Description
Sometimes resources are badly employed because of coordination failures. Actions by
decision makers that affect the likelihood of such failures are sometimes said to cause “systemic risk.”
This paper seeks to consider the externality in the choice of ex ante risk management policies by
individuals and firms, concerned with private risk, not with their contribution to systemic risk.
Journal of Financial Economic Policy
Systemic risk, coordination failures, and preparedness externalities: Applications to tax
and accounting policy
David Hirshleifer Siew Hong Teoh
Article information:
To cite this document:
David Hirshleifer Siew Hong Teoh, (2009),"Systemic risk, coordination failures, and preparedness
externalities", J ournal of Financial Economic Policy, Vol. 1 Iss 2 pp. 128 - 142
Permanent link to this document:
http://dx.doi.org/10.1108/17576380911010245
Downloaded on: 24 January 2016, At: 21:35 (PT)
References: this document contains references to 13 other documents.
To copy this document: [email protected]
The fulltext of this document has been downloaded 568 times since 2009*
Users who downloaded this article also downloaded:
Marc Vereecken, (1997),"Proposal for a Directive on systemic risk in payment systems", J ournal of
Financial Regulation and Compliance, Vol. 5 Iss 1 pp. 29-36 http://dx.doi.org/10.1108/eb024902
Iqbal Khadaroo, (2005),"Corporate reporting on the internet: some implications for the auditing profession",
Managerial Auditing J ournal, Vol. 20 Iss 6 pp. 578-591 http://dx.doi.org/10.1108/02686900510606074
Christian Herzig, J asmin Godemann, (2010),"Internet-supported sustainability reporting:
developments in Germany", Management Research Review, Vol. 33 Iss 11 pp. 1064-1082 http://
dx.doi.org/10.1108/01409171011085903
Access to this document was granted through an Emerald subscription provided by emerald-srm:115632 []
For Authors
If you would like to write for this, or any other Emerald publication, then please use our Emerald for
Authors service information about how to choose which publication to write for and submission guidelines
are available for all. Please visit www.emeraldinsight.com/authors for more information.
About Emerald www.emeraldinsight.com
Emerald is a global publisher linking research and practice to the benefit of society. The company
manages a portfolio of more than 290 journals and over 2,350 books and book series volumes, as well as
providing an extensive range of online products and additional customer resources and services.
Emerald is both COUNTER 4 and TRANSFER compliant. The organization is a partner of the Committee
on Publication Ethics (COPE) and also works with Portico and the LOCKSS initiative for digital archive
preservation.
*Related content and download information correct at time of download.
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Systemic risk, coordination
failures, and preparedness
externalities
Applications to tax and accounting policy
David Hirshleifer and Siew Hong Teoh
Paul Merage School of Business, University of California,
Irvine, California, USA
Abstract
Purpose – Sometimes resources are badly employed because of coordination failures. Actions by
decision makers that affect the likelihood of such failures are sometimes said to cause “systemic risk.”
This paper seeks to consider the externality in the choice of ex ante risk management policies by
individuals and ?rms, concerned with private risk, not with their contribution to systemic risk.
Design/methodology/approach – The implications for debates over fair value accounting are
considered.
Findings – One consequence is that individuals and ?rms become overleveraged from a social
viewpoint. The recent credit crisis exempli?es the importance of this problem. The US tax system
taxes equity more heavily than debt, and therefore exacerbates the bias toward overleveraging.
A possible solution is to reduce or eliminate taxation of corporate income and capital gains.
Preparedness externalities can also cause ?rms to become too transparent, and thereby subject to
?nancial runs.
Originality/value – The paper offers insights into systemic risk, coordination failures, and
preparedness externalities, focusing on tax and accounting policy.
Keywords Risk analysis, Taxation, Accounting policy, Economic conditions, United States of America
Paper type Research paper
1. Introduction
1.1 Preparedness externalities, coordination failures, and crises
According to macroeconomic theory, sometimes coordination failures cause resources
to be badly employed. During recessions, individuals have trouble ?nding jobs and,
despite the availability of cheap labor and other inputs, ?rms stockpile cash and refrain
from hiring and investing.
Discussions of ?nancial crises often consider “systemic risk” that arises from the
possibility that the failure of one or a few large ?nancial ?rms could create a chain
reaction that damages the entire ?nancial system. The recent ?nancial crisis has
highlighted the interdependence of ?nancial ?rms and the importance of systemic risk.
Failure of the ?nancial system in turn can cause coordination failures in the real
economy. If ?rms cannot raise capital easily, they need to reduce investment, cut
employment, and stockpile cash. So the problem of managing systemic risk is part of
the general problem of avoiding coordination failures.
The current issue and full text archive of this journal is available at
www.emeraldinsight.com/1757-6385.htm
The authors would like to thank Sheridan Titman for valuable insights and discussions in the
development of this paper.
JFEP
1,2
128
Journal of Financial Economic Policy
Vol. 1 No. 2, 2009
pp. 128-142
qEmerald Group Publishing Limited
1757-6385
DOI 10.1108/17576380911010245
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Problems of systemic risk are usually discussed from the viewpoint of regulators
limiting the excessive risk-taking of ?nancial ?rms. However, this begs the question of
why private ?rms would want to take excessive risk. There is an obvious explanation
in the case of banks, since deposit insurance provides a government subsidy to
risk-taking. But more generally, setting aside the prospect of ex post government
rescues or bailouts[1], it is not entirely clear why hedge funds, investment banks,
insurance companies, and other ?nancial entities – as well as industrial ?rms – would
want to take excessive risk.
There are of course, several speci?c explanations that have been offered. For
example, it is often said that the compensation contracts of fund managers or
investment bankers encourage excessive risk-taking. Corporate managers are said to be
driven to risk by stock option compensation. Alternatively, equity-based compensation
and reputational considerations are said to provide excessive incentives to focus on the
short-term, so that during a bubble, while the music is playing, ?nancial managers
feel they have to keep dancing.
Regardless of their validity, at least some of these arguments are not fundamental.
For example, if hedge fund compensation encourages excessive risk taking, it would be
feasible to offer a more concave compensation scheme that encourages less risk-taking.
Corporate managers could be given more salary and fewer options[2].
We argue that the more fundamental problem is a mismatch between the private
and social costs of risk-taking. An investor does not mind that the fund or ?rm he
invests in contributes to systemic risk, because the investor does not bear the full cost
of this contribution. It is this externality which allows compensation contracts and
institutional forms to persist that result in excessive risk from a social point of view.
In general, individuals and ?rms can take many kinds of prior actions that
prepare them for adverse macroeconomic shocks. They can undertake formal risk
management strategies such as a program for hedging, deleverage, and maintain a
higher cash buffer, follow disclosure and reporting policies that make it hard or easy
for stakeholders to see whether the ?rm is in trouble, and design contracts, institutions,
and incentive schemes that prevent agents from taking excessive risk. We call
choices that make the ?rm more resilient to bad states of the world preparedness
activities.
Our thesis is that because of the possibility of coordination failures, there is a
positive externality in preparing for bad times. An example makes this concrete.
Suppose that an automobile ?rm has low preparedness, i.e. it has high leverage and
low cash. Then in the event of a recession, the ?rm is more likely to need to lay off
workers and cut back on investment. This harms workers and suppliers, who in turn
will need to cut back on their demands from other ?rms. The ?rm’s cutbacks can
contribute to a chain reaction and to a general recession. Such a chain reaction could
instead start with a shock to ?nancial ?rms.
Ex ante, the owners of the automobile ?rm do not like the prospect that their ?rm
may have to cut back on needed investment. But they are not concerned about the fact
that should they lay off employees, those employees will cut their demands for the
products of other ?rms. Similarly, the owners of the auto ?rm do not care that their
suppliers will be hurt, and the suppliers of their suppliers. So in choosing its optimal
level of preparedness, the owners and managers of the auto ?rm will not take into
account the bene?t that preparedness confers upon others throughout the economy –
Tax and
accounting
policy
129
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
the bene?t of mitigating coordination failures. Our purpose here is to explore the policy
implications of these preparedness externalities.
Regardless of whether there is complementarity or substitutability in the preparedness
choices of different parties, it remains the case that the incentive to invest in preparedness
is, from a social viewpoint, insuf?cient. As argued above, a prepared party does not
absorb the full bene?t of preparedness, because the party does not bear the full cost of its
contribution to coordination failure.
1.2 Policy implications
We will mainly consider two applications of preparedness externalities, to tax policy
and to accounting policy. Minimizing systemic risk has been a key concern of ?nancial
regulators[3]. Regulators often choose extremely costly ex post responses to ?nancial
crises[4]. Given the high ex post costs of bubbles and panics, it is important to step back
and think about whether existing rules of the game contribute ex ante to their
occurrence. Just as it is often cheaper and more effective to prevent a disease than to
treat it, ex ante policy tools to prevent the formation of bubbles may be much cheaper
than ex post treatments.
We focus on how taxes and accounting policy can exacerbate systemic risk. Our
argument with respect to taxes focuses on ?nancial slack as a means of preparing for
bad times. Inversely, a key source of systemic risk is leverage. At the level of the
individual ?rm, an overhang of excessive debt creates costs of ?nancial distress. The
?rm is forced to cut investment, and engage in renegotiation with creditors. If a ?rm’s
stock price drops and it has an overhang of debt, it becomes hard to raise new capital
to cover needed expenditures. This adversely affects its employees and suppliers.
A ?nancial ?rm cuts back on supplying capital to other ?rms.
During bad times, problems of debt overhang, risk-shifting, and renegotiation
become severe, causing underinvestment, and risk-shifting by individual ?rms. For
the economy, owing to credit linkages between ?rms, a ?nancially distressed ?rm’s
default can trigger distress of its lender, and thereby infect other ?rms in turn.
Similarly, default can create distress for ?rms that have provided insurance to other
creditors through interest rate swaps. Chains of credit across ?rms, the complexity of
risk sharing arrangements, and intransparency in the reporting of who holds exactly
what risks and claims make valuation and renegotiating capital structure dif?cult[5].
In discussions of systemic risk, ?nancial crises, and macroeconomic ?uctuations,
tax policy has had a very low pro?le. Similarly, discussions of tax policy seldom
consider the effect on preparedness and systemic risk. For example, in an extensive
overview of what we know about corporate taxation and possible ways to reform it,
Auerbach et al. (2008) devote little attention to leverage, and do not discuss the effects
of tax policy on coordination failure or systemic risk.
We will argue that this is a major neglected issue in the design of tax policy. Owing
to preparedness externalities, individuals and ?rms become overleveraged from a
social viewpoint. Current tax policy perversely taxes equity more heavily than debt,
and therefore exacerbates the bias toward overleveraging. A possible solution is to
reduce or eliminate corporate income taxes and capital gains taxes.
In contrast, there has been a great deal of discussion about accounting rules as a
source of market crises. Many have placed a major part of the blame for the recent
?nancial crisis on fair value (mark to market) accounting. One argument is that fair
JFEP
1,2
130
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
value accounting actually adds noise to accounting reports, because market values are
hard to assess when markets are illiquid[6]. In contrast, another criticism in the popular
business press seems to assume that marking to market provides accurate information.
This argument is often made naively – that when the market environment turns bad,
marking-to-market makes ?nancial ?rms look bad, and that this creates a positive
feedback that makes things even worse. The conclusion drawn is that marking
to market is bad.
Such an argument is too simple. It focuses on a single ex post realization, and there
may be other possible realizations in which fair value accounting has good effects.
For example, during bad times, if people are on average unbiased, they will, in
realization, sometimes be too pessimistic about ?rms. Transparent reporting that
reveals that things are not as bad as they fear could help end a crisis.
Building on work of Teoh (1997), we will argue here that preparedness externalities
can make better sense out of the casual popular arguments. In other words, a
meaningful argument can be made that a degree of accounting opacity can sometimes
help ?rms be prepared for ?nancial crisis, and hence can help reduce systemic risk.
Opacity of course, has costs as well.
1.3 Bailout moral hazard problems
Two topics that have led to attention to preparedness problems are the moral hazard
problems of deposit insurance (creating a need for banking regulation), and the
too-big-to-fail moral hazard problem. The idea of too-big-to-fail is that ex post large
?nancial ?rms will be bailed out, so ex ante they will take imprudent actions.
More generally, the prospect of bailouts reduces the incentive for preparedness.
The problem is essentially the same if the ?rm is too regulated to fail, or too politically
connected to fail. We therefore refer to these as bailout moral hazard problems.
Owing to preparedness externalities, bailout moral hazard problems reduce ex ante
preparedness relative to a baseline that is already too low from society’s viewpoint.
Even in the absence of these moral hazard problems, private parties would choose too
little preparedness.
Given the attention to too-big-to-fail, it is surprising that most discussion takes size
as exogenous, so that the moral hazard problem is about risk-taking rather than about
choice of size. However, as we will discuss later, existing regulatory policy (even apart
from the prospect of bailouts) supports large-scale of ?rms. Given the ex post systemic
risk cost of ?rms that are too big to fail, a cost-bene?t evaluation of policies should take
into account whether such policies pressure ?rms (especially ?nancial ?rms) to expand.
1.4 Strategic complementarity versus substitutability of preparedness activity
A subtler issue is the extent to which there is strategic complementarity in
preparedness activity across individuals and ?rms. This issue is not critical for our
main conclusions, but would be relevant for developing the implications of
preparedness externalities for macroeconomics. For example, if general motors (GM)
were to choose low preparedness, that might increase the bene?t to Ford of choosing
high preparedness, so that Ford could clean up in the event of recession. This is not
entirely obvious, however.
For vertically placed ?rms, it also seems that preparedness activities could be either
strategic complements or substitutes. If an auto ?rm has low preparedness, it will have
Tax and
accounting
policy
131
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
low demand during a recession. If so, its suppliers may have little need for investment.
This suggests ex ante that a supplier to an unprepared ?rm has little bene?t to
preparedness. On the other hand, there may be a bene?t to the supplier of high
preparedness in order to be able to be able to make the investment needed to switch to a
more prepared buyer.
To the extent that preparedness investments are strategic substitutes, preparedness
mistakes by particular ?rms may have little effect on the overall preparedness of the
system. If GM prepares too little, in compensation its competitors and suppliers
prepare more. Overall, the resilience of the system with respect to shocks may be
similar.
If, on the other hand, there is strategic complementarity, there may be a multiplier
effect wherein the low preparedness of some ?rms drags down the preparedness of
others. There can even potentially be multiple equilibria with very different levels of
preparedness. In this scenario, when other ?rms are well-prepared, coordination
failures are unlikely to be severe, so that even in a recession my ?rm can bene?t from
holding on to employees and investing. So ex ante it pays for my ?rm to prepare as
well. But if other ?rms are ill-prepared, there is no point in my ?rm being prepared.
Focusing on the vertical channel, strategic substitutability is likely to result when
suppliers and buyers do not have to make large ?xed investments. In the event of
recession, if it is easy to switch to a better prepared supplier or buyer, then ex ante even
if my supplier or buyer is unprepared, it is valuable for me to prepare so that I am well
positioned to switch as needed. In contrast, if there was no way ever to switch, then
there is no bene?t to any ?rm along the vertical chain of being prepared for a state of
the world in which the chain will be broken anyway.
2. Tax policy
2.1 Anti-equity tax bias as a source of systemic risk
Existing regulation and tax law contributes to systemic risk by distorting the economy
toward excessive leverage in both general and ?nancial ?rms. At least from the time of
the classical economists (such as John Stuart Mill), business ?uctuations have been
attributed to expansions and contractions of credit (Kindleberger, 2005).
The recent credit crisis has focused attention on the ex post problems of leverage,
but much less to the incentives that create leverage in the ?rst place. Preserving
?nancial slack is a costly form of preparedness. Some basic features of tax law and
regulatory policy encourage individuals, ?rms, and ?nancial institutions to take on
high leverage.
At the individual level, the deductibility of interest (for a given interest rate)
encourages borrowing. At the corporate level, the tax asymmetry between debt and
equity, wherein interest payments are deductible whereas dividend payments are not,
discourages equity in favor of debt. Tax law also encourages high leverage of ?nancial
institutions[7]. The anti-equity bias of existing government policy increases systemic
risk. Shifting these policies would be win-win because it would reduce the ?rm-level
distortions toward excessive leverage, and at the same time reduce systemic risk.
In principle, the risks of high leverage can be offset by ?rms through derivative
hedging (Titman, 1985). This can reduce the risk and cost of ?nancial distress.
Unfortunately, this solution is imperfect for several reasons. First, systemic risk creates
an externality problem, even if only because of the expensive ex post government
JFEP
1,2
132
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
interventions in the economy following bubbles and crashes. Individual ?rms therefore
have insuf?cient incentive to hedge from a society’s viewpoint. This is especially the
case for potential bene?ciaries of government rescues. Second, liquid option markets do
not exist for all ?rms, especially for smaller ?rms. Third, there is an adverse selection
problem with ?rms trading in their own option, since other market participants may
suspect that the ?rm is trading on inside information. Fourth, hedging by trading the
?rm’s own stock creates moral hazard problems, since managers and large
shareholders have less incentive to operate ef?ciently or monitor when the ?rm is
generally hedged against declines in ?rm value, not just against external shocks.
Finally, empirically some ?rms choose not to hedge even when good hedging vehicles
are available (Tufano, 1996).
Although the importance of leverage for systemic risk is obvious, as discussed in
the introduction, policy discussions about systemic risk usually do not consider how
government policy, and especially the tax system affect leverage; and policy discussion
about the tax system in general also usually ignore possible effects on systemic risk.
In choosing their desired level of debt, individuals, and businesses do not have an
incentive to take into account the effect of their leverage on overall systemic risk.
Systemic risk is costly to society, because of coordination failures and because of costly
government bailouts and regulation taken in response to disasters. The party that
takes higher risk ex ante bears only a small part of the incremental ex post costs
resulting from that party’s decision.
Given this externality problem, it would be perverse to design a tax system to push
?rms toward high leverage. However, this is what the current US tax system does.
Equity ?nancing is penalized relative to debt ?nancing. Income that passes through
the corporate sector into the hands of equity holders is doubly taxed – at the corporate
level, and then the personal level. Income destined for debt holders is singly taxed, at
the personal level.
The magnitude of the anti-equity bias of the tax system is immense. The corporate
income tax in the USA is 35 percent (for income over $181/3 million). This imposes a
massive tax penalty on ?rms that ?nance with equity. In effect, the government forces
businesses to be in the business of leveraged speculation.
In equilibrium, theory predicts that this pressure will lead to a big increase in the
equilibrium level of debt in the economy. In the simplest models such as the Modigliani
and Miller model with corporate taxes, ?rms choose 100 percent debt ?nancing, i.e. a
level of leverage high enough that all payout will occur as interest payments. In more
complex models there is still a big shift[8].
There are some counterarguments to the argument we have made about general
?rms:
.
The tax system already offers an offsetting bene?t to equity over debt: equity
holders can defer their taxes by waiting to realize capital gains.
However, in equilibrium in a steady state this does not create any tax bene?t for equity,
because for people to consume, resources have to move out of the corporate sector and
into the hands of individuals. This requires some kind of payout by the ?rm, not just
deferred capital gains:
Tax and
accounting
policy
133
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
.
Firms can help equity investors avoid paying personal taxes on dividends by
paying out through repurchase of shares. There are two arguments here. One is
valid, the other has only a minor degree of validity.
The valid argument is that in the USA the capital gains rate is lower than the personal
income tax rate, and in some countries is zero.
The argument with only a slight degree of validity is that repurchase is more
attractive than dividends because the tax in a repurchase is only on the gain, not the
full amount of the repurchase. To see why this has a bit of validity, think of an
all-equity ?rm that is founded with an investment of $100, and makes pro?ts of $5
per year forever which it pays out. It can throw pro?ts off as dividends, or it can
continually repurchase $5 worth of shares each year[9]. Since the ?rm is not growing in
value, realized capital gains are always zero. So repurchase allows equity investors to
completely avoid personal taxation on their investment.
However, this argument is almost completely invalidated in a growing economy.
Suppose that a ?rm has a steady state growth path in which it makes 10 percent pro?ts
per year, and pays out half of them each year. Suppose the ?rm is originally ?nanced
with an investment of $100. Then over a long period of time dividends or repurchases
will rise to arbitrarily high levels. But the capital gains basis remains unchanged;
in any given repurchase, this will shield only some small fraction of the original
$100. Since the capital gains basis becomes a vanishingly small fraction of the dollar
value of the repurchase, it shields only a miniscule fraction of equity-holders’
investment income. So repurchases are taxed virtually as much as dividends.
It can be objected that when people buy and sell shares, the new owner gets a higher
tax basis, the purchase price. This is true, but is offset by the fact that the seller
immediately pays a capital gains tax on the price increase. So such trades do not reduce
the total tax paid in the repurchase scenario.
2.2 The effects of regulation and taxation on size of the corporate sector and ?rm scale
Afundamental regulation in modern capitalismcreates the limited liability corporation.
This creates a bias in favor of large-scale, and creates a corporate sector with
limited liability as distinct from the non-corporate sector. The regulatory creation of
the limited liability sector is presumably a solution to externality problems which make
it hard to spread risk and create large-scale enterprises. We do not consider in detail the
nature of these problems. But even taking as given that having a corporate sector and
large ?rms is desirable, in evaluating tax and regulatory policies, we still need to
take into account how they shift the size of the corporate sector and the scale of
individual ?rms.
More generally, existing tax and regulatory policy toward enterprise scale is
mixed. Limited liability encourages scale by making it feasible to have a broader
shareholder base, and too-big-to-fail implicit bailout policies subsidize size as well. On
the other hand, there are policies that discourage large-scale such as antitrust laws,
the progressive feature of the corporate income tax, and double taxation of corporate
income (which discourages incorporation)[10]. We argue that in evaluating alternative
regulations and policies, the effects on optimal enterprise scale, and thereby systemic
risk, should be taken into account.
JFEP
1,2
134
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
As discussed earlier, because of too-big-to-fail, large-scale of individual enterprises
can result in bailouts. This creates ex ante moral hazard problems in risk-taking that
exacerbate preparedness externalities. To the extent that the state (rightly or wrongly)
is implicitly committed to costly bailouts, there is a negatively externality in having
large ?rms, especially in the ?nancial sector.
A notable example of government promotion of large-scale is the creation of
gigantic government sponsored enterprises, Fannie Mae and Freddie Mac, to promote
home ownership. There are no obvious economies of scale that would require them to
be so large. This suggests that (if nothing else) they should be broken up.
Antitrust legislation is motivated by a negative externality of size, that high market
share can help a ?rm gain (temporary) monopoly power. Even in the absence of any
market power, large-scale creates a systemic risk externality. This is especially the
case for those ?nancial ?rms such as hedge funds that are in the business of taking
substantial risky bets. There has been little recognition that existing policy distorts the
scale of ?rms, and that such policy should be managed to reduce systemic risk. As far
as solutions go, the tools of antitrust (government imposed breakup of ?rms) seems too
blunt an instrument to use on private ?rms to address systemic risk. A less intrusive
approach is to modify the existing policies that create incentives for large-scale.
2.3 Solutions to anti-equity bias for general ?rms
Tax policy can be changed to avoid overleveraging. A number of different possible
changes in tax policies could reduce or eliminate the anti-equity/pro-debt bias. One is to
eliminate the capital gains tax. Indeed, many countries have a zero capital gains tax.
Firms would need to be permitted to use regular repurchase to perform regular payout.
Another solution might be to allow partial, equal deductibility at the corporate level
for both interest and dividends payout. For example, of every dollar paid as either
interest or dividend income, 50 cents (say) would be deductible. This would help level
the playing ?eld between debt and equity (though it would be an overcorrection in
favor of equity unless the capital gain rate were raised to match personal income tax
rate), and would result in a shift away from corporate leverage. The tax revenue could
be adjusted by calibrating the percentage deduction (instead of the arbitrary example
of 50 percent of the payout).
In addition to the distortion away from equity toward debt, double taxation also
discourages the net ?ow of capital into the corporate sector. Under the current system,
debt provides a back door route through which capital can ?ow into the corporate
sector without any double-taxation. This proposed remedy for the anti-equity bias, by
reducing deductibility of interest income would close this debt “back door.” So overall
we might expect it to further reduce investment in the corporate sector. This could be a
good or bad thing overall – again, we are not analyzing here the reasons for the
existence of limited liability and the existence of a corporate sector. From the viewpoint
of systemic risk, by making it harder to take advantage of limited liability, the scale of
?rms tends to be reduced, which will tend to reduce risk-taking based upon the
expectation of bailout.
On the other hand, the debt back door route is very costly to ?rms because of the
risk of ?nancial distress, and to managers because of the risk of being ?red. So even
though the back door route right now is a good deal from the tax viewpoint, it could be
Tax and
accounting
policy
135
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
that by making the equity route a better deal, the reform would in total increase
investment within the corporate sector.
There are other possible ways of address the tax system’s anti-equity bias. One is to
eliminate the penalty to equity ?nancing by making dividends tax deductible at the
corporate level. Or, more simply and with similar effect, eliminate the corporate income
tax. This solution solves the bias between debt and equity. Lower corporate
income taxes strongly encourages putting resources into the corporate sector, which
however, encourages larger scale of ?rms. The encouragement of larger scale could
increase systemic risk, though this would very likely be outweighed by the effect of
deleveraging.
In summary, double taxation has an immense hidden cost – systemic instability.
This cost can be reduced by changes in tax law to reduce its anti-equity/pro-debt bias.
Some possible changes are politically easier than others.
3. Leverage and regulation of ?nancial ?rms
In the aftermath of Bears Stearns’s demise there has been an increased interest in
regulating investment banks and hedge funds. The prospect that the Fed will bail out
?nancial institutions ex post creates a moral hazard problem in ex ante risk-taking.
This moral hazard problem exacerbates the ex ante negative preparedness externality.
If government is going to provide bailouts ex post, this suggests that the tax and
regulatory system should not be designed to bias ?rms toward greater risk-taking
ex ante.
It is well understood that high ?nancial leverage causes systemic propagation of a
?nancial crisis. Highly levered institutions that are hit by a negative shock to their
value are forced to sell assets. These ?re sales reduce the value of the assets that are
sold, further reducing the value of the selling institution as well as the values of other
highly levered institutions or portfolios that happen to hold similar assets. This will, in
turn, lead to additional forced sales, creating a systemic downward spiral. Such a
feedback loop is considered in Krishnamurthy (2000).
This feedback loop would be self-correcting if forced sales had no effect on the
fundamental value of the securities that were being sold. If there were no deterioration
of the fundamental values, then when the securities become suf?ciently cheap, better
capitalized individuals and institutions could pro?t by buying the cheap securities and
holding them until the crisis subsides.
However, there can be feedback from the ?nancial system to the real economy.
For example, a crisis in the mortgage market leads to less capital available for new
mortgages, which results in lower property prices. This affects the construction
industry, which in turn affects the aggregate economy, leading to a deterioration of the
fundamental value of the underlying real estate. Such negative feedbacks can explain
the poor economic performance of Japan in the 1990s.
Preparedness externalities and the systemic risk associated with high
leverage suggest that it is important to consider how regulatory policies affect the
?nancial structures of investment banks and hedge funds as well as commercial banks.
To evaluate leverage-in?uencing regulation, it is important to understand why
?nancial institutions tend to be so highly levered. For corporations like Bear Stearns,
Lehmann Brothers and Goldman Sachs, just as with non-?nancial ?rms, there is a clear
tax advantage associated with leverage. Given their lines of business, they require
JFEP
1,2
136
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
cheap capital to compete, and because of the tax deductibility of interest payments, the
USA. Tax code makes debt a cheap form of capital for public corporations.
To the extent that taxation is the source of leverage, the problem can be corrected by
allowing publicly held ?nancial institutions to organize as pass through investment
trusts, which pay no direct taxes and thus have no tax incentive to lever up. Examples
of such trusts include real estate investment trusts, royalty trusts, and mutual funds.
This, of course, would result in a reduction in tax revenues. However, the loss in tax
revenues would be small relative to the likely bene?ts of having lower leverage and a
more stable ?nancial system. Alternatively, as discussed for the case of non-?nancial
?rms, modi?cations to the tax code could be made that retain corporate taxation, but
level the playing ?eld between income paid out as interest versus dividends.
The tendency of ?nancial institutions to lever up does not seem to be driven solely
by tax considerations. For example, hedge funds do not have a tax advantage to higher
leverage. There are other possible explanations for leveraging in terms of agency
problems and misperceptions of investors. However, to the extent that leverage is a
problem, it is at least in principle possible for ?nancial contracts and institutions to
address it. More fundamentally, we believe that the key problem is that there are
negative preparedness externalities. A ?nancial institution that levers up takes into
account possible private costs of distress, but not the costs imposed on others in the
event of a general coordination failure. This results in socially super-optimal levels of
leverage.
4. Accounting policy, disclosure policy, and systemic risk
Accounting rules, such as requirements for marking-to-market of illiquid assets, and
tolerance toward off-balance sheet leverage, also contribute to preparedness or to
systemic risk. The effects of accounting rules on preparedness should be weighed in
designing accounting rules and regulation.
Accounting and disclosure policies emerge in part through a political process
involving regulators, politicians, voters, and lobbying activities. Owing to preparedness
externalities, ?rms will lobby for rules that confer private advantages upon them even
if this reduces preparedness of the overall ?nancial system. Similarly, to the extent
that disclosure policy is discretionary with the ?rm, it will tend to be chosen in a way that
results in suboptimal preparedness.
In some cases it is fairly clear how accounting rules affect preparedness. When
investors have limited attention, they neglect and misinterpret accounting information,
and neglect the strategic motives of managers to conceal adverse information
(Daniel et al., 2002; Hirshleifer and Teoh, 2003). For example, if investors have limited
attention, they may not discount suf?ciently for the presence of off-balance sheet
leverage, and therefore fail to suf?ciently impound the expected costs of ?nancial
distress. If so, this increases the incentive for ?rms to lever up, further contributing to
socially insuf?cient levels of preparedness.
Off-balance sheet ?nancing such as special purpose entities (SPEs) have been used
to obscure leverage or to circumvent regulatory requirements[11]. If investors have
limited attention, they may not be adequately skeptical about entities that are not
integrated into the balance sheet (Daniel et al., 2002). As a result, ?rms may not be
adequately penalized by the market for the risk of ?nancial distress that such entities
create[12]. This creates an incentive for excessive off-balance sheet leverage, which
Tax and
accounting
policy
137
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
adds to systemic risk. The externality costs of systemic risk therefore weigh in favor of
relatively tight restrictions to force integrated reporting instead of allowing
relationships with off-balance sheet entities.
Fair value accounting (marking to market) has received a great deal of criticism as
an alleged cause of the recent credit crisis. One of the issues of debate is whether fair
value accounting makes reporting more transparent or less so, and whether fair value
accounting induces systematic reporting biases. A second issue is whether, in the event
of crisis, greater transparency is a good or bad thing. A third issue (interrelated with
the ?rst two) is the effect of marking to market on the perceptions of irrational
investors. The desirability of requirements for marking-to-market of illiquid assets,
and their effects on ex ante preparedness, depends on the answers to these questions.
With respect to whether fair value accounting improves transparency, a key
problem is that it is hard to place a fair market value on an illiquid asset, and during
bad times assets tend to become much less liquid. Some assets will have to be marked
down severely not because their long-term fundamental value is low, but because a
distressed seller has been forced to unload the asset at an arti?cially low market price.
In other words, all ?rms that hold assets similar to that sold by a distressed seller will
suffer a similar devaluation. Therefore, mark to market accounting hardwires the
downside risks of one ?rm to other ?rms. If investors ?xate on accounting numbers
without regard to the fact that the long-run value of the illiquid assets is greater than
the short-term value, or if contracts in place are based on accounting numbers,
marking-to-market creates systemic risk.
This argument does not mean that fair value accounting deserves general
condemnation. For highly liquid assets this concern is much reduced. However,
moving down the scale toward less liquidity, at some point a line needs to be drawn at
which marking-to-market becomes marking to an unavailable or ?ctional price. On the
margin, the problem of systemic risk suggests placing this line more toward the side of
high liquidity.
With respect to the desirability of transparency when investors are irrational, even
if there is a good market price to mark to, marking-to-market causes ?rms’ accounting
measures to seem worse when ?nancial asset markets are doing badly simply by virtue
of re?ecting the bad news. If investors have limited attention, viable ?rms may be
revalued downward under marking-to-market than under an historical cost system
because of ?xation upon accounting numbers[13]. So fair value accounting can be said
to reduce ?rms’ preparedness for bad times.
Many other accounting rules affect transparency, especially when investors
with limited attention focus on earnings and are not able to fully “undo” the accounting
adjustments implied by different accounting rules or discretionary choices by
managers. Thus, the effects of accounting rules on preparedness and systemic risk
should be weighed in designing accounting rules and regulation.
If investors are rational in their assessments of accounting numbers, they will not
revalue a ?rm downward on average just because it has updated its accounting
numbers to fair market values. Consider a public state of the world in which asset
values have dropped. Investors do not know how much a given ?rm’s assets have
dropped, but their assessments are correct on average. Speci?cally, these assessments
already re?ect a good judgment of the losses incurred by the ?rm. If the ?rm then
reports new numbers based on fair market values, investors will sometimes be
JFEP
1,2
138
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
positively surprised and sometimes negatively surprised, but on average their
assessments are correct.
It might then be concluded that when investors are rational there is little effect of
having ?rms mark-to-market (assuming that there are meaningful market values to
mark to). However, this turns out to be incorrect. By providing investors with more
precise information about the status of the ?rm, marking-to-market can create ?nancial
market runs. For a general analysis of ?nancial market runs, Bernardo andWelch(2004).
We can consider creditors (or suppliers) to a distressed ?rm as engaged in creating a
public good. Each creditor can contribute to the survival of the ?rm by extending
further credit, making concessions, or forgiving part of the debt. However, this confers
a bene?t upon other creditors, so (if negotiation is imperfect) concessions will be too
small. Teoh (1997) analyzes the general effect of greater information disclosure on
equilibrium contributions in public goods games. She ?nds that bad news disclosures
can trigger complete collapse of cooperation, which in this context would mean
liquidation of the ?rm.
This does not mean that a policy of greater transparency or disclosure is, ex ante, a
bad thing. On the up side, a favorable disclosure can induce greater concessions and
greater success of the ?rm (because contributors believe that their concessions will not
be wasted). So, discussions in the popular business press about how marking to market
in some particular case contributed to ?rm failure are na? ¨ve in concluding that marking
to market must inherently be bad. Whether greater transparency is, on average, good
or bad depends on the speci?c shape of the payoff functions of the parties involved.
What this analysis does show, however, is that greater transparency can reduce
preparedness, and can set the ?rm up for a greater likelihood of ?nancial market runs.
The model indicates that for ?rms that are close to the edge of complete failure, the
balance of bene?ts can favor intransparency because of an asymmetry in the effects of
good versus bad news. A small amount of bad news may tip the ?rm to an equilibrium
in which creditors’ willingness to make concessions essentially ceases completely –
they just ?ght over the remains. On the upside, a small amount of good news only
increases concessions marginally. So overall, on average transparency reduces ?rm
value.
This suggests that even with rational investors, committing to a policy of
intransparency can be a means of preparedness. If so, regulatory policies that pressure
?rms into greater transparency, especially during crises periods, may have perverse
effects. Thus, the usual presumption of the desirability of transparency may need to be
quali?ed.
We do not wish to overstate the case, especially when fair value accounting is being
widely criticized in popular discussions in often-unsophisticated ways. There are also
well understood bene?ts to transparency for purposes of contracting and corporate
governance; the case of Enron, which was famous for intransparency even before
disaster hit is a case in point. We do not provide an overall assessment of the pros and
cons of transparency. We merely point out that it is possible to make a logical case that
a commitment to intransparency can help a ?rm be better prepared for ?nancial crisis.
5. Conclusion
We consider here an externality in the choice of ex ante risk management policies by
individuals and ?rms: they are concerned with private risk, not with their contribution
Tax and
accounting
policy
139
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
to systemic risk. This preparedness externality causes excessive risk-taking from a
social viewpoint. A general policy implication is that government should avoid
regulatory and tax policies that push individuals and ?rms toward greater risk-taking.
The recent credit crisis has focused attention on the ex post problems of leverage
and of having ?nancial ?rms that are “too big to fail,” but there has not been enough
attention paid to the incentives that create leverage and large ?rms in the ?rst place.
The anti-equity bias of the tax system encourages individuals and ?rms to borrow,
distorting the economy toward excessive leverage. Hedging can in principle help
reduce the effects on systemic risk, but does not solve the problem. By leveling the
playing ?eld of the tax system, the distortion away from high leverage and low
preparedness can be greatly reduced. We therefore suggest that discussion of policies
to average ?nancial crises should place more emphasis on taxation, and its effects on
leverage and systemic risk.
Existing government policy toward ?rm scale is more mixed. Overall, regulation
favors large ?rm scale through the creation of the limited liability corporation, but
various other regulations create biases in different directions. There is a moral hazard
problem in risk taking associated with the government’s tendency to bail out ?rms in
trouble; such bailouts are directed especially toward big ?rms. This moral hazard
problem further exacerbates preparedness externalities. Therefore, on the margin
regulatory policies that affect ?rm scale should take into account the adverse effect of
size on systemic risk.
Accounting rules, such as requirements for marking-to-market of illiquid assets,
and tolerance toward off-balance sheet leverage, also contribute to systemic risk. When
the asset market is so illiquid that there is not a good market price to mark to, fair value
accounting can add noise. Furthermore, investors with limited attention do not always
fully undo the effects of different accounting rules and choices. Since marking-
to-market will tend to mark down during a crisis, fair value accounting can exacerbate
crises, especially when there is feedback from market prices to the real economy. This
problem is exacerbated when contracts and regulations (such as bank reserve
requirements) are based on accounting numbers.
Furthermore, even when investors are rational, transparency can trigger ?nancial
market runs by reducing the average willingness of creditors to make concessions in
the event of distress. Na? ¨ve popular arguments along these lines implicitly focus only
on the adverse realizations in which marking to market leads to asset values that are
even worse than investors are expecting. However, even in a setting where creditors
are rational and on average foresee write-downs correctly, it can still be the case that
transparency increases the probability of ?rm failure.
The problem is that for a ?rm at the edge of disaster, there is an asymmetry
between the large loss in willingness of creditors to make concessions when a little
more bad news arrives, versus a relatively modest increase in willingness when a little
more good news arrives. It follows that a commitment to opacity either in reporting or
disclosure policies can help a ?rm be better prepared for a future crisis. We do not
claim that this consideration necessarily overrides the good reasons for transparency
in general. However, this shows that a logical case for greater opacity is possible to
make. More generally, we suggest that in designing accounting rules and regulation,
the effects on systemic risk should be considered.
JFEP
1,2
140
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Notes
1. Such prospects are quite important. But concerns about excessive risk-taking were raised
even before the recent crisis caused a general upward reassessment in beliefs about the
willingness of the USA Government to engage in bailouts of ?nancial ?rms.
2. Theoretical modeling indicates that the effects of reputational concerns of managers on
risk-taking can vary depending on regulation of ?nancial reporting. Furthermore, the effects
of optimal contracting to deal with these concerns are subtle and in some cases can reverse
the direct effect of reputational concerns (Holmstrom and Ricart i Costa, 1986; DeMarzo and
Duf?e, 1995).
3. The concern for systemic risk is the source of the too-big-to-fail agency problem. Ex ante,
large or important entities such as Fannie Mae, Freddie Mac, or American International
Group have reason to expect to be bailed out in the event of problems, creating an implicit
insurance policy. Ex ante this insurance creates a moral hazard problem.
4. For example, the Fed expands money supply to try to provide softer landing after bubbles.
Government spending is increased in the hope of stimulating the economy, and new regulation
is created.
5. We will later argue, however, that transparency is a double-edged blade in its effects on
systemic risk.
6. Under full rationality, however, individuals would be free to dismiss those component
numbers that they regard as unduly noisy and adjust their assessments accordingly.
7. Discussion often focuses on the deductibility of interest as the driving force, but the anomaly
is really the non-deductibility of dividends. In general, payouts to all providers of resources
to the ?rm (including providers of capital) are deductible as expenses at the corporate level,
except for equity holders.
8. The Miller (1977) model of corporate and personal taxes has capital structure irrelevance at
the ?rm level, but the aggregate equilibrium level of debt is still determinate and increases
when the corporate tax rate increases.
9. If repurchases are performed so regularly that they are viewed as a substitute for dividends,
the internal revenue service can treat the repurchase as a dividend payout. For simplicity we
set this issue aside.
10. Furthermore, limited liability does not always encourage large-scale, since a ?rm that is
subject to lawsuit (e.g. tobacco companies) may have an incentive to focus to prevent lawsuit
from extracting rents from its less lawsuit-prone divisions.
11. SPEs have been replaced by variable interest entities with stricter accounting reporting rules
that require consolidation with the parent company.
12. Zhang (2008) ?nds that a new accounting regulation (?nancial accounting standards board
Interpretation No. 46) that forced consolidation of SPEs on ?nancial statements resulted in
?rms taking actions to decrease book leverage by replacing conventional debt with equity.
Furthermore, after the adoption of this change, on average the S&P credit ratings of ?rms
with SPEs deteriorated, suggesting that prior to consolidation credit agencies did not
adequately discount for the leverage implied by SPEs.
13. Furthermore, some contracts and regulations (such as bank reserve requirements) are
written based upon accounting numbers, so even if investors are sophisticated enough to see
through the accounting numbers, marking-to-market can exacerbate the troubles of
distressed ?rms during a downturn.
Tax and
accounting
policy
141
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
References
Auerbach, A.J., Devereux, M.P. and Simpson, H. (2008), “Taxing corporate income”, NBER
Working Papers No. 14494, National Bureau of Economic Research, Cambridge, MA.
Bernardo, A.E. and Welch, I. (2004), “Liquidity and ?nancial market runs”, Quarterly Journal of
Economics, Vol. 119 No. 1, pp. 135-58.
Daniel, K., Hirshleifer, D. and Teoh, S.H. (2002), “Investor psychology in capital markets:
evidence and policy implications”, Journal of Monetary Economics, Vol. 49 No. 1,
pp. 139-209.
DeMarzo, P. and Duf?e, D. (1995), “Corporate incentives for hedging and hedge accounting”,
Review of Financial Studies, Vol. 8, pp. 743-71.
Hirshleifer, D. and Teoh, S.H. (2003), “Limited attention, information disclosure, and ?nancial
reporting”, Journal of Accounting and Economics, Vol. 36 Nos 1/3, pp. 337-86.
Holmstro¨m, B. and Ricart i Costa, J.E. (1986), “Managerial incentives and capital management”,
Quarterly Journal of Economics, Vol. 101, pp. 835-60.
Kindleberger, C.P. (2005), Manias, Panics, and Crashes: A History of Financial Crises, 5th ed.,
Wiley, New York, NY.
Krishnamurthy, A. (2000), “Ampli?cation mechanisms in liquidity crises”, NBER Working Paper
No. w15040, Kellogg School of Management, Northwestern University, Evanston, IL.
Miller, M.H. (1977), “Debt and taxes”, Journal of Finance, Vol. 32, pp. 261-75.
Teoh, S.H. (1997), “Information disclosure and voluntary contributions to public goods”, RAND
Journal of Economics, Vol. 28 No. 3, pp. 385-406.
Titman, S. (1985), “The effect of forward markets on the debt-equity mix of investor portfolios
and the optimal capital structure of ?rms”, Journal of Financial and Quantitative Analysis,
March.
Tufano, P. (1996), “Who manages risk? An empirical examination of risk management practices
in the gold mining industry”, Journal of Finance, Vol. 51 No. 4, pp. 1097-137.
Zhang, J.Li. (2008), “Economic consequences of recognizing off-balance sheet activities”,
Financial Accounting and Reporting Section (FARS), available at SSRN: http://ssrn.com/
abstract¼1266456 (accessed September 11, 2008).
Corresponding author
Siew Hong Teoh can be contacted at: [email protected]
JFEP
1,2
142
To purchase reprints of this article please e-mail: [email protected]
Or visit our web site for further details: www.emeraldinsight.com/reprints
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
doc_979299524.pdf
Sometimes resources are badly employed because of coordination failures. Actions by
decision makers that affect the likelihood of such failures are sometimes said to cause “systemic risk.”
This paper seeks to consider the externality in the choice of ex ante risk management policies by
individuals and firms, concerned with private risk, not with their contribution to systemic risk.
Journal of Financial Economic Policy
Systemic risk, coordination failures, and preparedness externalities: Applications to tax
and accounting policy
David Hirshleifer Siew Hong Teoh
Article information:
To cite this document:
David Hirshleifer Siew Hong Teoh, (2009),"Systemic risk, coordination failures, and preparedness
externalities", J ournal of Financial Economic Policy, Vol. 1 Iss 2 pp. 128 - 142
Permanent link to this document:
http://dx.doi.org/10.1108/17576380911010245
Downloaded on: 24 January 2016, At: 21:35 (PT)
References: this document contains references to 13 other documents.
To copy this document: [email protected]
The fulltext of this document has been downloaded 568 times since 2009*
Users who downloaded this article also downloaded:
Marc Vereecken, (1997),"Proposal for a Directive on systemic risk in payment systems", J ournal of
Financial Regulation and Compliance, Vol. 5 Iss 1 pp. 29-36 http://dx.doi.org/10.1108/eb024902
Iqbal Khadaroo, (2005),"Corporate reporting on the internet: some implications for the auditing profession",
Managerial Auditing J ournal, Vol. 20 Iss 6 pp. 578-591 http://dx.doi.org/10.1108/02686900510606074
Christian Herzig, J asmin Godemann, (2010),"Internet-supported sustainability reporting:
developments in Germany", Management Research Review, Vol. 33 Iss 11 pp. 1064-1082 http://
dx.doi.org/10.1108/01409171011085903
Access to this document was granted through an Emerald subscription provided by emerald-srm:115632 []
For Authors
If you would like to write for this, or any other Emerald publication, then please use our Emerald for
Authors service information about how to choose which publication to write for and submission guidelines
are available for all. Please visit www.emeraldinsight.com/authors for more information.
About Emerald www.emeraldinsight.com
Emerald is a global publisher linking research and practice to the benefit of society. The company
manages a portfolio of more than 290 journals and over 2,350 books and book series volumes, as well as
providing an extensive range of online products and additional customer resources and services.
Emerald is both COUNTER 4 and TRANSFER compliant. The organization is a partner of the Committee
on Publication Ethics (COPE) and also works with Portico and the LOCKSS initiative for digital archive
preservation.
*Related content and download information correct at time of download.
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Systemic risk, coordination
failures, and preparedness
externalities
Applications to tax and accounting policy
David Hirshleifer and Siew Hong Teoh
Paul Merage School of Business, University of California,
Irvine, California, USA
Abstract
Purpose – Sometimes resources are badly employed because of coordination failures. Actions by
decision makers that affect the likelihood of such failures are sometimes said to cause “systemic risk.”
This paper seeks to consider the externality in the choice of ex ante risk management policies by
individuals and ?rms, concerned with private risk, not with their contribution to systemic risk.
Design/methodology/approach – The implications for debates over fair value accounting are
considered.
Findings – One consequence is that individuals and ?rms become overleveraged from a social
viewpoint. The recent credit crisis exempli?es the importance of this problem. The US tax system
taxes equity more heavily than debt, and therefore exacerbates the bias toward overleveraging.
A possible solution is to reduce or eliminate taxation of corporate income and capital gains.
Preparedness externalities can also cause ?rms to become too transparent, and thereby subject to
?nancial runs.
Originality/value – The paper offers insights into systemic risk, coordination failures, and
preparedness externalities, focusing on tax and accounting policy.
Keywords Risk analysis, Taxation, Accounting policy, Economic conditions, United States of America
Paper type Research paper
1. Introduction
1.1 Preparedness externalities, coordination failures, and crises
According to macroeconomic theory, sometimes coordination failures cause resources
to be badly employed. During recessions, individuals have trouble ?nding jobs and,
despite the availability of cheap labor and other inputs, ?rms stockpile cash and refrain
from hiring and investing.
Discussions of ?nancial crises often consider “systemic risk” that arises from the
possibility that the failure of one or a few large ?nancial ?rms could create a chain
reaction that damages the entire ?nancial system. The recent ?nancial crisis has
highlighted the interdependence of ?nancial ?rms and the importance of systemic risk.
Failure of the ?nancial system in turn can cause coordination failures in the real
economy. If ?rms cannot raise capital easily, they need to reduce investment, cut
employment, and stockpile cash. So the problem of managing systemic risk is part of
the general problem of avoiding coordination failures.
The current issue and full text archive of this journal is available at
www.emeraldinsight.com/1757-6385.htm
The authors would like to thank Sheridan Titman for valuable insights and discussions in the
development of this paper.
JFEP
1,2
128
Journal of Financial Economic Policy
Vol. 1 No. 2, 2009
pp. 128-142
qEmerald Group Publishing Limited
1757-6385
DOI 10.1108/17576380911010245
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Problems of systemic risk are usually discussed from the viewpoint of regulators
limiting the excessive risk-taking of ?nancial ?rms. However, this begs the question of
why private ?rms would want to take excessive risk. There is an obvious explanation
in the case of banks, since deposit insurance provides a government subsidy to
risk-taking. But more generally, setting aside the prospect of ex post government
rescues or bailouts[1], it is not entirely clear why hedge funds, investment banks,
insurance companies, and other ?nancial entities – as well as industrial ?rms – would
want to take excessive risk.
There are of course, several speci?c explanations that have been offered. For
example, it is often said that the compensation contracts of fund managers or
investment bankers encourage excessive risk-taking. Corporate managers are said to be
driven to risk by stock option compensation. Alternatively, equity-based compensation
and reputational considerations are said to provide excessive incentives to focus on the
short-term, so that during a bubble, while the music is playing, ?nancial managers
feel they have to keep dancing.
Regardless of their validity, at least some of these arguments are not fundamental.
For example, if hedge fund compensation encourages excessive risk taking, it would be
feasible to offer a more concave compensation scheme that encourages less risk-taking.
Corporate managers could be given more salary and fewer options[2].
We argue that the more fundamental problem is a mismatch between the private
and social costs of risk-taking. An investor does not mind that the fund or ?rm he
invests in contributes to systemic risk, because the investor does not bear the full cost
of this contribution. It is this externality which allows compensation contracts and
institutional forms to persist that result in excessive risk from a social point of view.
In general, individuals and ?rms can take many kinds of prior actions that
prepare them for adverse macroeconomic shocks. They can undertake formal risk
management strategies such as a program for hedging, deleverage, and maintain a
higher cash buffer, follow disclosure and reporting policies that make it hard or easy
for stakeholders to see whether the ?rm is in trouble, and design contracts, institutions,
and incentive schemes that prevent agents from taking excessive risk. We call
choices that make the ?rm more resilient to bad states of the world preparedness
activities.
Our thesis is that because of the possibility of coordination failures, there is a
positive externality in preparing for bad times. An example makes this concrete.
Suppose that an automobile ?rm has low preparedness, i.e. it has high leverage and
low cash. Then in the event of a recession, the ?rm is more likely to need to lay off
workers and cut back on investment. This harms workers and suppliers, who in turn
will need to cut back on their demands from other ?rms. The ?rm’s cutbacks can
contribute to a chain reaction and to a general recession. Such a chain reaction could
instead start with a shock to ?nancial ?rms.
Ex ante, the owners of the automobile ?rm do not like the prospect that their ?rm
may have to cut back on needed investment. But they are not concerned about the fact
that should they lay off employees, those employees will cut their demands for the
products of other ?rms. Similarly, the owners of the auto ?rm do not care that their
suppliers will be hurt, and the suppliers of their suppliers. So in choosing its optimal
level of preparedness, the owners and managers of the auto ?rm will not take into
account the bene?t that preparedness confers upon others throughout the economy –
Tax and
accounting
policy
129
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
the bene?t of mitigating coordination failures. Our purpose here is to explore the policy
implications of these preparedness externalities.
Regardless of whether there is complementarity or substitutability in the preparedness
choices of different parties, it remains the case that the incentive to invest in preparedness
is, from a social viewpoint, insuf?cient. As argued above, a prepared party does not
absorb the full bene?t of preparedness, because the party does not bear the full cost of its
contribution to coordination failure.
1.2 Policy implications
We will mainly consider two applications of preparedness externalities, to tax policy
and to accounting policy. Minimizing systemic risk has been a key concern of ?nancial
regulators[3]. Regulators often choose extremely costly ex post responses to ?nancial
crises[4]. Given the high ex post costs of bubbles and panics, it is important to step back
and think about whether existing rules of the game contribute ex ante to their
occurrence. Just as it is often cheaper and more effective to prevent a disease than to
treat it, ex ante policy tools to prevent the formation of bubbles may be much cheaper
than ex post treatments.
We focus on how taxes and accounting policy can exacerbate systemic risk. Our
argument with respect to taxes focuses on ?nancial slack as a means of preparing for
bad times. Inversely, a key source of systemic risk is leverage. At the level of the
individual ?rm, an overhang of excessive debt creates costs of ?nancial distress. The
?rm is forced to cut investment, and engage in renegotiation with creditors. If a ?rm’s
stock price drops and it has an overhang of debt, it becomes hard to raise new capital
to cover needed expenditures. This adversely affects its employees and suppliers.
A ?nancial ?rm cuts back on supplying capital to other ?rms.
During bad times, problems of debt overhang, risk-shifting, and renegotiation
become severe, causing underinvestment, and risk-shifting by individual ?rms. For
the economy, owing to credit linkages between ?rms, a ?nancially distressed ?rm’s
default can trigger distress of its lender, and thereby infect other ?rms in turn.
Similarly, default can create distress for ?rms that have provided insurance to other
creditors through interest rate swaps. Chains of credit across ?rms, the complexity of
risk sharing arrangements, and intransparency in the reporting of who holds exactly
what risks and claims make valuation and renegotiating capital structure dif?cult[5].
In discussions of systemic risk, ?nancial crises, and macroeconomic ?uctuations,
tax policy has had a very low pro?le. Similarly, discussions of tax policy seldom
consider the effect on preparedness and systemic risk. For example, in an extensive
overview of what we know about corporate taxation and possible ways to reform it,
Auerbach et al. (2008) devote little attention to leverage, and do not discuss the effects
of tax policy on coordination failure or systemic risk.
We will argue that this is a major neglected issue in the design of tax policy. Owing
to preparedness externalities, individuals and ?rms become overleveraged from a
social viewpoint. Current tax policy perversely taxes equity more heavily than debt,
and therefore exacerbates the bias toward overleveraging. A possible solution is to
reduce or eliminate corporate income taxes and capital gains taxes.
In contrast, there has been a great deal of discussion about accounting rules as a
source of market crises. Many have placed a major part of the blame for the recent
?nancial crisis on fair value (mark to market) accounting. One argument is that fair
JFEP
1,2
130
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
value accounting actually adds noise to accounting reports, because market values are
hard to assess when markets are illiquid[6]. In contrast, another criticism in the popular
business press seems to assume that marking to market provides accurate information.
This argument is often made naively – that when the market environment turns bad,
marking-to-market makes ?nancial ?rms look bad, and that this creates a positive
feedback that makes things even worse. The conclusion drawn is that marking
to market is bad.
Such an argument is too simple. It focuses on a single ex post realization, and there
may be other possible realizations in which fair value accounting has good effects.
For example, during bad times, if people are on average unbiased, they will, in
realization, sometimes be too pessimistic about ?rms. Transparent reporting that
reveals that things are not as bad as they fear could help end a crisis.
Building on work of Teoh (1997), we will argue here that preparedness externalities
can make better sense out of the casual popular arguments. In other words, a
meaningful argument can be made that a degree of accounting opacity can sometimes
help ?rms be prepared for ?nancial crisis, and hence can help reduce systemic risk.
Opacity of course, has costs as well.
1.3 Bailout moral hazard problems
Two topics that have led to attention to preparedness problems are the moral hazard
problems of deposit insurance (creating a need for banking regulation), and the
too-big-to-fail moral hazard problem. The idea of too-big-to-fail is that ex post large
?nancial ?rms will be bailed out, so ex ante they will take imprudent actions.
More generally, the prospect of bailouts reduces the incentive for preparedness.
The problem is essentially the same if the ?rm is too regulated to fail, or too politically
connected to fail. We therefore refer to these as bailout moral hazard problems.
Owing to preparedness externalities, bailout moral hazard problems reduce ex ante
preparedness relative to a baseline that is already too low from society’s viewpoint.
Even in the absence of these moral hazard problems, private parties would choose too
little preparedness.
Given the attention to too-big-to-fail, it is surprising that most discussion takes size
as exogenous, so that the moral hazard problem is about risk-taking rather than about
choice of size. However, as we will discuss later, existing regulatory policy (even apart
from the prospect of bailouts) supports large-scale of ?rms. Given the ex post systemic
risk cost of ?rms that are too big to fail, a cost-bene?t evaluation of policies should take
into account whether such policies pressure ?rms (especially ?nancial ?rms) to expand.
1.4 Strategic complementarity versus substitutability of preparedness activity
A subtler issue is the extent to which there is strategic complementarity in
preparedness activity across individuals and ?rms. This issue is not critical for our
main conclusions, but would be relevant for developing the implications of
preparedness externalities for macroeconomics. For example, if general motors (GM)
were to choose low preparedness, that might increase the bene?t to Ford of choosing
high preparedness, so that Ford could clean up in the event of recession. This is not
entirely obvious, however.
For vertically placed ?rms, it also seems that preparedness activities could be either
strategic complements or substitutes. If an auto ?rm has low preparedness, it will have
Tax and
accounting
policy
131
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
low demand during a recession. If so, its suppliers may have little need for investment.
This suggests ex ante that a supplier to an unprepared ?rm has little bene?t to
preparedness. On the other hand, there may be a bene?t to the supplier of high
preparedness in order to be able to be able to make the investment needed to switch to a
more prepared buyer.
To the extent that preparedness investments are strategic substitutes, preparedness
mistakes by particular ?rms may have little effect on the overall preparedness of the
system. If GM prepares too little, in compensation its competitors and suppliers
prepare more. Overall, the resilience of the system with respect to shocks may be
similar.
If, on the other hand, there is strategic complementarity, there may be a multiplier
effect wherein the low preparedness of some ?rms drags down the preparedness of
others. There can even potentially be multiple equilibria with very different levels of
preparedness. In this scenario, when other ?rms are well-prepared, coordination
failures are unlikely to be severe, so that even in a recession my ?rm can bene?t from
holding on to employees and investing. So ex ante it pays for my ?rm to prepare as
well. But if other ?rms are ill-prepared, there is no point in my ?rm being prepared.
Focusing on the vertical channel, strategic substitutability is likely to result when
suppliers and buyers do not have to make large ?xed investments. In the event of
recession, if it is easy to switch to a better prepared supplier or buyer, then ex ante even
if my supplier or buyer is unprepared, it is valuable for me to prepare so that I am well
positioned to switch as needed. In contrast, if there was no way ever to switch, then
there is no bene?t to any ?rm along the vertical chain of being prepared for a state of
the world in which the chain will be broken anyway.
2. Tax policy
2.1 Anti-equity tax bias as a source of systemic risk
Existing regulation and tax law contributes to systemic risk by distorting the economy
toward excessive leverage in both general and ?nancial ?rms. At least from the time of
the classical economists (such as John Stuart Mill), business ?uctuations have been
attributed to expansions and contractions of credit (Kindleberger, 2005).
The recent credit crisis has focused attention on the ex post problems of leverage,
but much less to the incentives that create leverage in the ?rst place. Preserving
?nancial slack is a costly form of preparedness. Some basic features of tax law and
regulatory policy encourage individuals, ?rms, and ?nancial institutions to take on
high leverage.
At the individual level, the deductibility of interest (for a given interest rate)
encourages borrowing. At the corporate level, the tax asymmetry between debt and
equity, wherein interest payments are deductible whereas dividend payments are not,
discourages equity in favor of debt. Tax law also encourages high leverage of ?nancial
institutions[7]. The anti-equity bias of existing government policy increases systemic
risk. Shifting these policies would be win-win because it would reduce the ?rm-level
distortions toward excessive leverage, and at the same time reduce systemic risk.
In principle, the risks of high leverage can be offset by ?rms through derivative
hedging (Titman, 1985). This can reduce the risk and cost of ?nancial distress.
Unfortunately, this solution is imperfect for several reasons. First, systemic risk creates
an externality problem, even if only because of the expensive ex post government
JFEP
1,2
132
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
interventions in the economy following bubbles and crashes. Individual ?rms therefore
have insuf?cient incentive to hedge from a society’s viewpoint. This is especially the
case for potential bene?ciaries of government rescues. Second, liquid option markets do
not exist for all ?rms, especially for smaller ?rms. Third, there is an adverse selection
problem with ?rms trading in their own option, since other market participants may
suspect that the ?rm is trading on inside information. Fourth, hedging by trading the
?rm’s own stock creates moral hazard problems, since managers and large
shareholders have less incentive to operate ef?ciently or monitor when the ?rm is
generally hedged against declines in ?rm value, not just against external shocks.
Finally, empirically some ?rms choose not to hedge even when good hedging vehicles
are available (Tufano, 1996).
Although the importance of leverage for systemic risk is obvious, as discussed in
the introduction, policy discussions about systemic risk usually do not consider how
government policy, and especially the tax system affect leverage; and policy discussion
about the tax system in general also usually ignore possible effects on systemic risk.
In choosing their desired level of debt, individuals, and businesses do not have an
incentive to take into account the effect of their leverage on overall systemic risk.
Systemic risk is costly to society, because of coordination failures and because of costly
government bailouts and regulation taken in response to disasters. The party that
takes higher risk ex ante bears only a small part of the incremental ex post costs
resulting from that party’s decision.
Given this externality problem, it would be perverse to design a tax system to push
?rms toward high leverage. However, this is what the current US tax system does.
Equity ?nancing is penalized relative to debt ?nancing. Income that passes through
the corporate sector into the hands of equity holders is doubly taxed – at the corporate
level, and then the personal level. Income destined for debt holders is singly taxed, at
the personal level.
The magnitude of the anti-equity bias of the tax system is immense. The corporate
income tax in the USA is 35 percent (for income over $181/3 million). This imposes a
massive tax penalty on ?rms that ?nance with equity. In effect, the government forces
businesses to be in the business of leveraged speculation.
In equilibrium, theory predicts that this pressure will lead to a big increase in the
equilibrium level of debt in the economy. In the simplest models such as the Modigliani
and Miller model with corporate taxes, ?rms choose 100 percent debt ?nancing, i.e. a
level of leverage high enough that all payout will occur as interest payments. In more
complex models there is still a big shift[8].
There are some counterarguments to the argument we have made about general
?rms:
.
The tax system already offers an offsetting bene?t to equity over debt: equity
holders can defer their taxes by waiting to realize capital gains.
However, in equilibrium in a steady state this does not create any tax bene?t for equity,
because for people to consume, resources have to move out of the corporate sector and
into the hands of individuals. This requires some kind of payout by the ?rm, not just
deferred capital gains:
Tax and
accounting
policy
133
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
.
Firms can help equity investors avoid paying personal taxes on dividends by
paying out through repurchase of shares. There are two arguments here. One is
valid, the other has only a minor degree of validity.
The valid argument is that in the USA the capital gains rate is lower than the personal
income tax rate, and in some countries is zero.
The argument with only a slight degree of validity is that repurchase is more
attractive than dividends because the tax in a repurchase is only on the gain, not the
full amount of the repurchase. To see why this has a bit of validity, think of an
all-equity ?rm that is founded with an investment of $100, and makes pro?ts of $5
per year forever which it pays out. It can throw pro?ts off as dividends, or it can
continually repurchase $5 worth of shares each year[9]. Since the ?rm is not growing in
value, realized capital gains are always zero. So repurchase allows equity investors to
completely avoid personal taxation on their investment.
However, this argument is almost completely invalidated in a growing economy.
Suppose that a ?rm has a steady state growth path in which it makes 10 percent pro?ts
per year, and pays out half of them each year. Suppose the ?rm is originally ?nanced
with an investment of $100. Then over a long period of time dividends or repurchases
will rise to arbitrarily high levels. But the capital gains basis remains unchanged;
in any given repurchase, this will shield only some small fraction of the original
$100. Since the capital gains basis becomes a vanishingly small fraction of the dollar
value of the repurchase, it shields only a miniscule fraction of equity-holders’
investment income. So repurchases are taxed virtually as much as dividends.
It can be objected that when people buy and sell shares, the new owner gets a higher
tax basis, the purchase price. This is true, but is offset by the fact that the seller
immediately pays a capital gains tax on the price increase. So such trades do not reduce
the total tax paid in the repurchase scenario.
2.2 The effects of regulation and taxation on size of the corporate sector and ?rm scale
Afundamental regulation in modern capitalismcreates the limited liability corporation.
This creates a bias in favor of large-scale, and creates a corporate sector with
limited liability as distinct from the non-corporate sector. The regulatory creation of
the limited liability sector is presumably a solution to externality problems which make
it hard to spread risk and create large-scale enterprises. We do not consider in detail the
nature of these problems. But even taking as given that having a corporate sector and
large ?rms is desirable, in evaluating tax and regulatory policies, we still need to
take into account how they shift the size of the corporate sector and the scale of
individual ?rms.
More generally, existing tax and regulatory policy toward enterprise scale is
mixed. Limited liability encourages scale by making it feasible to have a broader
shareholder base, and too-big-to-fail implicit bailout policies subsidize size as well. On
the other hand, there are policies that discourage large-scale such as antitrust laws,
the progressive feature of the corporate income tax, and double taxation of corporate
income (which discourages incorporation)[10]. We argue that in evaluating alternative
regulations and policies, the effects on optimal enterprise scale, and thereby systemic
risk, should be taken into account.
JFEP
1,2
134
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
As discussed earlier, because of too-big-to-fail, large-scale of individual enterprises
can result in bailouts. This creates ex ante moral hazard problems in risk-taking that
exacerbate preparedness externalities. To the extent that the state (rightly or wrongly)
is implicitly committed to costly bailouts, there is a negatively externality in having
large ?rms, especially in the ?nancial sector.
A notable example of government promotion of large-scale is the creation of
gigantic government sponsored enterprises, Fannie Mae and Freddie Mac, to promote
home ownership. There are no obvious economies of scale that would require them to
be so large. This suggests that (if nothing else) they should be broken up.
Antitrust legislation is motivated by a negative externality of size, that high market
share can help a ?rm gain (temporary) monopoly power. Even in the absence of any
market power, large-scale creates a systemic risk externality. This is especially the
case for those ?nancial ?rms such as hedge funds that are in the business of taking
substantial risky bets. There has been little recognition that existing policy distorts the
scale of ?rms, and that such policy should be managed to reduce systemic risk. As far
as solutions go, the tools of antitrust (government imposed breakup of ?rms) seems too
blunt an instrument to use on private ?rms to address systemic risk. A less intrusive
approach is to modify the existing policies that create incentives for large-scale.
2.3 Solutions to anti-equity bias for general ?rms
Tax policy can be changed to avoid overleveraging. A number of different possible
changes in tax policies could reduce or eliminate the anti-equity/pro-debt bias. One is to
eliminate the capital gains tax. Indeed, many countries have a zero capital gains tax.
Firms would need to be permitted to use regular repurchase to perform regular payout.
Another solution might be to allow partial, equal deductibility at the corporate level
for both interest and dividends payout. For example, of every dollar paid as either
interest or dividend income, 50 cents (say) would be deductible. This would help level
the playing ?eld between debt and equity (though it would be an overcorrection in
favor of equity unless the capital gain rate were raised to match personal income tax
rate), and would result in a shift away from corporate leverage. The tax revenue could
be adjusted by calibrating the percentage deduction (instead of the arbitrary example
of 50 percent of the payout).
In addition to the distortion away from equity toward debt, double taxation also
discourages the net ?ow of capital into the corporate sector. Under the current system,
debt provides a back door route through which capital can ?ow into the corporate
sector without any double-taxation. This proposed remedy for the anti-equity bias, by
reducing deductibility of interest income would close this debt “back door.” So overall
we might expect it to further reduce investment in the corporate sector. This could be a
good or bad thing overall – again, we are not analyzing here the reasons for the
existence of limited liability and the existence of a corporate sector. From the viewpoint
of systemic risk, by making it harder to take advantage of limited liability, the scale of
?rms tends to be reduced, which will tend to reduce risk-taking based upon the
expectation of bailout.
On the other hand, the debt back door route is very costly to ?rms because of the
risk of ?nancial distress, and to managers because of the risk of being ?red. So even
though the back door route right now is a good deal from the tax viewpoint, it could be
Tax and
accounting
policy
135
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
that by making the equity route a better deal, the reform would in total increase
investment within the corporate sector.
There are other possible ways of address the tax system’s anti-equity bias. One is to
eliminate the penalty to equity ?nancing by making dividends tax deductible at the
corporate level. Or, more simply and with similar effect, eliminate the corporate income
tax. This solution solves the bias between debt and equity. Lower corporate
income taxes strongly encourages putting resources into the corporate sector, which
however, encourages larger scale of ?rms. The encouragement of larger scale could
increase systemic risk, though this would very likely be outweighed by the effect of
deleveraging.
In summary, double taxation has an immense hidden cost – systemic instability.
This cost can be reduced by changes in tax law to reduce its anti-equity/pro-debt bias.
Some possible changes are politically easier than others.
3. Leverage and regulation of ?nancial ?rms
In the aftermath of Bears Stearns’s demise there has been an increased interest in
regulating investment banks and hedge funds. The prospect that the Fed will bail out
?nancial institutions ex post creates a moral hazard problem in ex ante risk-taking.
This moral hazard problem exacerbates the ex ante negative preparedness externality.
If government is going to provide bailouts ex post, this suggests that the tax and
regulatory system should not be designed to bias ?rms toward greater risk-taking
ex ante.
It is well understood that high ?nancial leverage causes systemic propagation of a
?nancial crisis. Highly levered institutions that are hit by a negative shock to their
value are forced to sell assets. These ?re sales reduce the value of the assets that are
sold, further reducing the value of the selling institution as well as the values of other
highly levered institutions or portfolios that happen to hold similar assets. This will, in
turn, lead to additional forced sales, creating a systemic downward spiral. Such a
feedback loop is considered in Krishnamurthy (2000).
This feedback loop would be self-correcting if forced sales had no effect on the
fundamental value of the securities that were being sold. If there were no deterioration
of the fundamental values, then when the securities become suf?ciently cheap, better
capitalized individuals and institutions could pro?t by buying the cheap securities and
holding them until the crisis subsides.
However, there can be feedback from the ?nancial system to the real economy.
For example, a crisis in the mortgage market leads to less capital available for new
mortgages, which results in lower property prices. This affects the construction
industry, which in turn affects the aggregate economy, leading to a deterioration of the
fundamental value of the underlying real estate. Such negative feedbacks can explain
the poor economic performance of Japan in the 1990s.
Preparedness externalities and the systemic risk associated with high
leverage suggest that it is important to consider how regulatory policies affect the
?nancial structures of investment banks and hedge funds as well as commercial banks.
To evaluate leverage-in?uencing regulation, it is important to understand why
?nancial institutions tend to be so highly levered. For corporations like Bear Stearns,
Lehmann Brothers and Goldman Sachs, just as with non-?nancial ?rms, there is a clear
tax advantage associated with leverage. Given their lines of business, they require
JFEP
1,2
136
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
cheap capital to compete, and because of the tax deductibility of interest payments, the
USA. Tax code makes debt a cheap form of capital for public corporations.
To the extent that taxation is the source of leverage, the problem can be corrected by
allowing publicly held ?nancial institutions to organize as pass through investment
trusts, which pay no direct taxes and thus have no tax incentive to lever up. Examples
of such trusts include real estate investment trusts, royalty trusts, and mutual funds.
This, of course, would result in a reduction in tax revenues. However, the loss in tax
revenues would be small relative to the likely bene?ts of having lower leverage and a
more stable ?nancial system. Alternatively, as discussed for the case of non-?nancial
?rms, modi?cations to the tax code could be made that retain corporate taxation, but
level the playing ?eld between income paid out as interest versus dividends.
The tendency of ?nancial institutions to lever up does not seem to be driven solely
by tax considerations. For example, hedge funds do not have a tax advantage to higher
leverage. There are other possible explanations for leveraging in terms of agency
problems and misperceptions of investors. However, to the extent that leverage is a
problem, it is at least in principle possible for ?nancial contracts and institutions to
address it. More fundamentally, we believe that the key problem is that there are
negative preparedness externalities. A ?nancial institution that levers up takes into
account possible private costs of distress, but not the costs imposed on others in the
event of a general coordination failure. This results in socially super-optimal levels of
leverage.
4. Accounting policy, disclosure policy, and systemic risk
Accounting rules, such as requirements for marking-to-market of illiquid assets, and
tolerance toward off-balance sheet leverage, also contribute to preparedness or to
systemic risk. The effects of accounting rules on preparedness should be weighed in
designing accounting rules and regulation.
Accounting and disclosure policies emerge in part through a political process
involving regulators, politicians, voters, and lobbying activities. Owing to preparedness
externalities, ?rms will lobby for rules that confer private advantages upon them even
if this reduces preparedness of the overall ?nancial system. Similarly, to the extent
that disclosure policy is discretionary with the ?rm, it will tend to be chosen in a way that
results in suboptimal preparedness.
In some cases it is fairly clear how accounting rules affect preparedness. When
investors have limited attention, they neglect and misinterpret accounting information,
and neglect the strategic motives of managers to conceal adverse information
(Daniel et al., 2002; Hirshleifer and Teoh, 2003). For example, if investors have limited
attention, they may not discount suf?ciently for the presence of off-balance sheet
leverage, and therefore fail to suf?ciently impound the expected costs of ?nancial
distress. If so, this increases the incentive for ?rms to lever up, further contributing to
socially insuf?cient levels of preparedness.
Off-balance sheet ?nancing such as special purpose entities (SPEs) have been used
to obscure leverage or to circumvent regulatory requirements[11]. If investors have
limited attention, they may not be adequately skeptical about entities that are not
integrated into the balance sheet (Daniel et al., 2002). As a result, ?rms may not be
adequately penalized by the market for the risk of ?nancial distress that such entities
create[12]. This creates an incentive for excessive off-balance sheet leverage, which
Tax and
accounting
policy
137
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
adds to systemic risk. The externality costs of systemic risk therefore weigh in favor of
relatively tight restrictions to force integrated reporting instead of allowing
relationships with off-balance sheet entities.
Fair value accounting (marking to market) has received a great deal of criticism as
an alleged cause of the recent credit crisis. One of the issues of debate is whether fair
value accounting makes reporting more transparent or less so, and whether fair value
accounting induces systematic reporting biases. A second issue is whether, in the event
of crisis, greater transparency is a good or bad thing. A third issue (interrelated with
the ?rst two) is the effect of marking to market on the perceptions of irrational
investors. The desirability of requirements for marking-to-market of illiquid assets,
and their effects on ex ante preparedness, depends on the answers to these questions.
With respect to whether fair value accounting improves transparency, a key
problem is that it is hard to place a fair market value on an illiquid asset, and during
bad times assets tend to become much less liquid. Some assets will have to be marked
down severely not because their long-term fundamental value is low, but because a
distressed seller has been forced to unload the asset at an arti?cially low market price.
In other words, all ?rms that hold assets similar to that sold by a distressed seller will
suffer a similar devaluation. Therefore, mark to market accounting hardwires the
downside risks of one ?rm to other ?rms. If investors ?xate on accounting numbers
without regard to the fact that the long-run value of the illiquid assets is greater than
the short-term value, or if contracts in place are based on accounting numbers,
marking-to-market creates systemic risk.
This argument does not mean that fair value accounting deserves general
condemnation. For highly liquid assets this concern is much reduced. However,
moving down the scale toward less liquidity, at some point a line needs to be drawn at
which marking-to-market becomes marking to an unavailable or ?ctional price. On the
margin, the problem of systemic risk suggests placing this line more toward the side of
high liquidity.
With respect to the desirability of transparency when investors are irrational, even
if there is a good market price to mark to, marking-to-market causes ?rms’ accounting
measures to seem worse when ?nancial asset markets are doing badly simply by virtue
of re?ecting the bad news. If investors have limited attention, viable ?rms may be
revalued downward under marking-to-market than under an historical cost system
because of ?xation upon accounting numbers[13]. So fair value accounting can be said
to reduce ?rms’ preparedness for bad times.
Many other accounting rules affect transparency, especially when investors
with limited attention focus on earnings and are not able to fully “undo” the accounting
adjustments implied by different accounting rules or discretionary choices by
managers. Thus, the effects of accounting rules on preparedness and systemic risk
should be weighed in designing accounting rules and regulation.
If investors are rational in their assessments of accounting numbers, they will not
revalue a ?rm downward on average just because it has updated its accounting
numbers to fair market values. Consider a public state of the world in which asset
values have dropped. Investors do not know how much a given ?rm’s assets have
dropped, but their assessments are correct on average. Speci?cally, these assessments
already re?ect a good judgment of the losses incurred by the ?rm. If the ?rm then
reports new numbers based on fair market values, investors will sometimes be
JFEP
1,2
138
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
positively surprised and sometimes negatively surprised, but on average their
assessments are correct.
It might then be concluded that when investors are rational there is little effect of
having ?rms mark-to-market (assuming that there are meaningful market values to
mark to). However, this turns out to be incorrect. By providing investors with more
precise information about the status of the ?rm, marking-to-market can create ?nancial
market runs. For a general analysis of ?nancial market runs, Bernardo andWelch(2004).
We can consider creditors (or suppliers) to a distressed ?rm as engaged in creating a
public good. Each creditor can contribute to the survival of the ?rm by extending
further credit, making concessions, or forgiving part of the debt. However, this confers
a bene?t upon other creditors, so (if negotiation is imperfect) concessions will be too
small. Teoh (1997) analyzes the general effect of greater information disclosure on
equilibrium contributions in public goods games. She ?nds that bad news disclosures
can trigger complete collapse of cooperation, which in this context would mean
liquidation of the ?rm.
This does not mean that a policy of greater transparency or disclosure is, ex ante, a
bad thing. On the up side, a favorable disclosure can induce greater concessions and
greater success of the ?rm (because contributors believe that their concessions will not
be wasted). So, discussions in the popular business press about how marking to market
in some particular case contributed to ?rm failure are na? ¨ve in concluding that marking
to market must inherently be bad. Whether greater transparency is, on average, good
or bad depends on the speci?c shape of the payoff functions of the parties involved.
What this analysis does show, however, is that greater transparency can reduce
preparedness, and can set the ?rm up for a greater likelihood of ?nancial market runs.
The model indicates that for ?rms that are close to the edge of complete failure, the
balance of bene?ts can favor intransparency because of an asymmetry in the effects of
good versus bad news. A small amount of bad news may tip the ?rm to an equilibrium
in which creditors’ willingness to make concessions essentially ceases completely –
they just ?ght over the remains. On the upside, a small amount of good news only
increases concessions marginally. So overall, on average transparency reduces ?rm
value.
This suggests that even with rational investors, committing to a policy of
intransparency can be a means of preparedness. If so, regulatory policies that pressure
?rms into greater transparency, especially during crises periods, may have perverse
effects. Thus, the usual presumption of the desirability of transparency may need to be
quali?ed.
We do not wish to overstate the case, especially when fair value accounting is being
widely criticized in popular discussions in often-unsophisticated ways. There are also
well understood bene?ts to transparency for purposes of contracting and corporate
governance; the case of Enron, which was famous for intransparency even before
disaster hit is a case in point. We do not provide an overall assessment of the pros and
cons of transparency. We merely point out that it is possible to make a logical case that
a commitment to intransparency can help a ?rm be better prepared for ?nancial crisis.
5. Conclusion
We consider here an externality in the choice of ex ante risk management policies by
individuals and ?rms: they are concerned with private risk, not with their contribution
Tax and
accounting
policy
139
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
to systemic risk. This preparedness externality causes excessive risk-taking from a
social viewpoint. A general policy implication is that government should avoid
regulatory and tax policies that push individuals and ?rms toward greater risk-taking.
The recent credit crisis has focused attention on the ex post problems of leverage
and of having ?nancial ?rms that are “too big to fail,” but there has not been enough
attention paid to the incentives that create leverage and large ?rms in the ?rst place.
The anti-equity bias of the tax system encourages individuals and ?rms to borrow,
distorting the economy toward excessive leverage. Hedging can in principle help
reduce the effects on systemic risk, but does not solve the problem. By leveling the
playing ?eld of the tax system, the distortion away from high leverage and low
preparedness can be greatly reduced. We therefore suggest that discussion of policies
to average ?nancial crises should place more emphasis on taxation, and its effects on
leverage and systemic risk.
Existing government policy toward ?rm scale is more mixed. Overall, regulation
favors large ?rm scale through the creation of the limited liability corporation, but
various other regulations create biases in different directions. There is a moral hazard
problem in risk taking associated with the government’s tendency to bail out ?rms in
trouble; such bailouts are directed especially toward big ?rms. This moral hazard
problem further exacerbates preparedness externalities. Therefore, on the margin
regulatory policies that affect ?rm scale should take into account the adverse effect of
size on systemic risk.
Accounting rules, such as requirements for marking-to-market of illiquid assets,
and tolerance toward off-balance sheet leverage, also contribute to systemic risk. When
the asset market is so illiquid that there is not a good market price to mark to, fair value
accounting can add noise. Furthermore, investors with limited attention do not always
fully undo the effects of different accounting rules and choices. Since marking-
to-market will tend to mark down during a crisis, fair value accounting can exacerbate
crises, especially when there is feedback from market prices to the real economy. This
problem is exacerbated when contracts and regulations (such as bank reserve
requirements) are based on accounting numbers.
Furthermore, even when investors are rational, transparency can trigger ?nancial
market runs by reducing the average willingness of creditors to make concessions in
the event of distress. Na? ¨ve popular arguments along these lines implicitly focus only
on the adverse realizations in which marking to market leads to asset values that are
even worse than investors are expecting. However, even in a setting where creditors
are rational and on average foresee write-downs correctly, it can still be the case that
transparency increases the probability of ?rm failure.
The problem is that for a ?rm at the edge of disaster, there is an asymmetry
between the large loss in willingness of creditors to make concessions when a little
more bad news arrives, versus a relatively modest increase in willingness when a little
more good news arrives. It follows that a commitment to opacity either in reporting or
disclosure policies can help a ?rm be better prepared for a future crisis. We do not
claim that this consideration necessarily overrides the good reasons for transparency
in general. However, this shows that a logical case for greater opacity is possible to
make. More generally, we suggest that in designing accounting rules and regulation,
the effects on systemic risk should be considered.
JFEP
1,2
140
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
Notes
1. Such prospects are quite important. But concerns about excessive risk-taking were raised
even before the recent crisis caused a general upward reassessment in beliefs about the
willingness of the USA Government to engage in bailouts of ?nancial ?rms.
2. Theoretical modeling indicates that the effects of reputational concerns of managers on
risk-taking can vary depending on regulation of ?nancial reporting. Furthermore, the effects
of optimal contracting to deal with these concerns are subtle and in some cases can reverse
the direct effect of reputational concerns (Holmstrom and Ricart i Costa, 1986; DeMarzo and
Duf?e, 1995).
3. The concern for systemic risk is the source of the too-big-to-fail agency problem. Ex ante,
large or important entities such as Fannie Mae, Freddie Mac, or American International
Group have reason to expect to be bailed out in the event of problems, creating an implicit
insurance policy. Ex ante this insurance creates a moral hazard problem.
4. For example, the Fed expands money supply to try to provide softer landing after bubbles.
Government spending is increased in the hope of stimulating the economy, and new regulation
is created.
5. We will later argue, however, that transparency is a double-edged blade in its effects on
systemic risk.
6. Under full rationality, however, individuals would be free to dismiss those component
numbers that they regard as unduly noisy and adjust their assessments accordingly.
7. Discussion often focuses on the deductibility of interest as the driving force, but the anomaly
is really the non-deductibility of dividends. In general, payouts to all providers of resources
to the ?rm (including providers of capital) are deductible as expenses at the corporate level,
except for equity holders.
8. The Miller (1977) model of corporate and personal taxes has capital structure irrelevance at
the ?rm level, but the aggregate equilibrium level of debt is still determinate and increases
when the corporate tax rate increases.
9. If repurchases are performed so regularly that they are viewed as a substitute for dividends,
the internal revenue service can treat the repurchase as a dividend payout. For simplicity we
set this issue aside.
10. Furthermore, limited liability does not always encourage large-scale, since a ?rm that is
subject to lawsuit (e.g. tobacco companies) may have an incentive to focus to prevent lawsuit
from extracting rents from its less lawsuit-prone divisions.
11. SPEs have been replaced by variable interest entities with stricter accounting reporting rules
that require consolidation with the parent company.
12. Zhang (2008) ?nds that a new accounting regulation (?nancial accounting standards board
Interpretation No. 46) that forced consolidation of SPEs on ?nancial statements resulted in
?rms taking actions to decrease book leverage by replacing conventional debt with equity.
Furthermore, after the adoption of this change, on average the S&P credit ratings of ?rms
with SPEs deteriorated, suggesting that prior to consolidation credit agencies did not
adequately discount for the leverage implied by SPEs.
13. Furthermore, some contracts and regulations (such as bank reserve requirements) are
written based upon accounting numbers, so even if investors are sophisticated enough to see
through the accounting numbers, marking-to-market can exacerbate the troubles of
distressed ?rms during a downturn.
Tax and
accounting
policy
141
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
References
Auerbach, A.J., Devereux, M.P. and Simpson, H. (2008), “Taxing corporate income”, NBER
Working Papers No. 14494, National Bureau of Economic Research, Cambridge, MA.
Bernardo, A.E. and Welch, I. (2004), “Liquidity and ?nancial market runs”, Quarterly Journal of
Economics, Vol. 119 No. 1, pp. 135-58.
Daniel, K., Hirshleifer, D. and Teoh, S.H. (2002), “Investor psychology in capital markets:
evidence and policy implications”, Journal of Monetary Economics, Vol. 49 No. 1,
pp. 139-209.
DeMarzo, P. and Duf?e, D. (1995), “Corporate incentives for hedging and hedge accounting”,
Review of Financial Studies, Vol. 8, pp. 743-71.
Hirshleifer, D. and Teoh, S.H. (2003), “Limited attention, information disclosure, and ?nancial
reporting”, Journal of Accounting and Economics, Vol. 36 Nos 1/3, pp. 337-86.
Holmstro¨m, B. and Ricart i Costa, J.E. (1986), “Managerial incentives and capital management”,
Quarterly Journal of Economics, Vol. 101, pp. 835-60.
Kindleberger, C.P. (2005), Manias, Panics, and Crashes: A History of Financial Crises, 5th ed.,
Wiley, New York, NY.
Krishnamurthy, A. (2000), “Ampli?cation mechanisms in liquidity crises”, NBER Working Paper
No. w15040, Kellogg School of Management, Northwestern University, Evanston, IL.
Miller, M.H. (1977), “Debt and taxes”, Journal of Finance, Vol. 32, pp. 261-75.
Teoh, S.H. (1997), “Information disclosure and voluntary contributions to public goods”, RAND
Journal of Economics, Vol. 28 No. 3, pp. 385-406.
Titman, S. (1985), “The effect of forward markets on the debt-equity mix of investor portfolios
and the optimal capital structure of ?rms”, Journal of Financial and Quantitative Analysis,
March.
Tufano, P. (1996), “Who manages risk? An empirical examination of risk management practices
in the gold mining industry”, Journal of Finance, Vol. 51 No. 4, pp. 1097-137.
Zhang, J.Li. (2008), “Economic consequences of recognizing off-balance sheet activities”,
Financial Accounting and Reporting Section (FARS), available at SSRN: http://ssrn.com/
abstract¼1266456 (accessed September 11, 2008).
Corresponding author
Siew Hong Teoh can be contacted at: [email protected]
JFEP
1,2
142
To purchase reprints of this article please e-mail: [email protected]
Or visit our web site for further details: www.emeraldinsight.com/reprints
D
o
w
n
l
o
a
d
e
d
b
y
P
O
N
D
I
C
H
E
R
R
Y
U
N
I
V
E
R
S
I
T
Y
A
t
2
1
:
3
5
2
4
J
a
n
u
a
r
y
2
0
1
6
(
P
T
)
doc_979299524.pdf