawing out the
lessons for the institutional form of financial institutions regulation.
I. Financial
Characteristics
Institution Regulation: Rationales
and
Structural
Regulation governing financial markets generally seeks to address
macro- and microprudential issues.19 Macroprudential regulation focuses
on the financial system as a whole, seeking to minimize system-wide dis-
tress in order to avoid reductions in aggregate output (which may be
measured by GDP). By contrast, microprudential regulation seeks to min-
imize distress in individual institutions in order to protect depositors. In
short, macroprudential regulation focuses on common exposures across
financial systems and institutions rather than the entity-specific focus of
microprudential regulation.20
Understanding the differences between these two types of regulation
highlights the purposes of various regulators in a financial system. Ac-
cording to Herring and Carmassi, alternative models of financial supervi-
sion exist, with the two most common being a single or unified regula-
tor model on the one hand and a twin peaks or integrated model on the
other. Under the former, a countrys central bank conducts both macro-
and microprudential supervision, while under the latter an independent
authority external to the central bank is responsible for the micropruden-
tial function.21
While the United Kingdoms Financial Services Authority exemplified
the unified model (prior to recent reforms), Canadas regulatory structure
follows the twin-peaks approach, with OSFI and the Bank of Canada at
19 For a discussion, see Frank Milne, The Complexities of Financial Risk Management
and Systemic Risks [2009] 3 Bank of Canada Review 15.
20 See Claudio Borio, Towards a Macroprudential Framework for Financial Supervision
and Regulation? (2003) 49:2 CESifo Econ Stud 181 at 183-84.
21 See Richard J Herring & Jacopo Carmassi, The Structure of Cross-Sector Financial
Supervision (2008) 17:1 Financial Markets, Institutions & Instruments 51 at 57.
REGULATING FINANCIAL INSTITUTIONS 407
the centre of its financial market regulatory regime.22 OSFI is an arms-
length governmental agency that supervises individual financial institu-
tions to determine whether they are in sound financial condition and to
promote the adoption of sound risk management policies.23 By contrast,
the Bank of Canada regulates credit and currency, and controls national
monetary policy. It is also the lender of last resort and is ultimately
charged with promoting the economic and financial welfare of the country.
Despite the differences between the two models, the (obvious) over-
arching purpose of both is to ensure the stability of the financial system.
In the twin-peaks model, the focus of the independent regulatory body is
on financial institutions and the consequences of their behaviour on fi-
nancial markets.24 This body monitors compliance with standards and
guidelines and establishes new standards as necessary in order to protect
the economy at large.25 It also seeks to ensure that banks incentives are
consistent with the objective of systemic stability. We focus on this type of
institution in this article.
The interplay between the level of competition and economic stability
ultimately highlights the importance of the regulatory structure. In gen-
eral, studies suggest that Canadas financial regulation preserves compe-
tition between banks despiteor perhaps because ofgreater bank con-
centration.26 Further, Canadas combination of a competitive but concen-
trated sector is consistent both with international evidence that such sys-
22 Ibid at 57-58, 61. See also Claire Jones, Lord Turner: FSA Break-Up Necessary, Risk
Magazine (24 November 2010) online: Risk Magazine
Report on the UK reforms).
23 See OSFI Act, supra note 11.
24 See also Herring & Carmassi, supra note 21 at 57.
25 Allen & Gale report that the costs of financial instability are high (supra note 8 at 454).
They cite Glenn Hoggarth, Ricardo Reis & Victoria Saporta, who find that fiscal and
quasi-fiscal costs of banking crisis resolution averaged 16% of GDP in their sample
(Costs of Banking System Instability: Some Empirical Evidence (2002) 26:5 Journal of
Banking & Finance 825 at 831).
26 Allen & Gale, supra note 8 at 472; Klaus Schaeck, Martin Cihak & Simon Wolfe, Are
Competitive Banking Systems More Stable? (2009) 41:4 J Money, Credit, Banking 711
at 714, 719 (finding that Canada is relatively more stable than most other OECD coun-
tries); Michael D Bordo, Hugh Rockoff & Angela Redish, The U.S. Banking System
from a Northern Exposure: Stability Versus Efficiency (1994) 54:2 J of Econ Hist 325
(comparing the US and Canadian financial systems from 1920 to 1980 and finding that
Canadas branch-banking model was more competitive than the unitary banking model
that predominated in the US, which resulted in spatial monopolies. The branch model
also diversified Canadian banks across regions, lessening its susceptibility to region-
specific shocks).
408 (2012) 57:3 MCGILL LAW JOURNAL ~ REVUE DE DROIT DE MCGILL
tems are more stable27 and with the hypothesis that Canadas regulatory
structure strikes an appropriate balance between reducing risks of collu-
sion and preserving ease of monitoring for regulators.28 This literature
emphasizes the importance of regulatory structure in maintaining eco-
nomic stability.29 We turn now to examine regulatory structure from the
standpoint of institutional design before analyzing OSFI specifically.
II. Optimizing Administrative Design
Against this backdrop of the structure and purposes of financial insti-
tution regulation is the basic point that the regulatory structure, especial-
ly under the twin-peaks model sketched above, interfaces with adminis-
trative law. Administrative law is concerned in part with the appropriate
institutional structures by which government makes and enforces regula-
tions. These institutional structures include not only large-scale institu-
tional choices between governments, courts, and markets but also related
issues such as the degree of insulation of administrative bodies from the
legislature, the level of transparency, the nature of accountability, and the
processes for public participation.30 These structures may affect the power
of interest groups, the nature of deliberation, and the rationality of the
27 Alexandra Lai, Adi Mordel & Sheisha Kulkarni, Submission to the U.K. Independent Com-
mission of Banking (Bank of Canada, 2010), online: Independent Commission of Banking
Demirg-Kunt & Ross Levine, Bank Concentration, Competition, and Crises: First Re-
sults (2006) 30:5 Journal of Banking Finance 1581.
28 International Monetary Fund, Canada: Financial Stability AssessmentUpdate
(2008) IMF Country Report No 08/59 at 6-7, 35.
29 A further point to be drawn from the literature is that there is a public good character
to financial institution regulation. Ensuring the stability of financial institutions pre-
serves the integrity of the economy and its payment system. Further, there is an inter-
play between the financial institution regulator and other bodies, which provides checks
on the ability of the financial regulator to act in an unfettered manner. The central
banks role as the lender of last resort is one such example. Agencies that administer
deposit insurance (in Canada, the Canadian Deposit Insurance Corporation) are anoth-
er. These bodies contribute to systemic stability by backstopping risk taking by the fi-
nancial institution regulator. The incentives are perverse; knowing that such institu-
tions exist may render financial regulators more inclined to take on risks. Analogous
processes exist in other jurisdictions, namely the United States. This suggests that
there is some other factor(s) that is important to analyze in the Canadian context.
(Thanks to Grant Bishop for making this point.)
30 See Vermeule, Mechanisms of Democracy, supra note 18 (discussing the effects of dif-
ferent institutional features on legislative and administrative decision making); David
Markell, Slack in the Administrative State and Its Implications for Governance: The
Issue of Accountability (2005) 84:1 Or L Rev 1 (discussing the debate over the structure
of administrative decision making).
REGULATING FINANCIAL INSTITUTIONS 409
administrative decisions. Choosing among these structures requires tak-
ing account of their costs and benefits.31
Much of the administrative state operates through delegation: legisla-
tors delegate various powers to make rules and/or decisions to other bod-
ies such as agencies, boards, or tribunals. Powers may be delegated for a
range of reasons that include the delegates possession of greater exper-
tise, time, or information than the legislator, which may render the dele-
gate better able to make the decision at issue.32 However, this delegation
gives rise to two types of principal-agent problems. First, the legislators
(the principal) delegate decision-making power to another body (the
agent) which may or may not seek or be able to fulfill the legislators in-
terests adequately.33 Second, there is a deeper principal-agent problem, as
the ultimate principals (the citizens of the country or jurisdiction) have
granted authority to their agents (the legislators) the power to make deci-
sions on their behalf.
In part, the nature of and solutions for this principal-agent problem
depend on the underlying theory of the administrative state.34 On one
view, legislators and regulators are attempting to determine the optimal
outcome for society. This optimal outcome could, for example, be viewed
as maximizing welfare or satisfying the preferences of the majority of the
population. Pluralist theories identify the aggregation of the existing in-
terests or preferences of different groups within the relevant society. Civic
republican theories, by contrast, argue that there cannot be a mere aggre-
gation of preferences but that instead, there must be some informed de-
liberation behind policy decisions, as such deliberation helps inform the
decision and shape the preferences themselves.
On these theories, the principal-agent problem arises because the
agent may either incorrectly determine the principals preferences or seek
to implement its own preferences regarding the optimal welfare-
enhancing policy. Institutional design to address the principal-agent prob-
lem is concerned with ensuring that the delegated party is in a position to
obtain the appropriate information or to engage in an appropriate process
to make the decision that is in the public interest. The agent wishes to
31 See Vermeule, Mechanisms of Democracy, supra note 18 at 10 ff.
32 Andrew Green, Regulations and Rule-Making: The Dilemma of Delegation in Colleen
M Flood & Lorne Sossin, eds, Administrative Law in Context (Toronto: Edmond Mont-
gomery, 2008) 337 at 339-41.
33 Mathew D McCubbins, Roger G Noll & Barry R Weingast, Administrative Procedures
as Instruments of Political Control (1987) 3:2 JL Econ & Org 243 at 246-48.
34 For an overview of different theories of regulation, see Steven P Croley, Theories of
Regulation: Incorporating the Administrative Process (1998) 98:1 Colum L Rev 1.
410 (2012) 57:3 MCGILL LAW JOURNAL ~ REVUE DE DROIT DE MCGILL
make the decision that is in the interests of its principal and the design
question concerns how best to help it make this decision.
Alternatively, public choice theory holds that legislators and regula-
tors make decisions in their own interests. They may, for example, use
their powers to favour well-organized groups in exchange for benefits such
as electoral funding or future job opportunities. They may also simply at-
tempt to increase the scope of their power or the size of their budget. The
principal-agent problem under this theory is not based on the possibility
that, despite the agents desire to make a decision in the principals inter-
est, it fails to do so. Instead, the problem rests on the desire of the agent
to make decisions in its own interest rather than those of the principal.
A multitude of institutional design features may be used to address
the principal-agent problem in light of these different theories of the ad-
ministrative state. Vermeule argues that each design feature has costs
and benefits in terms of various democratic values such as impartiality,
accountability, and deliberation, and that the goal of institutional design
should be to optimize across these values.35 We focus on the effect and in-
teractions between two design features central to financial institution
regulation: insulation and transparency.
A. Insulation
The appropriate degree of insulation of executive bodies from the leg-
islature has been a long-standing subject of debate in administrative
law.36 By insulation, we mean the degree to which the executive body is
separated from the legislature. It is not a threshold variable (that is, insu-
lated or not insulated) but a sliding scale. At one end of the scale are deci-
sions made purely by legislators. Such decisions obviously include the en-
actment of statutes. At the other end of the spectrum are decisions made
by bodies that are removed from legislative control. Some administrative
tribunals, for example, may possess a level of independence from the leg-
islature and other executive bodies that approaches that of the judiciary.37
Indicia of insulation include fixed terms of appointment for the members
of the administrative bodies, limits on the power of the legislative or
35 Vermeule, Mechanisms of Democracy, supra note 18. There is a significant literature on
the trade-offs inherent in administrative design choices. See e.g. Richard B Stewart,
Administrative Law in the Twenty-First Century (2003) 78:2 NYUL Rev 437;
McCubbins, Noll & Weingast, supra note 34; David B Spence, Managing Delegation Ex
Ante: Using Law to Steer Administrative Agencies (1999) 28:2 J Legal Stud 413.
36 See Markell, supra note 30; Stephenson, supra note 18 at 54.
37 See Laverne Jacobs, Independence, Impartiality, and Bias in Flood & Sossin, supra
note 32, 139 at 153-55.
REGULATING FINANCIAL INSTITUTIONS 411
elected officials to remove the members of the regulatory body, lack of po-
litical control over the bodys structure and procedures, lack of political
review of the bodys decisions, and absence of a power for elected officials
to issue directives to the body.38
Insulation or independence of an executive body from the legislature
does not mean that the body is not accountable to the legislature or elect-
ed officials. First, as the legislature establishes or authorizes the continu-
ance of these bodies, elected officials are accountable to the electorate for
the decisions made by these bodies. The public may have difficulty moni-
toring exactly how the decisions of these bodies relate to legislative con-
trol, but there remains an element of accountability.39 Second, there are
varying mechanisms by which legislatures can control all executive bod-
ies. The most extreme control mechanism is, of course, abolishing execu-
tive bodies that make decisions that conflict with the legislators prefer-
ences. Less extreme control mechanisms include limiting the delegated
bodys budget, limiting the power to remove and appoint members, and
limiting the power to make regulations that overturn the bodys deci-
sions.40 There is also legislative review of decisions of these executive bod-
ies, though lack of constant supervision is the whole point of delegation
and undermines its benefits.41
The value of insulation (or independence) is that the delegated body
has the ability to use its expertise, time, and information to make a deci-
sion, to a greater or lesser extent, free of political interference.42 Such in-
terference may be negative to the extent that legislators are attempting to
alter a decision for self-interested reasons as opposed to welfare-
maximizing reasons. For example, legislators may serve their own inter-
ests by providing benefits to concentrated interest groupsa type of re-
sponsiveness that Vermeule labels bad accountability.43
38 Stephenson, supra note 18 at 68.
39 Ibid at 75-76, 80 (elected officials retain accountability for delegated decisions because
they have the ability to control the bodies or their decisions; monitoring costs should be
endogenous to the delegation). But see Justin Fox & Stuart V Jordan, Delegation and
Accountability (2009), online: Social Science Research Network
tion, but that under certain conditions the gains from expertise may outweigh the costs
in terms of democratic accountability).
40 See Jacobs, supra note 38; Stephenson, supra note 18 at 68; Green, supra note 32 at
343-56 (discussing mechanisms of controlling executive bodies).
41 Ibid at 345-46; Vermeule, Mechanisms of Democracy, supra note 18 at 81-82.
42 Insulation can have other values such as allowing faster decisions during a crisis.
43 Ibid at 184.
412 (2012) 57:3 MCGILL LAW JOURNAL ~ REVUE DE DROIT DE MCGILL
The level of insulation can be thought of in terms of the cost to the leg-
islature of altering the decisions of the delegated body; the higher the cost
of altering the decision, the more insulated the body.44 Control is costly
both directly (for instance, the costs of drafting and enacting regulations
to reverse a decision of an executive body) and indirectly (given the oppor-
tunity costs of the time used to monitor and control delegated decisions).
As Stephenson argues, the costs of control by the legislature tend to be in-
creasing in the sense that the larger the change needed to be made by the
legislature, the higher the costs of change, and there are increasing mar-
ginal costs of change.45
The costs of insulation itself can be thought of in terms of the differ-
ence between the decision of the delegated body and the decision of either
a fully informed legislature or informed voters.46 Insulation in such a case
may reduce good accountabilitythat of a well-meaning legislature at-
tempting to control the outcome of delegated decisions.47 Such a difference
may arise for a number of reasons. First, the delegated body (e.g., a regu-
lator of financial institutions) may be acting in its own self-interest, at-
tempting to obtain benefits from concentrated interests for non-welfare-
maximizing decisions. Second, the delegated body may be seeking to de-
cide in accordance with the preferences of the legislature but may be mis-
taken concerning those preferences. Third, the delegated body may wish
to make a decision that maximizes social welfare but seeks to implement
its own preferences in this regard (as opposed to those of the legislature).48
Finally, there may be something about the legislative process, such as
its deliberative nature, that shapes the preferences of the legislature as
distinct from the delegated body.49 Vermeule notes that there can be
44 Stephenson, supra note 18 at 69.
45 Ibid. In Stephensons model, the president (the executive) seeks to change the prefer-
ences of the bureaucracy, and the costs of control increase with the desired size of the
preference change.
46 Ibid. This difference may be termed slack: see Markell, supra note 30.
47 See Vermeule, Mechanisms of Democracy, supra note 18 at 81-2 (discussing the connec-
tion between delegation and accountability).
48 Stephenson models the legislative choice concerning a bureaucratic bodys degree of in-
sulation from control. Examining the difference between the bureaucratic decision and
the majoritarian decision, he finds that some insulation is always desirable. The benefit
derives from the reduction in the variance in policy outcomes. The degree of insulation,
therefore, relates to such factors as the responsiveness of the elected officials to majori-
tarian preferences and voter preference instability (Stephenson, supra note 18).
49 But see Vermeule, Mechanisms of Democracy, supra note 18 at 80-1 (discussing the re-
lationship between delegation and deliberation and noting that legislators may ensure
that delegated decisions have good deliberative features).
REGULATING FINANCIAL INSTITUTIONS 413
good deliberation and bad deliberation.50 Good deliberation is in-
formed argumentation that seeks to understand the issue and possibly to
come to a common understanding of what is at stake and the optimal de-
cision to be reached. Bad deliberation involves political posturing and
hardening of positions to score political advantage. To the extent that in-
sulation reduces good deliberation, it is costly; conversely, to the extent
that it avoids bad deliberation, it is beneficial.
B. Transparency and Opacity
Transparency is touted as a central feature of democratic governance.
It is argued, for example, that transparency will reduce the power of in-
terest groups, allow monitoring of decisions by the public, and increase de-
liberation in policy decisions.51 However, just as insulation is not an unal-
loyed good in a democracy, so too transparency has both benefits and
costs.52 For any particular policy, such as financial regulation, these bene-
fits and costs must be traded off to determine optimal institutional design.
We focus on transparency in the sense of an open process for decision
making in either policy or adjudication. Again, as with insulation, it is not
an all or nothing concept; transparency in decision making comes in de-
grees. Transparency includes the information provided to the public by
the decision-making body about its proposed decision and its implications,
the ability of the public or regulated parties to make oral or written sub-
missions concerning the decision, the information provided by the deci-
sion-making body about the decision process, and the extent to which the
resulting decision and its bases are made available to parties outside the
decision-making body. Thus, the indicia of transparency relate solely to
the decision-making processes (i.e., which parties contribute to the mak-
ing of a decision or set of decisions), including decisions about the sub-
stantive content of legislation.
The benefits of transparency centre on its connection to good and bad
deliberation and accountability, and the impact of these factors on the ra-
tionality of the ultimate decision as well as the connection of the decision
to the preferences of either the legislators or the public.53 In terms of ac-
countability, transparency lowers the monitoring costs for parties external
to the decision-making process. It therefore allows voters and legislators
to determine more easily if the party that was delegated the power is act-
50 Ibid at 11.
51 See Fenster, supra note 18 at 897, 899.
52 See e.g. Vermeule, Mechanisms of Democracy, supra note 18 at 11; Fenster, supra note
18.
53 Ibid at 186.
414 (2012) 57:3 MCGILL LAW JOURNAL ~ REVUE DE DROIT DE MCGILL
ing in accordance with its own interests or some version of social welfare
that does not accord with that of legislators or voters. Such good account-
ability lowers the risks and costs of capture and may shift the decision of
the decision-making body towards that of the principal.
Transparency may also enhance the rationality of the decision by
providing opportunities for various parties to provide the decision maker
with information. For example, a process that relies on a broad notice and
comment process or requires public hearings provides interested parties
with the ability to provide such information. The decision maker may not
decide in accordance with that information or even use the information in
all cases, but the information may change decisions on the margin. It may
also enhance good deliberation both because the transparency may pro-
vide the decision maker with information and because it allows parties to
debate the merits of any decision more openly.54
However, transparency also has potential costs because it might in-
crease both bad deliberation and bad accountability. As noted above, bad
deliberation involves political posturing and inflexibility in policy posi-
tions.55 Bad accountability arises because transparency allows not only
voters and legislators but also concentrated interests to monitor the ac-
tions of the decision maker. To the extent that the decision makers have
entered in non-social-welfare enhancing bargains with interest groups,
transparency allows interest groups to determine if the decision maker is
actually following through on the bargain. The more transparent the pro-
cess, the more difficult it is for a member of the decision-making body to
deceive the interest group about whether she held her end of the deal.56
The desirability and degree of transparency, therefore, will depend on
the balance of these costs and benefits in the particular policy context. To
the extent that there is a greater concern with enhancing good accounta-
bility or deliberation (because, for example, there may be less concern
about capture of legislators than members of the executive), transparency
should be enhanced. If there is more of a concern about bad accountability
of decision makers, some opacity may be warranted.
54 See Vermeule, Mechanisms of Democracy, supra note 18 at 180-81 (discussing the con-
nection between transparency, deliberation, and other values); contra Fenster, supra
note 18 at 908-909.
55 Ibid at 196.
56 Ibid at 197.
REGULATING FINANCIAL INSTITUTIONS 415
C. Combining Insulation and Opacity
Both transparency and insulation have costs and benefits that relate
to the connection between the ultimate decision and the preferences of the
relevant principal (either the voter or the legislator). They also connect to
the rationality of the decision, where rationality requires that the decision
be based on full information, with appropriate levels of expertise, and
with proper deliberation. The appropriate administrative structure in
each case will depend on an understanding of the nature of the underlying
principal-agent problem and the concerns that arise about the ability and
willingness of the delegated decision maker to enhance social welfare.
However, there is more to understanding institutional design. Differ-
ent institutional features, such as transparency and insulation, not only
give rise to particular costs and benefits, but they also interact with each
other.57 There can be various combinations of mechanisms. Figure 1, for
example, illustrates the broad possibilities relating to transparency and
insulation. Bodies exercising delegated powers that lie in Quadrant I are
both more transparent and insulated from political control than bodies in
the other quadrants. An administrative tribunal modeled on a court may
lie in this quadrantit is designed so that political control is weak (costly)
such as where an apolitical process appoints members for a lengthy, fixed
period of time. To the extent that the materials and hearings (and possi-
bly even deliberations) are open to the public, such tribunals can be very
transparent.
57 Robert E Goodin, Handy Gadgets for Institutional Design (2009) 18:1 The Good Socie-
ty 12 at 12; Adrian Vermeule, The Interaction of Democratic Mechanisms (2009) 18:1
The Good Society 21.
416 (2012) 57:3 MCGILL LAW JOURNAL ~ REVUE DE DROIT DE MCGILL
Figure 1: Transparency and Insulation in Institutional Design
Insulated
Transparent
I
III
II
IV
Opaque
Legislative
REGULATING FINANCIAL INSTITUTIONS 417
Decision-making bodies in Quadrant II are more insulated than other
bodies but are also more opaque. There may be little opportunity for pub-
lic, or interest, groups to be involved in the decision-making process. Even
the decisions themselves may be kept from public view. An example of a
body in Quadrant II is one that addresses issues requiring a high degree
of expertise as well as privacy such as a body determining mental health
issues or national security issues. As discussed in the next section, we
place OSFI in Quadrant II.
Bodies in Quadrants III and IV are subject to greater political control
over the ultimate decision or process. Bodies in Quadrant III are subject
to political control and are transparent. Such bodies could include com-
missions composed of elected officials, which typically hold open hearings.
Quadrant IV encompasses bodies that are subject to political control but
are more opaque than other bodies. Such bodies may be responsible for
highly political decisions that at the same time require secrecy. A tribunal
or commission making national security decisions is an example of such a
tribunal.
There are, therefore, four combinations relating to insulation and
transparency, with varying degrees of each characteristic (that is, the bod-
ies can lie anywhere in the Quadrants and they need not be at the ex-
tremes). The design of the particular institution will depend on how the
two features interact.58 They may be complements in the sense that an in-
crease in one feature (such as transparency) increases the benefits, or de-
creases the costs, of the other (insulation). Alternatively, transparency
and insulation may be substitutes in the sense that an increase in one de-
creases the benefits, or increases the costs, of the other.
To understand the relationship between transparency and insulation,
it is useful to consider two (related) aspects of policy decisionsthe extent
to which the decision accords with the preferences of the principal
(whether it be voters or legislators) and the rationality of the decision
(such as whether the decision accurately rests on and incorporates the
relevant costs and benefits of the policy options). The interaction of trans-
parency and insulation will be important to the extent that there is a con-
cern about matching the preferences of the decision maker with those of
the principal. If good accountability is important and possible, an increase
in transparency should reduce the costs of insulation in the form of
slackor the difference between the decision and preferencesby re-
ducing the monitoring and, therefore, the control costs. It requires that
the legislators or the public be both willing to obtain the information and
58 Ibid at 21-22 (discussing the possibility that mechanisms may be complements or sub-
stitutes).
418 (2012) 57:3 MCGILL LAW JOURNAL ~ REVUE DE DROIT DE MCGILL
able to understand it.59 However, if there is a concern about bad account-
ability in the form of interest group monitoring of bargains (to the detri-
ment of social welfare), transparency and insulation would be substi-
tutesan increase in transparency would increase the costs of insulation
(or the slack available given the costs of control). The nature of the rela-
tionship will then depend on the context and the nature of the concern
about accountability.
This same contextual concern holds for the impact of transparency
and insulation on the rationality of the decision. Transparency and insu-
lation are complements to the extent that transparency results in greater
information for the decision maker and enhances good deliberation about
the decision. An increase in transparency would increase the benefits of
insulation in that it enhances the ability of the decision maker to utilize
its expertise. However, to the extent that there is a concern about bad de-
liberation, transparency and insulation may be substitutes. The members
of the decision-making body may engage in political posturing or they may
become inflexible in their positions and rely less on their expertise.
Transparency would, therefore, decrease the benefits of insulation. Fur-
thermore, transparency and insulation may be substitutes if transparency
allows regulated parties, or others, to use their ability to provide infor-
mation to control the outcomes of the decision-making body.
The nature of the relationship between transparency and insulation
will, therefore, depend on the nature of the policy context. They may in
certain cases be complements pushing towards the desirability of creating
bodies that fall within Quadrants I or IV. Alternatively, they may be sub-
stitutes pointing towards creating a body lying within Quadrants II or III.
Of course, an administrative body need not lie solely within one quadrant
for all issues or even for all aspects of a particular issue. For example, in
discussing transparency, Vermeule proposes altering the budgetary pro-
cess of the US federal government to delay disclosure so that good delib-
eration is enhanced and bad accountability is reduced.60 Further research
is needed on how to design combinations within and across specific ad-
ministrative bodies to take advantage of transparency and insulation as
complements and substitutes.
III. Regulating Financial Institutions
The previous Part focused on general characteristics of administrative
agencies in terms of the extent of their transparency and insulation. This
59 See Fenster, supra note 18 at 940-41.
60 Vermeule, Mechanisms of Democracy, supra note 18 at 186.
REGULATING FINANCIAL INSTITUTIONS 419
Part discusses the choices concerning transparency and insulation that
underlie the design and structure of regulators of financial institutions.
As we noted in the introduction, we have chosen to study the Canadian
OSFI because it is seen as being an element in the relative success of Ca-
nadian banks during the recent financial crisis. This analysis forms the
basis for the discussion in Part IV, which points to aspects of OSFIs regu-
latory structure that may serve as a model for other countries to emulate.
OSFI is both insulated and opaque. Recall that indicia of insulation
include considerations relating to the extent to which decision making in
the administrative body is removed from legislative control or control by
elected officials. It depends on such factors as: the process for appointing
and removing the superintendent; the process for assessing the compe-
tence and performance of the administrative body; the degree of political
control over OSFIs structure and procedures; and the process for making
binding rules (this latter indicium overlaps with those that relate to
transparency, as we discuss below).
The minister of finance preside[s] over and is responsible [for] the
office.61 However, the locus of power in OSFI rests in one individual: the
superintendent of financial institutions. She has the powers, duties, and
functions assigned to her by various named statutes, including the Office
of the Superintendent of Financial Institutions Act as well as others like
the Bank Act and the Insurance Companies Act. The superintendent is
appointed by the Governor-in-Council (i.e., the federal Cabinet) to hold of-
fice for seven years. She can be reappointed for a further term and can on-
ly be removed for cause.62 The superintendent may appoint officers (called
deputy superintendents) and employees but all must act under her in-
structions.63
In terms of accountability provisions, the superintendent is required
to report to the Minister [of Finance] from time to time on all matters
connected with the administration of these acts.64 As noted above, the
OSFI Act specifically notes that the minister is responsible for the office.
Further, the superintendent is required to submit an annual report to
Parliament showing the operations of the office for that year.65 The lines
of accountability therefore flow through the minister to Parliament. This
accountability does not encompass budget control. Under the act, the su-
perintendent recovers the costs of her activities mainly through assess-
61 OSFI Act, supra note 11, ss 3, 4(1).
62 Ibid, s 5. The current superintendent is Julie Dickson.
63 Ibid, ss 8, 9, 11.
64 Ibid, s 6.
65 Ibid, s 40.
420 (2012) 57:3 MCGILL LAW JOURNAL ~ REVUE DE DROIT DE MCGILL
ments on regulated parties with small amounts coming from some user
pay programs.66 In addition, the minister is able to override OSFI deci-
sions in some instances.
In general, the Governor-in-Council has authority under the various
statutes that apply to OSFI to make regulations and the superintendent
is responsible for administering these regulations. The OSFI Act also con-
tains a broad provision that permits Cabinet to make regulations, pre-
scribing anything that is required or authorized by this Act to be pre-
scribed; and … prescribing the way in which anything that is required or
authorized by this Act to be prescribed shall be determined.67 In terms of
rule making, the federal government has put in place requirements for
prepublication of proposed regulations along with a form of cost-benefit
analysis. The guidelines also mandate a public consultation process. OSFI
itself may … provide input to the Department of Finance with respect to
regulations that support the various Acts. In some cases, OSFI will be the
sponsor of a regulation and will work with the Department of Finance and
Justice to complete the process of developing and issuing of a regula-
tion.68 Thus, communication between OSFI and federal ministries occurs
as a matter of practice and there is an open process for making regula-
tions.
While Cabinet may make regulations that OSFI must implement, the
enabling statute has left considerable room for OSFI to make policy
through various forms of guidance. These include: guidelines describing
OSFIs view of such matters as risk management; advisories setting out
OSFIs interpretation of different requirements; and guidelines and rul-
ings on specific matters.69 In addition, in its supervisory role, OSFI uses a
principle-based as opposed to a rules-based approach. This approach re-
lies heavily on the judgment of its supervisory staff.70
The guidelines can relate to crucial areas of OSFIs mandate. For ex-
ample, OSFI relies on a guideline (360 pages in length) to establish capi-
66 Ibid, ss 23, 23.1.
67 Ibid, s 38.
68 Email from Correspondence Officer, OSFI (4 October 2011) [Email correspondence].
69 Canada, Auditor General, 2010 Fall Report of the Auditor General of Canada (Ottawa:
Office of the Auditor General, 2010), ch 5, online: OAG
ditor General Report]. The report states that OSFI has issued 28 guidelines, 57 adviso-
ries and 23 rulings (ibid at 11).
70 Ibid, ch 5 at 19: [The Auditor-General] found that OSFI followed its supervisory ap-
proach in assessing the activities of the large banks for the year.
REGULATING FINANCIAL INSTITUTIONS 421
tal adequacy requirements.71 While the formulation of guidelines may fol-
low from participation in discussion around international regulatory
standards, such as the Basel III process, no formal consultation process is
prescribed for making guidelines. OSFI describes its process as follows:
The development process usually involves a research and drafting
phase followed by informal industry consultations with knowledgea-
ble parties where the draft is further revised and then approved for
posting in draft on our Web site for formal consultations. Comments
are received and analyzed, after which appropriate amendments are
made and a final version of the guidance is issued.72
OSFI seems, therefore, to have some interest in receiving comments from
the public or relevant stakeholders, but this is not a mandatory aspect of
the process.
OSFIs guideline and policy-making processes are opaque relative to
other financial regulatory bodies, both in the United States and in Cana-
da. The legislation contains no procedures that must accompany OSFIs
development of guidelines in marked contrast to rule-making procedures
that constrain other administrative agencies in the financial sector. For
the Ontario Securities Commission, for example, the rule-making process
requires proposed rules to be published for comment for ninety days;73 and
if there are material changes that are made to the proposed rule thereaf-
ter, the rule must be republished for a further reasonable comment pe-
riod as the Commission deems appropriate.74 The proposed rule must be
accompanied by a cost-benefit analysis. Once all comment periods have
expired, the proposed rule must be sent to the provincial minister of fi-
nance for approval.75 This process would fall on the transparent side of the
scale in Figure 1 above.
OSFIs process, on the other hand, requires no cost-benefit analysis
and no public comment periods for guidelines.76 There is only limited abil-
ity for members of the public to provide input to the rule-making process
through its informal notice and comment process while financial institu-
71 Canada, Officer of the Superintendent of Financial Institutions, Capital Adequacy Re-
quirements, No A-1, (Guideline) (Ottawa: Officer of the Superintendent of Financial In-
stitutions in Canada, 2007).
72 Email correspondence, supra note 71.
73 Securities Act, RSO 1990, c S.5, s 143.2(3).
74 Ibid, s 143.2(9).
75 Securities Act, RSO 1990, c S.5, ss 143.2, 143.3. The rule-making process in Ontario was
preceded by a policy-making phase that was challenged before the courts. See Ainsley
Financial Corp v Ontario Securities Commission (1994), 21 OR (3d) 104, 121 DLR (4th)
79 (CA).
76 Cabinet Directive applies to regulation making (supra note 15).
422 (2012) 57:3 MCGILL LAW JOURNAL ~ REVUE DE DROIT DE MCGILL
tions have greater input at least at the initial policy-formation stage.77
What is involved in these consultations is not clear, although OSFI like-
ly chooses with whom it will consult. The Auditor General notes that
[f]requent and open dialogue between OSFI and the banks and OSFIs
willingness to consider the banks views were also credited as promoting a
positive working relationship.78
Interestingly, OSFI extends this use of relatively opaque processes in
its enforcement role. The various statutes that OSFI administers provide
tools that can be used where OSFI determines that a financial institution
under its supervision is following unsafe or unsound practices. OSFIs
formal powers include: the ability to issue directions of compliance, dis-
qualification, or removal of directors or senior officers; taking control of a
financial institution; and a request for a winding up of a financial institu-
tion.79 In addition, OSFI has the power to issue administrative monetary
penalties under the OSFI Act. However, OSFI tends to employ a hands-on
style of negotiation and works with these institutions in order to ad-
dress areas of concern before they lead to a situation where more formal
mechanisms can be used.80 The enforcement process is thus based, to
some extent, on informal, nonpublic negotiations rather than a more
transparent, formal, penalty-based approach.
77 OSFI has indicated that it posts its proposed guidelines for comment. Email corre-
spondence, supra note 71, which states:
[g]uidance may be developed as a result of legislative changes or due to various envi-
ronmental factors that indicate new or revised guidance is needed. The development
process usually involves a research and drafting phase followed by informal industry
consultations with knowledgeable parties where the draft is further revised and then
approved for posting in draft on our Web site for formal consultations. Comments are
received and analysed, after which appropriate amendments are made and a final
version of the guidance is issued.
78 Auditor General, supra note 72 at 20. The OSFI has commissioned confidential consul-
tations with deposit-taking institutions: see The Strategic Counsel, supra note 17 at 8.
79 See Bank Act, supra note 12, Part XIII, ss 960 (directions of compliance), 620 (supervi-
sory intervention), 963 (removal of directors or senior officers), 621 (winding-up); Insur-
ance Companies Act, supra note 12, Part XV; Trust and Loan Companies Act, supra
note 12, Part XII, s 509.1 (disqualification from election or appointment), 509.2 (remov-
al of directors or senior officers), 510 (supervisory intervention), 515.1 (winding-up); Co-
operative Credit Associations Act, supra note 12, Part XIII, s 439 (directions of compli-
ance), 441.1 (disqualification from election or appointment), s 441.01 (removal of direc-
tors or senior officers).
80 Email correspondence, supra note 71.
REGULATING FINANCIAL INSTITUTIONS 423
IV. Policy Implications
In the previous Part, we set forth the way in which rules and policies
relating to financial institutions are made and OSFIs roles in these pro-
cesses. In this Part, we explore why OSFI appears to have been an effec-
tive regulator despite its insulated and opaque structure.
There are at least two possible stories.81 The first revolves around cap-
ture. Legislators wish to retain the high tax revenues and the primary
and secondary business activity from the banks (or merely succumb to
lobbying). The government therefore established a structure that enables
banks to operate without government or public intervention. The minister
of finance has oversight over the body and Cabinet can make regulations.
However, these powers are rarely used and are weak or costly (such as
the difficulty in removing a supervisor who is acting against the prefer-
ences of the government). OSFI operates as a true arms-length body with
a substantial portion of its administration accomplished though guide-
lines and negotiations with regulated entities.
This story is not entirely satisfactory. It is not clear why the govern-
ment would provide banks with such freedom. OSFI has an explicit public
interest role to play in protecting the rights and interests of depositors,
policyholders, and creditors of financial institutions.82 It is also ultimately
accountable to an elected official, who will presumably have an interest in
maintaining the integrity of the body he oversees if only to ensure that
the voting public will retain its confidence in him and the government he
represents. Permitting the banks such freedom may increase tax revenue
and economic activity but would likely undermine this objective. The story
would, therefore, likely have to be more than one of a desire to foster
growth to increase tax revenue and economic activity.
Furthermore, even if the federal government supported collusion, why
would it not provide itself with more direct oversight and control over
OSFI? For example, the seven-year term of the superintendent means
that one federal government is able to impose the preferences and direc-
tion of one superintendent on another (governments change at least every
five years). The government could have provided for shorter terms or
some bases on which the superintendent could be removed other than for
cause, expanding its ability to exercise control. Even if the government
wished to allow for collusion, it would seem advantageous to retain
81 As noted in the introduction, there are a number of other factors than the regulatory
structure that have been advanced regarding the relative success of Canadian banks at
weathering the economic crisis.
82 OSFI Act, supra note 11, s 4(3).
424 (2012) 57:3 MCGILL LAW JOURNAL ~ REVUE DE DROIT DE MCGILL
stronger forms of checks by the party in power. OSFIs highly insulated
and opaque structure does not provide for such controls.83
A second possible story for the existence of this structure rests in its
benefits. The benefits arise because of the nature of the regulatory subject
matter (i.e., financial institutions) and the nature of the financial sector in
Canada. In substantive terms, banking regulation can be highly dense
and technical. (The capital adequacy guidelines referred to above are a
case in point.) Rationality in regulation and related decision making re-
quires the exercise of considerable expertise to understand banks risk
management models and the application of policy to these models. In ad-
dition, stability of policy in this area is necessary in order to provide con-
fidence to the financial markets and to enable regulated entities, and
those who seek credit from regulated entities, to structure their business-
es efficiently.
As an insulated body with an expert superintendent and staff, OSFI is
effectively removed from legislative processes dominated by partisan poli-
tics and majoritarian preferences. Insulation ensures that the policy-
making process will be undertaken by those with expertise. It also allows
policy-making to be buffered from shifts in public opinion. This would not
be the case if elected officials were able to participate more broadly in the
policy-making process.
In this sense, the value of good accountability (i.e., a well-intentioned
legislature attempting to control the outcome of delegated decisions) is
low. These same factors also feed into the benefits of opacity in this area.
The fact that the process is opaque and does not allow for considerable
public input is beneficial to the extent that there is no need or desire to
closely tie banking regulation and policy to shifts in public preferences.
Further, given that financial institution stability and market confidence
are important, there are risks to regulation from bad deliberationthat
is, political posturing (such as vilifying the banks when it is politically ad-
vantageous to do so). The opacity and insulation of the process reduce the
negative impact of bad deliberation.
An insulated and opaque OSFI can also be effective because of the oli-
gopolistic nature of the banking sector, which is dominated by five big
banks. These banks engage in considerable lobbying of politicians who are
responsive to their concerns because they comprise approximately 85 per-
83 Further, at least in some cases, it appears that the minister does indeed take his super-
visory role seriously, as evidenced most recently with his rejection of the global bank tax
and support for the contingent capital proposal, regardless of OSFIs stance on these is-
sues: see Peter Foster, Flahertys Silver Bullet, Comment, The Financial Post (10
June 2010) online: The Financial Post
REGULATING FINANCIAL INSTITUTIONS 425
cent of the banking sector.84 The risks of bad accountability of OSFI to pol-
iticians as a result of such lobbying would be relatively high if the costs of
control were lower. But in an oligopoly, opacity in the policy-making pro-
cess is not as problematic. The existence of fewer banks reduces infor-
mation costs for the regulator, allowing greater control through less for-
mal processes.
In addition, an opaque structure is necessary and even desirable to
encourage co-operation by financial institutions. Where OSFIs intentions,
process, and exercise of power are not fully clear to banks, it is more diffi-
cult for banks to manage their regulator. OSFI profits from its infor-
mational advantage: it can effectively prevent banks from gaining infor-
mation about its views on particular conduct or policy and also effectively
prevent banks from colluding. OSFI may be able to keep the banks at bay
by providing them with little information about its intentions.
The insulated and opaque nature of the processes used by OSFI can be
contrasted with the more transparent and public approach that the gov-
ernment would need to take if it attempted to put in place regulations to
govern the banks at every turn (as in the area of capital adequacy re-
quirements). Such processes would not only be time-consuming but also
concluded only after legislative debate, stakeholder input and bank lobby-
ing. Highly sensitive policy, and the process of making such policy, would
be more closely tied to shifts in public preferences and to the effects of
lobbying of elected officials by the banks, potentially reducing the ration-
ality and effectiveness of the policy.
For their part, financial institutions do not want to fall out of favour
with OSFI. In particular, the five big banks comply not because they fear
enforcement but likely because of possible reputational effects among
their peers and the public at large of being sanctioned by OSFI (a fact that
could quickly become publicized). They also want to stave off mandatory
regulation, which would be more public. For this reason, banks voluntari-
ly comply with requests for information, meetings, transaction changes,
and internal policy amendments.85 The benefits of a formal legal system
84 Authors calculations using OSFI monthly banking data, available at Financial Data –
Banks Office of the Superintendent of Financial Institutions Canada, online: OSFI
Nova Scotia, Bank of Montreal and CIBC ) held 86% of the assets of all banks in Cana-
da (which includes Total Domestic Banks, Total Foreign Bank Subsidiaries and To-
tal Foreign Bank Branches).
85 The deferential view that financial institutions hold towards OSFI arose in our discus-
sions with Canadian banks. The interviews took place in May and June of 2009. One fi-
nancial institution stated that OSFI has voluminous information about all banks. It is,
therefore, difficult and unwise to challenge OSFI because it can penalize banks by, for
example, closing loopholes that may be open to them.
426 (2012) 57:3 MCGILL LAW JOURNAL ~ REVUE DE DROIT DE MCGILL
are reduced in part because of these broader forms of accountability or
control while the costs are potentially high.
Insulation and opacity, therefore, work as complements in this area.
In terms of preferences underlying policy, as opacity increases there is an
increase in the benefits of insulationstability in policy increases and the
influence of majoritarian preferences shrinks (since the value of good ac-
countability is low). At the same time, in terms of rationality of policy, an
increase in opacity reduces the risks of bad accountability from elected of-
ficials without significantly reducing the harm from less information (as
the banking sector is so concentrated).
A final question concerns regulatory capture of individuals working
within OSFI. If an institution like OSFI is opaque and insulated, what, if
any, controls exist to prevent capture of individual regulators such as the
superintendent or her deputies if they hope to secure employment oppor-
tunities and other benefits once they leave OSFI?86 Indeed, studies indi-
cate that regulatory capture exists in the banking industry, especially in
concentrated sectors.87 However, independence of a supervisory body is
argued to reduce capture in the financial sector.88 This conclusion is con-
sistent with our discussion of independence, transparency and accounta-
bility above. OSFI establishes fairly strict guidelines and closely monitors
financial institutions. One would expect more relaxed standards if OSFI
personnel were trying to curry favour with regulated entities in the long-
term. We recognize that contextand in particular the degree of inde-
pendence and transparency of the regulated entityis crucial, perhaps
determinative. We need to test our hypothesis empirically before pro-
nouncing on whether regulatory capture is an issue for the regulation of
Canadas financial institutions.
Conclusion
Since the global financial crisis, Canadas banking sector has been the
subject of favourable attention as it remained strong (e.g., free from fed-
eral government support in the form of bailouts), unlike banking sectors
86 On regulatory capture, see generally Jean-Jacques Laffont & Jean Tirole, The Politics
of Government Decision-Making: A Theory of Regulatory Capture (1991) 106:4 Quar-
terly Journal of Economics 1089; Jean-Jacques Laffont, Political Economy, Information
and Incentives (1999) 43:4-6 European Economic Review 649.
87 Daniel C Hardy, Regulatory Capture in Banking IMF Working Paper No 06/34
(2006), online: International Monetary Fund
88 Ross Levine, The Corporate Governance of Banks: A Concise Discussion of Concepts
and Evidence, World Bank Policy Research Working Paper No 3404 (2004), online:
World Bank Documents and Reports
REGULATING FINANCIAL INSTITUTIONS 427
in other developed countries. Part of this success has been attributed to
Canadas financial regulatory system including the effectiveness of its fi-
nancial services regulator, OSFI. Thus, we have sought to explore a ques-
tion of institutional design: What characteristics make a regulator effec-
tive? OSFI provides an ideal case to examine this question.
We have argued that OSFIs insulated and opaque structure provides
certain benefits that enhance its efficacy. In particular, it is able to regu-
late unhampered by political lobbying, partisan politics, and majoritarian
preferences. OSFI operates in a black box, keeping information about its
policy formation and enforcement confidential. With its informational ad-
vantage, it is able to undermine the possibility that banks will collude or
rent-seek. Our conclusions regarding the value of opacity cut against gen-
erally held views about the benefits of transparency in regulatory bodies.
We recognize that further research is warranted in this area. In par-
ticular, it would be beneficial to broaden the analysis to include other ad-
ministrative bodies in the financial area generally as well as other aspects
of OSFIs mandate. We might ask: Is it the case that the qualities of opac-
ity and insulation are advantageous in other legal settings, or is there
something specific about the financial services area that allows the bene-
fits that we have outlined to outweigh the costs (i.e., costs associated with
opacity and insulation)?
A related issue to explore further is the role of the oligopoly in this area. As
we noted above, it may be the case that OSFIs structure works well because
the regulated entities that it must supervise are relatively few in number (five
big banks). An opaque and insulated structure, under which the regulator
maintains close relationships with the regulated entities, holding closed-door
confidential meetings and negotiations, may not be as effective in a market
such as that of the United States where many more banks populate the finan-
cial markets. In practical terms, it certainly would be difficult for the regulator
to function as OSFI does now. In other words, there may be features particular
to Canadian financial markets that enable OSFIs structure to work well. Such
a structure may not be readily transferable to another jurisdiction.
While there may be other factors at play, OSFI has contributed to the
maintenance of a stable banking system. Other jurisdictions, in reforming
their own financial regulatory structure, would benefit from understanding the
connections between the insulation and opacity at the core of OSFIs structure.
Moreover, as we have discussed, an understanding of the costs and benefits of
opacity, transparency, and insulation in administrative bodies in general and
the trade-offs involved in structuring administrative bodies with any of these
characteristics could aid in designing regulatory institutions in other areas.
