Description
A topic of recent interest in accounting research has been the investigation of the role of
fair value accounting (FVA) in the global financial crisis. This research focused on finding
a link during the crisis time-period and often states that ‘‘accounting is only a messenger’’.
The model presented in this paper emphasises finding the link before the crisis and
‘‘accounting as money.’’ Use is made of an accounting model of the economy due to the
inability of standard models of monetary transmission to incorporate global financial crisis
characteristics such as feedback effects, systemic risk and the centrality of the financial
sector in the crisis. The model shows FVA in banks to be an accelerator that amplifies
the financial cycle upswing. Feedback effects noted in the model include changes in the
demand for financial instruments and changes in demand in the real economy.
Minsky-like, crisis is shown to be endogenous to the model, working through the fragility
of balance sheets in the real sector as well as in the financial sector. Bank balance sheet
fragility is caused by bad capital driving out good capital, banks reaching for yield and
the inversion of the yield curve.
Fair value accounting, fragile bank balance sheets and crisis:
A model
Phillip de Jager
?
Department of Finance & Tax and the African Collaboration for Quantitative Finance & Risk Research, University of Cape Town, South Africa
a b s t r a c t
A topic of recent interest in accounting research has been the investigation of the role of
fair value accounting (FVA) in the global ?nancial crisis. This research focused on ?nding
a link during the crisis time-period and often states that ‘‘accounting is only a messenger’’.
The model presented in this paper emphasises ?nding the link before the crisis and
‘‘accounting as money.’’ Use is made of an accounting model of the economy due to the
inability of standard models of monetary transmission to incorporate global ?nancial crisis
characteristics such as feedback effects, systemic risk and the centrality of the ?nancial
sector in the crisis. The model shows FVA in banks to be an accelerator that ampli?es
the ?nancial cycle upswing. Feedback effects noted in the model include changes in the
demand for ?nancial instruments and changes in demand in the real economy.
Minsky-like, crisis is shown to be endogenous to the model, working through the fragility
of balance sheets in the real sector as well as in the ?nancial sector. Bank balance sheet
fragility is caused by bad capital driving out good capital, banks reaching for yield and
the inversion of the yield curve. The model shows that the practice of not meeting rising
credit demand with increasing credit supply is an essential control mechanism in the
?nancial cycle.
Ó 2014 Elsevier Ltd. All rights reserved.
Introduction
In 2012, more than four years after the Lehman Broth-
ers failure, the aftermath of the global ?nancial crisis was
still felt, and stable, self-sustaining growth continued to
elude the world economy (Bank for International Settle-
ments, 2012:ix). Ryan (2008:1606) identi?ed the global
?nancial crisis as the ‘‘signal researchable-teachable
moment of my two-decade-plus career as an accounting
academic’’. Numerous authors have started to investigate
the link between accounting (usually FVA) and the crisis
(Badertscher, Burks, & Easton, 2012; Ball, 2008; Barth &
Landsman, 2010; Laux & Leuz, 2009, 2010; Magnan,
2009; Plantin, Sapra, & Shin, 2008; Pozen, 2009; Shaffer,
2010; United States Securities, 2008; Veron, 2008; Wal-
lace, 2009). These studies tend to focus on establishing
the link between FVA and the crisis during the time-period
of the crisis.
1
It is the argument of this paper that the focus
should rather be on the time-period before the crisis (the
upswing); this corresponds with one of Pinnuck’s (2012:5)
conclusions in his review of the role of ?nancial reporting
in the crisis that ‘‘the existing debate has focused on the role
of FVA during the crisis, but has ignored the possibility that
the illusory FV gains on subprime securities before the crisis
may have masked some of the underlying problems’’. Inhttp://dx.doi.org/10.1016/j.aos.2014.01.004
0361-3682/Ó 2014 Elsevier Ltd. All rights reserved.
?
Tel.: +27 216502296.
E-mail address: [email protected]
1
Readers are referred to Section ‘The global ?nancial crisis’ in the
literature review where the few arguments for FVA’s role during the
upswing are discussed.
Accounting, Organizations and Society 39 (2014) 97–116
Contents lists available at ScienceDirect
Accounting, Organizations and Society
j our nal homepage: www. el sevi er. com/ l ocat e/ aos
addition, some of the studies (Ball, 2008; Wallace, 2009)
argued that accounting was only a messenger and thus
should not be linked to the crisis. A number of the remaining
studies (Laux & Leuz, 2009:826; Magnan, 2009:191; Pozen,
2009:86; United States Securities and Exchange Commis-
sion, 2009:2) mention this pervasive argument; this paper
argues against this view.
The primary objective of this paper is to provide a
reasonable alternative perspective on the relationship be-
tween FVA and the global ?nancial crisis; a perspective
that focuses on ?nding the link before the crisis (during
the upswing) and one that reminds accountants that in
banking, accounting is more than just a messenger:
when bank deposits are created by accounting entries,
accounting is money. To this end, a model will be devel-
oped in this paper that demonstrates the link between
accounting and bank capital regulations and helps to
aid understanding of the global ?nancial crisis. The paper
answers the call by Arnold (2009:806) to ‘‘stimulate a
revival of accounting scholarship aimed at understanding
the relationship between accounting practice and the
macro political and economic environment in which it
operates’’.
The nexus of the model to be developed will be the im-
pact of FVA on the regulatory capital of banks. Not all FVA
entries impact banks’ regulatory capital. The general prin-
ciple in Basel II is that Tier 1 capital should be equity cap-
ital made up from permanent shareholders’ equity and
disclosed reserves that, importantly, must have been
posted through pro?t and loss (Bank for International
Settlements, 2006:245). Thus, only FVA entries related to
trading instruments or instruments designated at fair
value are relevant for the model when values increase.
FVA increases in the value of available-for-sale instru-
ments are not covered because those increases would be
excluded from regulatory capital. When values decrease,
other-than-temporary-impairments of available-for-sale
instruments become relevant as these are posted through
pro?t and loss. Shaffer (2010) investigated the role of FVA
in the global ?nancial crisis by looking only at FVA for
available-for-sale instruments. He argued for the exclu-
sion of FVA entries emanating from the trading book be-
cause ‘‘There are few who would argue that fair value is
inappropriate for measuring investments held for trading
purposes where deep and active markets exist’’ (Shaffer,
2010:10). However, if most instruments held for trading
purposes were traded on deep and active markets, then
they would have been shown at level 1 (mark-to-market)
on the fair value hierarchy. The United States Securities
and Exchange Commission (2008:62) found that in the
case of banks, most derivatives and trading assets were
at level 2 and not at level 1; this was especially true for
investment banks where almost all derivatives and trad-
ing assets were at level 2 and not level 1. The conclusion
is thus that trading assets and derivatives were not traded
on deep and active markets and that the exclusion of FVA
adjustments on those instruments was not justi?ed. In
addition to the exclusion of FVA entries emanating from
the trading book, Shaffer (2010:10) also argued for the
exclusion of FVA entries emanating from ‘‘designated at
fair value’’ instruments as this designation was ‘‘an
explicit decision by management to value these assets this
way and thus they are not included in the analysis’’.
The implicit assumption that management will not abuse
the fair value option is questionable. An example of the
impact of the preceding arguments will help to illustrate
this. Assume, therefore, that Bank X buys piecemeal $1
billion of AAA rated subprime securities yielding 5% whilst
‘‘normal’’ AAA securities yield 3%
2
with an assumed matu-
rity of 10 years (monthly). The portfolio of securities is then
designated at fair value and carried at level 2 of the fair
value hierarchy. The ?nancial result of the foregoing is an
immediate gain of $98 million (payments on $1 billion at
5% interest for 10 years present valued at 3% less the cost
of $1 billion) posted to net pro?t. Badertscher et al.
(2012) and Barth and Landsman (2010) continued this
exclusive focus on FVA for available-for-sale instruments
as the only possible avenue through which FVA could have
played a role in the global ?nancial crisis. In contrast, for
the reasons explained above, the model in this study is
based on the assumption that FVA materially impacts pro?t
and regulatory capital; this is one of the contributions that
this study makes to the accounting literature. Evidence for
the signi?cance of FVA entries that affect pro?t and loss can
be found in the study by the United States Securities and
Exchange Commission (2008:4) which observed that ‘‘fair
value measurements did signi?cantly affect ?nancial insti-
tutions’ reported income’’.
In addition, this study contributes to accounting liter-
ature in the following speci?c ways. Thus, several writers
have argued that FVA is procyclical (Enria et al., 2004:45;
Laux & Leuz, 2009:828); their views are repeated in this
paper but this time within a formal framework and
nuanced with the argument that the role of FVA during
a downturn is likely to be blunted by the ?exibility al-
lowed in FVA regulations. An implication of the foregoing
is that the role of FVA in the global ?nancial crisis should
rather be sought before or after the crisis and not during
the crisis. The model developed explains why McMahon
(2011:54) is accurate when stating ‘‘Mark-to-market feed-
back loops are not exclusive to declining periods – feed-
back loops also occur during upswings, although not
quite as fast’’ (this paper’s emphasis). The model shows
that the seeds for the coming downturn are sown during
the upswing with the purest role for FVA being the
replacement in bank capital of liquid assets with unrea-
lised (and thus risky) FVA gains; bad capital driving out
good capital. The simple running example used to demon-
strate the model that follows shows how a small change
in the value of fair valued assets can lead to a much larger
effect (a factor of 16 in the example but 30 is not uncom-
mon) on a bank’s balance sheet; this contradicts the argu-
ment that the fair value measurement of ?nancial
instruments must be pervasive to be dangerous (Geb-
hardt, 2012:267&271; United States Securities & Ex-
change Commission, 2008:4).
2
Even though instruments derived from securitised subprime mortgages
were often rated AAA (investment grade) by rating agencies prior to the
crisis, those instruments continued to offer higher returns than ‘‘normal’’
AAA securities (Lo, 2012:152).
98 P. de Jager / Accounting, Organizations and Society 39 (2014) 97–116
The model also makes a potential contribution to mon-
etary economics. Hyman Minsky’s ?nancial instability
hypothesis bene?ted from the global ?nancial crisis in
the sense that many saw the crisis as con?rming his pre-
dictions (see for example Krugman, 2009; Lahart, 2007;
Wolf, 2008). Minsky’s ‘‘notion of the ?rmas a balance sheet
of assets and liabilities, as opposed to the [traditional] no-
tion of the ?rm as an entrepreneur making production
decisions’’ (Toporowski, 2008:730) can be linked to the
type of model developed in this paper; a balance sheet cen-
tred model. The argument that fragility develops in the
balance sheets of the real economy is repeated in the mod-
el described in this paper, and is similar to the way the ‘‘ba-
sic Minsky cycle concerns the evolution of patterns of
?nancing arrangements and it captures the phenomenon
of emerging ?nancial fragility in business and household
balance sheets’’ (Palley, 2010:30). The model described in
this paper is different in the way the focus is not on this
fragility of individuals’ or businesses’ balance sheets but
rather on how fragility develops in the balance sheets of
banks. This aligns with the ‘‘complete reversal of the tradi-
tional monetary circuit, where the banking system is as-
sumed to ?nance business sector activity’’ and ‘‘it is the
dynamics of the banks/?nancial markets axis [. . .] which
has taken center stage’’ (Bello?ore & Passarella, 2010:10).
A second area where a contribution to monetary eco-
nomics is possible is in the presentation and exploration
of the impact of FVA to an audience of economists. Godley
and Lavoie (2012: Preface xxxvii) write that by not under-
standing in?ation accounting in the 1970s, economists
underestimated ‘‘the extent to which stocks of ?nancial as-
sets would rise in nominal terms’’ resulting in ‘‘some bad
projections’’. Similarly FVA can result in a signi?cant in-
crease in the nominal values of ?nancial assets during an
upswing. The model developed in this paper also has
important links with a recent paper by Claudio Borio, the
Deputy Head of the Monetary and Economic Department
and Director of Research and Statistics at the Bank for Inter-
national Settlements (Borio, 2012). He calls for the study of
the ‘‘?nancial cycle’’ as part of macroeconomics (Borio,
2012:1) and adds that a ‘‘?nancial boom should not just
precede the bust but cause it’’ (Borio, 2012:8); this call ties
in with this paper’s focus on the time-period before the
crisis. He reminds us that ‘‘the ?nancial system does not
just allocate, but also generates, purchasing power’’ (Borio,
2012:2) which ties back to accounting as money.
Bezemer (2010:676) referred pointed out that
‘‘‘accounting’ (or ?ow-of-funds) models of the economy
are the shared mindset of those analysts who worried
about a credit-cum-debt crisis followed by recession,
before the policy and academic establishment did. They
are ‘accounting’ models in the sense that they represent
households’, ?rms’ and governments’ balance sheets and
their interrelations.’’ Accounting models are also used in
two other streams of literature related to this paper.
Macroeconomic accounting (national accounting) is
argued to be both a measure of macroeconomic reality
and a constituent of that reality (Suzuki, 2003). Similarly
in this paper accounting both measures the extent of bank
balance sheets and indirectly impacts those balance
sheets with its effect on the money supply. Andersson,
Lee, Theodosopoulos, Yin, and Haslam (2014:79), within
the literature on the ?nancialization of the economy, argue
that ‘‘this process of ?nancialization can best be under-
stood within an augmented accounting framework . . .’’
Other reasons why a stock-?ow consistent accounting
framework of the economy
3
was chosen as a foundation
for the model developed in this paper include the role of bal-
ance sheets in past crises and the need to incorporate global
?nancial crisis characteristics such as feedback effects,
systemic risk and the centrality of the ?nancial sector in
the model. This choice of model is in contrast to the more
mainstream theories of the monetary transmission mecha-
nism. A difference between the typical stock-?ow consistent
accounting model and the model described in this paper is
that the supply of credit is not assumed simply to be a re-
sponse to the demand for credit; instead, this paper starts
with the supply of credit and its in?uence on the demand
for it. In this way it aims to be aligned better with what
was observed during the global ?nancial crisis; that is, an
oversupply of credit before the crisis and an undersupply
afterwards.
The model developed in this paper describes and ex-
plains many of the characteristics of the global ?nancial
crisis as well as aspects of other past crises. It is shown
how the change in banks’ capital ratios that result from
FVA can set in motion a procyclical process with the initial
aim being the restoration of the earlier capital ratio. A
debt-?nanced expansion of bank balance sheets and/or
the payment of remuneration and dividends results from
increases in fair value. The shrinkage of bank balance
sheets can result from a decrease in fair value but, argu-
ably, can be blunted by management action and the ?exi-
bility in FVA. Feedback effects in the model maintain the
momentum in the system once in motion. The feedback ef-
fects are the result of the additional (or decreased) demand
for ?nancial assets generated as well as the additional (or
decreased) spending power made available to the real
economy. In the instance where the additional spending
power (or decrease in spending power, as the case may
be) is made available solely to the banking sector the feed-
back effects are strongest. The model also highlights the
role of four processes that can result in a crisis/regime
change. The ?rst of these is the increased fragility of real
economy balance sheets during an expansion. The second
and third relate to two sources of bank balance sheet fra-
gility; bad capital driving out good capital and the reach
for yield. Lastly, interbank transactions lead to a yield
curve inversion effect, depriving banks of the easy pro?ts
available from maturity transformation; causing more
reach for yield. The crisis attributes explained by the model
include the booms in predatory lending and also the pay-
ment of dividends and remuneration during the upswing,
understanding the growth in the ?nancial sector relative
to the real sector, illuminating the temporary impact of
the reprieve obtained from FVA during the downturn and
explaining the occurrence of the Great Moderation. In the
3
According to Bello?ore and Passarella (2010:7) these models are
‘‘dynamic, consider the effects of stock magnitudes on ?ow variables, and
explicitly represent the role of the banking system’’. The model developed
in this paper can be described similarly.
P. de Jager / Accounting, Organizations and Society 39 (2014) 97–116 99
?nal discussion the bene?t of the perspective offered by
the model in this paper will brie?y be illustrated by using
it to understand the results of another study by Cabral
(2012) on banks and the global ?nancial crisis.
The paper proceeds as follows: Section ‘The historic role
of FVA in crises’ sets the scene by presenting FVA and
brie?y discussing the role of FVA in previous crises; Sec-
tion ‘Monetary transmission mechanism’ positions the
model developed within the monetary economics litera-
ture; Section ‘Model development’ develops and discusses
the model proposed; Section ‘Empirical example’ provides
an empirical example where the model is used to interpret
the information; Section ‘Implications, shortcomings and
opportunities for further research’ discusses the implica-
tions, shortcomings and opportunities for further research
resulting from the model and Section ‘Conclusion’ provides
the conclusion.
The historic role of FVA in crises
The purpose of this section is to overview the impact of
FVA during past crises. The information obtained will be
useful to benchmark the reasonableness of the model that
will be developed later in this paper. The crises included in
this overview are the Great Depression, the savings and
loans (S&L) crisis, the Enron crisis and the global ?nancial
crisis. But, ?rst it is necessary to de?ne FVA, to differentiate
it from mark-to-market accounting and to consider care-
fully how FVA impacts bank regulatory capital.
FVA is more complex than mark-to-market accounting.
Fair value is the exchange value in an idealised market and
can be determined in three ways, in order of preference
4
(the de?nitions that follow are based on the annual ?nancial
statements of Investec Bank Limited (Investec Bank Limited,
2012:148):
Mark-to-market accounting (level 1): Quoted (unad-
justed) prices in active markets for identical
instruments.
Mix of mark-to-market accounting and modelling (level
2): Model inputs other than quoted prices included
within level 1 that are observable for the instruments,
either directly (i.e. as prices) or indirectly (i.e. derived
from prices). This paper interprets this to mean the
modelling of inputs that are publically observable.
Mark-to-model (level 3): Signi?cant model inputs are not
based on observable market data.
FVA is thus more complicated than mark-to-market
accounting, but the two are closely related. For brevity this
paper will treat the two as equivalent.
The Great Depression
According to Scott (2009:3) the frequent upwards reval-
uation of capital assets was one of the practices of the
1920s to receive criticism after the crisis. This advanced
the case for historical cost accounting. Zeff (2007) concurs
and writes that the United States of America was a bastion
of predominantly historical cost accounting for inventories
and ?xed assets from 1934 to the 1970s. Pressure from the
Securities and Exchange Commission was a driving force
behind this insistence on historical cost accounting.
The use of mark-to-market accounting in banks during
the Great Depression can be found in the work of Friedman
and Schwartz (1971: chap. 7, 100–102). They describe and
provide evidence of how the fall in the market value of
bond portfolios was the most important source of bank
capital impairment during the Great Depression. They also
point out that, paradoxically, the most liquid assets were
the most serious threat to bank solvency due to the
mark-to-market regime for liquid assets rather than bonds
for ‘‘which there was no good market and few quotations’’
carried at cost. The United States Securities and Exchange
Commission (2008:34) wrote that mark-to-market
accounting for banks was changed in 1938 for reasons that
included the impact of mark-to-market accounting on
bank behaviour: ‘‘Further, prior to 1938, banking organiza-
tions were required for supervisory purposes to use market
value accounting for their investment securities portfolios.
Serious concerns on the part of the U.S. Treasury and the
bank regulators over how this affected the banks’ ?nancial
performance and investment decisions led the agencies to
abandon in that year the use of this accounting concept for
supervisory purposes’’.
Financial economist Fisher (1933) wrote that over-
indebtedness and de?ation were the two dominant factors
in all the great booms and depressions. He then proceeded
to describe a chain of consequences in a depression with
numerous feedback mechanisms driving distressed selling
of assets.
The ?ndings from the Great Depression indicate a role
for mark-to-market accounting in banks before and during
the crisis. The upward valuation of selected ?nancial
instruments during the upswing could have made banks
appear healthier, contributing to the over indebtedness of
banks and individuals. Moreover, the mark-to-market
accounting of liquid bond portfolios was clearly shown
by Friedman and Schwartz (1971) to have a negative effect
on bank lending during the downturn. Both of these effects
involve feedback processes.
The savings and loans crisis
The savings and loans (S&L) crisis of the 1980s and
1990s was caused by the failure of a signi?cant proportion
of the savings and loan associations in the United States of
America. A savings and loan association or ‘‘thrift’’ is a
?nancial institution that accepts deposits and makes mort-
gage, car and other personal loans to individual members.
According to Black (2005:xiii) ‘‘a wave of control fraud
ravaged the S&L industry’’. In a later publication he writes:
‘‘Control fraud occurs when the executives at a seemingly
legitimate ?rm use their control to loot the ?rm and its
shareholders and creditors’’ (Black, 2011). Accounting is
strongly linked by Black to control fraud: ‘‘Control frauds,
using accounting fraud as their primary weapon and shield,
4
A formal order of preference only entered accounting standards under
the International Accounting Standards Board (IASB) in October 2008 with
an amendment to International Accounting Standard (IAS) 39 which
adopted the American tiered system of fair value determination.
100 P. de Jager / Accounting, Organizations and Society 39 (2014) 97–116
typically report sensational pro?ts, followed by cata-
strophic failure. These ?ctitious pro?ts provide the means
for sophisticated, fraudulent CEOs to use common corpo-
rate mechanisms such as stock bonuses to convert ?rm
assets to their personal bene?t’’ (Black, 2005:xiv). Akerlof
and Romer (1993:2) ascribe what happened to looting;
‘‘to go broke for pro?t at society’s expense’’. They continue:
‘‘Bankruptcy for pro?t will occur if poor accounting, lax
regulation, or lowpenalties for abuse give owners an incen-
tive to pay themselves more than their ?rms are worth and
then default. . .’’ It is clear that accounting is important in
both instances to justify high corporate pay-outs.
The accounting profession focused attention on another
aspect of the crisis. Most thrifts extended loans on a ?xed
cost basis and obtained capital on a variable cost basis.
When interest rates rose in the 1980s, the industry found
itself in a dif?cult position: assets were generating less re-
turn than the cost of ?nancing them. Accountants argued
that historical cost accounting was hiding the bankruptcy
of the thrifts and that the fair valuation of the thrifts’ assets
would have revealed these problems (Young, 1995:72–75).
The ?ndings from the S&L crisis indicate a role for
accounting in thrifts before and during the crisis. Fictitious
pro?ts can be used to make an institution seem healthier
and to justify unreasonable pay-outs. During the downturn
FVA could have contributed towards revealing the bank-
ruptcy of the thrifts.
The Enron crisis
The Enron crisis was not nearly as systemic or impor-
tant as the other crises. Nevertheless, it does deserve a
mention in relation to the role of FVA in that crisis. Benston
(2006:465) ?nds that FVA was ‘‘substantially responsible’’
for the demise of Enron. FVA in Enron was used ‘‘opportu-
nistically to in?ate reported net income’’ and rationalise
excessive remuneration (Benston, 2006:466). FVA rarely
reduced income as ‘‘contrary to the way fair-value
accounting should be used, reductions in value rarely were
recognized and recorded because they either were ignored
or were assumed to be temporary.’’ (Benston, 2006:466).
FVA is also mentioned in another American scandal during
the same time period. Einhorn (2010:398) writes ‘‘. . .as the
crisis unfolded (it became apparent) that Allied’s abuse of
fair-value accounting was more prevalent in corporate
America than I had realized’’ in his telling of the Allied
scandal that he was involved in during the 2000s.
The global ?nancial crisis
The global ?nancial crisis was triggered by a sharp fall
in house prices in the United States of America (Posner,
2009:vii). It grew from the subprime crisis, into the credit
crisis, then into a ?nancial crisis and ?nally into a global
?nancial crisis. There are opinions that accounting, espe-
cially FVA, played a key role in causing or at least in wors-
ening the crisis. At the extreme end of this viewpoint is the
opinion held by Steve Forbes, chairman of Forbes Media,
that mark-to-market accounting was the ‘‘principal
reason’’ that the ?nancial system melted down in 2008
(Pozen, 2009:85). It was mentioned in the introduction
that most accounting research focused on the role of FVA
during the crisis and found limited evidence of a signi?cant
role at that time. This is to be expected as FVA contains at
least two built-in mechanisms by which banks can avoid
recognising losses; see Section ‘Implications of the basic
model’ for a detailed description.
Theoretical models by Allen and Carletti (2008) and by
Plantin et al. (2008) examined the impact of mark-to-
market accounting on capital markets and show that
mark-to-market accounting worsens a downturn. Their
studies have a shortcoming when attempting to explain
the global ?nancial crisis; mark-to-market accounting is
not the same as FVA and thus managers can use the ?exi-
bility inherent in fair value accounting to not accept mar-
ket values that they believe are incorrect. This is argued
by Laux and Leuz (2009:827) who also acknowledge that
FVA can be procyclical both during the boom and the bust
and that behavioural issues can make implementation of
FVA dif?cult. They argue that the increase in pro?t and
capital from asset value increases under FVA is also avail-
able as ‘‘gains trading’’ under historical cost accounting
and thus that the two types of accounting are similar dur-
ing the upswing, but this argument is not exactly correct.
FVA is a much faster process than gains trading that is
inherently lumpy. Bonuses and dividends paid from gains
trading are paid from increased cash resources whilst bo-
nuses and dividends from FVA gains only decrease cash
resources. The implication is that FVA is more able than
gains trading to be at the centre of a feedback process
and that FVA during a boom can leave the banking system
with reduced cash resources.
A role for FVA during the upswing is less clear from
accounting research but evidence from before the crisis is
starting to appear. Livne, Markarian, and Milne (2011) ?nd
that FVA gains are remuneration relevant. In the popular
press it seems to be common knowledge that FVA allowed
the payment of remuneration and dividends from imagi-
nary
5
pro?ts during the boom (Haldane, 2011; Kay, 2009,
2012; Mundy, 2012; Taylor, 2009; Wood, 2010). The effect
of such payments would be to weaken the banking system
as risk-free capital in the form of liquid assets left the banks
to be replaced by risky capital (potentially temporary value
increases). McMahon (2011) proposes that FVA plays an
integral role in feedback loops behind excessive lending
(and remuneration) during the boom and a lack of lending
during the bust. Magnan (2009:207) argues that a feedback
loop exists between FVA and market prices working through
the information channel. Khan (2010:5) does not ?nd sup-
port that the information channel is behind ‘‘the increased
bank contagion during crises associated with a more fair
value-oriented reporting regime’’. The focus of this paper is
on the money channel.
Conclusion from review of crises
One factor common to all of the four crises reviewed
was the role of ?rm (including banks) balance sheets. The
5
Terms used include ‘‘imaginary pro?ts’’, ‘‘spurious pro?ts’’ and ‘‘unre-
liable gains’’.
P. de Jager / Accounting, Organizations and Society 39 (2014) 97–116 101
crises have indicated a role for FVA both before and after
each crisis. Before a crisis, FVA can lead to ?rms or banks
appearing healthier than they are (Enron is a prime exam-
ple) and can thus facilitate further asset expansion
?nanced by debt (a feedback process). Based on what hap-
pened during the S&L crisis it could be expected that FVA
would reveal problems faster than historical cost account-
ing. But this revelation by FVA during the downturn can
lead to a downward spiral with forced sales of assets and
shrinking balance sheets (another feedback process);
alternatively this might not happen at all, as argued later
in ‘Implications of the basic model’.
Monetary transmission mechanism
The model to be developed inter alia describes the inter-
action between the real and the ?nancial economy. This
section starts by describing the mainstream explanations
for how money affects the real economy. Thereafter, argu-
ments are presented for a natural progression to Post-Key-
nesian stock-?ow consistent accounting models of the
economy. Similarities and differences between the model
developed in this paper and the standard stock-?ow con-
sistent accounting framework will round off this section.
According to Mankiw (2007:83) the quantity theory of
money provides the leading explanation of how money af-
fects the economy in the long run. According to this theory
the money supply has a direct, proportional relationship to
the price level. In the long run ‘‘monetary neutrality is
approximately correct’’ (Mankiw, 2007:109). The main dif-
?culty with this approach is that central banks cannot and
do not attempt directly to control the rate of growth of the
money supply (Howelss & Bain, 2008:273). A more realistic
approach should start with short-term interest rates as a
policy instrument. Howelss and Bain (2008:273) refer to
the interest rate control approach as an example of endog-
enous money and explain that money supply, via bank ac-
tions, adjusts to credit demand. The latter approach is
more realistic and, importantly, places commercial banks
in the middle of the money supply process, but the neutral-
ity of money assumption is maintained.
The neutrality of money argument is dif?cult to sustain
in the short run (Mankiw, 2007:536) and arguably even in
the long run when one considers that both the Great
Depression and the global ?nancial crisis started in the
?nancial sector and thereafter engulfed the real economy
for several years. In addition, accounting-orientated aca-
demic papers were published, after the global ?nancial cri-
sis, which encouraged researchers to ?nd links between
accounting and macroeconomic outcomes (Arnold, 2009;
Bezemer, 2010). The paper by Bezemer, amongst other
things, introduced accounting (or stock-?ow consistent)
macroeconomic models.
Authors such as Keen (2011) and Godley and Lavoie
(2012) present explicit accounting models of the economy
where accounting identities (not the equilibrium concept)
are the determinants of model outcomes in response to
shocks in the environment or in policy (Bezemer,
2010:679). In addition to this obvious link to accounting,
there are other reasons why this type of model forms a
?rm foundation for the development of a model of bank
behaviour, described in this paper, that links FVA, bank
capital regulation, management behaviour and real eco-
nomic outcomes. These models are not static or based on
equilibrium and can thus accommodate certain de?ning
characteristics of the global ?nancial crisis, feedback ef-
fects (Bank for International Settlements, 2012; McMahon,
2011; Soros, 2012) and the related concept of too-big-to-
fail banks (Haldane, 2011). The models are also not built
up from individual behaviour and summated to systemic
behaviour (microfoundations); they can therefore incorpo-
rate, easily, emergent properties such as systemic risk.
Finally, the models emphasise banks (and the provision
of credit) (Godley & Lavoie, 2012:17; Keen, 2011) which
is the sector most impacted by FVA.
Another author who emphasised the role of banks in
the economy and who rapidly gained prominence after
the global ?nancial crisis was Hyman Minsky (Bello?ore
& Passarella, 2010; Keen, 2011). Minsky’s Financial Insta-
bility Hypothesis is based on the idea that in times of
?nancial stability the actors in the system and the system
itself move towards ?nancial fragility. His focus on balance
sheets and the idea that the key to the downturn is to be
found during the preceding upswing will be incorporated
into the suggested model. A typical schematic overview
of a stock-?ow consistent model is shown in Fig. 1.
Fig. 1. Typical stock-?ow consistent model of the economy. Source: Bezemer (2010:683)
102 P. de Jager / Accounting, Organizations and Society 39 (2014) 97–116
The model to be presented in this paper is much simpler
and its advantage is that the concepts under investigation
are emphasised. An arguably negative impact is that the
model is a partial one in the sense that it does not describe
everything about a modern economy. However, this partial
perspective does meet the Friedman signi?cance level of
‘‘explain much by little’’ (Friedman, 1953:14). In the model
the real economy including consumers, producers and
government, is collapsed into one and the ?nancial sector
consists of a combination of two banks (see Fig. 2).
The model developed and described below is Post-Key-
nesian in the sense that the principle of effective demand is
assumed to hold in the real markets; that is, markets do
not automatically clear. An important difference from the
typical ?ow-of-funds model is that credit supply is not as-
sumed just to be a response to credit demand (endogenous
money supply) (Godley & Lavoie, 2012:127–128). On the
contrary, bank behaviour is assumed to in?uence credit
supply. This assumption is consistent with what was found
during and after the global ?nancial crisis; an oversupply
of credit before the crisis and an undersupply after the
crisis (Bank for International Settlements, 2012; Bank of
England, 2009:43:ix&1; International Monetary Fund,
2012:xi&12; The Independent Commission on Banking,
2011:8&16). This primacy of credit supply is also sup-
ported by a letter from Dr. John Whiteman in the Financial
Times (Whiteman, 2012) where he states that ‘‘. . .borrow-
ing is not driven by the price of credit (as common sense
might assume), but rather by the sheer availability of cred-
it in the ?rst place’’. The argument is perhaps best made
with a quote from the CEO of a micro-lender (African Bank
Limited) in South Africa: ‘‘A credit cycle is always going to
be driven from a supply dynamic, not a demand dynamic
and markets overheat because suppliers push too much
credit into the system’’ (Rees, 2013).
Model development
Basic model
The initial model is based on the work of Masaki Kusano
(Kusano, 2011). The familiar accounting equation of Assets
(A) – Liabilities (L) = Equity (E) is the starting point in the
development of his model that links accounting capital
with bank capital regulation and bank asset expansion.
This section will then add the payment of remuneration
and dividends from FVA gains to his model as well as con-
sider the results of the model when FVA generates losses.
Bank management are incentivised to minimise the
capital ratio C
0
¼
AÀL
A
À Á
as this will maximise return on
equity, arguably the most popular bank performance met-
ric as it is a measure of the rate of return ?owing to share-
holders (Rose & Hudgins, 2008:167). Of course, bank
regulators do not allow banks to leverage to in?nity and
they demand that banks keep a minimum capital buffer
to protect depositors in the case of default. The calculation
of
@C
0
@A
shows that the capital ratio de?ned above is directly
related to the value of assets:
@C
0
@A
¼
L
A
2
> 0 ð1Þ
The implication is that an increase in asset values will
increase the capital ratio (which is bad for return on equi-
ty); management will want somehow to manage or neu-
tralise this increase in the capital ratio. The capital ratios
6
of the ?ve largest South African universal banks for the per-
iod before the global ?nancial crisis illustrate this point
Fig. 3 (based on data provided by the South African Reserve
Bank). The period was one of high bank pro?ts (by fair value
accounting and other means) and yet capital ratios stayed
steady or declined; high pro?ts, ceteris paribus, should have
increased the capital ratio.
The Basel capital accords de?ne their own version of the
capital ratio, where accounting capital is expressed as a ra-
tio of risk-weighted assets:
C
1
¼
A À L
P
r
i
a
i
; ð2Þ
where A ¼
P
a
i
or the total asset value of the bank is the
sum of the individual assets, L is the total value of the lia-
bilities of the bank and r
i
is the risk-weight applicable to
each asset a
i
according to the Basel Accords.
Financial sector
Real economy
Bank 2
Bank 1
Credit
Investment
Investment
Remuneration &
dividends
Debt repayment
Credit
Fig. 2. Schematic representation of the model developed in this paper (this is the ?nal completed model; the schematic representations depicting the
development of the model is illustrated graphically with an accompanying example in Appendix A and numbered Figs. A-E).
6
This is the capital ratio as de?ned immediately above (C
0
¼
AÀL
A
) and not
the Basel ‘‘risk-weighted’’ capital ratio. In effect this method risk-weights
all assets at 100%.
P. de Jager / Accounting, Organizations and Society 39 (2014) 97–116 103
Suppose that the market value of all assets
7
increases by
a constant k, where k = k(t) and k > 0.
8
The total asset value A
is now equal to (1 + k)A. If assets are measured at fair value
then during an upswing the capital ratio can be written as:
C
2
¼
ð1 þ kÞA À L
ð1 þlÞ
P
r
i
a
i
; ð3Þ
where l is the increase in the total of risk-weighted assets.
As the risk-weights under the Basel Accords
9
in general
meets 0 6 r
i
6 1 the following can be assumed: 0 < l 6 k.
10
Assuming that management is planning to neutralise
the increase in the capital ratio by originating additional
assets denoted by B and fully ?nanced by debt
11
the follow-
ing equation is obtained:
.03
.04
.05
.06
.07
.08
00 01 02 03 04 05 06 07 08 09 10
Absa
.03
.04
.05
.06
.07
.08
00 01 02 03 04 05 06 07 08 09 10
Fnb
.05
.06
.07
.08
.09
.10
.11
.12
00 01 02 03 04 05 06 07 08 09 10
Investec
.04
.05
.06
.07
.08
.09
.10
00 01 02 03 04 05 06 07 08 09 10
Nedbank
.03
.04
.05
.06
.07
.08
00 01 02 03 04 05 06 07 08 09 10
Standard
Fig. 3. Monthly capital ratio per bank January 2000–August 2010 (shaded areas indicate post-crisis time period).
7
All the assets of a bank are not normally fair valued but the complexity
of modelling different types of assets is not necessary. Refer to the example
in the appendix where a bank is assumed to have $30 of fair valued assets
(through pro?t and loss) and $70 of historical cost assets. If the market
value of the fair valued assets increases by 10% and the tax rate is 30% then
the value of k is (30 * (10% * (1 À 30%)))/(30 + 70) = 2.1%.
8
It may seem as if the possibility that liabilities can also experience an
increase in market value when assets increase in market value is ignored.
Given that banks mostly own interest-rate sensitive assets of longer
maturity than their interest-rate sensitive liabilities, it is to be expected
that DA PDL when interest rates decrease. This state of affairs can be just
as effectively modelled with DA and L kept constant.
9
Under Basel I the risk-weightings for most categories of assets were
below 100%. Basel II and III moved away from rougher asset categories to
risk-weightings for individual assets or portfolios of assets, based on
individual bank history, that re?ect the ‘‘expected unexpected’’ loss (not
routine, high frequency events) (Gebhardt & Novotny-Farkas, 2011:295) on
that asset or portfolio. The planned-for unexpected losses on bank assets
should be less than 100% as banks are pro?t seeking organisations.
10
An anonymous reviewer pointed out that over time (not currently in
the model) there is the tendency for l to increase asymptotically to k. This
aspect is not explored further in the model.
11
The goal of the exercise is to bring the capital ratio down again after the
increase in equity caused by the increase in asset values and thus an
increase in debt is required.
104 P. de Jager / Accounting, Organizations and Society 39 (2014) 97–116
C
2
¼
ð1 þ kÞA þ B À ðL þ BÞ
ð1 þlÞ
P
r
i
a
i
þ r
j
B
¼
A À L
P
r
i
a
i
¼ C
1
; ð4Þ
where r
j
is the risk-weight of the originated/purchased,
debt-?nanced assets. It should be noted that although
the B added to the asset base and the B added to liabilities
are of the same value, the Bs are not different sides of the
same ?nancial instrument.
12
An increase in deposits
increases the money supply or the spending power available
in the economy. Solving for B:
B
1
¼
ðk ÀlÞA þlL
r
j
C
1
: ð5Þ
Fig. A in the Appendix provides a schematic representa-
tion followed by a narrative description of the model up to
this point. Management neutralised the increase in the
capital ratio by investing in additional ?nancial assets fully
?nanced by debt; this increases the demand for ?nancial
assets in the ?nancial sector. Bank deposits created are
money in the banking systemand so this increases demand
in the real economy. If time were already part of the model
then a few comments regarding feedback processes would
be possible at this stage. However, time will only be
included in the model at a later stage of its development.
The increase in the market value of ?nancial assets con-
tributes to k (an increase in the market value of all bank as-
sets) and thus provides an impulse for the asset expansion
cycle to start again. The additional demand for ?nancial as-
sets, ceteris paribus, increases the market value of ?nancial
assets and thus reduces the return on those assets. To keep
returns constant, bank management is incentivised to take
on more risk; for example, to invest in riskier assets or
originate riskier loans. The other side to additional demand
for ?nancial assets is the provision of additional spending
power (credit) to the real economy. Over the longer term,
the extension of credit to the real economy can only be
sustained if the additional capital is invested in productiv-
ity enhancing real assets. With increased productivity the
real economy might be in a position to pay back the capital
and interest. If not, the only way for the real economy to
keep a constant level of demand is to receive more and
more credit fromthe ?nancial sector. In a world of decreas-
ing marginal returns it is safe to conclude that over the
longer term, the increased demand in the real economy
due to advancement of more credit is likely to be followed
by a period of decreased demand when loans are being re-
paid; in other words, less spending in the real economy.
This simple model only allows for banks to expand their
asset base and debt in reaction to an increase in the market
values of all assets within the boundary of the original
capital ratio. The model will now be extended to include
the possibility of remuneration and dividend payments
from the increase in asset values.
13
The new capital ratio
that results after an increase in the market values of all
assets is:
C
3
¼
ð1 þXkÞA À L
ð1 þXlÞ
P
r
i
a
i
; ð6Þ
where X is the portion of the increase in asset values re-
tained by the bank (after appropriation of the increase by
management and shareholders via remuneration and/or
dividends) (0 < X6 1).
Again, if one assumes that management is planning to
neutralise the increase in the capital ratio by investing in
additional assets denoted by B and fully ?nanced by debt,
then the following equation is obtained:
C
3
¼
ð1 þXkÞA þ B À ðL þ BÞ
ð1 þXlÞ
P
r
i
a
i
þ r
j
B
¼
A À L
P
r
i
a
i
¼ C
1
: ð7Þ
Solving for B:
B
2
¼
Xðk ÀlÞA þXlL
r
j
C
1
: ð8Þ
If all of the increase is retained (X= 1) the result simpli?es
to the previous solution B
1
.
Comparing the solution of B
2
with that of B
1
it is appar-
ent that B
2
must always smaller than B
1
because the two
terms in the numerator are both a portion of the numera-
tor terms in B
1
and the denominator in both instances is
the same. The implication is that the need to increase as-
sets and debt to restore the previous capital ratio, after
an increase in the market values of all assets, is reduced
when a portion of the value increase is paid out.
A similar logic applies when calculating
@B
@X
:
@B
@X
¼
ðk ÀlÞA þlL
r
j
C
1
> 0: ð9Þ
As the retention ratio (X) increases, the need to expand the
asset base and debt also increases. The more of the increase
in asset values is paid out (smaller X) the less the need for
an expansion in assets and debt.
Fig. B in the Appendix provides a schematic representa-
tion followed by a narrative description of the model up to
this point. Thus, the inclusion in the model of the payment
of remuneration and dividends (1 À X) (for positive k) sig-
ni?cantly changes the effects previously noted. The need to
increase assets and debt to restore the previous capital ra-
tio is reduced when a portion of the value increase is paid
out. It is implied that the additional demand for ?nancial
assets will be less with a smaller price effect and the in-
creased spending power transferred to the real economy
will be less. The money paid to the real economy as remu-
neration and dividends, ceteris paribus, does increase de-
mand in the real economy. This increase in demand will
be less, however, than the increase in demand that would
have resulted if the remuneration and dividends were not
paid out, with the capital ratio being restored solely by the
debt-?nanced expansion of the bank balance sheet. This is
due to the leverage that is applied to the capital on the
bank’s balance sheet. Of course, not all of the remuneration
and dividends paid to the real economy will be consumed;
a portion of these will return to the ?nancial sector in the
form of additional demand for ?nancial assets, which will
12
It is often argued that the ?nancial sector can be ignored in models as
any loan is owned by somebody else and thus the debit and credit can be
collapsed without mishap. This argument ignores the fact that bank
deposits are money and bank loans are not.
13
Both management and shareholders have a strong incentive to reward
themselves during the business cycle upswing despite the possibility of
future losses during the downturn as elucidated by Taylor (2009) and
Haldane (2011).
P. de Jager / Accounting, Organizations and Society 39 (2014) 97–116 105
increase their market values and decrease their returns. If
time were already part of the model (note previous com-
ment that time will only be formally included in the model
at a later stage) then a new comment regarding feedback
processes would be possible at this stage.
The investment demand for ?nancial assets increases
their market values and contributes to an increase in the
average market value of all assets. This provides an im-
pulse for the assets expansion cycle to start again. Here,
one should also note the replacement effect that took place
in capital; liquid assets were paid out as remuneration and
dividends and were replaced by unrealised and riskier cap-
ital gains on ?nancial instruments. This is bad capital driv-
ing out good capital; this is explained in more detail in
Section ‘Further development of the basic model’.
The described model does not allow for a general reduc-
tion in asset values (0 < l 6 k was assumed). However,
such a general reduction in asset values would almost cer-
tainly be experienced during a recession. If the above
assumption is relaxed and negative growth is allowed with
the only proviso that l should always represent a portion
of k, then banks will reduce their asset base and pay off
debt in response to a general negative value adjustment.
The solution for B remains:
B ¼
ðk ÀlÞA þlL
r
j
C
1
: ð10Þ
Fig. C in the Appendix provides a schematic representa-
tion followed by a narrative description of the model up to
this point. In the case of a general reduction in asset values,
X is not added to the model as management and share-
holders will not pay in any money.
In summary, the solution for B:
B ¼
XðkÀlÞAþXlL
r
j
C
1
; if k > 0:
0; if k ¼ 0:
ðkÀlÞAþlL
r
j
C
1
; if k < 0:
8
>
>
>
>
<
>
>
>
>
:
Implications of the basic model
The basic model described above shows how the inter-
action of FVA gains and bank capital regulation motivates
banks to expand their asset base. In the context of that
basic model this expansion was ?nanced by debt. Such a
feedback process can help to explain why credit advances
in the United States grew at such a high rate (from $3 tril-
lion total debt to $36 trillion total debt in 2007 (NCCFEC,
2011:xvii)). The following Fig. 4 provides data from the
United Kingdom’s Independent Commission on Banking
(2011) and illustrates how credit extension in the years
preceding the global ?nancial crisis exceeded GDP and typ-
i?es a debt-?nanced boom.
The basic model also illustrates how this incentive to
expand the asset base is tempered by the payment of div-
idends and remuneration. These dividend and remunera-
tion payments would not have been necessary if FVA did
not increase accounting equity (the argument can be found
in Kay, 2009; Taylor, 2009; Turner, 2010; Wood, 2010).
The model shows that when allowance is made for a
negative valuation shock to asset values then the asset
base decrease required, to restore the capital ratio, must
be greater than the asset base increase required for a posi-
tive valuation shock of the same size. It is here that a sim-
ple model that assumes all price changes are passed into
bank equity by FVA is not descriptive of reality; at least,
not over the short-term. Banks have ‘‘shock absorbers’’
available to them in the FVA rules to avoid these write-
downs; they can argue that market prices are not correct
and then move to mark-to-model values and/or reclassify
items held at ‘‘fair value’’ to ’’held at cost’’: in October
2008, accounting standards were changed to allow for
the reclassi?cation of ?nancial instruments from ‘‘carried
at fair value’’ to ‘‘carried at cost’’. Bischof, Brüggemann,
and Daske (2010) as well as Fiechter and Meyer (2011)
found that banks made ample use of these opportunities.
In view of this, it is not surprising that empirical studies
by Shaffer (2010) and by Badertscher et al. (2012) found
Fig. 4. Total loans to different sectors of the economy as a percentage of GDP. Source: Vickers Report (2011:51).
106 P. de Jager / Accounting, Organizations and Society 39 (2014) 97–116
little evidence of bank capital being depleted by FVA losses.
The possibility of long-term FVA losses impacting bank
capital cannot be ascertained from these writers’ evidence
as they only considered data up to the year 2008. Over the
longer term, it is unlikely that banks will be able to avoid
recognising actual losses, permanently, because the instru-
ments will mature and a fundamental loss in value will
then realise. This postponement of FVA losses by banks
possibly explains why banks have been unwilling to lend
following the global ?nancial crisis even though they have
enough capital according to their of?cial regulatory ?g-
ures; bank management know that the capital on the
books is not high quality (risky) and they act accordingly;
thus, they act as if they have less capital than stated in
the of?cial numbers. Japanese banks in the 1990s had a
similar problem where the banks of?cially had enough
capital and despite the of?cial numbers, were not writing
down non-performing assets and neglected to advance
the necessary credit to the economy.
It can also be argued that the increase in bank capital
that was called for under Basel III (Bank for International
Settlements, 2011:2), in reaction to the global ?nancial cri-
sis, works to counter this bene?t of the ‘‘shock absorbers’’
and that the net effect would be a need to shrink bank bal-
ance sheets. Reducing the size of bank balance sheets is a
current global occurrence.
The main shortcomings of the basic model can be sum-
marised as follows: ?rstly, it depends on external valuation
shocks for regime change between bank balance sheet
expansion and shrinkage; secondly, feedback effects are
implied but not expressly modelled and thirdly, it does
not explain the faster growth in loans to ?nancial compa-
nies compared to loans to non-?nancial companies; the
latter is shown in Fig. 4 which indicates information pro-
vided by the Independent Commission on Banking.
Further development of the basic model
The increase in the general market value of bank assets
(a positive k) during the upswing that preceded the global
?nancial crisis was the product of complex interconnected
developments in the ?nancial system. These developments
included progressively lower interest rates globally that,
all else equal, increased the value of especially ?xed rate
?nancial instruments. A portfolio of ?xed rate loans desig-
nated at fair value through pro?t and loss would have
increased in value and the increase would have increased
regulatory capital. Financial instruments exposed to credit
risk would also have increased in value during the upswing
as default data improved and con?dence increased.
Another possible source of value increase was the increase
in ?nancial instrument trading within the ?nancial sector.
This increase in trading within the ?nancial sector resulted
from a combination of ?nancial innovation, deregulation
and a search for yield in a low interest rate environment.
Increased demand for ?nancial instruments led to price
rises.
Given an initial increase in the general market value of
bank assets, it is relatively simple to expand the basic mod-
el to incorporate feedback effects that will maintain bank
balance sheet expansion once started; this is done by
formally incorporating time into the model. The following
intellectual device is useful for the discussion that follows;
to see equations C as re?ecting the capital ratio of the
banking system as a whole in the one instance and as
re?ecting the capital ratio of an individual bank in the
other instance.
In the instance when equations C re?ect the capital
ratio of the banking system as a whole it is important to
realise that the expansion of the bank sector balance sheet
is only possible by advancing further credit to the real
economy (external to banking system). The expansion of
the balance sheet simultaneously increases the money
supply with the concurrent increase in bank deposits. Real
economic actors will spend a portion of this increased
money supply on the purchase of ?nancial assets, with
the additional demand causing the values of those assets
to increase. At the same time the money paid out by the
banking system as remuneration or dividends (1 À X) will
not all be consumed and will, at least to some extent, be
used to purchase ?nancial assets, increasing their market
values.
In basic model mathematical terms this can be repre-
sented as follows:
@kðtÞ
@t
¼ f ðBðkÞ; XðkÞ; UÞ; ð11Þ
where U represents exogenous factors that might im-
pact on k The implication is that an initial value shock
can initiate a feedback loop; this can arise through the de-
mand for ?nancial instruments from the ?nancial sector
(B) or from the demand for ?nancial instruments from
the real sector (portion of B + portion of X) driving contin-
uous bank balance sheet expansion or shrinkage. Fig. D in
the Appendix provides a schematic representation fol-
lowed by a narrative description of the model up to this
point.
The question that arises is why input to the model will
change from a positive k to a negativek. Events external to
the model are obviously one answer, but the model up to
this point holds three endogenous answers to this ques-
tion; wealth effects, bad capital driving out good capital
and reaching for yield. Note that the ?rst of these answers,
the ‘‘wealth effect’’ is related to the balance sheets of indi-
viduals becoming more fragile while the two latter an-
swers are related to increasingly fragile bank balance
sheets.
Individuals in the real economy have more credit avail-
able to them during the upswing. This credit availability is
driven by individuals becoming seemingly more credit
worthy as the upswing advances; the market value of their
security increases and default rates improve. In other
words, as the asset side of individuals’ balance sheets in-
creased (improved), the liability side of their balance
sheets also increased as they took on more debt. In this
manner the individuals’ leverage ratio might stay constant
but, relative to their income, their balance sheets are more
and more fragile as the upswing continues.
The payment of remuneration and or dividends (1 À X)
from FVA pro?ts leads to a replacement effect where the
most liquid asset (money) leaves the ?nancial sector as
payment and is replaced in bank capital by FVA-derived
P. de Jager / Accounting, Organizations and Society 39 (2014) 97–116 107
value increases in less liquid assets. This can be called a
special case of Gresham’s law
14
where ‘‘bad capital drives
out good capital’’. The result is a banking system much more
sensitive to external disturbances (or U).
The additional demand for risky assets during the bal-
ance sheet expansion process drives down the returns to
be earned on those risky assets and thus incentivises bank
managers to take more risk in the search for returns.
15
The
implication for bank balance sheets is, even just to show
constant returns, to become riskier. The result is a banking
system much more sensitive to external disturbances (or
U). This reach for yield can explain the decline in underwrit-
ing standards in the U.S. (NCCFEC, 2011:4) as well as the
observation that ?nancial ?rms changed from staid to risk
seeking during the boom (NCCFEC, 2011:xvii). Evidence of
predatory lending highlighted by the Commission (NCCFEC,
2011:10) calls into question the dissenting report’s conclu-
sion that credit demand emanating from the U.S. Govern-
ment’s housing policies was to blame for the
overexpansion in U.S. housing. Predatory lending points to
the role of credit supply in the process that ties back to
the reach for yield argument from the model.
A further argument for regime change is possible in the
instance when equations C re?ect the capital ratio of an
individual bank rather than that of the total banking sys-
tem. In addition, one of the shortcomings of the basic mod-
el highlighted earlier in this paper was that it did not
explain the growth in loans between ?nancial companies.
By adding more features to the basic model and by seeing
equations C as re?ecting the capital ratio of an individual
bank, endogenous regime change as well as the growth
in loans between ?nancial companies can be explained.
If we recall from previous discussion that:
C
2
¼
ð1 þ kÞA þ B À ðL þ BÞ
ð1 þlÞ
P
r
i
a
i
þ r
j
B
; ð12Þ
and that was used to determine the balance sheet expan-
sion B required for a given positive value shock, it is now
necessary to emphasise that a bank will only expand its
asset base, ?nanced by debt, if the following condition is
met:
R
asset
À R
liability
> 0; ð13Þ
where R represents the return on the ?nancial instrument.
Bank debtors want to spend their newly raised debt and
thus the asset side of the transaction will tend to be of
longer maturity than the liability side of the transaction
due to the required ability to spend. When the banking sys-
tem experiences a positive value shock and banks are
simultaneously expanding their asset bases, then two
effects can be noted. First, the supply of credit can over-
whelm the real economy’s demand for credit and lead to
increased inter-bank lending that will lead to a stronger
(compared to the additional demand emanating from the
real economy described previously) positive feedback loop.
This is because banks will invest all of the new bank depos-
its created in ?nancial instruments, increasing k Second,
arbitrage will drive down (R
asset
À R
liability
) until, at the limit
and ignoring other factors, no more margin is available and
the positive feedback loop will come to an end. This helps
to explain the ‘‘exponential growth’’ in trading activities
observed by the NCCFEC (2011:xvii)) as well as the growth
of the ?nancial sector from 5% of gross domestic product to
8% (NCCFEC, 2011:64). Finally, the feedback process de-
scribed in the model between the ?nancial system and
the real economy explains why, parallel to the rapid
expansion of credit in the U.S. economy, the Commission
observed overdone real economic activity (NCCFEC,
2011:5).
The margin effect noted above should be related to the
literature on the ability of an inverted yield curve to pre-
dict recessions (see for example Ang, Piazzesi, & Wei,
2006; Wright, 2006). Banks, generally speaking, invest in
?nancial instruments with longer maturity ?nanced by
?nancial instruments of shorter maturity and realise a
return spread as elucidated above when the yield curve
is normal and upward sloping. If this investment in longer
maturity instruments ?nanced by shorter maturity instru-
ments became systemic, it will drive down the returns on
longer maturity instruments: greater demand drives up
prices leading to lower returns. At the same time it will
drive up returns on shorter maturity instruments: greater
supply drives down prices leading to higher returns.
Table 1 summarises the feedback effects:
These feedback effects, without inhibitors (designated
‘‘regime change processes’’ below), will lead to continuous
expansion or continuous contraction. Table 2 summarises
the endogenous regime change processes identi?ed in
the model:
Empirical example
It is not possible to directly observe the macro-economy
and thus the economic data that we do have are selected
and made sense of by theory (Suzuki, 2003:484). Accord-
ingly, no attempt will be made to demonstrate the model
proposed in this paper in totality. Rather, data from two
international and systemic universal banks will be used
to demonstrate the following: the slow (controlled) recog-
nition of FVA losses; the materiality of fair value account-
ing entries that impact pro?t and loss and the crowding
out of risk-free capital with risky unrealised FVA gains.
The model developed in this paper can then be used to
make sense of the information.
The two banks that data will be gathered from are one
of the largest banks in the Netherlands, ING Bank N.V.,
and the largest bank in South Africa, the Standard Bank
of South Africa Limited. Apart from size and systemic
importance further similarities between the two banks
are that they both report under International Financial
Reporting Standards (IFRS) and are supervised from coun-
tries where Basel II was implemented in 2008. Both banks
14
Gresham’s law is commonly stated as: ‘‘Bad money drives out good’’,
but is more accurately stated: ‘‘Bad money drives out good if their exchange
rate is set by law’’ (Rolnick & Weber, 1986:185).
15
A more subtle possibility should also be noted; that this process of
riskier asset accumulation is possible even without an actual decision to
invest in higher risk assets. During the upswing default data as well as loan
to value ratios (the market value of security tends to increase) improves
and this provides the rationale for further credit extension. In such a
process, risk increases as more credit is now backed by the same cash ?ows.
108 P. de Jager / Accounting, Organizations and Society 39 (2014) 97–116
are large (globally important) according to Shaffer’s
(2010:9) de?nition (>$100 billion of assets) with ING Bank.
N.V. about ten times larger than the Standard Bank of
South Africa Limited in 2012. Very important for this paper
is that both banks operate in an income tax regime where
?nancial instruments are taxed on realisation and not on a
mark-to-market basis; this makes it possible to identify
the unrealised portion of fair value accounting gains and
losses through pro?t and loss and will be explained further
below.
IFRS 7 (IASB, 2005) does not require disclosure of the
realised and unrealised portions of items of income, ex-
pense, gains, and losses related to ?nancial instruments;
only the combined total is required. The deferred tax notes
to the ?nancial statements of both ING Bank N.V and the
Standard Bank of South Africa Limited contain information
that makes possible the derivation of the portion of net
pro?t that is due to unrealised fair value adjustments on
?nancial instruments. Deferred tax arises when a differ-
ence exists between accounting net pro?t and taxable in-
come. Accounting net pro?t contains realised and
unrealised items related to ?nancial instruments whereas
current Dutch and South African taxable income contains
only realised items. An increase in the deferred tax liability
due to ?nancial instruments thus re?ects the tax portion of
an unrealised gain on ?nancial instruments. Fig. 5 and
Fig. 6 are excerpts from the banks’ 2012 annual ?nancial
statements where the red ovals indicate the relevant
movements in the deferred tax liability due to ?nancial
instruments for that year:
Table three that follows summarises the available infor-
mation on deferred tax movements. To gain an under-
standing of the signi?cance of these movements it is
important to note that what is observed on the deferred
tax line is only the income tax effect of the transactions.
For example, the current corporate income tax rate for
the Netherlands is 25% and thus 25% is the movement on
deferred tax and 75% is the impact on pro?t. It is the 75%
that we are interested in. In table three a few simplifying
assumptions have been made: that the current corporate
income tax rates in the Netherlands of 25% and in South
Africa of 28% was constant over the whole period; that
the debit entries against deferred tax for ING from 2008
onwards reversed previous credit entries and that retained
income is a suitable proxy for bank capital.
When looking at the movement in the deferred tax line
in table one it is important to note that for ING Bank N.V.
the reversal from credit entries to debit entries (debit en-
tries imply unrealised losses on ?nancial instruments)
against deferred tax happened in 2008 as would be
Table 1
Summary of feedback effects in the model.
Feedback process Description Relative strength
Bank balance sheet expansion Additional demand for ?nancial instruments impacting k Medium (leverage ampli?es effect
but leakage to real economy)
Real economy investment in ?nancial instruments Additional spending power not all consumed Weak (no leverage effect)
Intra bank lending and investment Balance sheet expansion creates additional demand for
?nancial instruments impacting k. Spending power
created is within the banking system and not in the real
economy
Strong (additional spending
power created within ?nancial
sector and subject to leverage)
Fig. 5. Changes in deferred tax for ING Bank N.V. due to FVA on ?nancial instruments through pro?t and loss. Source: ING Bank N.V Annual Report
(2012:64).
P. de Jager / Accounting, Organizations and Society 39 (2014) 97–116 109
expected with a global ?nancial crisis that reached its peak
in September 2008 with the failure of Lehman Brothers.
Strangely, the reversal for the Standard Bank of South Afri-
ca Limited only takes place in 2009. For ING Bank N.V. the
reversal that started in 2008 continued until 2012,
whereas for the Standard Bank of South Africa Limited
the reversal only lasted two years and then turned again.
What the information shows is possibly the use of account-
ing discretion to avoid/manage FVA losses: The Standard
Bank of South Africa Limited only shows losses for the ?rst
time in 2009 whilst the crisis started in 2008 and for both
banks the FVA losses did not come through as one ‘‘hit’’ but
took time.
The rows that relate the movements on deferred tax to
pro?t show just how material these FVA entries (that im-
pact pro?t and loss and thus bank capital) are. For ING
Bank N.V very few entries are available when FVA gener-
ated unrealised gains, but unrealised losses against pro?t
between À180% and À34% show how large these entries
can be. The entries for the Standard Bank of South Africa
Limited peak at an unrealised FVA gain portion of net in-
come of 33% in 2008 with the largest reversal of À58%
the next year.
Arguably the most important part of Tier 1 bank capital
is retained income. The idea is that this capital should
always be available to absorb unexpected bank losses.
Table 3 shows that the unrealised FVA gain portions of
retained income are not as stable as would be expected.
For both banks the peak is during the ?nancial crisis with
a sharp reduction thereafter. This is indicative of unrea-
lised FVA gains in bank capital being risky and during the
upturn displacing less risky bank capital as predicted by
the model. Back in South Africa, the CEO of one of the four
large banks, Nedbank Limited, admitted during an inter-
view on his bank’s ?nancial performance in 2013 that
FVA gains through pro?t and loss are risky: ‘‘We did bene-
?t during the ?rst six months from some fair value gains,
which are lower quality income and harder to forecast. . .’’
(Tarrant, 2013).
Implications, shortcomings and opportunities for
further research
The model demonstrates the expansionary impulse
generated when FVA increases bank regulatory capital. It
can be argued that the same expansionary impulse will re-
sult from other accounting entries that also increase regu-
latory capital. These entries could include realised pro?ts
fromthe selective selling-off of investments (gains trading)
or incurred loan loss provisioning that ‘‘enables banks to
present higher earnings and (regulatory) equity capital,
which allow the bank to extend more credit’’ (Gebhardt
& Novotny-Farkas, 2011:302). FVA’s impact differs in two
important aspects. First, FVA impacts regulatory capital
much faster and more materially than other accounting
entries that need time to impact retained earnings. It is
the speed of the feedback process that matters. The second
way in which the capital increase from FVA differs from in-
creases caused by other accounting entries is that these
alternatives will not result in the replacement of liquid
assets in bank capital with riskier FVA gains (bad capital
driving out good capital); FVA pro?ts do not provide
Table 2
Summary of endogenous regime change processes in the model.
Regime change
process
Description
Balance sheet fragility
– real economy
Real economy has more assets with more
debt but the same income
Balance sheet fragility
– banks
Bad capital driving out good capital:
Liquid assets in capital replaced by illiquid
capital gains
Balance sheet fragility
– banks
Reach for yield: Riskier assets originated
to keep income constant; the other side of
the fragile real economy balance sheet
Yield curve inversion Arbitrage drives down yield on longer
term ?nancial instruments and drives up
yield on shorter term instruments
Fig. 6. Changes in deferred tax for the Standard Bank of South Africa Limited due to FVA on ?nancial instruments (including derivatives). Source: Standard
Bank of South Africa (2012:153)
110 P. de Jager / Accounting, Organizations and Society 39 (2014) 97–116
liquidity that can fund dividends and remuneration
payments.
Another characteristic of the model is that it is incom-
plete. Banks do not just lend money due to the interaction
between FVA and their capital ratios. The real economy has
a need for credit before the supply of credit is considered.
In the same way it is certainly not argued here that FVA is
the primary or sole reason behind the cyclicality of ?nan-
cial capitalism. The business and ?nancial cycles existed
before the adoption of FVA and booms and busts were even
evident during the heyday of historical cost accounting in
the period after the Great Depression. The caveat is that
those booms and busts were less damaging in their impact;
they were less extreme than either the Great Depression or
the global ?nancial crisis.
Toporowski (2010:224) argues that those who view the
capitalist system as being prone to crisis have dif?culty in
explaining the relative stability of ?nancial capitalism in
the decades before the global ?nancial crisis. This paper ar-
gues that the absence of FVA from the banking system dur-
ing that time can explain the relative calm. A system needs
control mechanisms in the presence of feedback loops to
keep the process from running away. One such ?nancial
feedback loop was highlighted by Minsky and describes
how stable economic periods improve the balance sheets
of individuals by increasing the values of their assets and
enabling them to borrow more because they now have
more security; this in turn increases systemic fragility as
more debt is now supported by the same cash ?ows. This
situation can be summarised by stating that the individ-
ual’s (borrower’s) balance sheet was ‘‘fair valued’’ in the
credit evaluation process. A control mechanism that will
inhibit this expansionary impulse is not to fair value the
balance sheets of banks as well; in other words do not
meet rising credit demand with increasing credit supply.
These checks and balances prevailed in the relatively stable
economic period that preceded the global ?nancial crisis;
as a result, the feedback loop was not allowed to run away
due to the presence of a control mechanism. However,
with the introduction of formal FVA into accounting stan-
dards this control mechanism was removed, leading to
the over-indebtedness of individuals and ?nancial busi-
nesses. FVA did not cause the crisis but ampli?ed it.
The model explains the conclusion of the NCCFEC
(2011:xix) that ‘‘a combination of excessive borrowing,
risky investments, and lack of transparency put the ?nan-
cial system on a collision course with crisis’’. Banks and
individuals borrowed excessively as explained in the
paragraph above and risky investments were made in the
process as there was a search for yield. The bene?t of the
perspective offered by the model will brie?y be illustrated
by using it to understand the results of the study by Cabral
(2012). Cabral introduced a novel model of individual bank
pro?tability and used the model to explain paradoxes such
as very high bank pro?tability immediately prior to the cri-
sis even though intermediation margins were very low
(the return to banking activities was low). It is argued that
balance sheet growth and the taking on of additional risk
prior to the crisis explains this paradox. In addition it is
argued that low levels of liquidity prior to the crisis played
an important role in the crisis. First, bank pro?ts immedi-
ately prior to the crisis were very high, possibly due to
the effect of FVA. Second, the returns to banking activities
were low prior to the crisis because of the search for yield
necessitated by the bank balance sheet expansion that
simultaneously increased the money supply. Cabral’s
(2012:114) explanation for the low returns is to argue that
it is the result of increased competition. The model in this
paper explains why there was an increase in competition.
Cabral’s model does not explain where the additional cred-
it demand comes from when banks, according to his mod-
el, expand their balance sheets to protect pro?tability; the
model in this chapter does explain. Finally, Cabral uses the
exogenous impact of bank regulations to explain the low
levels of liquidity prior to the crisis. In contrast the model
in this chapter explains the low levels of liquidity
endogenously.
Controls unrelated to FVA can be imagined, to counter-
act this ampli?cation process. One such alternative control
is the introduction of a countercyclical capital buffer as
part of the changes made to the global bank regulatory
infrastructure following the global ?nancial crisis; thus,
during the upswing, banks might be required to increase
their capital ratios as described in the following quote from
Basel III: ‘‘It will be deployed by national jurisdictions
when excess aggregate credit growth is judged to be
associated with a build-up of system-wide risk to ensure
the banking system has a buffer of capital to protect it
against future potential losses’’ (Bank for International
Settlements, 2011:57). In terms of the model this increase
in the minimum capital required, if of the exact size
necessary, will eliminate the impulse for balance sheet
expansion resulting from k, halting the feedback process.
However, there are reasons why this might not be an ideal
hedge for the expansionary impulse created by FVA. First,
it will require central bank decision makers to get the tim-
Table 3
Summary of deferred tax movements on ?nancial instruments (credit entries are positive and debit entries negative).
2004 2005 2006 2007 2008 2009 2010 2011 2012
Movement in deferred tax due to ?nancial instruments – ING
(Euro million)
NA NA NA 72 À301 À325 À192 À162 À366
Movement in deferred tax due to ?nancial instruments – Standard
(rand million)
83 486 903 700 1132 À1837 À1747 514 9
Unrealised portion of net pro?t after tax – ING NA NA NA 6% À128% À180% À13% À12% À34%
Unrealised portion of net pro?t after tax – Standard 4% 19% 29% 19% 33% À58% À57% 14% 0%
Unrealised portion of retained income – ING NA NA NA 30% 32% 21% 11% 7% 0%
Unrealised portion of retained income – Standard 43% 54% 81% 88% 83% 41% 14% 17% 19%
P. de Jager / Accounting, Organizations and Society 39 (2014) 97–116 111
ing and extent of any increase exactly right. Second, banks
as a body have become very effective lobbyists in some
countries and can be expected to try and in?uence the set-
ting of an increased capital ratio. Lastly and perhaps most
important is the argument that a control mechanism
implemented on a regulatory level cannot take into ac-
count the unique circumstances of each individual bank.
Numerous opportunities for future research can be
identi?ed:
More and better descriptive evidence of the co-
movement of FVA in banks, with the business cycle.
Do bank managers pay dividends out of unrealised
FVA pro?ts, rewarding shareholders as well, for
managers’ excessive risk-taking and by doing so,
weakening the ?nancial system?
Augmentation of standard stock-?ow consistent
accounting models of the economy with this credit
supply model.
Do bank management pay out remuneration from
FVA gains?
Was the avoidance of FVA write-downs during the
global ?nancial crisis temporary or permanent?
Causality studies on the relationship between credit
demand and credit supply from a real economic per-
spective and from a ?nancial perspective; and
Studies of the class struggle that determines the
split of the economic pie between wages, pro?ts
and, in addition to the standard production view
of capitalism, interest.
Conclusion
The focus by accounting researchers on the role of
accounting and FVA during the global ?nancial crisis and
not on a role before the crisis, as well as the claim that
accounting is only a messenger are two factors that moti-
vate the need for the development of the model described
in this paper, that links FVA, bank regulatory capital,
money supply, remuneration, dividends and economic
activity. The requirement to incorporate global ?nancial
crisis characteristics (such as feedback effects, systemic
risk and the primacy of banks in the crisis) into that model
rationalised the use of a stock-?ow consistent accounting
type model of the economy rather than standard economic
models. In contrast to the usual stock-?ow consistent
accounting model of the economy, the model developed
in this paper reverses the causality between credit demand
and credit supply; in other words, the starting point is
credit supply, not credit demand, which gives banks the
centre stage in the account.
The model developed in this paper describes and then
links together many of the characteristics of the global
?nancial crisis and previous crises. The procyclical expan-
sion of bank balance sheets during the upswing followed
by a contraction (over the long-term because during the
short-term, banks can avoid some FVA write-downs)
during the downturn is explained by the need to maintain
regulatory capital ratios and/or feedback effects both with-
in the ?nancial sector and between the ?nancial and real
sectors. In accordance with the doctrine of Minsky, a state
of crisis is shown to be endogenous to the model’s opera-
tion, because of systemic fragility. The focus is on bank bal-
ance sheet fragility that is caused by bad capital driving out
good capital, banks reaching for yield and the inversion of
the yield curve. It is argued that not having FVA pro?ts in
bank capital during booms is an essential control mecha-
nism that inhibits the feedback effects emanating from
the real sector.
An empirical example is given where the evolution of
the unrealised FVA gains through pro?t and loss of two
universal and systemic banks (ING Bank N.V & the Stan-
dard Bank of South Africa) can be better understood using
the model in this paper.
In conclusion, the model offers a simple explanation of
why the world, for so long after the height of the global
?nancial crisis, is still mired in slow growth. During a pro-
longed and excessive boom bank pro?ts and capital were
materially increased by unrealised FVA pro?ts. These prof-
its justi?ed the pay-out of liquid assets weakening the
?nancial system. The additional capital further justi?ed
more debt ?nanced asset expansion. With the crisis bank
management realised that these unrealised capital items
were not permanent. Management was able to postpone
the recognition of FVA losses by using the ?exibility inher-
ent in FVA regulations, but lending was slowed to a point
re?ective of ‘‘safe’’ (capital excluding unrealised items)
capital levels. Lending activity will stay subdued until all
of these marked-up items have been worked off banks’ bal-
ance sheets.
Appendix A
A.1. Derivation of mathematical results
If C
0
¼
AÀL
A
then:
Let u = A–L and v = A; then
du
dA
¼ 1 and
dm
dA
¼ 1
@C
0
@A
¼
v
du
dA
À u
dv
dA
v
2
¼
A:1 À ðA À LÞ:1
A
2
¼
L
A
2
If C
2
¼
ð1þkÞAþBÀðLþBÞ
ð1þlÞ
P
r
i
a
i
þr
j
B
¼
AÀL
P
r
i
a
i
¼ C
1
then:
A
X
r
i
a
i
þ kA
X
r
i
a
i
þ B
X
r
i
a
i
À L
X
r
i
a
i
À B
X
r
i
a
i
¼ A
X
r
i
a
i
þlA
X
r
i
a
i
þ ABr
j
À L
X
r
i
a
i
ÀlL
X
r
i
a
i
À BLr
j
BðÀAr
j
þ Lr
j
Þ ¼ Àðk ÀlÞA
X
r
i
a
i
ÀlL
X
r
i
a
i
B ¼
ðk ÀlÞA
P
r
i
a
i
þlL
P
r
i
a
i
ðA À LÞr
j
¼
ðk ÀlÞA þlL
r
j
C
1
If C
3
¼
ð1þXkÞAþBÀðLþBÞ
ð1þXlÞ
P
r
i
a
i
þr
j
B
¼
AÀL
P
r
i
a
i
¼ C
1
then:
A
X
r
i
a
i
þXkA
X
r
i
a
i
þ B
X
r
i
a
i
À L
X
r
i
a
i
À B
X
r
i
a
i
¼ A
X
r
i
a
i
þXlA
X
r
i
a
i
ABr
j
À L
X
r
i
a
i
ÀXlL
X
r
i
a
i
À BLr
j
BðÀAr
j
þ Lr
j
Þ ¼ Àðk ÀlÞXA
X
r
i
a
i
ÀXlL
X
r
i
a
i
112 P. de Jager / Accounting, Organizations and Society 39 (2014) 97–116
B ¼
Xðk ÀlÞA þXlL
r
j
C
1
If B ¼
XðkÀlÞAþXlL
r
j
C
1
then:
Let u ¼ Xðk ÀlÞA þXlL and m = r
j
C
1
then
du
dX
¼ ðk ÀlÞA þlL and
dv
dX
¼ 0
@B
@X
¼
v
du
dX
À u
dv
dX
v
2
¼
r
j
C
1
Á ½ðk ÀlÞA þlL?
ðr
j
C
1
Þ
2
¼
ðk ÀlÞA þlL
r
j
C
1
A.2. Graphical representation of the development of Fig. 2
The following example will illustrate the main points. A
bank (or the total banking system) has $100 of total assets
(A) with 30% of those assets subject to FVA and 70% of those
assets subject to historical cost accounting. The bank has
equity capital of $6 (A–L) and deposit liabilities of $94 (L).
The capital ratio is thus 6/100 = 6% and the effective tax rate
is 30%. If the market value of the fair valued assets increase
by 10% then the value of k is (30 * (10% * (1–30%)))/
(30 + 70) = 2.1%. An increase in debt-?nanced assets of 2.1/
6% = $35 is required. After this increase the original capital
ratio is restored
AÀL
A
¼ ðð100 þ2:1 þ35Þ À ð94 þ35ÞÞ=
À
ð100 þ2:1 þ35Þ ¼ 6%Þ. Additional demand for ?nancial
assets is $35 and $35 of additional spending power is made
available to the real economy (see Fig. A).
The previous example of the bank above is developed
further in this paragraph. Suppose that the remuneration
and dividend policy of the bank is that 50% of after tax
pro?ts is paid as remuneration and that 25% of after tax
pro?ts is paid as dividends; in this case, the retention rate
of the bank is 25% (X). If only 25% of the 2.1% increase in
average assets is retained (100 * 2.1% * 25% = $0.525 is left
as additional capital and 100 * 2.1% * 75% = $1.575 is paid
out), then the debt-?nanced increase in ?nancial assets re-
quired to restore the 6% capital ratio is (2.1 * 25%)/
6% = $8.75. The asset expansion (B) and the remuneration
and dividend payments to the real economy (1 À X) in-
crease the spending power available to the real economy
by 1.575 + 8.75 = $10.325; substantially less than the $35
previously when X was equal to one. The new demand
for ?nancial assets is only $8.75 versus the $35 previously.
The reason for the differences is the gearing that takes
place inside this banking system (see Fig. B).
The previous example of the bank now continues. Given
k = À2.1% the reduction in ?nancial assets and deposits
required to restore the 6% original capital ratio is equal
to $-35. This reduces the demand for ?nancial assets as
well as the spending power available to the real economy.
In our example a 2.1% increase led to a $10.325 increase in
spending power for the real economy and increased de-
mand for ?nancial assets of $8.75. A decrease of 2.1% leads
to a reduction in spending power available to the real
economy of $35 and a reduction in the demand for
?nancial instruments of $35. The reason for this difference
between the increase in balance sheet required and the
Financial sector
Real economy
Credit
Fig. A. C
2
¼
ð1þkÞAþBÀðLþBÞ
ð1þlÞ
P
r
i
a
i
þr
j
B
.
Investment
Remuneration &
dividends
Financial sector
Real economy
Credit
Fig. B. C
3
¼
ð1þXkÞAþBÀðLþBÞ
ð1þXlÞ
P
r
i
a
i
þr
j
B
.
P. de Jager / Accounting, Organizations and Society 39 (2014) 97–116 113
decrease in balance sheet required to restore the original
capital ratio is because no contribution towards alleviating
the capital shortage can be expected from employees
and shareholders (X= 1) with the 2.1% decrease (see
Fig. C).
The previous example is now continued: thus, it has
already been shown that with k = 2.1% and X= 25%,
the increase in ?nancial assets and deposits required to
restore the 6% original capital ratio is equal to $8.75
and the increase in spending power available to the real
economy is $10.325. Let us now suppose that that the real
economy spends 50% of any additional spending power
on the purchase of ?nancial assets and that the demand
function for ?nancial assets is a simple linear function
of the form Q = 10 + 90P. The additional demand for ?nan-
cial assets is 10.325/2 + 8.75 = $13.9125. Using the
demand equation the initial price was 1 as the total of
?nancial assets available then equated $100. The total le-
vel of demand is now 100 + 13.9125 = $113.9125 and
solving the demand equation the new price equals 1.16.
As 30% of the assets are subject to fair value accounting
then 30% of that 16% increase will enter capital via FVA
and set the whole process in motion again (see Figs. D
and E).
Financial sector
Real economy
Credit
Investment
Remuneration &
dividends
Debt repayment
Fig. D.
@kðtÞ
@t
¼ f ðBðkÞ; XðkÞ; UÞ.
Financial sector
Real economy
Bank 2
Bank 1
Credit
Investment
Investment
Remuneration &
dividends
Debt repayment
Credit
Fig. E. (R
asset
À R
liability
) > 0.
Debt repayment
Financial sector
Real economy
Fig. C. k < 0.
114 P. de Jager / Accounting, Organizations and Society 39 (2014) 97–116
References
Akerlof, G. A., & Romer, P. M. (1993). Looting: The economic underworld
of bankruptcy for pro?t. Brookings Papers on Economic Activity,
1993(2), 1–73.
Allen, F., & Carletti, E. (2008). Mark-to-market accounting and liquidity
pricing. Journal of Accounting and Economics, 45(2), 358–378.
Andersson, T., Lee, E., Theodosopoulos, G., Yin, Y. A., & Haslam, C. (2014).
Accounting for the ?nancialized UK and US national business model.
Critical Perspectives on Accounting, 25(1), 78–91.
Ang, A., Piazzesi, M., & Wei, M. (2006). What does the yield curve tell us
about GDP growth? Journal of Econometrics, 131(1), 359–403.
Arnold, P. J. (2009). Global ?nancial crisis: The challenge to accounting
research. Accounting, Organizations and Society, 34(6–7), 803–809.
Badertscher, B. A., Burks, J. J., & Easton, P. D. (2012). A convenient
scapegoat: Fair value accounting by commercial banks during the
?nancial crisis. The Accounting Review, 87(1), 59–90.
Ball, R. (2008). Don’t blame the messenger ... or ignore the message. (22.06.09).
Bank for International Settlements (2006). International convergence of
capital measurement and capital standards: A Revised framework. Basel:
BIS. (10.02.13).
Bank for International Settlements (2011). Basel III: A global regulatory
framework for more resilient banks and banking systems – Revised
version, June 2011. Basel: BIS.
Bank for International Settlements (2012). 82nd Annual report. Basel: BIS.
Bank of England (2009). Financial stability report. Section 3. United
Kingdom: Bank of England.
Barth, M., & Landsman, W. (2010). How did ?nancial reporting contribute
to the ?nancial crisis? European Accounting Review, 19(3), 399–423.
Bello?ore, R., & Passarella, M. (2010). Minsky, the monetary circuit and
the current crisis: A sfc monetary accounting framework. Paper
presented at the conference ‘‘Can it happen again?’’, 1–2 October 2010,
University of Macerata, Italy.
Benston, G. J. (2006). Fair-value accounting: A cautionary tale from Enron.
The Journal of Accounting and Public Policy, 25(4), 465–484.
Bezemer, D. J. (2010). Understanding ?nancial crisis through accounting
models. Accounting, Organizations and Society, 35(7), 676–688.
Bischof, J., Brüggemann, U., & Daske, H. (2010). Relaxation of fair value
rules in times of crisis: An analysis of economic bene?ts and costs of the
amendment to IAS 39. or (06.02.13).
Black, W. K. (2005). The best way to rob a bank is to own one: How corporate
executives and politicians looted the S&L industry. Austin: University of
Texas Press.
Black, W. K. (2011). Still banking on fraud. Dollars & Sense, 1 June 2011.
(06.02.13).
Borio, C. (2012). The ?nancial cycle and macroeconomics: What have we
learnt?. BIS working papers no. 395. Basel: BIS.
Cabral, R. (2012). A perspective on the symptoms and causes of the
?nancial crisis. Journal of Banking & Finance, 37(2013), 103–117.
Einhorn, D. (2010). Fooling some of the people all of the time, a long short
(and now complete) story, updated with new epilogue. Hoboken, New
Jersey: Wiley.
Enria, A., Cappiello, L., Dierick, F., Grittini, S., Haralambous, A., Maddaloni,
A., et al. (April 2004). Fair value accounting and ?nancial stability.
Occasional paper series 13. European Central Bank.
Fiechter, P., & Meyer, C. (2011). Discretion in fair value measurement of
banks during the 2008 ?nancial crisis. or
(06.02.13).
Fisher, I. (1933). The debt-de?ation theory of great depressions.
Econometrica, 1(4), 337–357.
Friedman, M. (1953). The methodology of positive economics. In Essays in
positive economics. Chicago: University of Chicago Press.
Friedman, M., & Schwartz, A. J. (1971). A monetary history of the United
States, 1867–1960. New Jersey: Princeton University Press.
Gebhardt, G. (2012). Financial instruments in non-?nancial ?rm: What do
we know? Accounting and Business Research, 42(3), 267–289.
Gebhardt, G., & Novotny-Farkas, Z. (2011). Mandatory IFRS adoption and
accounting quality of European Banks. Journal of Business Finance &
Accounting, 38(3&4), 289–333.
Godley, W., & Lavoie, M. (2012). Monetary economics (2nd ed.).
Basingstoke: Palgrave Macmillan.
Haldane, A. G. (2011). Control rights (and wrongs). 24 October 2011.
Wincott Annual Memorial Lecture, Westminster, London. (12.04.13).
Howelss, P., & Bain, K. (2008). The economics of money, banking and
?nance: A European text (4th ed.). Gosport, Hampshire: Pearson
Education Limited.
ING Bank N.V. (2012). Annual ?nancial statements.
International Accounting Standards Board (2005). International ?nancial
reporting standard (IFRS) 7 ?nancial instruments: Disclosure. London:
IASB.
International Monetary Fund (2012). Global ?nancial stability report.
Washington: International Monetary Fund.
Investec Bank Limited (2012). Investec Bank Limited Group and company
annual ?nancial statements.
Kay, J. (2009). A boom based on little more than a bezzle. Financial Times.
(16.05.09).
Kay, J. (2012). Investors should resist the tyranny of quarterly earnings.
Financial Times (London) (29.02.12).
Keen, S. (2011). Debunking economics: Revised and expanded edition: The
naked emperor dethroned? London: Zed Books.
Khan, U. (2010). Does fair value accounting contribute to systemic risk in the
banking industry? Working paper. (22.07.13).
Krugman, P. (2009). The night they reread Minsky [Blog, 17 May]. (17.05.09).
Kusano, M. (2011). Fair value accounting and procyclicality: Accounting for
securitization.
(22.02.12).
Lahart, J. (2007). In time of tumult, obscure economist gains currency.
Wall Street Journal, 18(August), 1.
Laux, C., & Leuz, C. (2010). Did fair-value accounting contribute to the
?nancial crisis? Journal of Economic Perspectives, 24(1), 93–118.
Laux, C., & Leuz, C. (2009). The crisis of fair-value accounting: Making
sense of the recent debate. Accounting, Organizations and Society,
34(6–7), 826–834.
Livne, G., Markarian, G., & Milne, A. (2011). Bankers’ compensation and
fair value accounting. Journal of Corporate Finance, 17(4), 1096–1115.
Lo, A. W. (2012). Reading about the ?nancial crisis: A twenty-one-book
review. Journal of Economic Literature, 50(1), 151–178.
Magnan, M. L. (2009). Fair value accounting and the ?nancial crisis:
Messenger or contributor? Accounting Perspectives, 8(3), 189–213.
Mankiw, N. G. (2007). Macroeconomics (6th ed.). New York: Worth
Publishers.
McMahon, G. (2011). Mark-to-market’s real role in the crisis. The CPA
Journal (February 2011).
Mundy, S. (2012). Albuthnot chief blames accounting rules for loss.
Financial Times (London). 18 March.
Palley, T. I. (2010). The limits of Minsky’s ?nancial instability
hypothesis as an explanation of the crisis. Monthly Review, 61(11),
28–43.
Pinnuck, M. (2012). A review of the role of ?nancial reporting in the global
?nancial crisis. Australian Accounting Review, 60(2), 1–14.
Plantin, G., Sapra, H., & Shin, H. S. (2008). Marking to market: Panacea or
Pandora’s box? Journal of Accounting Research, 46(2), 435–460.
Posner, R. A. (2009). A failure of capitalism: The crisis of ‘08 and the descent
into depression. United States of America: Harvard University Press.
Pozen, R. C. (2009). Is it fair to blame fair value accounting for the
?nancial crisis? Harvard Business Review, 87(11), 84–92.
Rees, M. (2013). The unsecured boom may be over. (06.02.13).
Rolnick, A. J., & Weber, W. E. (1986). Gresham’s law or Gresham’s fallacy.
Journal of Political Economy, 94(1), 185–199.
Rose, P. S., & Hudgins, S. C. (2008). Bank management & ?nancial services
(7th ed.). New York: McGraw-Hill Irwin.
Ryan, S. G. (2008). Accounting in and for the subprime crisis. The
Accounting Review, 83(6), 1605–1638.
Scott, W. R. (2009). Financial accounting theory (5th ed.). United States of
America: Pearson Prentice Hall.
Shaffer, S. (2010). Fair value accounting: Villain or innocent victim: Exploring
the links between fair value accounting, bank regulatory capital and the
recent ?nancial crisis. Working paper, Federal Reserve Bank of Boston
Quantitative Analysis Unit. (12.04.13).
P. de Jager / Accounting, Organizations and Society 39 (2014) 97–116 115
Soros, G. (2012). George soros on the Euro crisis. Moneyweb (6 June).
Standard Bank of South Africa (2012). Annual ?nancial statements.
Suzuki, T. (2003). The accounting ?guration of business statistics as a
foundation for the spread of economic ideas. Accounting, Organizations
and Society, 28(2003), 65–95.
Tarrant, H. (2013). Nedbank’s Mike Brown on First Strut, personal loans and
Africa. (20.01.14).
Taylor, M. (2009). Innumerate bankers were ripe for a reckoning. Financial
Times (London). 16 December.
The Independent Commission on Banking (2011). Final report of the
independent commission on banking. London: Stationery Of?ce.
The National Commission on the Causes of the Financial and Economic
Crisis in the United States (2011). The ?nancial crisis inquiry report.
U.S. Government Printing Of?ce.
Toporowski, J. (2008). Minsky’s induced investment and business cycles.
Cambridge Journal of Economics, 32(5), 725–737.
Toporowski, J. (2010). Excess debt and asset de?ation. In S. Kates (Ed.),
Macroeconomic theory and its failings (pp. 221–234). Cheltenham, UK:
Edward Elgar Publishing Limited.
Turner, A. (2010). Banks are different: should accounting re?ect that fact?
Speech by Adair Turner, Chairman, FSA to the Institute of Chartered
Accountants in England and Wales (ICAEW), London. 21 January
2010.
(12.04.13).
United States Securities and Exchange Commission (2008). Report and
recommendations pursuant to section 133 of the emergency economic
stabilisation act of 2008: Study on mark-to-market accounting.
(05.08.10).
Veron, N. (2008). Fair value accounting is the wrong scapegoat for this crisis.
Bruegel Policy Contribution/May 2008.
(12.04.13).
Wallace, M. (2009). Is fair value accounting responsible for the ?nancial
crisis? Bank Accounting and Finance, 22(1), 9–18.
Whiteman, J. (2012). No assumptions please – We Brits like to consult the
facts. Financial Times, letter published on 22 December 2012.
Wolf, M. (2008). An inevitable earthquake. Business spectator.
(18.09.08).
Wood, D. (2010). Unrealised gains out of their hands. Risk.net. 4 February.
Wright, J. (2006). The yield curve and predicting recessions. FEDs Working
Paper No. 2006-7. Board of Governors of the Federal Reserve System –
Trade and Financial Studies Section.
Young, J. J. (1995). Getting the accounting ‘‘right’’: Accounting and the
savings and loan crisis. Accounting, Organizations and Society, 20(1),
55–80.
Zeff, S. A. (2007). The SEC rules historical cost accounting: 1934 to the
1970s. Accounting and Business Research, 37(3), 49–62.
116 P. de Jager / Accounting, Organizations and Society 39 (2014) 97–116
doc_676917743.pdf
A topic of recent interest in accounting research has been the investigation of the role of
fair value accounting (FVA) in the global financial crisis. This research focused on finding
a link during the crisis time-period and often states that ‘‘accounting is only a messenger’’.
The model presented in this paper emphasises finding the link before the crisis and
‘‘accounting as money.’’ Use is made of an accounting model of the economy due to the
inability of standard models of monetary transmission to incorporate global financial crisis
characteristics such as feedback effects, systemic risk and the centrality of the financial
sector in the crisis. The model shows FVA in banks to be an accelerator that amplifies
the financial cycle upswing. Feedback effects noted in the model include changes in the
demand for financial instruments and changes in demand in the real economy.
Minsky-like, crisis is shown to be endogenous to the model, working through the fragility
of balance sheets in the real sector as well as in the financial sector. Bank balance sheet
fragility is caused by bad capital driving out good capital, banks reaching for yield and
the inversion of the yield curve.
Fair value accounting, fragile bank balance sheets and crisis:
A model
Phillip de Jager
?
Department of Finance & Tax and the African Collaboration for Quantitative Finance & Risk Research, University of Cape Town, South Africa
a b s t r a c t
A topic of recent interest in accounting research has been the investigation of the role of
fair value accounting (FVA) in the global ?nancial crisis. This research focused on ?nding
a link during the crisis time-period and often states that ‘‘accounting is only a messenger’’.
The model presented in this paper emphasises ?nding the link before the crisis and
‘‘accounting as money.’’ Use is made of an accounting model of the economy due to the
inability of standard models of monetary transmission to incorporate global ?nancial crisis
characteristics such as feedback effects, systemic risk and the centrality of the ?nancial
sector in the crisis. The model shows FVA in banks to be an accelerator that ampli?es
the ?nancial cycle upswing. Feedback effects noted in the model include changes in the
demand for ?nancial instruments and changes in demand in the real economy.
Minsky-like, crisis is shown to be endogenous to the model, working through the fragility
of balance sheets in the real sector as well as in the ?nancial sector. Bank balance sheet
fragility is caused by bad capital driving out good capital, banks reaching for yield and
the inversion of the yield curve. The model shows that the practice of not meeting rising
credit demand with increasing credit supply is an essential control mechanism in the
?nancial cycle.
Ó 2014 Elsevier Ltd. All rights reserved.
Introduction
In 2012, more than four years after the Lehman Broth-
ers failure, the aftermath of the global ?nancial crisis was
still felt, and stable, self-sustaining growth continued to
elude the world economy (Bank for International Settle-
ments, 2012:ix). Ryan (2008:1606) identi?ed the global
?nancial crisis as the ‘‘signal researchable-teachable
moment of my two-decade-plus career as an accounting
academic’’. Numerous authors have started to investigate
the link between accounting (usually FVA) and the crisis
(Badertscher, Burks, & Easton, 2012; Ball, 2008; Barth &
Landsman, 2010; Laux & Leuz, 2009, 2010; Magnan,
2009; Plantin, Sapra, & Shin, 2008; Pozen, 2009; Shaffer,
2010; United States Securities, 2008; Veron, 2008; Wal-
lace, 2009). These studies tend to focus on establishing
the link between FVA and the crisis during the time-period
of the crisis.
1
It is the argument of this paper that the focus
should rather be on the time-period before the crisis (the
upswing); this corresponds with one of Pinnuck’s (2012:5)
conclusions in his review of the role of ?nancial reporting
in the crisis that ‘‘the existing debate has focused on the role
of FVA during the crisis, but has ignored the possibility that
the illusory FV gains on subprime securities before the crisis
may have masked some of the underlying problems’’. Inhttp://dx.doi.org/10.1016/j.aos.2014.01.004
0361-3682/Ó 2014 Elsevier Ltd. All rights reserved.
?
Tel.: +27 216502296.
E-mail address: [email protected]
1
Readers are referred to Section ‘The global ?nancial crisis’ in the
literature review where the few arguments for FVA’s role during the
upswing are discussed.
Accounting, Organizations and Society 39 (2014) 97–116
Contents lists available at ScienceDirect
Accounting, Organizations and Society
j our nal homepage: www. el sevi er. com/ l ocat e/ aos
addition, some of the studies (Ball, 2008; Wallace, 2009)
argued that accounting was only a messenger and thus
should not be linked to the crisis. A number of the remaining
studies (Laux & Leuz, 2009:826; Magnan, 2009:191; Pozen,
2009:86; United States Securities and Exchange Commis-
sion, 2009:2) mention this pervasive argument; this paper
argues against this view.
The primary objective of this paper is to provide a
reasonable alternative perspective on the relationship be-
tween FVA and the global ?nancial crisis; a perspective
that focuses on ?nding the link before the crisis (during
the upswing) and one that reminds accountants that in
banking, accounting is more than just a messenger:
when bank deposits are created by accounting entries,
accounting is money. To this end, a model will be devel-
oped in this paper that demonstrates the link between
accounting and bank capital regulations and helps to
aid understanding of the global ?nancial crisis. The paper
answers the call by Arnold (2009:806) to ‘‘stimulate a
revival of accounting scholarship aimed at understanding
the relationship between accounting practice and the
macro political and economic environment in which it
operates’’.
The nexus of the model to be developed will be the im-
pact of FVA on the regulatory capital of banks. Not all FVA
entries impact banks’ regulatory capital. The general prin-
ciple in Basel II is that Tier 1 capital should be equity cap-
ital made up from permanent shareholders’ equity and
disclosed reserves that, importantly, must have been
posted through pro?t and loss (Bank for International
Settlements, 2006:245). Thus, only FVA entries related to
trading instruments or instruments designated at fair
value are relevant for the model when values increase.
FVA increases in the value of available-for-sale instru-
ments are not covered because those increases would be
excluded from regulatory capital. When values decrease,
other-than-temporary-impairments of available-for-sale
instruments become relevant as these are posted through
pro?t and loss. Shaffer (2010) investigated the role of FVA
in the global ?nancial crisis by looking only at FVA for
available-for-sale instruments. He argued for the exclu-
sion of FVA entries emanating from the trading book be-
cause ‘‘There are few who would argue that fair value is
inappropriate for measuring investments held for trading
purposes where deep and active markets exist’’ (Shaffer,
2010:10). However, if most instruments held for trading
purposes were traded on deep and active markets, then
they would have been shown at level 1 (mark-to-market)
on the fair value hierarchy. The United States Securities
and Exchange Commission (2008:62) found that in the
case of banks, most derivatives and trading assets were
at level 2 and not at level 1; this was especially true for
investment banks where almost all derivatives and trad-
ing assets were at level 2 and not level 1. The conclusion
is thus that trading assets and derivatives were not traded
on deep and active markets and that the exclusion of FVA
adjustments on those instruments was not justi?ed. In
addition to the exclusion of FVA entries emanating from
the trading book, Shaffer (2010:10) also argued for the
exclusion of FVA entries emanating from ‘‘designated at
fair value’’ instruments as this designation was ‘‘an
explicit decision by management to value these assets this
way and thus they are not included in the analysis’’.
The implicit assumption that management will not abuse
the fair value option is questionable. An example of the
impact of the preceding arguments will help to illustrate
this. Assume, therefore, that Bank X buys piecemeal $1
billion of AAA rated subprime securities yielding 5% whilst
‘‘normal’’ AAA securities yield 3%
2
with an assumed matu-
rity of 10 years (monthly). The portfolio of securities is then
designated at fair value and carried at level 2 of the fair
value hierarchy. The ?nancial result of the foregoing is an
immediate gain of $98 million (payments on $1 billion at
5% interest for 10 years present valued at 3% less the cost
of $1 billion) posted to net pro?t. Badertscher et al.
(2012) and Barth and Landsman (2010) continued this
exclusive focus on FVA for available-for-sale instruments
as the only possible avenue through which FVA could have
played a role in the global ?nancial crisis. In contrast, for
the reasons explained above, the model in this study is
based on the assumption that FVA materially impacts pro?t
and regulatory capital; this is one of the contributions that
this study makes to the accounting literature. Evidence for
the signi?cance of FVA entries that affect pro?t and loss can
be found in the study by the United States Securities and
Exchange Commission (2008:4) which observed that ‘‘fair
value measurements did signi?cantly affect ?nancial insti-
tutions’ reported income’’.
In addition, this study contributes to accounting liter-
ature in the following speci?c ways. Thus, several writers
have argued that FVA is procyclical (Enria et al., 2004:45;
Laux & Leuz, 2009:828); their views are repeated in this
paper but this time within a formal framework and
nuanced with the argument that the role of FVA during
a downturn is likely to be blunted by the ?exibility al-
lowed in FVA regulations. An implication of the foregoing
is that the role of FVA in the global ?nancial crisis should
rather be sought before or after the crisis and not during
the crisis. The model developed explains why McMahon
(2011:54) is accurate when stating ‘‘Mark-to-market feed-
back loops are not exclusive to declining periods – feed-
back loops also occur during upswings, although not
quite as fast’’ (this paper’s emphasis). The model shows
that the seeds for the coming downturn are sown during
the upswing with the purest role for FVA being the
replacement in bank capital of liquid assets with unrea-
lised (and thus risky) FVA gains; bad capital driving out
good capital. The simple running example used to demon-
strate the model that follows shows how a small change
in the value of fair valued assets can lead to a much larger
effect (a factor of 16 in the example but 30 is not uncom-
mon) on a bank’s balance sheet; this contradicts the argu-
ment that the fair value measurement of ?nancial
instruments must be pervasive to be dangerous (Geb-
hardt, 2012:267&271; United States Securities & Ex-
change Commission, 2008:4).
2
Even though instruments derived from securitised subprime mortgages
were often rated AAA (investment grade) by rating agencies prior to the
crisis, those instruments continued to offer higher returns than ‘‘normal’’
AAA securities (Lo, 2012:152).
98 P. de Jager / Accounting, Organizations and Society 39 (2014) 97–116
The model also makes a potential contribution to mon-
etary economics. Hyman Minsky’s ?nancial instability
hypothesis bene?ted from the global ?nancial crisis in
the sense that many saw the crisis as con?rming his pre-
dictions (see for example Krugman, 2009; Lahart, 2007;
Wolf, 2008). Minsky’s ‘‘notion of the ?rmas a balance sheet
of assets and liabilities, as opposed to the [traditional] no-
tion of the ?rm as an entrepreneur making production
decisions’’ (Toporowski, 2008:730) can be linked to the
type of model developed in this paper; a balance sheet cen-
tred model. The argument that fragility develops in the
balance sheets of the real economy is repeated in the mod-
el described in this paper, and is similar to the way the ‘‘ba-
sic Minsky cycle concerns the evolution of patterns of
?nancing arrangements and it captures the phenomenon
of emerging ?nancial fragility in business and household
balance sheets’’ (Palley, 2010:30). The model described in
this paper is different in the way the focus is not on this
fragility of individuals’ or businesses’ balance sheets but
rather on how fragility develops in the balance sheets of
banks. This aligns with the ‘‘complete reversal of the tradi-
tional monetary circuit, where the banking system is as-
sumed to ?nance business sector activity’’ and ‘‘it is the
dynamics of the banks/?nancial markets axis [. . .] which
has taken center stage’’ (Bello?ore & Passarella, 2010:10).
A second area where a contribution to monetary eco-
nomics is possible is in the presentation and exploration
of the impact of FVA to an audience of economists. Godley
and Lavoie (2012: Preface xxxvii) write that by not under-
standing in?ation accounting in the 1970s, economists
underestimated ‘‘the extent to which stocks of ?nancial as-
sets would rise in nominal terms’’ resulting in ‘‘some bad
projections’’. Similarly FVA can result in a signi?cant in-
crease in the nominal values of ?nancial assets during an
upswing. The model developed in this paper also has
important links with a recent paper by Claudio Borio, the
Deputy Head of the Monetary and Economic Department
and Director of Research and Statistics at the Bank for Inter-
national Settlements (Borio, 2012). He calls for the study of
the ‘‘?nancial cycle’’ as part of macroeconomics (Borio,
2012:1) and adds that a ‘‘?nancial boom should not just
precede the bust but cause it’’ (Borio, 2012:8); this call ties
in with this paper’s focus on the time-period before the
crisis. He reminds us that ‘‘the ?nancial system does not
just allocate, but also generates, purchasing power’’ (Borio,
2012:2) which ties back to accounting as money.
Bezemer (2010:676) referred pointed out that
‘‘‘accounting’ (or ?ow-of-funds) models of the economy
are the shared mindset of those analysts who worried
about a credit-cum-debt crisis followed by recession,
before the policy and academic establishment did. They
are ‘accounting’ models in the sense that they represent
households’, ?rms’ and governments’ balance sheets and
their interrelations.’’ Accounting models are also used in
two other streams of literature related to this paper.
Macroeconomic accounting (national accounting) is
argued to be both a measure of macroeconomic reality
and a constituent of that reality (Suzuki, 2003). Similarly
in this paper accounting both measures the extent of bank
balance sheets and indirectly impacts those balance
sheets with its effect on the money supply. Andersson,
Lee, Theodosopoulos, Yin, and Haslam (2014:79), within
the literature on the ?nancialization of the economy, argue
that ‘‘this process of ?nancialization can best be under-
stood within an augmented accounting framework . . .’’
Other reasons why a stock-?ow consistent accounting
framework of the economy
3
was chosen as a foundation
for the model developed in this paper include the role of bal-
ance sheets in past crises and the need to incorporate global
?nancial crisis characteristics such as feedback effects,
systemic risk and the centrality of the ?nancial sector in
the model. This choice of model is in contrast to the more
mainstream theories of the monetary transmission mecha-
nism. A difference between the typical stock-?ow consistent
accounting model and the model described in this paper is
that the supply of credit is not assumed simply to be a re-
sponse to the demand for credit; instead, this paper starts
with the supply of credit and its in?uence on the demand
for it. In this way it aims to be aligned better with what
was observed during the global ?nancial crisis; that is, an
oversupply of credit before the crisis and an undersupply
afterwards.
The model developed in this paper describes and ex-
plains many of the characteristics of the global ?nancial
crisis as well as aspects of other past crises. It is shown
how the change in banks’ capital ratios that result from
FVA can set in motion a procyclical process with the initial
aim being the restoration of the earlier capital ratio. A
debt-?nanced expansion of bank balance sheets and/or
the payment of remuneration and dividends results from
increases in fair value. The shrinkage of bank balance
sheets can result from a decrease in fair value but, argu-
ably, can be blunted by management action and the ?exi-
bility in FVA. Feedback effects in the model maintain the
momentum in the system once in motion. The feedback ef-
fects are the result of the additional (or decreased) demand
for ?nancial assets generated as well as the additional (or
decreased) spending power made available to the real
economy. In the instance where the additional spending
power (or decrease in spending power, as the case may
be) is made available solely to the banking sector the feed-
back effects are strongest. The model also highlights the
role of four processes that can result in a crisis/regime
change. The ?rst of these is the increased fragility of real
economy balance sheets during an expansion. The second
and third relate to two sources of bank balance sheet fra-
gility; bad capital driving out good capital and the reach
for yield. Lastly, interbank transactions lead to a yield
curve inversion effect, depriving banks of the easy pro?ts
available from maturity transformation; causing more
reach for yield. The crisis attributes explained by the model
include the booms in predatory lending and also the pay-
ment of dividends and remuneration during the upswing,
understanding the growth in the ?nancial sector relative
to the real sector, illuminating the temporary impact of
the reprieve obtained from FVA during the downturn and
explaining the occurrence of the Great Moderation. In the
3
According to Bello?ore and Passarella (2010:7) these models are
‘‘dynamic, consider the effects of stock magnitudes on ?ow variables, and
explicitly represent the role of the banking system’’. The model developed
in this paper can be described similarly.
P. de Jager / Accounting, Organizations and Society 39 (2014) 97–116 99
?nal discussion the bene?t of the perspective offered by
the model in this paper will brie?y be illustrated by using
it to understand the results of another study by Cabral
(2012) on banks and the global ?nancial crisis.
The paper proceeds as follows: Section ‘The historic role
of FVA in crises’ sets the scene by presenting FVA and
brie?y discussing the role of FVA in previous crises; Sec-
tion ‘Monetary transmission mechanism’ positions the
model developed within the monetary economics litera-
ture; Section ‘Model development’ develops and discusses
the model proposed; Section ‘Empirical example’ provides
an empirical example where the model is used to interpret
the information; Section ‘Implications, shortcomings and
opportunities for further research’ discusses the implica-
tions, shortcomings and opportunities for further research
resulting from the model and Section ‘Conclusion’ provides
the conclusion.
The historic role of FVA in crises
The purpose of this section is to overview the impact of
FVA during past crises. The information obtained will be
useful to benchmark the reasonableness of the model that
will be developed later in this paper. The crises included in
this overview are the Great Depression, the savings and
loans (S&L) crisis, the Enron crisis and the global ?nancial
crisis. But, ?rst it is necessary to de?ne FVA, to differentiate
it from mark-to-market accounting and to consider care-
fully how FVA impacts bank regulatory capital.
FVA is more complex than mark-to-market accounting.
Fair value is the exchange value in an idealised market and
can be determined in three ways, in order of preference
4
(the de?nitions that follow are based on the annual ?nancial
statements of Investec Bank Limited (Investec Bank Limited,
2012:148):
Mark-to-market accounting (level 1): Quoted (unad-
justed) prices in active markets for identical
instruments.
Mix of mark-to-market accounting and modelling (level
2): Model inputs other than quoted prices included
within level 1 that are observable for the instruments,
either directly (i.e. as prices) or indirectly (i.e. derived
from prices). This paper interprets this to mean the
modelling of inputs that are publically observable.
Mark-to-model (level 3): Signi?cant model inputs are not
based on observable market data.
FVA is thus more complicated than mark-to-market
accounting, but the two are closely related. For brevity this
paper will treat the two as equivalent.
The Great Depression
According to Scott (2009:3) the frequent upwards reval-
uation of capital assets was one of the practices of the
1920s to receive criticism after the crisis. This advanced
the case for historical cost accounting. Zeff (2007) concurs
and writes that the United States of America was a bastion
of predominantly historical cost accounting for inventories
and ?xed assets from 1934 to the 1970s. Pressure from the
Securities and Exchange Commission was a driving force
behind this insistence on historical cost accounting.
The use of mark-to-market accounting in banks during
the Great Depression can be found in the work of Friedman
and Schwartz (1971: chap. 7, 100–102). They describe and
provide evidence of how the fall in the market value of
bond portfolios was the most important source of bank
capital impairment during the Great Depression. They also
point out that, paradoxically, the most liquid assets were
the most serious threat to bank solvency due to the
mark-to-market regime for liquid assets rather than bonds
for ‘‘which there was no good market and few quotations’’
carried at cost. The United States Securities and Exchange
Commission (2008:34) wrote that mark-to-market
accounting for banks was changed in 1938 for reasons that
included the impact of mark-to-market accounting on
bank behaviour: ‘‘Further, prior to 1938, banking organiza-
tions were required for supervisory purposes to use market
value accounting for their investment securities portfolios.
Serious concerns on the part of the U.S. Treasury and the
bank regulators over how this affected the banks’ ?nancial
performance and investment decisions led the agencies to
abandon in that year the use of this accounting concept for
supervisory purposes’’.
Financial economist Fisher (1933) wrote that over-
indebtedness and de?ation were the two dominant factors
in all the great booms and depressions. He then proceeded
to describe a chain of consequences in a depression with
numerous feedback mechanisms driving distressed selling
of assets.
The ?ndings from the Great Depression indicate a role
for mark-to-market accounting in banks before and during
the crisis. The upward valuation of selected ?nancial
instruments during the upswing could have made banks
appear healthier, contributing to the over indebtedness of
banks and individuals. Moreover, the mark-to-market
accounting of liquid bond portfolios was clearly shown
by Friedman and Schwartz (1971) to have a negative effect
on bank lending during the downturn. Both of these effects
involve feedback processes.
The savings and loans crisis
The savings and loans (S&L) crisis of the 1980s and
1990s was caused by the failure of a signi?cant proportion
of the savings and loan associations in the United States of
America. A savings and loan association or ‘‘thrift’’ is a
?nancial institution that accepts deposits and makes mort-
gage, car and other personal loans to individual members.
According to Black (2005:xiii) ‘‘a wave of control fraud
ravaged the S&L industry’’. In a later publication he writes:
‘‘Control fraud occurs when the executives at a seemingly
legitimate ?rm use their control to loot the ?rm and its
shareholders and creditors’’ (Black, 2011). Accounting is
strongly linked by Black to control fraud: ‘‘Control frauds,
using accounting fraud as their primary weapon and shield,
4
A formal order of preference only entered accounting standards under
the International Accounting Standards Board (IASB) in October 2008 with
an amendment to International Accounting Standard (IAS) 39 which
adopted the American tiered system of fair value determination.
100 P. de Jager / Accounting, Organizations and Society 39 (2014) 97–116
typically report sensational pro?ts, followed by cata-
strophic failure. These ?ctitious pro?ts provide the means
for sophisticated, fraudulent CEOs to use common corpo-
rate mechanisms such as stock bonuses to convert ?rm
assets to their personal bene?t’’ (Black, 2005:xiv). Akerlof
and Romer (1993:2) ascribe what happened to looting;
‘‘to go broke for pro?t at society’s expense’’. They continue:
‘‘Bankruptcy for pro?t will occur if poor accounting, lax
regulation, or lowpenalties for abuse give owners an incen-
tive to pay themselves more than their ?rms are worth and
then default. . .’’ It is clear that accounting is important in
both instances to justify high corporate pay-outs.
The accounting profession focused attention on another
aspect of the crisis. Most thrifts extended loans on a ?xed
cost basis and obtained capital on a variable cost basis.
When interest rates rose in the 1980s, the industry found
itself in a dif?cult position: assets were generating less re-
turn than the cost of ?nancing them. Accountants argued
that historical cost accounting was hiding the bankruptcy
of the thrifts and that the fair valuation of the thrifts’ assets
would have revealed these problems (Young, 1995:72–75).
The ?ndings from the S&L crisis indicate a role for
accounting in thrifts before and during the crisis. Fictitious
pro?ts can be used to make an institution seem healthier
and to justify unreasonable pay-outs. During the downturn
FVA could have contributed towards revealing the bank-
ruptcy of the thrifts.
The Enron crisis
The Enron crisis was not nearly as systemic or impor-
tant as the other crises. Nevertheless, it does deserve a
mention in relation to the role of FVA in that crisis. Benston
(2006:465) ?nds that FVA was ‘‘substantially responsible’’
for the demise of Enron. FVA in Enron was used ‘‘opportu-
nistically to in?ate reported net income’’ and rationalise
excessive remuneration (Benston, 2006:466). FVA rarely
reduced income as ‘‘contrary to the way fair-value
accounting should be used, reductions in value rarely were
recognized and recorded because they either were ignored
or were assumed to be temporary.’’ (Benston, 2006:466).
FVA is also mentioned in another American scandal during
the same time period. Einhorn (2010:398) writes ‘‘. . .as the
crisis unfolded (it became apparent) that Allied’s abuse of
fair-value accounting was more prevalent in corporate
America than I had realized’’ in his telling of the Allied
scandal that he was involved in during the 2000s.
The global ?nancial crisis
The global ?nancial crisis was triggered by a sharp fall
in house prices in the United States of America (Posner,
2009:vii). It grew from the subprime crisis, into the credit
crisis, then into a ?nancial crisis and ?nally into a global
?nancial crisis. There are opinions that accounting, espe-
cially FVA, played a key role in causing or at least in wors-
ening the crisis. At the extreme end of this viewpoint is the
opinion held by Steve Forbes, chairman of Forbes Media,
that mark-to-market accounting was the ‘‘principal
reason’’ that the ?nancial system melted down in 2008
(Pozen, 2009:85). It was mentioned in the introduction
that most accounting research focused on the role of FVA
during the crisis and found limited evidence of a signi?cant
role at that time. This is to be expected as FVA contains at
least two built-in mechanisms by which banks can avoid
recognising losses; see Section ‘Implications of the basic
model’ for a detailed description.
Theoretical models by Allen and Carletti (2008) and by
Plantin et al. (2008) examined the impact of mark-to-
market accounting on capital markets and show that
mark-to-market accounting worsens a downturn. Their
studies have a shortcoming when attempting to explain
the global ?nancial crisis; mark-to-market accounting is
not the same as FVA and thus managers can use the ?exi-
bility inherent in fair value accounting to not accept mar-
ket values that they believe are incorrect. This is argued
by Laux and Leuz (2009:827) who also acknowledge that
FVA can be procyclical both during the boom and the bust
and that behavioural issues can make implementation of
FVA dif?cult. They argue that the increase in pro?t and
capital from asset value increases under FVA is also avail-
able as ‘‘gains trading’’ under historical cost accounting
and thus that the two types of accounting are similar dur-
ing the upswing, but this argument is not exactly correct.
FVA is a much faster process than gains trading that is
inherently lumpy. Bonuses and dividends paid from gains
trading are paid from increased cash resources whilst bo-
nuses and dividends from FVA gains only decrease cash
resources. The implication is that FVA is more able than
gains trading to be at the centre of a feedback process
and that FVA during a boom can leave the banking system
with reduced cash resources.
A role for FVA during the upswing is less clear from
accounting research but evidence from before the crisis is
starting to appear. Livne, Markarian, and Milne (2011) ?nd
that FVA gains are remuneration relevant. In the popular
press it seems to be common knowledge that FVA allowed
the payment of remuneration and dividends from imagi-
nary
5
pro?ts during the boom (Haldane, 2011; Kay, 2009,
2012; Mundy, 2012; Taylor, 2009; Wood, 2010). The effect
of such payments would be to weaken the banking system
as risk-free capital in the form of liquid assets left the banks
to be replaced by risky capital (potentially temporary value
increases). McMahon (2011) proposes that FVA plays an
integral role in feedback loops behind excessive lending
(and remuneration) during the boom and a lack of lending
during the bust. Magnan (2009:207) argues that a feedback
loop exists between FVA and market prices working through
the information channel. Khan (2010:5) does not ?nd sup-
port that the information channel is behind ‘‘the increased
bank contagion during crises associated with a more fair
value-oriented reporting regime’’. The focus of this paper is
on the money channel.
Conclusion from review of crises
One factor common to all of the four crises reviewed
was the role of ?rm (including banks) balance sheets. The
5
Terms used include ‘‘imaginary pro?ts’’, ‘‘spurious pro?ts’’ and ‘‘unre-
liable gains’’.
P. de Jager / Accounting, Organizations and Society 39 (2014) 97–116 101
crises have indicated a role for FVA both before and after
each crisis. Before a crisis, FVA can lead to ?rms or banks
appearing healthier than they are (Enron is a prime exam-
ple) and can thus facilitate further asset expansion
?nanced by debt (a feedback process). Based on what hap-
pened during the S&L crisis it could be expected that FVA
would reveal problems faster than historical cost account-
ing. But this revelation by FVA during the downturn can
lead to a downward spiral with forced sales of assets and
shrinking balance sheets (another feedback process);
alternatively this might not happen at all, as argued later
in ‘Implications of the basic model’.
Monetary transmission mechanism
The model to be developed inter alia describes the inter-
action between the real and the ?nancial economy. This
section starts by describing the mainstream explanations
for how money affects the real economy. Thereafter, argu-
ments are presented for a natural progression to Post-Key-
nesian stock-?ow consistent accounting models of the
economy. Similarities and differences between the model
developed in this paper and the standard stock-?ow con-
sistent accounting framework will round off this section.
According to Mankiw (2007:83) the quantity theory of
money provides the leading explanation of how money af-
fects the economy in the long run. According to this theory
the money supply has a direct, proportional relationship to
the price level. In the long run ‘‘monetary neutrality is
approximately correct’’ (Mankiw, 2007:109). The main dif-
?culty with this approach is that central banks cannot and
do not attempt directly to control the rate of growth of the
money supply (Howelss & Bain, 2008:273). A more realistic
approach should start with short-term interest rates as a
policy instrument. Howelss and Bain (2008:273) refer to
the interest rate control approach as an example of endog-
enous money and explain that money supply, via bank ac-
tions, adjusts to credit demand. The latter approach is
more realistic and, importantly, places commercial banks
in the middle of the money supply process, but the neutral-
ity of money assumption is maintained.
The neutrality of money argument is dif?cult to sustain
in the short run (Mankiw, 2007:536) and arguably even in
the long run when one considers that both the Great
Depression and the global ?nancial crisis started in the
?nancial sector and thereafter engulfed the real economy
for several years. In addition, accounting-orientated aca-
demic papers were published, after the global ?nancial cri-
sis, which encouraged researchers to ?nd links between
accounting and macroeconomic outcomes (Arnold, 2009;
Bezemer, 2010). The paper by Bezemer, amongst other
things, introduced accounting (or stock-?ow consistent)
macroeconomic models.
Authors such as Keen (2011) and Godley and Lavoie
(2012) present explicit accounting models of the economy
where accounting identities (not the equilibrium concept)
are the determinants of model outcomes in response to
shocks in the environment or in policy (Bezemer,
2010:679). In addition to this obvious link to accounting,
there are other reasons why this type of model forms a
?rm foundation for the development of a model of bank
behaviour, described in this paper, that links FVA, bank
capital regulation, management behaviour and real eco-
nomic outcomes. These models are not static or based on
equilibrium and can thus accommodate certain de?ning
characteristics of the global ?nancial crisis, feedback ef-
fects (Bank for International Settlements, 2012; McMahon,
2011; Soros, 2012) and the related concept of too-big-to-
fail banks (Haldane, 2011). The models are also not built
up from individual behaviour and summated to systemic
behaviour (microfoundations); they can therefore incorpo-
rate, easily, emergent properties such as systemic risk.
Finally, the models emphasise banks (and the provision
of credit) (Godley & Lavoie, 2012:17; Keen, 2011) which
is the sector most impacted by FVA.
Another author who emphasised the role of banks in
the economy and who rapidly gained prominence after
the global ?nancial crisis was Hyman Minsky (Bello?ore
& Passarella, 2010; Keen, 2011). Minsky’s Financial Insta-
bility Hypothesis is based on the idea that in times of
?nancial stability the actors in the system and the system
itself move towards ?nancial fragility. His focus on balance
sheets and the idea that the key to the downturn is to be
found during the preceding upswing will be incorporated
into the suggested model. A typical schematic overview
of a stock-?ow consistent model is shown in Fig. 1.
Fig. 1. Typical stock-?ow consistent model of the economy. Source: Bezemer (2010:683)
102 P. de Jager / Accounting, Organizations and Society 39 (2014) 97–116
The model to be presented in this paper is much simpler
and its advantage is that the concepts under investigation
are emphasised. An arguably negative impact is that the
model is a partial one in the sense that it does not describe
everything about a modern economy. However, this partial
perspective does meet the Friedman signi?cance level of
‘‘explain much by little’’ (Friedman, 1953:14). In the model
the real economy including consumers, producers and
government, is collapsed into one and the ?nancial sector
consists of a combination of two banks (see Fig. 2).
The model developed and described below is Post-Key-
nesian in the sense that the principle of effective demand is
assumed to hold in the real markets; that is, markets do
not automatically clear. An important difference from the
typical ?ow-of-funds model is that credit supply is not as-
sumed just to be a response to credit demand (endogenous
money supply) (Godley & Lavoie, 2012:127–128). On the
contrary, bank behaviour is assumed to in?uence credit
supply. This assumption is consistent with what was found
during and after the global ?nancial crisis; an oversupply
of credit before the crisis and an undersupply after the
crisis (Bank for International Settlements, 2012; Bank of
England, 2009:43:ix&1; International Monetary Fund,
2012:xi&12; The Independent Commission on Banking,
2011:8&16). This primacy of credit supply is also sup-
ported by a letter from Dr. John Whiteman in the Financial
Times (Whiteman, 2012) where he states that ‘‘. . .borrow-
ing is not driven by the price of credit (as common sense
might assume), but rather by the sheer availability of cred-
it in the ?rst place’’. The argument is perhaps best made
with a quote from the CEO of a micro-lender (African Bank
Limited) in South Africa: ‘‘A credit cycle is always going to
be driven from a supply dynamic, not a demand dynamic
and markets overheat because suppliers push too much
credit into the system’’ (Rees, 2013).
Model development
Basic model
The initial model is based on the work of Masaki Kusano
(Kusano, 2011). The familiar accounting equation of Assets
(A) – Liabilities (L) = Equity (E) is the starting point in the
development of his model that links accounting capital
with bank capital regulation and bank asset expansion.
This section will then add the payment of remuneration
and dividends from FVA gains to his model as well as con-
sider the results of the model when FVA generates losses.
Bank management are incentivised to minimise the
capital ratio C
0
¼
AÀL
A
À Á
as this will maximise return on
equity, arguably the most popular bank performance met-
ric as it is a measure of the rate of return ?owing to share-
holders (Rose & Hudgins, 2008:167). Of course, bank
regulators do not allow banks to leverage to in?nity and
they demand that banks keep a minimum capital buffer
to protect depositors in the case of default. The calculation
of
@C
0
@A
shows that the capital ratio de?ned above is directly
related to the value of assets:
@C
0
@A
¼
L
A
2
> 0 ð1Þ
The implication is that an increase in asset values will
increase the capital ratio (which is bad for return on equi-
ty); management will want somehow to manage or neu-
tralise this increase in the capital ratio. The capital ratios
6
of the ?ve largest South African universal banks for the per-
iod before the global ?nancial crisis illustrate this point
Fig. 3 (based on data provided by the South African Reserve
Bank). The period was one of high bank pro?ts (by fair value
accounting and other means) and yet capital ratios stayed
steady or declined; high pro?ts, ceteris paribus, should have
increased the capital ratio.
The Basel capital accords de?ne their own version of the
capital ratio, where accounting capital is expressed as a ra-
tio of risk-weighted assets:
C
1
¼
A À L
P
r
i
a
i
; ð2Þ
where A ¼
P
a
i
or the total asset value of the bank is the
sum of the individual assets, L is the total value of the lia-
bilities of the bank and r
i
is the risk-weight applicable to
each asset a
i
according to the Basel Accords.
Financial sector
Real economy
Bank 2
Bank 1
Credit
Investment
Investment
Remuneration &
dividends
Debt repayment
Credit
Fig. 2. Schematic representation of the model developed in this paper (this is the ?nal completed model; the schematic representations depicting the
development of the model is illustrated graphically with an accompanying example in Appendix A and numbered Figs. A-E).
6
This is the capital ratio as de?ned immediately above (C
0
¼
AÀL
A
) and not
the Basel ‘‘risk-weighted’’ capital ratio. In effect this method risk-weights
all assets at 100%.
P. de Jager / Accounting, Organizations and Society 39 (2014) 97–116 103
Suppose that the market value of all assets
7
increases by
a constant k, where k = k(t) and k > 0.
8
The total asset value A
is now equal to (1 + k)A. If assets are measured at fair value
then during an upswing the capital ratio can be written as:
C
2
¼
ð1 þ kÞA À L
ð1 þlÞ
P
r
i
a
i
; ð3Þ
where l is the increase in the total of risk-weighted assets.
As the risk-weights under the Basel Accords
9
in general
meets 0 6 r
i
6 1 the following can be assumed: 0 < l 6 k.
10
Assuming that management is planning to neutralise
the increase in the capital ratio by originating additional
assets denoted by B and fully ?nanced by debt
11
the follow-
ing equation is obtained:
.03
.04
.05
.06
.07
.08
00 01 02 03 04 05 06 07 08 09 10
Absa
.03
.04
.05
.06
.07
.08
00 01 02 03 04 05 06 07 08 09 10
Fnb
.05
.06
.07
.08
.09
.10
.11
.12
00 01 02 03 04 05 06 07 08 09 10
Investec
.04
.05
.06
.07
.08
.09
.10
00 01 02 03 04 05 06 07 08 09 10
Nedbank
.03
.04
.05
.06
.07
.08
00 01 02 03 04 05 06 07 08 09 10
Standard
Fig. 3. Monthly capital ratio per bank January 2000–August 2010 (shaded areas indicate post-crisis time period).
7
All the assets of a bank are not normally fair valued but the complexity
of modelling different types of assets is not necessary. Refer to the example
in the appendix where a bank is assumed to have $30 of fair valued assets
(through pro?t and loss) and $70 of historical cost assets. If the market
value of the fair valued assets increases by 10% and the tax rate is 30% then
the value of k is (30 * (10% * (1 À 30%)))/(30 + 70) = 2.1%.
8
It may seem as if the possibility that liabilities can also experience an
increase in market value when assets increase in market value is ignored.
Given that banks mostly own interest-rate sensitive assets of longer
maturity than their interest-rate sensitive liabilities, it is to be expected
that DA PDL when interest rates decrease. This state of affairs can be just
as effectively modelled with DA and L kept constant.
9
Under Basel I the risk-weightings for most categories of assets were
below 100%. Basel II and III moved away from rougher asset categories to
risk-weightings for individual assets or portfolios of assets, based on
individual bank history, that re?ect the ‘‘expected unexpected’’ loss (not
routine, high frequency events) (Gebhardt & Novotny-Farkas, 2011:295) on
that asset or portfolio. The planned-for unexpected losses on bank assets
should be less than 100% as banks are pro?t seeking organisations.
10
An anonymous reviewer pointed out that over time (not currently in
the model) there is the tendency for l to increase asymptotically to k. This
aspect is not explored further in the model.
11
The goal of the exercise is to bring the capital ratio down again after the
increase in equity caused by the increase in asset values and thus an
increase in debt is required.
104 P. de Jager / Accounting, Organizations and Society 39 (2014) 97–116
C
2
¼
ð1 þ kÞA þ B À ðL þ BÞ
ð1 þlÞ
P
r
i
a
i
þ r
j
B
¼
A À L
P
r
i
a
i
¼ C
1
; ð4Þ
where r
j
is the risk-weight of the originated/purchased,
debt-?nanced assets. It should be noted that although
the B added to the asset base and the B added to liabilities
are of the same value, the Bs are not different sides of the
same ?nancial instrument.
12
An increase in deposits
increases the money supply or the spending power available
in the economy. Solving for B:
B
1
¼
ðk ÀlÞA þlL
r
j
C
1
: ð5Þ
Fig. A in the Appendix provides a schematic representa-
tion followed by a narrative description of the model up to
this point. Management neutralised the increase in the
capital ratio by investing in additional ?nancial assets fully
?nanced by debt; this increases the demand for ?nancial
assets in the ?nancial sector. Bank deposits created are
money in the banking systemand so this increases demand
in the real economy. If time were already part of the model
then a few comments regarding feedback processes would
be possible at this stage. However, time will only be
included in the model at a later stage of its development.
The increase in the market value of ?nancial assets con-
tributes to k (an increase in the market value of all bank as-
sets) and thus provides an impulse for the asset expansion
cycle to start again. The additional demand for ?nancial as-
sets, ceteris paribus, increases the market value of ?nancial
assets and thus reduces the return on those assets. To keep
returns constant, bank management is incentivised to take
on more risk; for example, to invest in riskier assets or
originate riskier loans. The other side to additional demand
for ?nancial assets is the provision of additional spending
power (credit) to the real economy. Over the longer term,
the extension of credit to the real economy can only be
sustained if the additional capital is invested in productiv-
ity enhancing real assets. With increased productivity the
real economy might be in a position to pay back the capital
and interest. If not, the only way for the real economy to
keep a constant level of demand is to receive more and
more credit fromthe ?nancial sector. In a world of decreas-
ing marginal returns it is safe to conclude that over the
longer term, the increased demand in the real economy
due to advancement of more credit is likely to be followed
by a period of decreased demand when loans are being re-
paid; in other words, less spending in the real economy.
This simple model only allows for banks to expand their
asset base and debt in reaction to an increase in the market
values of all assets within the boundary of the original
capital ratio. The model will now be extended to include
the possibility of remuneration and dividend payments
from the increase in asset values.
13
The new capital ratio
that results after an increase in the market values of all
assets is:
C
3
¼
ð1 þXkÞA À L
ð1 þXlÞ
P
r
i
a
i
; ð6Þ
where X is the portion of the increase in asset values re-
tained by the bank (after appropriation of the increase by
management and shareholders via remuneration and/or
dividends) (0 < X6 1).
Again, if one assumes that management is planning to
neutralise the increase in the capital ratio by investing in
additional assets denoted by B and fully ?nanced by debt,
then the following equation is obtained:
C
3
¼
ð1 þXkÞA þ B À ðL þ BÞ
ð1 þXlÞ
P
r
i
a
i
þ r
j
B
¼
A À L
P
r
i
a
i
¼ C
1
: ð7Þ
Solving for B:
B
2
¼
Xðk ÀlÞA þXlL
r
j
C
1
: ð8Þ
If all of the increase is retained (X= 1) the result simpli?es
to the previous solution B
1
.
Comparing the solution of B
2
with that of B
1
it is appar-
ent that B
2
must always smaller than B
1
because the two
terms in the numerator are both a portion of the numera-
tor terms in B
1
and the denominator in both instances is
the same. The implication is that the need to increase as-
sets and debt to restore the previous capital ratio, after
an increase in the market values of all assets, is reduced
when a portion of the value increase is paid out.
A similar logic applies when calculating
@B
@X
:
@B
@X
¼
ðk ÀlÞA þlL
r
j
C
1
> 0: ð9Þ
As the retention ratio (X) increases, the need to expand the
asset base and debt also increases. The more of the increase
in asset values is paid out (smaller X) the less the need for
an expansion in assets and debt.
Fig. B in the Appendix provides a schematic representa-
tion followed by a narrative description of the model up to
this point. Thus, the inclusion in the model of the payment
of remuneration and dividends (1 À X) (for positive k) sig-
ni?cantly changes the effects previously noted. The need to
increase assets and debt to restore the previous capital ra-
tio is reduced when a portion of the value increase is paid
out. It is implied that the additional demand for ?nancial
assets will be less with a smaller price effect and the in-
creased spending power transferred to the real economy
will be less. The money paid to the real economy as remu-
neration and dividends, ceteris paribus, does increase de-
mand in the real economy. This increase in demand will
be less, however, than the increase in demand that would
have resulted if the remuneration and dividends were not
paid out, with the capital ratio being restored solely by the
debt-?nanced expansion of the bank balance sheet. This is
due to the leverage that is applied to the capital on the
bank’s balance sheet. Of course, not all of the remuneration
and dividends paid to the real economy will be consumed;
a portion of these will return to the ?nancial sector in the
form of additional demand for ?nancial assets, which will
12
It is often argued that the ?nancial sector can be ignored in models as
any loan is owned by somebody else and thus the debit and credit can be
collapsed without mishap. This argument ignores the fact that bank
deposits are money and bank loans are not.
13
Both management and shareholders have a strong incentive to reward
themselves during the business cycle upswing despite the possibility of
future losses during the downturn as elucidated by Taylor (2009) and
Haldane (2011).
P. de Jager / Accounting, Organizations and Society 39 (2014) 97–116 105
increase their market values and decrease their returns. If
time were already part of the model (note previous com-
ment that time will only be formally included in the model
at a later stage) then a new comment regarding feedback
processes would be possible at this stage.
The investment demand for ?nancial assets increases
their market values and contributes to an increase in the
average market value of all assets. This provides an im-
pulse for the assets expansion cycle to start again. Here,
one should also note the replacement effect that took place
in capital; liquid assets were paid out as remuneration and
dividends and were replaced by unrealised and riskier cap-
ital gains on ?nancial instruments. This is bad capital driv-
ing out good capital; this is explained in more detail in
Section ‘Further development of the basic model’.
The described model does not allow for a general reduc-
tion in asset values (0 < l 6 k was assumed). However,
such a general reduction in asset values would almost cer-
tainly be experienced during a recession. If the above
assumption is relaxed and negative growth is allowed with
the only proviso that l should always represent a portion
of k, then banks will reduce their asset base and pay off
debt in response to a general negative value adjustment.
The solution for B remains:
B ¼
ðk ÀlÞA þlL
r
j
C
1
: ð10Þ
Fig. C in the Appendix provides a schematic representa-
tion followed by a narrative description of the model up to
this point. In the case of a general reduction in asset values,
X is not added to the model as management and share-
holders will not pay in any money.
In summary, the solution for B:
B ¼
XðkÀlÞAþXlL
r
j
C
1
; if k > 0:
0; if k ¼ 0:
ðkÀlÞAþlL
r
j
C
1
; if k < 0:
8
>
>
>
>
<
>
>
>
>
:
Implications of the basic model
The basic model described above shows how the inter-
action of FVA gains and bank capital regulation motivates
banks to expand their asset base. In the context of that
basic model this expansion was ?nanced by debt. Such a
feedback process can help to explain why credit advances
in the United States grew at such a high rate (from $3 tril-
lion total debt to $36 trillion total debt in 2007 (NCCFEC,
2011:xvii)). The following Fig. 4 provides data from the
United Kingdom’s Independent Commission on Banking
(2011) and illustrates how credit extension in the years
preceding the global ?nancial crisis exceeded GDP and typ-
i?es a debt-?nanced boom.
The basic model also illustrates how this incentive to
expand the asset base is tempered by the payment of div-
idends and remuneration. These dividend and remunera-
tion payments would not have been necessary if FVA did
not increase accounting equity (the argument can be found
in Kay, 2009; Taylor, 2009; Turner, 2010; Wood, 2010).
The model shows that when allowance is made for a
negative valuation shock to asset values then the asset
base decrease required, to restore the capital ratio, must
be greater than the asset base increase required for a posi-
tive valuation shock of the same size. It is here that a sim-
ple model that assumes all price changes are passed into
bank equity by FVA is not descriptive of reality; at least,
not over the short-term. Banks have ‘‘shock absorbers’’
available to them in the FVA rules to avoid these write-
downs; they can argue that market prices are not correct
and then move to mark-to-model values and/or reclassify
items held at ‘‘fair value’’ to ’’held at cost’’: in October
2008, accounting standards were changed to allow for
the reclassi?cation of ?nancial instruments from ‘‘carried
at fair value’’ to ‘‘carried at cost’’. Bischof, Brüggemann,
and Daske (2010) as well as Fiechter and Meyer (2011)
found that banks made ample use of these opportunities.
In view of this, it is not surprising that empirical studies
by Shaffer (2010) and by Badertscher et al. (2012) found
Fig. 4. Total loans to different sectors of the economy as a percentage of GDP. Source: Vickers Report (2011:51).
106 P. de Jager / Accounting, Organizations and Society 39 (2014) 97–116
little evidence of bank capital being depleted by FVA losses.
The possibility of long-term FVA losses impacting bank
capital cannot be ascertained from these writers’ evidence
as they only considered data up to the year 2008. Over the
longer term, it is unlikely that banks will be able to avoid
recognising actual losses, permanently, because the instru-
ments will mature and a fundamental loss in value will
then realise. This postponement of FVA losses by banks
possibly explains why banks have been unwilling to lend
following the global ?nancial crisis even though they have
enough capital according to their of?cial regulatory ?g-
ures; bank management know that the capital on the
books is not high quality (risky) and they act accordingly;
thus, they act as if they have less capital than stated in
the of?cial numbers. Japanese banks in the 1990s had a
similar problem where the banks of?cially had enough
capital and despite the of?cial numbers, were not writing
down non-performing assets and neglected to advance
the necessary credit to the economy.
It can also be argued that the increase in bank capital
that was called for under Basel III (Bank for International
Settlements, 2011:2), in reaction to the global ?nancial cri-
sis, works to counter this bene?t of the ‘‘shock absorbers’’
and that the net effect would be a need to shrink bank bal-
ance sheets. Reducing the size of bank balance sheets is a
current global occurrence.
The main shortcomings of the basic model can be sum-
marised as follows: ?rstly, it depends on external valuation
shocks for regime change between bank balance sheet
expansion and shrinkage; secondly, feedback effects are
implied but not expressly modelled and thirdly, it does
not explain the faster growth in loans to ?nancial compa-
nies compared to loans to non-?nancial companies; the
latter is shown in Fig. 4 which indicates information pro-
vided by the Independent Commission on Banking.
Further development of the basic model
The increase in the general market value of bank assets
(a positive k) during the upswing that preceded the global
?nancial crisis was the product of complex interconnected
developments in the ?nancial system. These developments
included progressively lower interest rates globally that,
all else equal, increased the value of especially ?xed rate
?nancial instruments. A portfolio of ?xed rate loans desig-
nated at fair value through pro?t and loss would have
increased in value and the increase would have increased
regulatory capital. Financial instruments exposed to credit
risk would also have increased in value during the upswing
as default data improved and con?dence increased.
Another possible source of value increase was the increase
in ?nancial instrument trading within the ?nancial sector.
This increase in trading within the ?nancial sector resulted
from a combination of ?nancial innovation, deregulation
and a search for yield in a low interest rate environment.
Increased demand for ?nancial instruments led to price
rises.
Given an initial increase in the general market value of
bank assets, it is relatively simple to expand the basic mod-
el to incorporate feedback effects that will maintain bank
balance sheet expansion once started; this is done by
formally incorporating time into the model. The following
intellectual device is useful for the discussion that follows;
to see equations C as re?ecting the capital ratio of the
banking system as a whole in the one instance and as
re?ecting the capital ratio of an individual bank in the
other instance.
In the instance when equations C re?ect the capital
ratio of the banking system as a whole it is important to
realise that the expansion of the bank sector balance sheet
is only possible by advancing further credit to the real
economy (external to banking system). The expansion of
the balance sheet simultaneously increases the money
supply with the concurrent increase in bank deposits. Real
economic actors will spend a portion of this increased
money supply on the purchase of ?nancial assets, with
the additional demand causing the values of those assets
to increase. At the same time the money paid out by the
banking system as remuneration or dividends (1 À X) will
not all be consumed and will, at least to some extent, be
used to purchase ?nancial assets, increasing their market
values.
In basic model mathematical terms this can be repre-
sented as follows:
@kðtÞ
@t
¼ f ðBðkÞ; XðkÞ; UÞ; ð11Þ
where U represents exogenous factors that might im-
pact on k The implication is that an initial value shock
can initiate a feedback loop; this can arise through the de-
mand for ?nancial instruments from the ?nancial sector
(B) or from the demand for ?nancial instruments from
the real sector (portion of B + portion of X) driving contin-
uous bank balance sheet expansion or shrinkage. Fig. D in
the Appendix provides a schematic representation fol-
lowed by a narrative description of the model up to this
point.
The question that arises is why input to the model will
change from a positive k to a negativek. Events external to
the model are obviously one answer, but the model up to
this point holds three endogenous answers to this ques-
tion; wealth effects, bad capital driving out good capital
and reaching for yield. Note that the ?rst of these answers,
the ‘‘wealth effect’’ is related to the balance sheets of indi-
viduals becoming more fragile while the two latter an-
swers are related to increasingly fragile bank balance
sheets.
Individuals in the real economy have more credit avail-
able to them during the upswing. This credit availability is
driven by individuals becoming seemingly more credit
worthy as the upswing advances; the market value of their
security increases and default rates improve. In other
words, as the asset side of individuals’ balance sheets in-
creased (improved), the liability side of their balance
sheets also increased as they took on more debt. In this
manner the individuals’ leverage ratio might stay constant
but, relative to their income, their balance sheets are more
and more fragile as the upswing continues.
The payment of remuneration and or dividends (1 À X)
from FVA pro?ts leads to a replacement effect where the
most liquid asset (money) leaves the ?nancial sector as
payment and is replaced in bank capital by FVA-derived
P. de Jager / Accounting, Organizations and Society 39 (2014) 97–116 107
value increases in less liquid assets. This can be called a
special case of Gresham’s law
14
where ‘‘bad capital drives
out good capital’’. The result is a banking system much more
sensitive to external disturbances (or U).
The additional demand for risky assets during the bal-
ance sheet expansion process drives down the returns to
be earned on those risky assets and thus incentivises bank
managers to take more risk in the search for returns.
15
The
implication for bank balance sheets is, even just to show
constant returns, to become riskier. The result is a banking
system much more sensitive to external disturbances (or
U). This reach for yield can explain the decline in underwrit-
ing standards in the U.S. (NCCFEC, 2011:4) as well as the
observation that ?nancial ?rms changed from staid to risk
seeking during the boom (NCCFEC, 2011:xvii). Evidence of
predatory lending highlighted by the Commission (NCCFEC,
2011:10) calls into question the dissenting report’s conclu-
sion that credit demand emanating from the U.S. Govern-
ment’s housing policies was to blame for the
overexpansion in U.S. housing. Predatory lending points to
the role of credit supply in the process that ties back to
the reach for yield argument from the model.
A further argument for regime change is possible in the
instance when equations C re?ect the capital ratio of an
individual bank rather than that of the total banking sys-
tem. In addition, one of the shortcomings of the basic mod-
el highlighted earlier in this paper was that it did not
explain the growth in loans between ?nancial companies.
By adding more features to the basic model and by seeing
equations C as re?ecting the capital ratio of an individual
bank, endogenous regime change as well as the growth
in loans between ?nancial companies can be explained.
If we recall from previous discussion that:
C
2
¼
ð1 þ kÞA þ B À ðL þ BÞ
ð1 þlÞ
P
r
i
a
i
þ r
j
B
; ð12Þ
and that was used to determine the balance sheet expan-
sion B required for a given positive value shock, it is now
necessary to emphasise that a bank will only expand its
asset base, ?nanced by debt, if the following condition is
met:
R
asset
À R
liability
> 0; ð13Þ
where R represents the return on the ?nancial instrument.
Bank debtors want to spend their newly raised debt and
thus the asset side of the transaction will tend to be of
longer maturity than the liability side of the transaction
due to the required ability to spend. When the banking sys-
tem experiences a positive value shock and banks are
simultaneously expanding their asset bases, then two
effects can be noted. First, the supply of credit can over-
whelm the real economy’s demand for credit and lead to
increased inter-bank lending that will lead to a stronger
(compared to the additional demand emanating from the
real economy described previously) positive feedback loop.
This is because banks will invest all of the new bank depos-
its created in ?nancial instruments, increasing k Second,
arbitrage will drive down (R
asset
À R
liability
) until, at the limit
and ignoring other factors, no more margin is available and
the positive feedback loop will come to an end. This helps
to explain the ‘‘exponential growth’’ in trading activities
observed by the NCCFEC (2011:xvii)) as well as the growth
of the ?nancial sector from 5% of gross domestic product to
8% (NCCFEC, 2011:64). Finally, the feedback process de-
scribed in the model between the ?nancial system and
the real economy explains why, parallel to the rapid
expansion of credit in the U.S. economy, the Commission
observed overdone real economic activity (NCCFEC,
2011:5).
The margin effect noted above should be related to the
literature on the ability of an inverted yield curve to pre-
dict recessions (see for example Ang, Piazzesi, & Wei,
2006; Wright, 2006). Banks, generally speaking, invest in
?nancial instruments with longer maturity ?nanced by
?nancial instruments of shorter maturity and realise a
return spread as elucidated above when the yield curve
is normal and upward sloping. If this investment in longer
maturity instruments ?nanced by shorter maturity instru-
ments became systemic, it will drive down the returns on
longer maturity instruments: greater demand drives up
prices leading to lower returns. At the same time it will
drive up returns on shorter maturity instruments: greater
supply drives down prices leading to higher returns.
Table 1 summarises the feedback effects:
These feedback effects, without inhibitors (designated
‘‘regime change processes’’ below), will lead to continuous
expansion or continuous contraction. Table 2 summarises
the endogenous regime change processes identi?ed in
the model:
Empirical example
It is not possible to directly observe the macro-economy
and thus the economic data that we do have are selected
and made sense of by theory (Suzuki, 2003:484). Accord-
ingly, no attempt will be made to demonstrate the model
proposed in this paper in totality. Rather, data from two
international and systemic universal banks will be used
to demonstrate the following: the slow (controlled) recog-
nition of FVA losses; the materiality of fair value account-
ing entries that impact pro?t and loss and the crowding
out of risk-free capital with risky unrealised FVA gains.
The model developed in this paper can then be used to
make sense of the information.
The two banks that data will be gathered from are one
of the largest banks in the Netherlands, ING Bank N.V.,
and the largest bank in South Africa, the Standard Bank
of South Africa Limited. Apart from size and systemic
importance further similarities between the two banks
are that they both report under International Financial
Reporting Standards (IFRS) and are supervised from coun-
tries where Basel II was implemented in 2008. Both banks
14
Gresham’s law is commonly stated as: ‘‘Bad money drives out good’’,
but is more accurately stated: ‘‘Bad money drives out good if their exchange
rate is set by law’’ (Rolnick & Weber, 1986:185).
15
A more subtle possibility should also be noted; that this process of
riskier asset accumulation is possible even without an actual decision to
invest in higher risk assets. During the upswing default data as well as loan
to value ratios (the market value of security tends to increase) improves
and this provides the rationale for further credit extension. In such a
process, risk increases as more credit is now backed by the same cash ?ows.
108 P. de Jager / Accounting, Organizations and Society 39 (2014) 97–116
are large (globally important) according to Shaffer’s
(2010:9) de?nition (>$100 billion of assets) with ING Bank.
N.V. about ten times larger than the Standard Bank of
South Africa Limited in 2012. Very important for this paper
is that both banks operate in an income tax regime where
?nancial instruments are taxed on realisation and not on a
mark-to-market basis; this makes it possible to identify
the unrealised portion of fair value accounting gains and
losses through pro?t and loss and will be explained further
below.
IFRS 7 (IASB, 2005) does not require disclosure of the
realised and unrealised portions of items of income, ex-
pense, gains, and losses related to ?nancial instruments;
only the combined total is required. The deferred tax notes
to the ?nancial statements of both ING Bank N.V and the
Standard Bank of South Africa Limited contain information
that makes possible the derivation of the portion of net
pro?t that is due to unrealised fair value adjustments on
?nancial instruments. Deferred tax arises when a differ-
ence exists between accounting net pro?t and taxable in-
come. Accounting net pro?t contains realised and
unrealised items related to ?nancial instruments whereas
current Dutch and South African taxable income contains
only realised items. An increase in the deferred tax liability
due to ?nancial instruments thus re?ects the tax portion of
an unrealised gain on ?nancial instruments. Fig. 5 and
Fig. 6 are excerpts from the banks’ 2012 annual ?nancial
statements where the red ovals indicate the relevant
movements in the deferred tax liability due to ?nancial
instruments for that year:
Table three that follows summarises the available infor-
mation on deferred tax movements. To gain an under-
standing of the signi?cance of these movements it is
important to note that what is observed on the deferred
tax line is only the income tax effect of the transactions.
For example, the current corporate income tax rate for
the Netherlands is 25% and thus 25% is the movement on
deferred tax and 75% is the impact on pro?t. It is the 75%
that we are interested in. In table three a few simplifying
assumptions have been made: that the current corporate
income tax rates in the Netherlands of 25% and in South
Africa of 28% was constant over the whole period; that
the debit entries against deferred tax for ING from 2008
onwards reversed previous credit entries and that retained
income is a suitable proxy for bank capital.
When looking at the movement in the deferred tax line
in table one it is important to note that for ING Bank N.V.
the reversal from credit entries to debit entries (debit en-
tries imply unrealised losses on ?nancial instruments)
against deferred tax happened in 2008 as would be
Table 1
Summary of feedback effects in the model.
Feedback process Description Relative strength
Bank balance sheet expansion Additional demand for ?nancial instruments impacting k Medium (leverage ampli?es effect
but leakage to real economy)
Real economy investment in ?nancial instruments Additional spending power not all consumed Weak (no leverage effect)
Intra bank lending and investment Balance sheet expansion creates additional demand for
?nancial instruments impacting k. Spending power
created is within the banking system and not in the real
economy
Strong (additional spending
power created within ?nancial
sector and subject to leverage)
Fig. 5. Changes in deferred tax for ING Bank N.V. due to FVA on ?nancial instruments through pro?t and loss. Source: ING Bank N.V Annual Report
(2012:64).
P. de Jager / Accounting, Organizations and Society 39 (2014) 97–116 109
expected with a global ?nancial crisis that reached its peak
in September 2008 with the failure of Lehman Brothers.
Strangely, the reversal for the Standard Bank of South Afri-
ca Limited only takes place in 2009. For ING Bank N.V. the
reversal that started in 2008 continued until 2012,
whereas for the Standard Bank of South Africa Limited
the reversal only lasted two years and then turned again.
What the information shows is possibly the use of account-
ing discretion to avoid/manage FVA losses: The Standard
Bank of South Africa Limited only shows losses for the ?rst
time in 2009 whilst the crisis started in 2008 and for both
banks the FVA losses did not come through as one ‘‘hit’’ but
took time.
The rows that relate the movements on deferred tax to
pro?t show just how material these FVA entries (that im-
pact pro?t and loss and thus bank capital) are. For ING
Bank N.V very few entries are available when FVA gener-
ated unrealised gains, but unrealised losses against pro?t
between À180% and À34% show how large these entries
can be. The entries for the Standard Bank of South Africa
Limited peak at an unrealised FVA gain portion of net in-
come of 33% in 2008 with the largest reversal of À58%
the next year.
Arguably the most important part of Tier 1 bank capital
is retained income. The idea is that this capital should
always be available to absorb unexpected bank losses.
Table 3 shows that the unrealised FVA gain portions of
retained income are not as stable as would be expected.
For both banks the peak is during the ?nancial crisis with
a sharp reduction thereafter. This is indicative of unrea-
lised FVA gains in bank capital being risky and during the
upturn displacing less risky bank capital as predicted by
the model. Back in South Africa, the CEO of one of the four
large banks, Nedbank Limited, admitted during an inter-
view on his bank’s ?nancial performance in 2013 that
FVA gains through pro?t and loss are risky: ‘‘We did bene-
?t during the ?rst six months from some fair value gains,
which are lower quality income and harder to forecast. . .’’
(Tarrant, 2013).
Implications, shortcomings and opportunities for
further research
The model demonstrates the expansionary impulse
generated when FVA increases bank regulatory capital. It
can be argued that the same expansionary impulse will re-
sult from other accounting entries that also increase regu-
latory capital. These entries could include realised pro?ts
fromthe selective selling-off of investments (gains trading)
or incurred loan loss provisioning that ‘‘enables banks to
present higher earnings and (regulatory) equity capital,
which allow the bank to extend more credit’’ (Gebhardt
& Novotny-Farkas, 2011:302). FVA’s impact differs in two
important aspects. First, FVA impacts regulatory capital
much faster and more materially than other accounting
entries that need time to impact retained earnings. It is
the speed of the feedback process that matters. The second
way in which the capital increase from FVA differs from in-
creases caused by other accounting entries is that these
alternatives will not result in the replacement of liquid
assets in bank capital with riskier FVA gains (bad capital
driving out good capital); FVA pro?ts do not provide
Table 2
Summary of endogenous regime change processes in the model.
Regime change
process
Description
Balance sheet fragility
– real economy
Real economy has more assets with more
debt but the same income
Balance sheet fragility
– banks
Bad capital driving out good capital:
Liquid assets in capital replaced by illiquid
capital gains
Balance sheet fragility
– banks
Reach for yield: Riskier assets originated
to keep income constant; the other side of
the fragile real economy balance sheet
Yield curve inversion Arbitrage drives down yield on longer
term ?nancial instruments and drives up
yield on shorter term instruments
Fig. 6. Changes in deferred tax for the Standard Bank of South Africa Limited due to FVA on ?nancial instruments (including derivatives). Source: Standard
Bank of South Africa (2012:153)
110 P. de Jager / Accounting, Organizations and Society 39 (2014) 97–116
liquidity that can fund dividends and remuneration
payments.
Another characteristic of the model is that it is incom-
plete. Banks do not just lend money due to the interaction
between FVA and their capital ratios. The real economy has
a need for credit before the supply of credit is considered.
In the same way it is certainly not argued here that FVA is
the primary or sole reason behind the cyclicality of ?nan-
cial capitalism. The business and ?nancial cycles existed
before the adoption of FVA and booms and busts were even
evident during the heyday of historical cost accounting in
the period after the Great Depression. The caveat is that
those booms and busts were less damaging in their impact;
they were less extreme than either the Great Depression or
the global ?nancial crisis.
Toporowski (2010:224) argues that those who view the
capitalist system as being prone to crisis have dif?culty in
explaining the relative stability of ?nancial capitalism in
the decades before the global ?nancial crisis. This paper ar-
gues that the absence of FVA from the banking system dur-
ing that time can explain the relative calm. A system needs
control mechanisms in the presence of feedback loops to
keep the process from running away. One such ?nancial
feedback loop was highlighted by Minsky and describes
how stable economic periods improve the balance sheets
of individuals by increasing the values of their assets and
enabling them to borrow more because they now have
more security; this in turn increases systemic fragility as
more debt is now supported by the same cash ?ows. This
situation can be summarised by stating that the individ-
ual’s (borrower’s) balance sheet was ‘‘fair valued’’ in the
credit evaluation process. A control mechanism that will
inhibit this expansionary impulse is not to fair value the
balance sheets of banks as well; in other words do not
meet rising credit demand with increasing credit supply.
These checks and balances prevailed in the relatively stable
economic period that preceded the global ?nancial crisis;
as a result, the feedback loop was not allowed to run away
due to the presence of a control mechanism. However,
with the introduction of formal FVA into accounting stan-
dards this control mechanism was removed, leading to
the over-indebtedness of individuals and ?nancial busi-
nesses. FVA did not cause the crisis but ampli?ed it.
The model explains the conclusion of the NCCFEC
(2011:xix) that ‘‘a combination of excessive borrowing,
risky investments, and lack of transparency put the ?nan-
cial system on a collision course with crisis’’. Banks and
individuals borrowed excessively as explained in the
paragraph above and risky investments were made in the
process as there was a search for yield. The bene?t of the
perspective offered by the model will brie?y be illustrated
by using it to understand the results of the study by Cabral
(2012). Cabral introduced a novel model of individual bank
pro?tability and used the model to explain paradoxes such
as very high bank pro?tability immediately prior to the cri-
sis even though intermediation margins were very low
(the return to banking activities was low). It is argued that
balance sheet growth and the taking on of additional risk
prior to the crisis explains this paradox. In addition it is
argued that low levels of liquidity prior to the crisis played
an important role in the crisis. First, bank pro?ts immedi-
ately prior to the crisis were very high, possibly due to
the effect of FVA. Second, the returns to banking activities
were low prior to the crisis because of the search for yield
necessitated by the bank balance sheet expansion that
simultaneously increased the money supply. Cabral’s
(2012:114) explanation for the low returns is to argue that
it is the result of increased competition. The model in this
paper explains why there was an increase in competition.
Cabral’s model does not explain where the additional cred-
it demand comes from when banks, according to his mod-
el, expand their balance sheets to protect pro?tability; the
model in this chapter does explain. Finally, Cabral uses the
exogenous impact of bank regulations to explain the low
levels of liquidity prior to the crisis. In contrast the model
in this chapter explains the low levels of liquidity
endogenously.
Controls unrelated to FVA can be imagined, to counter-
act this ampli?cation process. One such alternative control
is the introduction of a countercyclical capital buffer as
part of the changes made to the global bank regulatory
infrastructure following the global ?nancial crisis; thus,
during the upswing, banks might be required to increase
their capital ratios as described in the following quote from
Basel III: ‘‘It will be deployed by national jurisdictions
when excess aggregate credit growth is judged to be
associated with a build-up of system-wide risk to ensure
the banking system has a buffer of capital to protect it
against future potential losses’’ (Bank for International
Settlements, 2011:57). In terms of the model this increase
in the minimum capital required, if of the exact size
necessary, will eliminate the impulse for balance sheet
expansion resulting from k, halting the feedback process.
However, there are reasons why this might not be an ideal
hedge for the expansionary impulse created by FVA. First,
it will require central bank decision makers to get the tim-
Table 3
Summary of deferred tax movements on ?nancial instruments (credit entries are positive and debit entries negative).
2004 2005 2006 2007 2008 2009 2010 2011 2012
Movement in deferred tax due to ?nancial instruments – ING
(Euro million)
NA NA NA 72 À301 À325 À192 À162 À366
Movement in deferred tax due to ?nancial instruments – Standard
(rand million)
83 486 903 700 1132 À1837 À1747 514 9
Unrealised portion of net pro?t after tax – ING NA NA NA 6% À128% À180% À13% À12% À34%
Unrealised portion of net pro?t after tax – Standard 4% 19% 29% 19% 33% À58% À57% 14% 0%
Unrealised portion of retained income – ING NA NA NA 30% 32% 21% 11% 7% 0%
Unrealised portion of retained income – Standard 43% 54% 81% 88% 83% 41% 14% 17% 19%
P. de Jager / Accounting, Organizations and Society 39 (2014) 97–116 111
ing and extent of any increase exactly right. Second, banks
as a body have become very effective lobbyists in some
countries and can be expected to try and in?uence the set-
ting of an increased capital ratio. Lastly and perhaps most
important is the argument that a control mechanism
implemented on a regulatory level cannot take into ac-
count the unique circumstances of each individual bank.
Numerous opportunities for future research can be
identi?ed:
More and better descriptive evidence of the co-
movement of FVA in banks, with the business cycle.
Do bank managers pay dividends out of unrealised
FVA pro?ts, rewarding shareholders as well, for
managers’ excessive risk-taking and by doing so,
weakening the ?nancial system?
Augmentation of standard stock-?ow consistent
accounting models of the economy with this credit
supply model.
Do bank management pay out remuneration from
FVA gains?
Was the avoidance of FVA write-downs during the
global ?nancial crisis temporary or permanent?
Causality studies on the relationship between credit
demand and credit supply from a real economic per-
spective and from a ?nancial perspective; and
Studies of the class struggle that determines the
split of the economic pie between wages, pro?ts
and, in addition to the standard production view
of capitalism, interest.
Conclusion
The focus by accounting researchers on the role of
accounting and FVA during the global ?nancial crisis and
not on a role before the crisis, as well as the claim that
accounting is only a messenger are two factors that moti-
vate the need for the development of the model described
in this paper, that links FVA, bank regulatory capital,
money supply, remuneration, dividends and economic
activity. The requirement to incorporate global ?nancial
crisis characteristics (such as feedback effects, systemic
risk and the primacy of banks in the crisis) into that model
rationalised the use of a stock-?ow consistent accounting
type model of the economy rather than standard economic
models. In contrast to the usual stock-?ow consistent
accounting model of the economy, the model developed
in this paper reverses the causality between credit demand
and credit supply; in other words, the starting point is
credit supply, not credit demand, which gives banks the
centre stage in the account.
The model developed in this paper describes and then
links together many of the characteristics of the global
?nancial crisis and previous crises. The procyclical expan-
sion of bank balance sheets during the upswing followed
by a contraction (over the long-term because during the
short-term, banks can avoid some FVA write-downs)
during the downturn is explained by the need to maintain
regulatory capital ratios and/or feedback effects both with-
in the ?nancial sector and between the ?nancial and real
sectors. In accordance with the doctrine of Minsky, a state
of crisis is shown to be endogenous to the model’s opera-
tion, because of systemic fragility. The focus is on bank bal-
ance sheet fragility that is caused by bad capital driving out
good capital, banks reaching for yield and the inversion of
the yield curve. It is argued that not having FVA pro?ts in
bank capital during booms is an essential control mecha-
nism that inhibits the feedback effects emanating from
the real sector.
An empirical example is given where the evolution of
the unrealised FVA gains through pro?t and loss of two
universal and systemic banks (ING Bank N.V & the Stan-
dard Bank of South Africa) can be better understood using
the model in this paper.
In conclusion, the model offers a simple explanation of
why the world, for so long after the height of the global
?nancial crisis, is still mired in slow growth. During a pro-
longed and excessive boom bank pro?ts and capital were
materially increased by unrealised FVA pro?ts. These prof-
its justi?ed the pay-out of liquid assets weakening the
?nancial system. The additional capital further justi?ed
more debt ?nanced asset expansion. With the crisis bank
management realised that these unrealised capital items
were not permanent. Management was able to postpone
the recognition of FVA losses by using the ?exibility inher-
ent in FVA regulations, but lending was slowed to a point
re?ective of ‘‘safe’’ (capital excluding unrealised items)
capital levels. Lending activity will stay subdued until all
of these marked-up items have been worked off banks’ bal-
ance sheets.
Appendix A
A.1. Derivation of mathematical results
If C
0
¼
AÀL
A
then:
Let u = A–L and v = A; then
du
dA
¼ 1 and
dm
dA
¼ 1
@C
0
@A
¼
v
du
dA
À u
dv
dA
v
2
¼
A:1 À ðA À LÞ:1
A
2
¼
L
A
2
If C
2
¼
ð1þkÞAþBÀðLþBÞ
ð1þlÞ
P
r
i
a
i
þr
j
B
¼
AÀL
P
r
i
a
i
¼ C
1
then:
A
X
r
i
a
i
þ kA
X
r
i
a
i
þ B
X
r
i
a
i
À L
X
r
i
a
i
À B
X
r
i
a
i
¼ A
X
r
i
a
i
þlA
X
r
i
a
i
þ ABr
j
À L
X
r
i
a
i
ÀlL
X
r
i
a
i
À BLr
j
BðÀAr
j
þ Lr
j
Þ ¼ Àðk ÀlÞA
X
r
i
a
i
ÀlL
X
r
i
a
i
B ¼
ðk ÀlÞA
P
r
i
a
i
þlL
P
r
i
a
i
ðA À LÞr
j
¼
ðk ÀlÞA þlL
r
j
C
1
If C
3
¼
ð1þXkÞAþBÀðLþBÞ
ð1þXlÞ
P
r
i
a
i
þr
j
B
¼
AÀL
P
r
i
a
i
¼ C
1
then:
A
X
r
i
a
i
þXkA
X
r
i
a
i
þ B
X
r
i
a
i
À L
X
r
i
a
i
À B
X
r
i
a
i
¼ A
X
r
i
a
i
þXlA
X
r
i
a
i
ABr
j
À L
X
r
i
a
i
ÀXlL
X
r
i
a
i
À BLr
j
BðÀAr
j
þ Lr
j
Þ ¼ Àðk ÀlÞXA
X
r
i
a
i
ÀXlL
X
r
i
a
i
112 P. de Jager / Accounting, Organizations and Society 39 (2014) 97–116
B ¼
Xðk ÀlÞA þXlL
r
j
C
1
If B ¼
XðkÀlÞAþXlL
r
j
C
1
then:
Let u ¼ Xðk ÀlÞA þXlL and m = r
j
C
1
then
du
dX
¼ ðk ÀlÞA þlL and
dv
dX
¼ 0
@B
@X
¼
v
du
dX
À u
dv
dX
v
2
¼
r
j
C
1
Á ½ðk ÀlÞA þlL?
ðr
j
C
1
Þ
2
¼
ðk ÀlÞA þlL
r
j
C
1
A.2. Graphical representation of the development of Fig. 2
The following example will illustrate the main points. A
bank (or the total banking system) has $100 of total assets
(A) with 30% of those assets subject to FVA and 70% of those
assets subject to historical cost accounting. The bank has
equity capital of $6 (A–L) and deposit liabilities of $94 (L).
The capital ratio is thus 6/100 = 6% and the effective tax rate
is 30%. If the market value of the fair valued assets increase
by 10% then the value of k is (30 * (10% * (1–30%)))/
(30 + 70) = 2.1%. An increase in debt-?nanced assets of 2.1/
6% = $35 is required. After this increase the original capital
ratio is restored
AÀL
A
¼ ðð100 þ2:1 þ35Þ À ð94 þ35ÞÞ=
À
ð100 þ2:1 þ35Þ ¼ 6%Þ. Additional demand for ?nancial
assets is $35 and $35 of additional spending power is made
available to the real economy (see Fig. A).
The previous example of the bank above is developed
further in this paragraph. Suppose that the remuneration
and dividend policy of the bank is that 50% of after tax
pro?ts is paid as remuneration and that 25% of after tax
pro?ts is paid as dividends; in this case, the retention rate
of the bank is 25% (X). If only 25% of the 2.1% increase in
average assets is retained (100 * 2.1% * 25% = $0.525 is left
as additional capital and 100 * 2.1% * 75% = $1.575 is paid
out), then the debt-?nanced increase in ?nancial assets re-
quired to restore the 6% capital ratio is (2.1 * 25%)/
6% = $8.75. The asset expansion (B) and the remuneration
and dividend payments to the real economy (1 À X) in-
crease the spending power available to the real economy
by 1.575 + 8.75 = $10.325; substantially less than the $35
previously when X was equal to one. The new demand
for ?nancial assets is only $8.75 versus the $35 previously.
The reason for the differences is the gearing that takes
place inside this banking system (see Fig. B).
The previous example of the bank now continues. Given
k = À2.1% the reduction in ?nancial assets and deposits
required to restore the 6% original capital ratio is equal
to $-35. This reduces the demand for ?nancial assets as
well as the spending power available to the real economy.
In our example a 2.1% increase led to a $10.325 increase in
spending power for the real economy and increased de-
mand for ?nancial assets of $8.75. A decrease of 2.1% leads
to a reduction in spending power available to the real
economy of $35 and a reduction in the demand for
?nancial instruments of $35. The reason for this difference
between the increase in balance sheet required and the
Financial sector
Real economy
Credit
Fig. A. C
2
¼
ð1þkÞAþBÀðLþBÞ
ð1þlÞ
P
r
i
a
i
þr
j
B
.
Investment
Remuneration &
dividends
Financial sector
Real economy
Credit
Fig. B. C
3
¼
ð1þXkÞAþBÀðLþBÞ
ð1þXlÞ
P
r
i
a
i
þr
j
B
.
P. de Jager / Accounting, Organizations and Society 39 (2014) 97–116 113
decrease in balance sheet required to restore the original
capital ratio is because no contribution towards alleviating
the capital shortage can be expected from employees
and shareholders (X= 1) with the 2.1% decrease (see
Fig. C).
The previous example is now continued: thus, it has
already been shown that with k = 2.1% and X= 25%,
the increase in ?nancial assets and deposits required to
restore the 6% original capital ratio is equal to $8.75
and the increase in spending power available to the real
economy is $10.325. Let us now suppose that that the real
economy spends 50% of any additional spending power
on the purchase of ?nancial assets and that the demand
function for ?nancial assets is a simple linear function
of the form Q = 10 + 90P. The additional demand for ?nan-
cial assets is 10.325/2 + 8.75 = $13.9125. Using the
demand equation the initial price was 1 as the total of
?nancial assets available then equated $100. The total le-
vel of demand is now 100 + 13.9125 = $113.9125 and
solving the demand equation the new price equals 1.16.
As 30% of the assets are subject to fair value accounting
then 30% of that 16% increase will enter capital via FVA
and set the whole process in motion again (see Figs. D
and E).
Financial sector
Real economy
Credit
Investment
Remuneration &
dividends
Debt repayment
Fig. D.
@kðtÞ
@t
¼ f ðBðkÞ; XðkÞ; UÞ.
Financial sector
Real economy
Bank 2
Bank 1
Credit
Investment
Investment
Remuneration &
dividends
Debt repayment
Credit
Fig. E. (R
asset
À R
liability
) > 0.
Debt repayment
Financial sector
Real economy
Fig. C. k < 0.
114 P. de Jager / Accounting, Organizations and Society 39 (2014) 97–116
References
Akerlof, G. A., & Romer, P. M. (1993). Looting: The economic underworld
of bankruptcy for pro?t. Brookings Papers on Economic Activity,
1993(2), 1–73.
Allen, F., & Carletti, E. (2008). Mark-to-market accounting and liquidity
pricing. Journal of Accounting and Economics, 45(2), 358–378.
Andersson, T., Lee, E., Theodosopoulos, G., Yin, Y. A., & Haslam, C. (2014).
Accounting for the ?nancialized UK and US national business model.
Critical Perspectives on Accounting, 25(1), 78–91.
Ang, A., Piazzesi, M., & Wei, M. (2006). What does the yield curve tell us
about GDP growth? Journal of Econometrics, 131(1), 359–403.
Arnold, P. J. (2009). Global ?nancial crisis: The challenge to accounting
research. Accounting, Organizations and Society, 34(6–7), 803–809.
Badertscher, B. A., Burks, J. J., & Easton, P. D. (2012). A convenient
scapegoat: Fair value accounting by commercial banks during the
?nancial crisis. The Accounting Review, 87(1), 59–90.
Ball, R. (2008). Don’t blame the messenger ... or ignore the message. (22.06.09).
Bank for International Settlements (2006). International convergence of
capital measurement and capital standards: A Revised framework. Basel:
BIS. (10.02.13).
Bank for International Settlements (2011). Basel III: A global regulatory
framework for more resilient banks and banking systems – Revised
version, June 2011. Basel: BIS.
Bank for International Settlements (2012). 82nd Annual report. Basel: BIS.
Bank of England (2009). Financial stability report. Section 3. United
Kingdom: Bank of England.
Barth, M., & Landsman, W. (2010). How did ?nancial reporting contribute
to the ?nancial crisis? European Accounting Review, 19(3), 399–423.
Bello?ore, R., & Passarella, M. (2010). Minsky, the monetary circuit and
the current crisis: A sfc monetary accounting framework. Paper
presented at the conference ‘‘Can it happen again?’’, 1–2 October 2010,
University of Macerata, Italy.
Benston, G. J. (2006). Fair-value accounting: A cautionary tale from Enron.
The Journal of Accounting and Public Policy, 25(4), 465–484.
Bezemer, D. J. (2010). Understanding ?nancial crisis through accounting
models. Accounting, Organizations and Society, 35(7), 676–688.
Bischof, J., Brüggemann, U., & Daske, H. (2010). Relaxation of fair value
rules in times of crisis: An analysis of economic bene?ts and costs of the
amendment to IAS 39. or (06.02.13).
Black, W. K. (2005). The best way to rob a bank is to own one: How corporate
executives and politicians looted the S&L industry. Austin: University of
Texas Press.
Black, W. K. (2011). Still banking on fraud. Dollars & Sense, 1 June 2011.
(06.02.13).
Borio, C. (2012). The ?nancial cycle and macroeconomics: What have we
learnt?. BIS working papers no. 395. Basel: BIS.
Cabral, R. (2012). A perspective on the symptoms and causes of the
?nancial crisis. Journal of Banking & Finance, 37(2013), 103–117.
Einhorn, D. (2010). Fooling some of the people all of the time, a long short
(and now complete) story, updated with new epilogue. Hoboken, New
Jersey: Wiley.
Enria, A., Cappiello, L., Dierick, F., Grittini, S., Haralambous, A., Maddaloni,
A., et al. (April 2004). Fair value accounting and ?nancial stability.
Occasional paper series 13. European Central Bank.
Fiechter, P., & Meyer, C. (2011). Discretion in fair value measurement of
banks during the 2008 ?nancial crisis. or
(06.02.13).
Fisher, I. (1933). The debt-de?ation theory of great depressions.
Econometrica, 1(4), 337–357.
Friedman, M. (1953). The methodology of positive economics. In Essays in
positive economics. Chicago: University of Chicago Press.
Friedman, M., & Schwartz, A. J. (1971). A monetary history of the United
States, 1867–1960. New Jersey: Princeton University Press.
Gebhardt, G. (2012). Financial instruments in non-?nancial ?rm: What do
we know? Accounting and Business Research, 42(3), 267–289.
Gebhardt, G., & Novotny-Farkas, Z. (2011). Mandatory IFRS adoption and
accounting quality of European Banks. Journal of Business Finance &
Accounting, 38(3&4), 289–333.
Godley, W., & Lavoie, M. (2012). Monetary economics (2nd ed.).
Basingstoke: Palgrave Macmillan.
Haldane, A. G. (2011). Control rights (and wrongs). 24 October 2011.
Wincott Annual Memorial Lecture, Westminster, London. (12.04.13).
Howelss, P., & Bain, K. (2008). The economics of money, banking and
?nance: A European text (4th ed.). Gosport, Hampshire: Pearson
Education Limited.
ING Bank N.V. (2012). Annual ?nancial statements.
International Accounting Standards Board (2005). International ?nancial
reporting standard (IFRS) 7 ?nancial instruments: Disclosure. London:
IASB.
International Monetary Fund (2012). Global ?nancial stability report.
Washington: International Monetary Fund.
Investec Bank Limited (2012). Investec Bank Limited Group and company
annual ?nancial statements.
Kay, J. (2009). A boom based on little more than a bezzle. Financial Times.
(16.05.09).
Kay, J. (2012). Investors should resist the tyranny of quarterly earnings.
Financial Times (London) (29.02.12).
Keen, S. (2011). Debunking economics: Revised and expanded edition: The
naked emperor dethroned? London: Zed Books.
Khan, U. (2010). Does fair value accounting contribute to systemic risk in the
banking industry? Working paper. (22.07.13).
Krugman, P. (2009). The night they reread Minsky [Blog, 17 May]. (17.05.09).
Kusano, M. (2011). Fair value accounting and procyclicality: Accounting for
securitization.
(22.02.12).
Lahart, J. (2007). In time of tumult, obscure economist gains currency.
Wall Street Journal, 18(August), 1.
Laux, C., & Leuz, C. (2010). Did fair-value accounting contribute to the
?nancial crisis? Journal of Economic Perspectives, 24(1), 93–118.
Laux, C., & Leuz, C. (2009). The crisis of fair-value accounting: Making
sense of the recent debate. Accounting, Organizations and Society,
34(6–7), 826–834.
Livne, G., Markarian, G., & Milne, A. (2011). Bankers’ compensation and
fair value accounting. Journal of Corporate Finance, 17(4), 1096–1115.
Lo, A. W. (2012). Reading about the ?nancial crisis: A twenty-one-book
review. Journal of Economic Literature, 50(1), 151–178.
Magnan, M. L. (2009). Fair value accounting and the ?nancial crisis:
Messenger or contributor? Accounting Perspectives, 8(3), 189–213.
Mankiw, N. G. (2007). Macroeconomics (6th ed.). New York: Worth
Publishers.
McMahon, G. (2011). Mark-to-market’s real role in the crisis. The CPA
Journal (February 2011).
Mundy, S. (2012). Albuthnot chief blames accounting rules for loss.
Financial Times (London). 18 March.
Palley, T. I. (2010). The limits of Minsky’s ?nancial instability
hypothesis as an explanation of the crisis. Monthly Review, 61(11),
28–43.
Pinnuck, M. (2012). A review of the role of ?nancial reporting in the global
?nancial crisis. Australian Accounting Review, 60(2), 1–14.
Plantin, G., Sapra, H., & Shin, H. S. (2008). Marking to market: Panacea or
Pandora’s box? Journal of Accounting Research, 46(2), 435–460.
Posner, R. A. (2009). A failure of capitalism: The crisis of ‘08 and the descent
into depression. United States of America: Harvard University Press.
Pozen, R. C. (2009). Is it fair to blame fair value accounting for the
?nancial crisis? Harvard Business Review, 87(11), 84–92.
Rees, M. (2013). The unsecured boom may be over. (06.02.13).
Rolnick, A. J., & Weber, W. E. (1986). Gresham’s law or Gresham’s fallacy.
Journal of Political Economy, 94(1), 185–199.
Rose, P. S., & Hudgins, S. C. (2008). Bank management & ?nancial services
(7th ed.). New York: McGraw-Hill Irwin.
Ryan, S. G. (2008). Accounting in and for the subprime crisis. The
Accounting Review, 83(6), 1605–1638.
Scott, W. R. (2009). Financial accounting theory (5th ed.). United States of
America: Pearson Prentice Hall.
Shaffer, S. (2010). Fair value accounting: Villain or innocent victim: Exploring
the links between fair value accounting, bank regulatory capital and the
recent ?nancial crisis. Working paper, Federal Reserve Bank of Boston
Quantitative Analysis Unit. (12.04.13).
P. de Jager / Accounting, Organizations and Society 39 (2014) 97–116 115
Soros, G. (2012). George soros on the Euro crisis. Moneyweb (6 June).
Standard Bank of South Africa (2012). Annual ?nancial statements.
Suzuki, T. (2003). The accounting ?guration of business statistics as a
foundation for the spread of economic ideas. Accounting, Organizations
and Society, 28(2003), 65–95.
Tarrant, H. (2013). Nedbank’s Mike Brown on First Strut, personal loans and
Africa. (20.01.14).
Taylor, M. (2009). Innumerate bankers were ripe for a reckoning. Financial
Times (London). 16 December.
The Independent Commission on Banking (2011). Final report of the
independent commission on banking. London: Stationery Of?ce.
The National Commission on the Causes of the Financial and Economic
Crisis in the United States (2011). The ?nancial crisis inquiry report.
U.S. Government Printing Of?ce.
Toporowski, J. (2008). Minsky’s induced investment and business cycles.
Cambridge Journal of Economics, 32(5), 725–737.
Toporowski, J. (2010). Excess debt and asset de?ation. In S. Kates (Ed.),
Macroeconomic theory and its failings (pp. 221–234). Cheltenham, UK:
Edward Elgar Publishing Limited.
Turner, A. (2010). Banks are different: should accounting re?ect that fact?
Speech by Adair Turner, Chairman, FSA to the Institute of Chartered
Accountants in England and Wales (ICAEW), London. 21 January
2010.
(12.04.13).
United States Securities and Exchange Commission (2008). Report and
recommendations pursuant to section 133 of the emergency economic
stabilisation act of 2008: Study on mark-to-market accounting.
(05.08.10).
Veron, N. (2008). Fair value accounting is the wrong scapegoat for this crisis.
Bruegel Policy Contribution/May 2008.
(12.04.13).
Wallace, M. (2009). Is fair value accounting responsible for the ?nancial
crisis? Bank Accounting and Finance, 22(1), 9–18.
Whiteman, J. (2012). No assumptions please – We Brits like to consult the
facts. Financial Times, letter published on 22 December 2012.
Wolf, M. (2008). An inevitable earthquake. Business spectator.
(18.09.08).
Wood, D. (2010). Unrealised gains out of their hands. Risk.net. 4 February.
Wright, J. (2006). The yield curve and predicting recessions. FEDs Working
Paper No. 2006-7. Board of Governors of the Federal Reserve System –
Trade and Financial Studies Section.
Young, J. J. (1995). Getting the accounting ‘‘right’’: Accounting and the
savings and loan crisis. Accounting, Organizations and Society, 20(1),
55–80.
Zeff, S. A. (2007). The SEC rules historical cost accounting: 1934 to the
1970s. Accounting and Business Research, 37(3), 49–62.
116 P. de Jager / Accounting, Organizations and Society 39 (2014) 97–116
doc_676917743.pdf