Delegation and the Regulation of Financial Markets∗
Thomas Groll†
Columbia University
Sharyn O’Halloran‡
Columbia University
Geraldine McAllister§
Columbia University
June 2016
Abstract
We explore the politico-institutional determinants of U.S. financial market regulation with a
formal model of the policy-making process in which government regulates financial risk at both
the firm and systemic levels, and test these predictions with a novel, comprehensive data set
of financial regulatory laws enacted since 1950. We find that political factors impact Congress’
decision to delegate regulatory authority to executive agencies, which in turn impacts the stringency of financial market regulation. In particular, Congress delegates authority to regulators
when: 1) policy preferences between Congress and executive officials become more similar; 2)
Firms’ investment risks become more uncertain; and 3) Congress’ concerns about a bailout are
greater. As a result, financial markets are more heavily regulated when firm-specific and systemic risks are uncertain. However, when inter-branch preferences differ or perceived systemic
risk is low, Congress may allow risky investments to be made that, ex post, it wished it had
regulated.
Keywords: Delegation, Financial Regulation, Systemic Risk, Partisan Conflict
JEL Codes: D78, K2, L5, P16
∗
The authors are grateful to participants at the Brussels APE-X Meeting on The Rise and Fall of Democracy, the
Columbia Business School Finance Seminar, the University of Michigan Political Science Department and Business
School, the University of Rochester, Washington University, the Princeton Political Institutions and Economic Policy
Conference, and the Midwest Political Science Association Conference for helpful comments. All errors are our own.
†
School of International and Public Affairs, Columbia University, New York, NY; [email protected].
‡
Department of Political Science and School of International and Public Affairs, Columbia University, New York,
NY; [email protected].
§
Columbia University, New York, NY; [email protected].
1
Introduction
The recent financial crisis has been called a “perfect storm” of failures. Wall Street, hedge funds,
traditional banks, mortgage lenders, ratings agencies, and the media all played their roles, but
regulatory failure on the part of the executive branch should perhaps head the list, as a host of
agencies did not carry out their mandates to oversee the financial institutions that brought down
the economy.
Even so, debate about the nature of this failure persists. Some claim that the problem lies in
the past, regulators had too little authority, especially when it came to overseeing newer forms of
financial organization such as hedge funds, private equity firms, and other elements of the “shadow
banking system.” Despite grave worries about the financial stability of Lehman Brothers, for
instance, the Federal Reserve lacked the necessary resolution authority to liquidate the firm in an
orderly manner. As a result, Lehman’s sudden collapse marked the beginning of the most turbulent
period of the crisis. Under this view, reforms need to concentrate on giving executive agencies more
power, bringing more institutions under regulatory scrutiny, and making sure that regulators have
the resources and authority to assure the safety and soundness of our banking system.
Others claim that executive officials did have the authority to regulate exotic new firms and
financial instruments, but that they failed to fully use the policy tools at their disposal due to
intensive lobbying efforts by the regulated sector. Regulators of Haven Trust Bank, for example,
knew that its real estate holdings were four to six times over the recommended limit, and when they
finally took steps to address the bank’s lending practices it was too late. “Hindsight is a wonderful
thing,” said the chief bank examiner for the OCC. “At the height of the economic boom, to take
an aggressive supervisory approach and tell people to stop lending is hard to do.”1
If this is true, then reforms will mean little unless they are coupled with strong incentives for
regulators to use the power delegated to them and resist being influenced by the industry they
are supposed to be overseeing. Otherwise, giving more power to bureaucrats will only hasten their
capture and make regulatory outcomes even more inefficient.
We are agnostic toward these views of regulatory design. Yet we note that to know what will
work in the future, it would be helpful to know what has been tried in the past, and to what
effect. This paper thus seeks to understand the determinants of financial market regulation in the
United States over time. Toward this end, we review the relevant literature on financial regulatory
1
See “Post-Mortems Reveal Obvious Risk at Banks,” New York Times, Nov. 19, 2009.
1
design underlying this debate. We then build a formal model of the policy-making process in which
government regulates financial risk at both the firm and systemic levels. The analysis shows that
Congress delegates more discretionary authority: 1) as the policy preferences between Congress
and the executive become more similar; 2) as firm investments become riskier; and 3) as Congress
becomes more uncertain of the level of systemic risk. As a result, financial markets are more heavily
regulated when firm-specific and systemic risks are high. However, when inter-branch preferences
diverge, Congress may allow risky investments to be made that, ex post, it wished it had regulated.
To test these predictions, we assemble a comprehensive data set for all financial regulatory laws
passed by Congress since 1950. We code each financial law for the amount of authority delegated to
executive agencies and the administrative procedures that constrain regulators use of these policy
setting powers. The empirical analysis finds that political dimensions impact regulatory design.
In particular, regulation responds to partisan control of Congress, but the party controlling the
presidency may or may not matter.
When discussing government regulation of financial markets, a threshold question is the necessity of regulation in the first place. The regulation of financial intermediaries identifies two levels of
risk: investment risk for individual firms and systemic risk for significant parts of the sector. At the
firm level, it has been long recognized that banks and bank-like institutions borrow short-term and
lend long-term; thus, they may encounter liquidity crises leading to bank runs and panics. Since
the Great Depression, the antidote to this phenomenon has been to regulate individual institutions
for safety and soundness by establishing capital reserve requirements, enforcing prudential lending
practices, and instituting direct oversight by regulators, as well as providing insurance funds to
repay investors in the case of bank failure.2
At the industry level, the case for regulation has been made stronger by the spectacular failure of
the efficient market hypothesis (EMH) over the past decade and more.3 Beginning with the failure
of Long-Term Capital Management, extending through the Asian and Argentinean financial crises,
and on through the great mortgage-fueled global meltdown of 2008, there is abundant evidence that
asset prices are far more volatile than EMH would predict, leading to bubbles whose collapse can
2
See Goodhart, Hartmann, Llewellyn, Rojas-Suarez, and Weisbrod (1998) for the standard theory of financial
regulation to prevent bank runs and panics.
3
The efficient market hypothesis (e.g., Fama 1965; 1970) suggests that regulation is not only unnecessary but is
necessarily harmful. A stronger form states that asset prices reflect all public and private information concerning the
asset; hence, price changes reflect changes in information, and asset bubbles do not exist. Whereas weaker versions of
EMH assert merely that future prices cannot be determined from past prices, so future earnings conform to a random
walk. But asymmetric information and herd behavior can lead to deviations between an asset’s current price and its
long-term underlying value. See Harris (2003, 240) for an overview, and Fox (2009) more generally for a discussion
of EMH.
2
affect national and global economic activity.4 Even if individual financial institutions are properly
regulated for safety and soundness, the cumulative risk present in the system as a whole may
lead to large-scale market failure.5 Nevertheless, the EMH had significant influence on lawmakers’
perception of the firm-level and systemic risk of financial products such as mortgages, derivatives,
and mortgage-backed securities.
Systematic underestimation of correlated industry-wide risk is illustrated by then Federal Reserve Chairman Alan Greenspan’s optimism when he stated at a Futures Industry Association
conference in March 1999 that the “fact that the OTC [over the counter] markets function quite
effectively without the benefits of [CFTC regulation] provides a strong argument for development
of a less burdensome regime for exchange-traded financial derivatives.”6
Similarly, former Secretary of the Treasury Larry Summers acknowledged in The Financial
Crisis Inquiry Report [p. 49] that “by 2008 our regulatory framework with respect to derivatives
was manifestly inadequate,” and that “the derivatives that proved to be by far the most serious,
those associated with credit default swaps, increased 100 fold between 2000 and 2008.”7
Two classic theories examine government’s role in the regulation of markets. The public interest
approach focuses on potential market failures caused by market power, externalities and informational asymmetries that can be regulated by government agencies (e.g. via a Pigouvian tax) that
act in the public’s interest.8 The other view recognizes the private interest of policymakers as well
as regulators. In this view, administrative agencies may be designed to protect the very industries
that they regulate and distort market outcomes—this is the “original sin” hypothesis, advanced by
Stigler (1971).9
Despite these conflicting views, it is clear that the financial sector is indeed regulated, often
quite extensively, and that in many cases these regulations contribute to the smooth functioning of
financial markets. Without insider trading statutes and anti-fraud protections, for example, stock
exchanges and the mutual fund industry that they support could not have such broad participation.
4
A recent article reviewing evidence against the strong form of the EMH can be found in “Efficiency and Beyond,”
The Economist, July 16, 2009, online at http://www.economist.com/displaystory.cfm?story_id=14030296.
5
For a discussion of systemic, as opposed to firm-level, risk in finance, see Epstein and O’Halloran (2009).
6
Quoted in Financial Crisis Inquiry Report, p.48 http://www.gpo.gov/fdsys/pkg/GPO-FCIC/pdf/GPO-FCIC.pdf.
7
It should also be noted that Larry Summers was the architect of the 2000 Commodities Modernization Act, which
prohibited the regulation of derivatives and SWAP contracts by the CFTC and SEC. See Pub. L. 106-554.
8
The public interest theory of regulation is often cited but rarely advanced as a realistic description of government
actions and motivations. In reality, it is more of a restatement of welfare economics and the possibility of efficient
government intervention. See Hantke-Domas (2003) for a review of the intellectual history of public interest theory
and Shleifer (2005) for an excellent review of the regulatory literature.
9
Similarly, “iron triangles” or subgovernments may form, in which congressional committees, interest groups, and
bureaucrats combine in an unholy trinity to deliver benefits to the interest group’s members at public expense. This
line of reasoning, then, dovetails with Lowi’s (1979) critique on delegation as an abdication of legislative responsibility.
3
The question, then, becomes not whether governments are capable of regulating financial markets,
but under which circumstances they will have the incentive and ability to do so effectively. That
is, one needs a theory relating the structure of political and economic institutions to regulatory
outcomes and, further, to market development and performance.
In a foundational essay, La Porta, Lopez-de-Silanes, Shleifer, and Vishny (1998) assert that
the extent of a country’s financial development can be explained by its inheritance of the English
common law system as opposed to the French civil law system. The authors argue that common law
systems afford greater protection to minority shareholders and better protect creditor rights, thus
fostering broader financial market development.10 In a related study, Keefer (2008) asserts, to the
contrary, that legal origin has less explanatory power than a country’s governmental institutions.
In particular, Keefer shows that separation of powers and competitive elections are correlated with
strong investor protection and lending to the private sector. Competitive governmental structures,
in this view, are linked with competitive markets.11
Kroszner and Strahan (1999) engage in a different sort of cross-sectional exercise, examining
the timing of branching deregulation across states in the United States. Their results support
what they term an “interest group” theory: the relative strength of winners from deregulation
(large banks and smaller, bank-dependent firms) and the losers (small banks and insurance firms)
explains patterns in the data better than does public interest theory.12
Historical studies of financial development in the United States tell similar stories. Haber
(2008), for instance, argues that governments free from outside political competition will do little
to implement regulations in the banking sector. He uses the examples of Mexico and the United
States to argue that institutionalized competition through electoral suffrage, political parties, separation of powers, and/or federalism is a necessary precursor to broad financial development. Wallis
(2008) reviews the history of early banking in the United States, showing that the system of statechartered banks that predominated throughout most of the nineteenth century was by and large a
failure: states acted like traditional despotic regimes, using banks as sources of revenue and limiting
competition.
The overall lesson from this literature is that effective financial regulation evolves in two stages:
10
These findings are also supported by Beck, Demirgüç-Kunt, and Levine (2001), who argue that legal origin offers
a stronger explanation of financial development than do theories based on the extent of political checks and balances.
11
Barth, Caprio, and Levine (2006) use a comprehensive data set on banking regulation in over 150 countries and
show that those developing countries that regulate by encouraging private enforcement of banking laws (e.g., through
litigation) rather than direct control or no regulation at all, saw the highest rates of financial sector development.
12
Similarly, Rajan and Zingales (2003) argue that financial sector development across countries is impeded when
incumbent banking firms lobby government actors in an attempt to restrain competition.
4
first, the political institutions of national governments must be strong enough so that executive
branch actors face electoral competition and can be actively monitored; and second, legislatures
must resist capture by the financial industry and delegate authority to regulatory agencies. As the
United States in the postwar era has cleared the first of these hurdles, effective banking regulation
should be correlated with authority delegated to executive branch actors. Our comprehensive data
set of U.S. financial regulatory laws enacted since 1950 allows us to specifically focus on the first
stage of the regulatory evolution in financial markets.
Another recent line of research seeks to understand the political economy of financial regulation
and the recent crisis. Mian, Sufi, and Trebbi (2010) focus on the political economy of U.S. mortgages
in the 2000s that followed the mortgage default crisis. They focus on the influence of constituents,
special interests and ideology and the policymaking process. Legislators representing districts with
sharper increases in mortgage defaults were more likely to vote for the American Housing Rescue and
and Foreclosure Prevention Act (2008); whereas campaign contributions from the financial industry
made votes for the Emergency Economic Stabilization Act (2008) more likely. Overall, they found
that legislators were more sensitive to their voters than special interests. Similarly, Mian, Sufi, and
Trebbi (2013) illustrate how special interests by the mortgage industry and constituent interests by
homeowners and aspiring homeowners fueled the mortgage credit expansion in the mid-2000s. Our
work is complimentary as we focus on the evolution of regulation and delegation with preferences
for Congress and the executive over financial rescue activities and financial products.
In our model of the policymaking process Congress can delegate authority to the better informed
executive that can oversee both firm-level and systemic risk in financial markets. It also incorporates
a potential conflict between policymakers and regulators by considering differences in the salience
of systemic failure and potential cost of bailing out the finance and banking sector. We find that
Congress delegates more discretionary authority when: 1) the policy preferences between Congress
and the executive become more similar; 2) firms’ investment risks become more uncertain; and
3) Congress’ bailout salience and costs become more severe. As a result, financial markets are
more heavily regulated when firm-specific and systemic risks are uncertain. But when inter-branch
preferences differ or perceived systemic risk is low, Congress may allow risky investments to be
made that, ex post, it wished it had regulated. As a result, financial markets are more heavily
regulated when firm-specific and systemic risks are high and uncertain. But when inter-branch
preferences differ, Congress may allow risky investments to be made that, ex post, it wished it had
regulated. Our comprehensive data set contains on all financial regulatory laws passed by Congress
5
since 1950. We code each financial law for the amount of authority delegated to executive branch
actors and the administrative procedures constraining this delegated authority.13
Our model of delegation of financial regulation thus emphasizes the tradeoff of executive branch
expertise against the principal-agent problems of imperfect control. This tradeoff is the subject
of earlier work (Epstein and O’Halloran 1994; 1996; 1999), and has since been elaborated in a
series of interesting studies examining the politics of delegation with an executive veto (Volden,
2002), civil service protections for bureaucrats (Gailmard and Patty, 2007), and executive review of
proposed regulations (Wiseman, 2009), among others.14 Banking is certainly a complex area where
bureaucratic expertise would be valuable; Morgan (2002), for instance, shows that rating agencies
disagree significantly more over banks and insurance companies than over other types of firms.
Furthermore, continual innovation in the financial sector means that older regulations become less
effective, or “decay,” over time. If it did not delegate authority in this area, Congress would have to
continually pass new legislation to deal with the new forms of financial firms and products, which
it has shown neither the ability nor inclination to do.
Our work also overlaps with the economic literature on the location of policy making, as in
Maskin and Tirole (2004) and Alesina and Tabellini (2007), both of which emphasize the benefits
of delegation to bureaucrats or other non-accountable officials (like courts) when presented with
technical policy issues about which the public would have pay high costs to become informed.15 We
also draw parallels with Hiriart and Martimort (2012), who study the regulation of risky markets
and show that when firms cannot be held individually responsible for the consequences of their
actions ex post, regulators will be faced with an ex ante moral hazard problem of firms’ engaging in
overly risky behavior. Finally, we draw inspiration from agency-based models of corporate finance,
as summarized in Tirole (2006).
2
Model
Our model examines strategic interactions among two firms, Congress, and the executive that are
all risk-neutral. The firms can each make an investment requiring an up-front cost of 1, which will
13
See Epstein and O’Halloran (1999). Recent work by O’Halloran, Maskey, McAllister, Park, and Chen (2015) uses
natural language processing and machine algorithm to measure delegation and discretion.
14
See also Bendor and Meirowitz (2004) for contributions to the spatial model of delegation and Volden and
Wiseman (2009) for an overview of the development of this literature.
15
Other studies have focused on optimal interval delegation when preferences of a principal and agent are or are
not aligned, and have been applied to the regulation of a monopoly (Alonso and Matouschek, 2008) and monopolistic
competition (Amador and Bagwell, 2013).
6
return r with probability p and 0 otherwise. The return of an investment is initially unknown, but
ex ante known to be uniformly distributed with r ∈ [1/p, 1/p + 1].16 These returns across firms may
be correlated; in particular, with ex ante probability θ the correlation ρ between investments is 1,
and with probability (1 − θ) they are uncorrelated such that ρ = 0. Should either or both ventures
succeed, then there is no systemic failure and no social cost. If, however, both investments fail,
then there is systemic failure that causes a social cost 2c to bailing out the sector.17 Our structure
implies uncertainty over firm investments as well as systemic risk that is greater when investments
are correlated, both key characteristics of the banking and finance industry. When firms learn
about the specific return r of an investment, each firm undertakes the specific investment as long
as the expected returns are nonnegative, which implies r ≥ 1/p ≡ rm , but does not internalize the
systemic risk of a banking crisis in the absence of regulation.
Congress assumes that investments are correlated with probability θ, which could be factually
correct or a subjective belief, and can regulate the industry on its own or delegate this decision
to executive agencies, who are better informed as to whether projects are correlated or not.18
This expertise comes at a cost, as executive agents’ preferences may differ from those of Congress.
Neither Congress nor the executive knows the specific returns of firms’ investments but both can
set tax rates or capital requirements for all investments that de facto discourage firms, which know
the investments’ returns, to undertake investments with returns below politically desired levels.
The bailout cost are weighted by Congress relative to firms’ expected market profits by α > 0.
Congress thus has expected payoffs of
EU Cρ=0 = 2p2 r + 2p(1 − p)r − 2(1 − p)2 αc − 2
(2.1)
with four possible outcomes when both investments are uncorrelated, ρ = 0, and
EU Cρ=1 = 2pr − 2(1 − p)αc − 2
(2.2)
16
For analytical convenience, we focus on the interesting range of investments. Firms would not make investments
with r < 1/p and policymakers would allow all projects with return 1/p + 1, even if ρ = 1.
17
For instance, a central depository institution may have reserves R = 2 − 2c, where c < 1/2. Banks can costlessly
borrow from the central bank if needed, up to the total reserves in the bank. Then a double failure will necessitate
costly borrowing of 2c to insure the depositors at both institutions.
18
One may argue that regulators did not possess full information about systemic risk and each of them had only a
small piece of the puzzle. Our assumption is for technical convenience and our theoretical predictions also hold for
the case when the executive believes that investments are correlated with some probability θ and Congress would
possess less information or no information. In other words, important for our results is that there is an information
symmetry and the executive possesses at least some advantage in expertise.
7
with two possible outcomes when both investments are correlated, ρ = 1. The executive’s expected
payoff is the same as Congress’ except, the executive’s bailout salience, derived from its national
constituency, is greater, β ≥ α. When the financial system experiences a shock, that is, voters
are more likely to hold the President and executive agencies accountable than any one member of
Congress.19 Furthermore, the absolute difference in bailout salience is also affected by whether the
government is unified or divided with respect to party control over the executive and Congress as
voters may hold the executive’s party representatives more accountable.
The government can de facto allow the investment r, to be made or discourage it through, for
instance, capital reserve requirements or Pigouvian tax rates that would make the investment too
expensive for the firms to undertake. Congress can either make policy itself, which we denote with
D = 0, or delegate the regulation to the better informed executive, which we denote with D = 1.
2.1
Regulation by Congress
Congress’ expected payoff from making policy itself is 0 if it discourages any investment and
EU CD=0 = θEU Cρ=1 + (1 − θ)EU Cρ=0 = 2 (p(r + αc(2 + p(θ − 1) − θ)) − αc − 1)
if it allows the investments.
(2.3)
The expected payoff from allowing investments is decreasing in
Congress’ bailout salience and bailout cost as well as the likelihood of correlated investments.
The risk of individual returns has positive effects on Congress’ expected payoff. Hence, Congress
- though not observing returns directly - would set tax rates or capital requirements and only
encourage investments with return r if
r ≥ r̃C ≡
1 + αc(1 − p)(1 + p(θ − 1))
,
p
(2.4)
which is increasing in Congress’ bailout salience and bailout cost as well as the probability of
correlated investments but decreasing in the likelihood of successful investments.20 Congress can
19
Similar arguments are made in trade policy, where presidents are assumed to be more free-trade oriented than
legislators because they are held more responsible for the overall state of the economy. See O’Halloran (1994) for a
discussion.
20
Congress may not know the return for each potential investment but can set reserve requirements or Pigouvian
tax rates such that firms internalize the social cost of systemic risk – i.e., Congress can set a tax of t̃C = r̃C − rm to
discourage firms from investments if their returns fall below Congress’ desired threshold.
8
therefore always ensure itself an expected payoff from making policy itself of
Z
EU CD=0,r̃C =
1/p+1
EU CD=0 dr =
r̃C
(αc − p(1 + αc(2 − p − θ + pθ)))2
,
p
(2.5)
which also depends on Congress’ bailout salience and expected likelihood of correlated investments.
2.2
Delegation of Regulation to Executive Agencies
Alternatively, Congress can delegate regulation of the financial sector to an executive agency that
is better informed but differs from Congress in its bailout salience. The executive’s expected payoff
can be represented as
EU Eρ=i =



2p2 r + 2p(1 − p)r − 2(1 − p)2 βc − 2
for i = 0


2pr − 2(1 − p)βc − 2
for i = 1
(2.6)
The executive knows whether investments are correlated and can set appropriate regulation
reflecting the state of the world of ρ = 0 or ρ = 1. If Congress delegates regulatory authority, then
the executive encourage only those investments that ensure a nonnegative expected payoff – i.e.,
EU ED=1 ≥ 0. Hence, the regulator would only allow investments with return r if
r≥



rE ≡
1+βc(1−p)2
p
if ρ = 0


rE ≡
1+βc(1−p)
p
if ρ = 1.
(2.7)
As both thresholds depend on the executive’s bailout salience, they differ from Congress’ desired
return thresholds if it would possess the executive’s expertise with rC ≤ rE and rC ≤ rE as α ≤ β.
In other words, the executive puts greater weight on the bailout cost and, therefore, would prefer a
stricter regulation than Congress, which is illustrated in Figure 1. Congress can clearly gain from
the executive’s expertise by delegating to the executive, and will do so for those values of r in which
their desired regulation overlaps. The delegation range follows then from rE and rC such that
rC − rE =
c(1 − p)(α − (1 − p)β)
.
p
(2.8)
Disagreements arise when Congress has lower bailout costs than the executive and would prefer to
allow uncorrelated low-return investments, but the executive could always claim, illustrating the
9
Allow
Disallow
Disallow
Allow
rM
B
Delegation
A
rE
rC
rE
rC
r
Return on Investment
Figure 1: Regulation and Delegation.
moral hazard problem, that these investments are correlated and discourage them, as illustrated
by region A in Figure 1, and correlated high-return investments, as illustrated by region B. Note
that the delegation range does not have to be positive and depends on Congress’ and the executive’
salience. We summarize these results in Proposition 1.
Proposition 1. For the delegation game in financial regulation,
1. Congress delegates if (1 − p)β ≤ α and r ∈ rE , rC .
2. Otherwise, Congress sets policy itself and allows the investments if r ≥ r̃C .
When we consider the ability of Congress to delegate regulatory authority with some amount
of discretion to the better-informed executive, which we denote with d, we find that in fact the
optimal amount of discretion, which is the executive’s ability to set rates freely, is identical to the
delegation range. To see this, suppose Congress would allow the executive agency to deviate from
a baseline (status quo) policy determined by Congress and set a threshold above the baseline value
incorporating its better knowledge of investment returns. To determine the optimal amount of
discretion, note that the executive would implement its preferred return requirements, rE and rE ,
but Congress can limit the executive’s ability to implement rE when the discretion binds. Then
for a given level of discretion d, Congress’ expected utility is
Z
1/p+1
EU Cd = (1 − θ)
Z
1/p+1
EU Cρ=0 dr + θ
rE
EU Cρ=1 dr.
(2.9)
rE +d
To solve this for the optimal amount of discretion, we set the derivative of equation 2.9 with respect
to d to zero, yielding
d∗ =
c(1 − p)(α − (1 − p)β)
,
p
(2.10)
which is identical to equation 2.8. We can summarize the implications of delegation and optimal
discretion as follows.
10
Preference Difference
β
α
1
1−p
Allow
Disallow
Delegation
1
rM
rE
r̃C
rC
r
Return on Investment
Figure 2: Delegation and Preference Differences.
Proposition 2. Congress delegates more discretionary authority to executive agencies when preference differences are small but bailout costs are high. Holding preference differences constant,
the optimal amount of discretion is increasing in Congress’ bailout salience. Investment risk has
ambiguous effects.
All else being equal, the largest delegation range and therefore largest discretion occurs when
β = α, which implies that there is no preference difference between Congress and the executive
agency and both consider identical regulation levels. As the gap between preferred outcomes
increases, the delegation range of rE , rC and optimal discretion d∗ shrink. Figure 2 illustrates this
relationship between preference differences and the delegation range. The shaded area illustrates
cases where Congress prefers to delegate to the better informed executive, while outside this area
Congress will regulate on its own and make policy according to equation 2.8 above. However, as
the cost of bailing out the banking sector, c, increases, Congress values the executive’s expertise
more and delegation and discretion increase. Furthermore, holding preference differences β/α
constant, delegation and discretion increase when Congress’ bailout salience increases. The effect
of investment risk p is ambiguous and depends on the preference difference and investment risk.21 If
the preference difference is relatively small and therefore disagreement low, Congress delegates more
discretion to the executive when investments are riskier. But when disagreement is high, Congress
delegates less to the executive when investments are riskier. We show in Figure 10 in the appendix
that, all else equal, optimal discretion tends to be greater for intermediate levels of the investment
risk p and, therefore, when there is more uncertainty about investment success. Technically, when
p goes to 1, projects are most likely to be successful and the benefit of delegation shrinks as the
21
The effect depends on β/α R 1/(1 − p2 ). If β/α < 1/(1 − p2 ), then the delegation range shrinks as p increases
and therefore investment risk decreases. If β/α > 1/(1 − p2 ), then the delegation range expands as p increases.
11
informational advantage of the executive disappears. Similarly, when p approaches 0, projects are
most likely failures and informational gains from delegation are negligible. In other words, when
Congress is less informed about firm-specific investment risk and overall systemic risk, it delegates
greater discretion.
2.3
Gains from Regulation
We next compare outcomes under the two regulatory regimes, when Congress regulates financial
markets itself and when Congress delegates discretionary authority to executive agencies. We
describe Congress’ expected net payoff with EU Cnet = EU Cd∗ − EU CD=0,r̃C , which follows from
equations 2.9 with 2.10 and 2.5, and leads to the following proposition.
Proposition 3. Congress realizes the greatest expected net payoff from delegation with discretion
when preference differences are low. When preference differences are sufficiently large and the
probability of correlated investments and investment risk are sufficiently low, Congress’ expected
net payoff from delegation is decreasing in bailout salience and cost.
The intuition for the first part of Proposition 3 follows straightforwardly from the previous
analysis. Whenever policy disagreement is low, Congress gains relatively more from the executive’s
expertise and will delegate regulatory authority. However, when preference differences dominate
the gains from executive expertise to reduce the uncertainty over investment risk and systematic
risk, Congress regulates on its own even when bailout costs and salience are high.
Furthermore, comparing the two regulation outcomes with respect to the other variables of
interest, we can show that Congress prefers to regulate on its own when investments are very risky
though uncorrelated, but prefers to delegate when investments and correlations are uncertain and
bailout costs and salience are high. Figure 3 shows the difference in Congress’ expected payoffs
from delegation with discretion and Congress’ expected payoffs when it regulates itself. To see why
these differences in payoffs occur, suppose first that investments are risky but uncorrelated – that
is, low p and θ that are in the lower left corners in Figure 3. Recall that whenever investments are
uncorrelated, Congress prefers a lower return threshold than the executive, rC ≤ rE . Therefore,
when Congress perceives correlation risk as low, executive expertise adds little and Congress prefers
to regulate on its own and set the policy threshold directly.
When Congress is uncertain about investment risk and correlated investments – that is, intermediate p and θ that are in the centers in Figure 3 – then its gains from delegation increases in
12
(a) α = 0.65, β/α = 1.2, c = 1
(b) α = 1.25, β/α = 1.2, c = 1.5
Figure 3: Expected Net Payoff for Congress.
the severity of a bailout. For example, if Congress assigns no value to the cost of bailing out the
industry or the social costs are low, then there is little for Congress to gain from the executive’s
expertise. However, as Congress faces greater costs of bailing out the industry, in absolute and perceived terms, Congress gains most from the executive’s expertise when Congress’s own knowledge
is low and perceived systemic risk is high.
2.4
Summarizing Hypotheses
Overall, then, we can summarize the analysis above with the following hypotheses.
1. Congress delegates more discretion when:
(a)
(b)
(c)
(d)
The preferences of the executive and Congress are more similar;
Congress’ bailout salience is high;
There is more uncertainty about investment risk; and
The costs of a bailout are higher.
2. Congress prefers to regulate when investments are risky and systemic risk is low, but to
delegate when investments and systemic risk are uncertain.
3. The more Congress cares about bailout costs, the higher are
(a) Congress’ preferred level of regulation;
(b) The executive’s discretion and regulation; and
(c) Overall levels of regulation.
3
Data and Results
Although many excellent histories of financial regulation are available,22 and despite the popular
argument that deregulation of the financial sector played a key role in the recent economic crisis,
22
See for example, Macey, Miller and Carnell (2001).
13
there is as yet no consistent measure of banking regulatory policy over time.23
We therefore create a new dataset on financial regulation since 1950.24 The unit of analysis is
a law governing financial markets. We define the universe of finance and financial institutions as
state- and federally-chartered banks, bank holding companies, thrifts/savings and loan associations,
credit unions, investment banks, financial holding companies, securities, commodities, and mortgage
lending institutions.
Consistent with the findings of the literature reviewed above, for each law we code the amount
of authority delegated to executive agencies and the procedural constraints circumscribing the executive’s use of delegated authority. Together, these define a regulatory agency’s discretionary
authority. In addition, we list the agencies receiving delegated authority (e.g., Securities and Exchange Commission, Commodity Futures Trading Commission, U.S. Department of the Treasury)
and the location of the agency within the administrative hierarchy (e.g., Executive Office of the
President, cabinet, independent agency, or government corporation).
3.1
Selecting the Relevant Laws
We identify relevant legislation in a three-sweep process. First, we include all laws mentioned in the
Congressional Quarterly’s policy trackers for the categories of Banking, Savings and Loan Industry,
Federal Reserve, Stock Market and Financial Services, Insurance, and Mortgages. In the second
sweep, we reviewed a comprehensive survey of banking law (Macey, Miller and Carnell, 2001) as
well as the websites of the federal banking regulators and added any laws not already included in
the first sweep. The third sweep included recent laws not yet available in CQ, where THOMAS ’
“Legislation in Current Congress” provides summaries of post-2005 financial laws.25
In total, we identified 112 federal laws that meet these criteria. For each law we recorded the
23
Several comparative studies construct measures from the annual International Monetary Fund report on laws
that restrict either the payment or receipt of payment for economic exchanges. For example, Quinn (1997) classifies
each country’s laws governing current and capital accounts into a series of quantitative indicators that measure a
country’s degree of International Financial Openness and the liberalization of capital. The exercise, different from
ours here, is to correlate changes in international financial openness with economic indicators such as corporate tax
rates, government expenditures and income inequality overtime. In addition, Philippon and Reshef (2012) in their
study of U.S. financial sector wage differentials from 1909 to 2006, create a deregulation index calculated as: state
branching - separation of commercial and investment banks - interest rate ceilings - separation of banks and insurance
companies. However, there are a number of weaknesses in the deregulation index as calculated here, as well as a
number of limitations relative to our research, where our focus is on regulation of the financial sector at the federal
level, and does not include the insurance industry, which is regulated at the state level. See O’Halloran, Maskey,
McAllister, Park, and Chen (2015) for a review of this measure.
24
The analysis begins in 1950 because in that year the Congressional Quarterly began providing consistent reviews
of the key provisions of enacted legislation.
25
The Library of Congress’s THOMAS database is available at: http://thomas.loc.gov/. It is announced to be
replaced by https://www.congress.gov/ in 2016.
14
10
Financial Bills Passed per Congress, 1950-2009
1950
1952
1954
1956
1958
1960
1962
1964
1966
1968
1970
1972
1974
1976
1978
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
2000
2002
2004
2006
2008
0
2
4
6
8
Average Laws per Congress = 3.58
Average Under Unified = 3.54
Average Under Divided = 3.61
Unified
Divided
Figure 4: Number of Laws Regulating Finance Passed per Congress.
year in which it was passed, the Congress number, the public law number, the subject area it
pertained to, the agencies receiving delegated authority and their position in the executive branch
hierarchy, whether the law as a whole regulated or deregulated the industry, and all recorded votes
taken on the law in the House and Senate. The distribution of laws by Congress is illustrated
in Figure 4, with unified and divided governments shown in different colors. At first blush, the
figure does not indicate the influence of partisan factors in passing financial legislation; the average
number of laws per Congress is almost identical under periods of unified and divided government.
The figure also shows that, on average, Congress enacts approximately three and half financial
regulations each Congress.
3.2
Coding Discretion
The primary source for coding each bill’s contents was the Congressional Quarterly Almanac’s yearend summary of major legislation (80 laws). Where data was not available from the CQ Almanac,
we referred to the Library of Congress’s THOMAS database (27 laws). Where neither source
contained detailed data on a specific law, we referred to the U.S. Statutes (4 laws) and the Federal
Deposit Insurance Corporation website (1 law). Each law was classified as belonging to one or
more categories: depository institutions; securities; commodities; insurance; interest rate controls;
consumer protection; mortgage lending/government sponsored enterprises; and state-federal issues.
Each law was read independently, its provisions numbered, and all provisions that delegated
substantive authority to the executive were identified.26 Delegation was defined as giving discre26
In omnibus legislation, when there was a financial subpart, we code only the relevant provisions. This occurred
in 10 cases.
15
Number of Constraint Categories per Law
0
0
.05
.5
.1
Density
1
Density
.15
1.5
.2
2
.25
Histogram of Delegation Ratio
0
.2
.4
.6
Delegation Ratio
.8
1
0
(a) Delegation Ratio.
2
4
6
Constraint Categories
8
10
(b) Constraint Ratio.
Figure 5: Delegation and Constraint Ratios.
tionary authority to an executive entity to move policy away from the status quo. The executive
includes the president, cabinet, independent agencies and commissions, government corporations,
and federally-mandated private corporations.
For each law in our database, we defined the delegation ratio as the ratio of the number of
provisions that delegate to the executive over the total number of provisions. The histogram of
delegation ratios is shown in Figure 5(a). As indicated, the distribution follows a more or less normal
pattern, with a slight spike for those laws with 100% delegation (these usually had a relatively small
number of provisions).
Executive discretion depends not only on the amount of authority delegated but also on the
administrative procedures that constrain executive actions.27 Accordingly, we identified fourteen
possible types of procedural constraints associated with the delegation of authority. These constraints on discretion, along with their definition, are given in Table 3 in the appendix.
We counted the number of times that each category of constraint appeared in the summary
of each law. The histogram of constraint ratios is shown in Figure 5(b). Including each of the
fourteen categories in our analysis would be unwieldy, so we investigated the feasibility of using
factor analysis to analyze the correlation matrix of constraint categories.
As only one factor was significant, first dimension factor scores for each law were calculated and
termed the constraint index. Third, total discretion was defined as delegation minus constraints –
that is, the amount of unconstrained authority delegated to executive actors. Fourth, the average
number of regulators per law was determined; this shows the degree to which authority is being
27
See McCubbins, Noll, and Weingast (1987).
16
Regulate
0
1
Density
2
3
Deregulate
0
.5
1
0
.5
1
Discretion
Figure 6: Distribution of Discretion Index for Laws that Deregulate and Laws that Regulate.
divided across executive branch actors. And fifth, the degree of autonomy of regulators is measured
by the relative mix of independent regulatory actors receiving authority, as opposed to actors and
executive agencies under more direct presidential control.
Coding each law and its assigned discretion, we have created a measure of the stock of discretion
in a given time period that is consistent with our theoretical model and used in our later empirical
analysis. However, to understand the transitory flow of discretion, each law was also coded for
whether it increased, decreased or left unchanged the regulatory stringency on financial market
participants. This was done by noting disclosure rules, capital requirements, and/or increased
oversight of products and firms. This enables us to construct a regulation-deregulation index,
beginning in 1950 and running to the present day.
3.3
Discretion Analysis
As a basic check on our coding of delegation and regulation, we compare the distribution of the
discretion index for laws that regulated the financial industry overall, and laws that deregulated.
We would expect that laws regulating the industry would delegate more discretionary authority,
and Figure 6 shows that this is indeed the case. The average discretion index for the 24 laws that
deregulate is 0.24, as opposed to 0.35 for the 85 laws that regulate, and both illustrated distributions
are not statistically equal.28
Do differences in executive discretion, in turn, affect the financial industry? It is in general
difficult to determine a measure of the degree to which regulation is successful, but Philippon and
28
The hypothesis that both distributions are equal were rejected by the KolmogorovSmirnov test with a corrected
p-value of 0.016 and by WilcoxonMannWhitney rank-sum test with a p-value of 0.0156.
17
1
2
3
4
Executive Discretion
Financial Sector Share of GDP
.04
.05
.06
.07
.08
1950
1960
1970
1980
year
Excess Wage
1990
2000
2010
0
0
.03
0
1
.1
Excess Wage
.2
.3
2
3
4
Executive Discretion
.4
5
Financial Sector Size and Regulation
5
Financial Sector Wages and Regulation
1950
Discretion
1960
1970
1980
Year
Finance % of GDP
(a) Discretion and financial wages.
1990
2000
2010
Discretion
(b) Discretion and importance of financial sector.
Figure 7: Regulatory Discretion and Financial Sector.
Reshef (2012)’s measure of excess wages in the financial industry serves as a useful proxy; the
greater the degree of regulation, the lower excess wages should be.
Accordingly, Figure 7(a) overlays total discretion delegated each year with Phillipon and Reshef’s
excess wage measure. The trends seem to be in the correct direction: excess wages began to increase
in the 1960s, then a spate of regulation drove them down. They increased again the late 1980s,
and again were lowered by a spike in regulation. Finally wages started rising precipitously during
the 1990s and on to the first decade of the 2000s, but there was no corresponding rise in regulation
to meet them, so they just kept increasing to the levels we see today. The trends in Figure 7(a)
thus accord with the notion that delegating discretionary authority to executive branch officials
does constrain the level of excess wages in the financial industry.29 They also pose a puzzle: why
was the spate of financial innovation in the 1960s — this decade saw the explosion of credit in the
economy, including the widespread use of credit cards and the creation of credit unions — met
with a regulatory response, while the most recent innovations — derivatives, non-bank lenders and
the rise of the shadow banking system — were not? We postpone our suggested answer until the
concluding section of this paper.
Figure 7(a) also indicates that the trend in recent decades has been for Congress to give executive
branch actors less discretion in financial regulation. Since the Great Society era of the 1960’s, and
on into the early 1970s, the total amount of new executive branch authority to regulate the financial
29
Further, the patterns also seem consistent with the notion that regulations “decay” over time as new financial
instruments appear to replace the old one. If one estimates a Koyck distributed lag model yt = α + βxt + βφxt−1 +
βφ2 xt−2 + . . . + t via the usual instrumental variables technique, then β = −0.025 and φ = 0.49, indicating that
regulations lose roughly half their effectiveness each year. See Wooldridge (2006), pp. 635–637 for details of the
estimation technique.
18
Delegation, Constraints, and Agencies Receiving Authority
Agency Independence
3.3
5
.5
Financial Regulation, 1950-2009
Index of Agency Indepencence
2.7
2.9
3.1
Delegation
1
.1
2.5
2
# of Agencies
3
4
Number of Agencies
Delegation/Constraint Index
.2
.3
.4
Constraints
1950
1960
1970
1980
Year
1990
2000
2010
1950
(a) Delegation, Constraints, and Number of Agencies.
1960
1970
1980
Year
1990
2000
2010
(b) Independence of Agencies.
Figure 8: Trends in Constraints and Delegation.
sector has generally declined. The exceptions are a few upticks in discretion which coincide with the
aftermaths of well-publicized financial crises and scandals, including the Savings and Loan crisis,
the Asian crisis, and the Enron scandal. Otherwise, the government has been given steadily less
authority over time to regulate financial firms, even as innovations in that sector have made the
need for regulation greater than ever, and even as the importance of the financial sector in the
national economy has greatly increased as illustrated in Figure 7(b).
What is the source of this decrease in discretion? As shown in Figure 8(a), the amount of
authority delegated to oversee the financial sector has remained fairly constant over time, perhaps
decreasing slightly in the past decade. The trends in Figure 7, then, are due mainly to a large
and significant increase in the number of constraints placed on the regulators’ use of this authority.
In addition, we find that the number of actors receiving authority has risen significantly over the
time period studied, as also shown in Figure 8(a). However, the location of these agencies in
the executive hierarchy has changed as well, away from more independent agencies to those more
directly under the president’s control as illustrated in the scatter plot in Figure 8(b).
Overall, then, our preliminary analysis suggests that the current rules defining financial regulation may create a web of interlocking and conflicting mandates, making it difficult for regulators to
innovate the rules and standards governing the financial industry, while at the same time opening
regulatory agencies to industry capture. The problem is not lack of regulation, then, but that
regulators have little discretion. Modern laws delegate less, constrain more and split authority
across more agencies than their predecessors. This has created a situation where many areas of
financial activity are heavily regulated by the Federal government, but those charged with oversight
19
r
Cong
R
R
Cong
P res
D
D
P res
(a) Divided Government
r
Cong
R
R
P res
Cong
D
D
P res
(b) Cross-Party Coalitions
Figure 9: Partisan Effects.
are hamstrung by overlapping jurisdictions, the need for other actors to sign off on their policies,
or outright prohibitions on regulatory actions by Congress.
Testing Hypothesis 1a generated from our theoretical model above, we would expect Congress to
delegate greater levels of authority to executive branch actors with preferences closer to their own.
As Barth, Caprio, and Levine (2006) report, policymaking in this area tends to be uni-dimensional,
separating actors with more pro-industry preferences from those placing more emphasis on consumer
protection. In the United States over the period studied, Republicans have represented the former
viewpoint and Democrats, the latter.30 We also posit that presidents will tend to be less proindustry than legislators, as their national constituency would lead them to weigh more heavily
consumer interests and the stability of the banking system at large.
As Figure 9 shows, however, two patterns of delegation are consistent with these constraints.
If partisan differences are stronger than inter-branch differences, as in panel (a), then delegation
should be higher under unified government as opposed to divided government; this was the pattern
of delegation found in Epstein and O’Halloran (1999). If interbranch differences predominate,
as in panel (b), though, then delegation will actually be highest from a Democratic Congress
to a Republican president, lowest from a Republican Congress to a Democratic president, and
intermediate for the other two combinations. Furthermore, in this “cross-party coalition” case,
delegation should increase when Congress is controlled by Democrats as opposed to Republicans,
and when the presidency is controlled by Republicans as opposed to Democrats.
We thus have the particular prediction that, when regressing discretion on partisan control
of the branches, we should obtain a positive and significant coefficient on Democratic control of
Congress and Republican control of the presidency. Further, Hypothesis 3 predicts that regulation
will respond to partisan control of Congress, so it should increase when Democrats control Congress
30
This is consistent with the findings of Kroszner and Strahan (1999), who analyze rollcall votes on bank branching
deregulation.
20
Cross-Party
Disc.
(1)
-.084
Disc.
(2)
Disc.
(3)
Reg/Dereg
(5)
Reg/Dereg
(6)
-.066
-.573
-.637
(.037)∗
(.283)∗∗
(.712)
(.029)∗∗∗
Divided
Disc.
(4)
-.080
(.037)∗∗
.043
(.039)
Dem. President
Dem. Congress
.065
.764
1.546
(.024)∗∗∗
(.173)∗∗∗
(.566)∗∗∗
112
.169
23
.425
Start of Term
.041
(.042)
Activist Mood
.012
(.048)
Budget Deficit
.027
(.320)
∆ DJIA
.077
(.133)
Obs.
R2
112
.071
112
.011
112
.091
108
.074
Models 1-4 are OLS regressions with discretion as the dependent variable. Models 5 and 6 are
ordered probits with regulation/deregulation as the dependent variable. In Model 6, only those
laws with discretion indices under 0.2 are included in the sample.
Table 1: Estimation Results.
as opposed to Republicans, but the party controlling the presidency may or may not matter.
The estimation results are given in Table 1. The crossparty partisan conflict variable is constructed to equal 1 when Republicans control Congress and Democrats control the presidency, -1
when Democrats hold Congress and the president is Republican, and 0 otherwise. As predicted,
this variable is consistently negative and significant in predicting discretion, while the usual divided
government variable in Model 2 is not significant.31 The signs on Democratic control of Congress
and the presidency are also as predicted, as shown in Model 3,32 and the crossparty effects hold
constant even when a number of control variables are added to the regression in Model 4.
Models 5 and 6 indicate that when predicting whether a given law will regulate, deregulate, or
leave unchanged the level of regulation of the financial industry, the coefficient on partisan control
of Congress is significant in all cases, and in the predicted direction. The coefficient on control of
the executive is significant in Model 5 as well. Model 6 includes only those cases with a discretion
index of 0.2 or under, as the regulation/deregulation relationship should hold most clearly when
31
Similarly, the effect of seat share margins is not significant as well, which implies that interbranch conflicts rather
than innerbranch conflicts seem to be the driving factor.
32
The addition of an interaction term for Dem.P resident × Dem.Congress does not affect those signs and the
interaction term itself is not significant.
21
Congress does not delegate to the executive. Indeed, in these cases the coefficient on Congress
remains positive and significant, while the coefficient on control of the presidency is no longer
significant.
4
Conclusion
This paper analyzed the political and economic determinants of financial regulation. We first
developed a formal model where legislators can either regulate firms directly or delegate authority
to executive branch actors who have greater expertise, and are therefore better able to assess the
true social costs of investment activities. The analysis indicated that agencies have more discretion,
and regulation is more effective, when the preferences of Congress and the executive more nearly
overlap. Conversely, interbranch conflict leads to less regulation, or to regulation in which the
executive is so tightly constrained that they have little latitude to respond to changing market
conditions. We also showed that Congress prefers to regulate risky investments on its own when
systemic risk is perceived to be low.
We tested the implications of this model with data drawn from the regulation of the financial sector in the United States since 1950. We identified every law regulating state and federally
chartered banks, bank holding companies, thrifts/savings and loan associations, credit unions,
investment banks, financial holding companies, securities, commodities, or mortgage lending institutions.
For each law, we coded the amount of authority delegated to executive branch actors, the
location of these actors in the executive branch hierarchy, and the administrative procedures that
constrain agencies’ use of delegated authority. This allowed us to test the proposition that domestic
political factors affect the degree of financial regulation, and the patterns in the data so far are
consistent with our delegation model, assuming that interbranch differences outweigh partisan
differences in these policy areas.
As to the answer to the puzzle posed above — why were the financial innovations of the 1960s
met with new regulations, while those of the past two decades were not — we can answer that
the configurations of partisan control of government were favorable in the first instance, but not
the second. The mid-1960s and early 1970s saw first unified government under the Democrats,
and then divided government with a Democratic Congress and Republican president. Both these
alignments should exhibit higher levels of regulation; Democrats are less indebted to the holders
22
of capital as opposed to labor, and Democratic Congresses and Republican presidents are close
together in the cross-party hypothesis. To the contrary, the 1990s and 2000s saw divided government
with a Democratic president and Republican Congress, followed by unified Republican control of
government. These should each lead to low regulation, as they are the opposite of the first two
situations.
As it expands, this project will hopefully cast light on current policy debates concerning the
revision of national structures of financial market regulation. On a larger scale, we hope to advance
our understanding of the interaction between governments and markets. In particular, under which
conditions will regulation interfere with the smooth operation of markets, and when is regulation
market-enhancing? And furthermore, as to address the second element of financial regulation,
how can legislatures resist capture by the financial industry and delegate authority to regulatory
agencies in order to avoid influence on the law-making process? These questions remain for future
research.
References
[1] Alesina, Alberto and Guido Tabellini. 2007. “Bureaucrats or Politicians? Part I: A Single
Policy Task.” The American Economic Review 97(1): 169-179.
[2] Alonso, Ricardo and Niko Matouschek. 2008. “Optimal Delegation.” Review of Economic Studies 75: 259-293.
[3] Amador, Manuel and Kyle Bagwell. 2013. “The Theory of Optimal Delegation with an Application to Tariff Caps.” Econometrica 81(4): 1541-1599.
[4] Barth, James R., Gerard Caprio, Jr., and Ross Levine. 2006. Rethinking Banking Regulation:
Till Angels Govern. New York: Cambridge University Press.
[5] Beck, Thorsten, Asli Demirgüç-Kunt, and Ross Levine. 2001. “Legal Theories of Financial
Development.” Oxford Review of Economic Policy 17(4): 483-501.
[6] Bendor, Jonathon and Adam Meirowitz. 2004. “Spatial Models of Delegation.” American Political Science Review 98(2): 293-310.
[7] Black, Sandra E. and Philip E. Strahan. 2001. “The Division of Spoils: Rent-Sharing and
Discrimination in a Regulated Industry” The American Economic Review 91(4): 814-831.
[8] Epstein, David and Sharyn O’Halloran. 1994. “Administrative Procedures, Information, and
Agency Discretion: Slack vs. Flexibility.” American Journal of Political Science 38: 697-722.
[9] Epstein, David and Sharyn O’Halloran. 1996. “A Theory of Strategic Oversight: Congress,
Lobbyists, and the Bureaucracy.” Journal of Law, Economics and Organization 11: 227-55.
23
[10] Epstein, David and Sharyn O’Halloran. 1999. Delegating Powers New York: Cambridge University Press.
[11] Epstein, David and Sharyn O’Halloran. 2009. “Avoiding Financial Katrinas: Systemic Risk as
a Common Pool Problem.” Working Paper, Columbia University.
[12] Fama, Eugene. 1965. “The Behavior of Stock Market Prices”. Journal of Business 38: 34-105.
[13] Fama, Eugene. 1970. “Efficient Capital Markets: A Review of Theory and Empirical Work”.
Journal of Finance 25: 383-417.
[14] Fox, Justin. 2009. The Myth of the Rational Market. New York: HarperBusiness.
[15] Freixas, Xavier and Jean-Charles Rochet. 1997. Microeconomics of Banking. Cambridge: MIT
Press.
[16] Gilbert, R. Alton. 1986. “Requiem for Regulation Q: What it did and why it passed away.”
Federal Reserve Bank of St. Louis Review February, 22-37.
[17] Gailmard, Sean and John W. Patty. 2007. “Slackers and Zealots: Civil Service, Policy Discretion, and Bureaucratic Expertise.” American Journal of Political Science 51(4): 873-889.
[18] Goodhart, Charles, Phlipp Hartmann, David Llewellyn, Lilian Rojas-Suarez, and Steven Weisbrod. 1998. Financial Regulation: Why, how and where now? London: Routledge.
[19] Haber, Stephen, Douglass C. North, and Barry R. Weingast, eds. 2008. Political Institutions
and Financial Development. Stanford: Stanford University Press.
[20] Haber, Stephen. 2008. “Political Institutions and Financial Development: Evidence from the
Political Economy of Bank Regulation in Mexico and the United States, in Political Institutions
and Financial Development, eds Haber, North, and Weingast. Stanford: Stanford University
Press.
[21] Hantke-Domas, Michael. 2003. “The Public Interest Theory of Regulation: Non-existence or
Misinterpretation?” European Journal of Law and Economics 15: 165-94.
[22] Harris, Larry. 2003. Trading and Exchanges: Market Microstructure for Practitioners. New
York: Oxford University Press.
[23] Hiriart, Yolande and David Martimort. 2012. “How Much Discretion for Risk Regulators?”
RAND Journal of Economics 43(2): 283-314.
[24] Keefer, Philip. 2008. “Beyond Legal Origin and Checks and Balances: Political Credibility, Citizen Information, and Financial Sector Development,” in Political Institutions and Financial
Development, eds. Haber, North, and Weingast. Stanford: Stanford University Press.
[25] Krozner, Randall and Philip Strahan. 1999. “What Drives Deregulation? Economics and
Politics of the Relaxation of Bank Branching Restrictions.” Quarterly Journal of Economics,
Volume 114(4): 1437-67.
[26] La Porta, Rafael, Florencio Lopez-de-Silanes, Andrei Shleifer, and Robert Vishny. 1998. “Law
and Finance.” Journal of Political Economy 106(6): 1113-1155.
[27] Lapinski, John S. 2008. “Policy Substance and Performance in American Lawmaking, 18771994.” American Journal of Political Science 52(2): 235-251
[28] Lowi, Theodore. 1979. The End of Liberalism: The Second Republic of the United States. 2nd
ed. New York: WW Norton.
24
[29] Macey, Jonathan R., Geoffrey P. Miller and Richard Scott Carnell. 2001. Banking Law and
Regulation. New York: Aspen Publishers.
[30] Maskin, Eric and Jean Tirole. 2004. “The Politician and the Judge: Accountability in Government.” The American Economic Review 94(4): 1034-1054.
[31] McCubbins, Mathew, Roger Noll and Barry Weingast. 1987. “Administrative Procedures as
Instruments of Political Control.” Journal of Law, Economics and Organization 3: 243-77.
[32] Mian, Atif, Amir Sufi and Francesco Trebbi. 2010. “The Political Economy of the US Mortgage
Default Crisis.” American Economic Review 100(5): 1967-1998.
[33] Mian, Atif, Amir Sufi and Francesco Trebbi. 2013. “The Political Economy of the Subprime
Mortgage Credit Expansion.” Quarterly Journal of Political Science 8: 373-408.
[34] Morgan, Donald. 2002. “Rating Banks: Risk and Uncertainty in an Opaque Industry.” American Economic Review 92(4): 874-88.
[35] O’Halloran, Sharyn. 1994. Politics, Process and Trade Policy. Ann Arbor: University of Michigan Press.
[36] O’Halloran, Sharyn, Sameer Maskey, Geraldine McAllister, David K. Park, and Kaiping Chen.
2015. “Data Science and Political Economy: Application to Financial Regulatory Structure.”
Working Paper, Columbia University.
[37] Philippon, Thomas and Ariell Reshef. 2012. “Wages and Human Capital in the U.S. Financial
Industry: 1909-2006.” Quarterly Journal of Economics 112(4): 1551-1609.
[38] Quinn, Dennis. 1997. “The Correlates of Change in International Financial Regulation.” American Political Science Review 91(3), 531-551.
[39] Rajan, Raghuram G. and Luigi Zingales. 2003. “The great reversals: the politics of financial
development in the twentieth century. Journal of Financial Economics 69: 5-50.
[40] Shleifer, Andrei. 2005. “Understanding Regulation.” European Financial Management 11(4):
439-451.
[41] Stigler, G. J. 1971. “The theory of economic regulation.” Bell Journal of Economics, Vol. 2:
pp. 3-21.
[42] Tirole, Jean. 2006. The Theory of Corporate Finance. Princeton: Princeton University Press.
[43] Volden, Craig. 2002. “A Formal Model of the Politics of Delegation in a Separation of Powers
System.” American Journal of Political Science 46(1): 111-133.
[44] Volden, Craig, and Alan Wiseman. 2011. “Formal Approaches to the Study of Congress.”
In Eric Schickler and Frances Lee, eds. Oxford Handbook of the American Congress. Oxford:
Oxford University Press pp. 36-65.
[45] Wallis, John Joseph. 2008. “Answering Mary Shirley’s Question, or: What Can the World
Bank Learn from American History.” in Political Institutions and Financial Development, eds.
Haber, North, and Weingast. Stanford: Stanford University Press.
[46] Wiseman, Alan E. 2009. “Delegation and Positive-Sum Bureaucracies.” The Journal of Politics
71(3): 998-1014.
[47] Wooldridge, Jeffrey. 2006. Introductory Econometrics. Mason, OH: Thompson.
25
A
Proofs and Additional Illustrations
Proof of Proposition 1
If Congress would regulate investments, then it would set capital requirements that encourage
investments with return r ≥ r̃C and disallow investments with r < r̃C by making them prohibitively
unattractive, which is shown by the derivation of (2.4). Now consider Figure 1 and the following
returns:
1. For any r < rC , both Congress and the executive would discourage uncorrelated and correlated investments. Hence, Congress would be indifferent.
2. For any rC ≤ r ≤ rE , Congress would allow uncorrelated investments but discourage correlated investments. The executive would discourage both uncorrelated and correlated investments and could claim, as an issue of moral hazard, that uncorrelated investments would be
correlated as a matter of cheap talk. Hence, Congress would not gain from delegation.
3. For any rE ≤ r ≤ rC , both Congress and the executive would allow or discourage investments
depending on the correlation of investments. Hence, Congress would prefer to delegate to the
better informed executive.
4. For any rC < r < rE , Congress would allow uncorrelated and correlated investments but
the executive would discourage correlated investments. Hence, Congress would not gain from
delegation.
5. For any rE ≤ r, both Congress and the executive would allow uncorrelated and correlated
investments. Hence, Congress would be indifferent.
We can conclude, and assuming that Congress makes policy itself when indifferent, that Congress
delegates for any rE ≤ r ≤ rC but would regulate on its own with decision rule r̃C .
Proof of Proposition 2
Consider d∗ > 0 with α > (1 − p)β. Define k = β/α as preference difference and rewrite (2.10) in
terms of k. We then have
d∗ (k) =
c(1 − p)(α − (1 − p)αk)
p
26
(A.1)
Taking the first derivative with respect to k, we can show that the optimal discretion is decreasing
in the preference difference:
∂d∗ (k)
c(1 − p)2 α
=−
< 0.
∂k
p
(A.2)
Following the same reasoning, we can derive that optimal discretion is increasing in the bailout
cost:
(1 − p)(α − (1 − p)αk)
∂d∗ (k)
=
> 0,
∂c
p
(A.3)
which also holds for d∗ with ∂d∗ /∂c > 0. Further, applying d∗ (k) and holding k constant, optimal
discretion is increasing in Congress’ bailout salience:
∂d∗ (k)
c(1 − p)(1 − (1 − p)k)
=−
< 0,
∂α
p
(A.4)
The first derivative for d∗ and d∗ (k) with respect to p are ambiguous:
∂d∗
∂p
∗
∂d (k)
∂p
=
=
c(α − (1 − p2 )β
R 0,
p2
c(k − 1 − kp2 )α
R 0.
p2
(A.5)
(A.6)
Illustration of Optimal Discretion and Investment Uncertainty
In Figure 10 we illustrate the optimal amount of discretion described in (2.10) by varying Congress’
bailout salience, the executive’s salience and the bailout cost. We can see that optimal discretion
seems to be highest for intermediate values of p when investment risk is more uncertain.
Opt Discr, d *
Opt Discr, d *
0.6
0.6
0.4
0.4
0.2
0.2
0.4
0.6
0.8
1.0
p
0.4
-0.2
-0.2
-0.4
-0.4
-0.6
-0.6
(a) α = 0.55, β = 0.8, c = 75
0.6
0.8
(b) α = 1, β = 1.25, c = 1.1
Figure 10: Optimal Discretion and Investment Risk.
27
1.0
p
Proof of Proposition 3
Congress’ expected net payoff from delegation with optimal discretion is
EU Cnet = EU Cd∗ − EU CD=0,r̃C
c2 (p − 1)2 (1 − θ) p2 α2 θ − (p − 1)2 (α − β)2
=
.
p
(A.7)
Defining k = β/α as preference difference, we can write
c2 (p − 1)2 α2 (θ − 1) (k − 1)2 (p − 1)2 − p2 θ
EU Cnet (k) =
.
p
(A.8)
Taking the first derivative with respect to k, we can show that the expected net payoff is decreasing
in the preference difference:
2α2 c2 (θ − 1)(k − 1)(p − 1)4
∂EU Cnet (k)
=
< 0.
∂k
p
(A.9)
Taking the first derivative with respect to α and c, we get
∂EU Cnet (k)
∂α
∂EU Cnet (k)
∂c
=
=
2αc2 (θ − 1)(p − 1)2 (k − 1)2 (p − 1)2 − θp2
,
p
2α2 c(θ − 1)(p − 1)2 (k − 1)2 (p − 1)2 − θp2
.
p
(A.10)
(A.11)
The sign of each derivative depends on θp2 −(k −1)2 (p−1)2 R 0. If k and p are sufficiently high and
θ sufficiently low, then both derivatives are negative, which implies that net payoffs are decreasing
in Congress’ bailout salience and cost. And vice versa.
28
B
Data Appendix
This appendix details the coding rules used to compile the financial regulation database. From the
theory detailed above in Section 2, the unit of analysis is a law that impacts financial regulation
enacted during a given year and Congress. This data structure enables us to construct a comprehensive database tracing the legislative process of a given law from introduction to passage, to
regulatory decision making to rules promulgated by agencies and, ultimately, to impact on financial
markets. For each financial law, we collected data on committee action, floor voting, legislative
details, agency resource, interest group action and agency rulemaking and enforcement. Embedded in each of these individual datasets are further dimensions or data cubes on voting patterns,
committee types, full text summaries of committee hearings, agency rulemaking details and so on.
Financial Regulatory Structure
The focus of our analysis is on the determinants of financial regulatory structure over time. In
particular, we analyze the authority delegated to executive branch agencies and the administrative
procedures that curtail the use of that authority. In compiling our dataset, we restrict the analysis
to the years 1950 to 2010 for a number of reasons. First, during the late 1940s, Congressional
Quarterly began to publish digests of significant legislation in a common format, allowing for
consistent textual coding. Second, any financial market disruptions precipitated by World War II
are mitigated. Finally, most of the primary financial regulatory agencies were extant.
Defining the Universe of Finance and Financial institutions
We define financial regulation as a principle, rule, or law designed to control or govern the conduct
of financial institutions or markets. A number of institutions or market participants are subject
to financial market regulation. While distinctions between these different types of entities have
become less clear over time, for the purposes of this research, we define the universe of financial
institutions as follows:
• National and state banks (member and non-member, FDIC insured and non-FDIC insured);
• Bank holding companies and financial holding companies;
• Federal and state saving associations (including thrifts);
• Federal and state credit unions; investment banks, investment bank holding companies, and
investment companies;
29
• Hedge funds;
• Broker-dealers;
• Non-bank lenders;
• Asset backed commercial paper conduits; and
• Government sponsored enterprises: Federal National Mortgage Association (Fannie Mae),
Federal Home Loan Mortgage Corporation (Freddie Mac), Government National Mortgage
Association (Ginnie Mae), and the Federal Home Loan Banks.
In the McCarran-Ferguson Act of 1945 (P.L.15), Congress declared that “The business of insurance, and every person engaged therein, shall be subject to the laws of the several States which
relate to the regulation or taxation of such business.” Our focus is on the regulation at the federal
level and delegation at the executive level and we, therefore, included insurance only as it relates
to banks’ insurance activities and financial holding companies (in light of the Gramm-Leach-Bliley
Act of 1999, P.L.106-102), and to the insurance of mortgage loans by the Federal Housing Administration. In addition, The Fraud Enforcement and Recovery Act of 2009 (P.L. 111-21) broadened
the institutions subject to regulation. The Act amended federal criminal code to include within
the definition of “financial institution” a mortgage lending business or any person or entity that
makes, in whole or in part, a federally related mortgage loan. Furthermore, it defined a“mortgage
lending business” as an organization that finances or refinances any debt secured by an interest in
real estate, including private mortgage companies and their subsidiaries, and whose activities affect
interstate or foreign commerce.
Selecting the Relevant Laws
Having defined the universe of finance, we next identify the relevant legislation governing the
financial sector as follows:
1. We conducted a keyword search using the Congressional Quarterly’s policy tracker. We
specified the search query to restrict summaries to post-1950 and to legislative summaries
containing the words: Banking; Savings and Loan Industry; Federal Reserve; Stock Market
and Financial Services; Insurance; Credit and Consumer Debt; and Mortgages. From this
search we compiled a first list of critical legislation, which contained 69 laws covering the
sample period.
2. A review of the secondary literature including Banking Law and Regulation (Macey, Miller
and Carnell, 2001) and reports by the Congressional Research Service (CRS) identified additional significant legislation. This search added 43 laws to the list.
30
3. We augmented this list with publicly available websites of the federal banking regulators.
These online sources provided details of the principal laws and regulations that govern their
activities, were a further valuable source of information on key legislation. An additional 29
laws were added to the list.
4. THOMAS ’s “Legislation in Current Congress” enabled up to update our data set as new laws
were enacted, most notably the Dodd-Frank Act, along with eight other laws .
5. Finally, we compared this list of key financial regulation laws against John Lapinski’s (2008)
“1000 most significant U.S. laws” to ensure that we did not omit any critical financial regulation legislation, adding five laws to our list.
In total, we identified 155 critical federal laws that address the regulation of the financial sector
in the United States from 1950 to 2010. We catalogued these laws: the year of passage; the Congress
that passed the bill; the public law number; the title title (the short title as enacted into law); and
a short description of the law’s content.
The role played by mortgage lending in the financial crisis of 2008, combined with the redefinition, in 2009, of what constitutes a financial institution led us to include mortgage lending and
legislation around this, including community development, in the data set. However, this legislation
behaves quite differently to legislation impacting regulation of a financial institution more narrowly
defined.
For the purposes of this project, therefore, we excluded those laws focused primarily on housing
and mortgage lending that do not address rules governing financial institutions. We retained in the
data set the Participation Sales Act of 1966 (P.L.89-429), legislation which authorized the federal
government to pool mortgages held by six agencies and in essence securitize these assets. An early
instance of securitization, this law constituted an important principle and therefore was included
in the analysis. We retained also Title VII from the Housing and Community Development Act of
1974 (P.L.93-383), which amended real estate lending restrictions for savings and loans, national
banks and credit unions. Finally, we maintained in the data set Title I from the Home Mortgage
Disclosure Act of 1975 (P.L.94-200), which addressed Regulation Q. Our final dataset includes 121
laws.
Data Sources for Coding Laws
Having identified the relevant legislation, the next step was to code this legislation. The Congressional Quarterly Almanac 33 was our primary source of data on laws addressing regulation of the
financial sector in the United States (81 laws). Where data was not available from the CQ Almanac,
33
See: http://library.cqpress.com.monstera.cc.columbia.edu:2048/cqalmanac/index.php
31
Figure 11: An Example of a Coding Sheet.
we referred to THOMAS 34 (Library of Congress) (35 laws); and where neither source contained
detailed data on a specific law we referred to U.S. Statute35 (5 laws).
Coding Method
As a general rule, each of the data sets was coded independently by two different researchers, and
then checked over by a third where any discrepancies were resolved. Upon final entry, each law was
then checked a fourth and final time by the authors.
On average, a well-trained researcher could code a single law in approximately one hour. Obviously, there is considerable variation across laws depending on the length and complexity of the
legislation. For example, the size of the legislative summaries ranged from one page in P.L.86-746
to 79 pages for the P.L.111-203, the Dodd Frank Wall Street Reform and Consumer Protection Act
of 2010. An example of a coding sheet used to collect the data is provided below in Figure 11.
In the field of financial regulation, particularly given that we examined regulation over the
six decades from 1950 to 2010, a number of initial pieces of regulation were revised and updated
34
35
See: http://thomas.loc.gov/bss/d111/d111laws.html
See: http://heinonline.org.monstera.cc.columbia.edu:2048/HOL/Index?collection=statute
32
over this period and, for example, the upper limits placed on spending, borrowing, insured, or
exempted amounts were increased over time. It is important to note, therefore, that each law
was coded as a stand-alone law, where any provision that shifted policy away from the status quo
was coded as delegation. The Federal Deposit Insurance Act of 1950 (P.L.81-797), for example,
authorized the Federal Deposit Insurance Corporation to insure individual bank deposits up to
$10,000 increased from $5,000 previously. Under our coding system this was, therefore, considered
as increased delegation of authority.
In the case of “omnibus bills”, such as omnibus budget bills, incorporating multiple unrelated
titles, we identified those titles addressing regulation of the financial sector and coded only these
relevant titles.
Major Provisions
For the most part, counting the number of major provisions in a particular piece of legislation,
whether coded from the Congressional Quarterly’s legislative summaries or from the Library of
Congress’ THOMAS summaries was straightforward: each new paragraph counted as a provision.
In difficult cases, we followed these rules:
• Bullets and paragraphs count as separate provisions;
• Sub-bullets do not count if they merely elaborate on the previous paragraph;
• Sub-bullets do count if they include new substantive authority;
• Unbulleted paragraphs count as a separate provision if they are substantively distinct from
the previous, bulleted paragraph; and
• If a paragraph is followed by a colon and a list of elements, and if the elements of the list
merely elaborate on the main point of the paragraph, then we count the paragraph and
accompanying list as one provision.
In rare cases where a detailed summary of a public law was not available from either the CQ
Almanac or THOMAS, we coded from U.S. statute. Here, each section (SEC.101, SEC.102, etc.)
counts as one provision, unless it contained subsections ((a), (b), (c), etc.), in which case each
subsection counts as one provision. See Table B below.
Category of Financial Product
Each law was assigned a “category,” according to the financial products / stitutions it affects: depository institution; securities; commodities; insurance; interest rate controls; consumer protection;
33
U.S. Statute/Section
SEC 101:
(a)
(b)
Sec 102: (no sub-sections)
SEC 103: (no subsections)
Provision numbering
——————————->
——————————->
——————————->
——————————->
Coding Delegation
1
2
3
4
Table 2: Delegation Coding Sheet.
mortgage lending / government sponsored enterprises and State-Federal issues.
Delegation
Each law was read independently, its provisions numbered and all provisions that delegated substantive authority to the executive were identified. Delegation was defined as giving discretionary
authority to an executive entity to move policy away from the status quo. Our definition of delegation is any major provision that gives another governmental body the authority to move policy away
from the status quo. To maintain consistency across laws, we developed the following guidelines.
Examples of what delegation is:
• The authorization of a new program with some discretionary powers;
• Discretion to make or modify decision-making criteria;
• Extension of discretionary authority that would otherwise expire;
• The creation of a new commission, board, or agency;
• Demonstration projects;
• Grants and loans where the agency determines the size of the award and/or the recipients;
• The right to issue subpoenas;
• The right to bring suit or intervene in an existing suit;
• The right to issue waivers;
• The ability to enter into contracts;
• The ability to undertake a study;
• The removal of previously existing constraints;
• The repeal of exemptions;
• The authority to exempt;
34
• The increase of existing authority through an increase in, for example, a spending limit /
lending limit / borrowing limit etc. imposed on an agency (beyond merely indexing this to
inflation or adjusting for inflation); or
• The possession of “residual authority”.
Examples of what delegation is not:
• Authorizing appropriations or funds for a program (including the creation of a revolving fund:
the revolving fund is simply the vehicle and does not constitute a new program);
• Requiring reports or publication of information;
• The hiring of staff or personnel;
• Transferring delegated authority from one executive branch actor to another without increasing the scope of that authority;
• Evaluations, recommendations, and assessments that do not directly alter policy; or
• Audits, which are considered constraints and not delegation to, for instance, the GAO.
Some additional, specific examples of what constitutes delegation:
• Directing an agency/representative “to approve to disapprove.” For example, the Federal
Financing Bank Act (P.L. 93-224). “Directed the [Treasury] secretary to approve to disapprove a securities issue within 120 days. If the secretary had not acted after 60 days, he was
required to report his reasons for delay to Congress.”
• The authority to impose penalties: this is coded as executive delegation when such penalties
are imposed directly by the agency in the form of fines. When penalties are imposed by
the courts this is coded as non-executive delegation. An increase in penalties beyond merely
indexing these to inflation is considered additional delegation.
• The inability of one agency to implement change without the “prior approval” of another
agency or body constitutes de facto delegation to that other agency/body.
Agencies and Agency location
• We identified which agency received authority. In our sample, 175 agencies were mentioned
in the legislation.
• We noted where an agency resided in the executive hierarchy. The five categories in ascending
order of independence:
– The Office of the President, including the president;
– Cabinet level departments;
– Independent agencies;
– Independent regulatory commissions;
– Government corporations and federally mandated private corporations.
• We also defined if the agency was newly created by the law. This was a simple, binary
indicator.
35
Constraints
The next step is to define how we coded the fourteen categories of administrative procedures
associated with each law. These categories are all defined as constraints on executive action above
and beyond those specified in the Administrative Procedure Act of 1946 (P.L.79-404). In this
project, the compensation constraint was not found, and ultimately there were a total of thirteen
types of constraints identified.
Category
Description
Appointment Power Limits
Are there any constraints on appointment powers that go
beyond the advice and consent of the Senate?
Are there sunset limits? That is, does the delegated authority
expire after a certain fixed time period?
Does the act define a maximum amount that the agency can
allocate to any activity or set of activities, either stated explicitly or in a formula?
Do agency determinations require the action of Congress to
take effect?
Do agency determinations require the action of another
agency or the President to take effect?
Does Congress retain an ex post veto (of some kind) over the
enactment of agency regulation?
What specific reporting requirements are imposed on agency
rulemaking? Studies are included in this ategory as is making
information publicly available through a website/ free phone
line etc.
Are consultations with any non-agency actor required prior
to final agency actions?
Are public hearings explicitly required?
Is there a procedure explicitly stated in the act for a party
adversely affected by agency actions to appeal?
Do explicit mandates require rulemaking or adjudicatory processes to be carried out in a certain manner (beyond the requirements of the Administrative Procedure Act?
Is there a procedure defined in the implementing legislation
by which a non-agency actor reviews agency’s activities–i.e.
a GAO audit of the agency, or congressional hearings?
Is any particular group, product, or affected interest exempt
from any aspect of regulation for a given period of time?
Are any groups, industries, or states given a specific compensation? In particular, does the act mention any group as
receiving either additional time to adjust to the new regulations or some concession because of the costs that may be
imposed?
Time Limits
Spending Limits
Legislative Action Required
Executive Action Required
Legislative Veto
Reporting Requirements
Consultation Requirements
Public Hearings
Appeals Procedures
Rulemaking Requirements
Direct Oversight
Exemptions
Compensations
Table 3: Categories of Constraints on Executive Discretion.
36
Regulation Index
Finally, each law was coded for whether it increased, decreased or left unchanged the regulatory
stringency on market participants. This was done by noting disclosure rules, capital requirements,
increased oversight of products and firms et cetera. This will enable us to construct a regulationderegulation index, beginning in 1950 and running to 2010.
Although many excellent histories of financial regulation are available,36 and despite the popular
argument that deregulation of the financial sector played a key role in the recent economic crisis,
there is as yet no consistent measure of banking regulatory policy over time.
Several comparative studies construct measures from the annual International Monetary Fund
report on laws that restrict either the payment or receipt of payment for economic exchanges.
For example, Quinn (1997) classifies each country’s laws governing current and capital accounts
into a series of quantitative indicators that measure a country’s degree of International Financial
Openness and the liberalization of capital. The exercise, different from ours here, is to correlate
changes in international financial openness with economic indicators such as corporate tax rates,
government expenditures and income inequality overtime.
Philippon and Reshef (2012) in their study of U.S. financial sector wage differentials from 1909
to 2006, create a deregulation index calculated as: state branching - separation of commercial
and investment banks - interest rate ceilings - separation of banks and insurance companies. And
interestingly the authors’ Figure 8(C) (Financial Deregulation and Relative Wage) is not dissimilar
to Figure 7(a) (Discretion and Financial Wages). However, there are a number of weaknesses in the
deregulation index as calculated here, as well as a number of limitations relative to our research,
where our focus is on regulation of the financial sector at the federal level, and does not include the
insurance industry, which is regulated at the state level. Several additional limitations regarding
the application of this deregulation measure to our current study are worth noting.
The first element in the deregulation index is “Branching,” specifically intrastate bank branching, calculated as the share of the U.S. population living in states that have removed branching
restrictions via mergers and acquisitions, using data from Black and Strahan (2001). This is a
rather narrow index focusing on state level regulation, and the fact that rural and Midwestern
states - generally with smaller populations - were among the last to remove restrictions on intrastate banking will affect this index. Furthermore this excludes significant areas of banking such
36
See for example, Macey, Miller and Carnell (2001).
37
as Savings and Loans and Credit Unions.
The second element in the deregulation index is “separation of commercial and investment
banks.” As noted by the authors, the formerly strict separations were weakened gradually over a
number of years - the result of a combination of legal challenge and regulatory reform including:
• 1987 Competitive Equality Banking Act (P.L.100-86);
• 1989 Financial Institutions Reform Recovery and Enforcement Act (P.L.101-73);
• 1996 Economic Growth and Regulatory Paperwork Reduction Act (P.L.104-208); and
• 1999 Gramm-Leach-Bliley Act. (P.L.106-102).
From the passage of Glass Steagall in 1933, to the Competitive Equality Banking Act in 1987,
this is a rather static measure. Consequently, the measure is really an indicator for post-1988.
The third element in the deregulation index is interest rate ceiling (Regulation Q). Introduced in
1933 and phased out in the early 1980s (originally to be phased out over the six years that followed
passage of the Depository Institutions Deregulation and Monetary Control Act of 1980 (P.L.96221), the Garn -St. Germain Act of 1982 (P.L.97-320) accelerated the phase out), Regulation Q
established ceilings on the interest rates that commercial banks could pay on time and savings
deposits. However, over the more or less 60 years of its existence, Regulation Q was applied
to different degrees and in very different circumstances, but overall with little or no success in
achieving its prime objectives. For instance, Gilbert (1986: 22) notes that “The Banking Act of
1933 established controls over deposit rates for commercial banks that were members of the Federal
Reserve System. Nonmember commercial banks became subject to the same controls in the Banking
Act of 1935. Mutual Savings banks and Savings and loan associations were exempt from the ceiling
interest deposits until the fall of 1966.” In short, the number of financial sector institutions covered
under Regulation Q increased during a third of the sample period and decreased slowly until 1986
when all interest rate ceilings had been eliminated except for the ban on demand deposits, which
was effectively eliminated by the Dodd-Frank Act. In addition, throughout the time period the
ceiling rates on Time and Savings Deposits only increased as did the highest rates on time deposits
at commercial banks. (See Gilbert, 1986: 29, Chart 3).
The fourth and final element in the deregulation index is the separation of banks and insurance
companies: introduced with the Bank Holding Company Act of 1956, this separation ended in 1999
with the passage of the Gramm-Leach Bliley Act. If this measure does indeed seek to capture
38
restrictions on the investment opportunities of insurance companies and banks, it should also
consider the important amendments to the Bank Holding Company Act passed in 1966 and 1970.
What is most important for the current analysis is that this measure fails to capture instances
of re-regulation, including: (i) Sarbanes-Oxley Act of 2002 (P.L.107-204): this resulted in increased
oversight of the accounting industry with the creation of the Public Company Accounting Oversight
Board (PCAOB), overseen by the SEC. It also saw increased penalties for securities fraud; (ii) Credit
Rating Agency Reform Act of 2006 (P.L.109-291) this required nationally recognized statistical
rating organizations (NRSROs) to register with the SEC.
As important as some of the measures included in the deregulation index, a more robust index
would incorporate at least two other critical areas of regulation: Regulation of commodity futures,
and consumer protection.
39
40
Table 4: Descriptive Statistics and Data Sources on Key Variables.