Journal of Banking & Finance 58 (2015) 327342

Contents lists available at ScienceDirect

Journal of Banking & Finance

journal homepage: www.elsevier.com/locate/jbf

Fair value disclosure, liquidity risk and stock returns

Oliviero Roggi a, Alessandro Giannozzi b,
a
b

University of Florence and NYU Stern School of Business, via delle Pandette, 9, 50127 Florence, Italy
Link Campus University Rome, via Nomentana, 335, 00162 Rome, Italy

a r t i c l e

i n f o

Article history:
Received 31 October 2013
Accepted 10 April 2015
Available online 5 May 2015
JEL classication:
G18
M40
G30
Keywords:
Corporate liquidity risk
Fair value disclosure
Stock returns
Financial crisis

a b s t r a c t
This paper aims to investigate the impact of company liquidity risk on the stock prices of nancial and
non-nancial companies by analyzing investors reactions to 106 crisis events over the period from
2008 to 2010. Companies liquidity risk shows up in the three levels of fair value information (level
1-mark to market, level 2-market observable input and level 3-mark to model) disclosed in their balance
sheets, with level 3 illiquid assets representing a greater liquidity risk and resulting in a greater company
liquidity risk. The role played by liquidity risk information in investors decision-making is explored by
analyzing their reactions to liquidity-constraining events, capital injections and bank bailouts for 313
European companies. The empirical evidence is based on the xed effects model and Partial Least
Squares regressions. These ndings demonstrate that investors reactions to the crisis events are affected
by the liquidity risk conveyed by the levels of fair value hierarchy in both nancial and non-nancial
rms. During liquidity-constraining events, investors have stronger negative reactions to companies with
more level 3 illiquid assets and liabilities on their balance sheets. During liquidity-expanding events,
investors react more positively to companies with more illiquid assets.
2015 Elsevier B.V. All rights reserved.

1. Introduction
The period from 2007 to 2010 featured major economic upheavals that have profoundly affected the nancial industry. Rare negative events with high economic impact have occurred. These
events have led to high volatility in the nancial markets, leading
to losses and spreading a fear of a depression among investors
comparable to that of 1929. This framework provides an opportunity to deepen our understanding of the relationship between the
information disclosed by companies to the markets on fair value
assets/liabilities held on their balance sheets and stock returns
during crisis events.
According to international accounting standards (IAS 39 and
IFRS 7), the value of nancial securities is determined using the fair
value principle, by which asset values are assessed using three valuation methodologies (Fair Value Hierarchy). The rst of these is
the market price, where the security is listed in a liquid and active
market (mark-to-market or 1st level). If the market price is not
available, then the fair value of the security is estimated using
the price of a similar nancial instrument listed in a liquid and
Corresponding author.
E-mail addresses: oroggi@stern.nyu.edu (O. Roggi), a.giannozzi@unilink.it,
alessandro.giannozzi@uni.it (A. Giannozzi).
http://dx.doi.org/10.1016/j.jbankn.2015.04.011
0378-4266/ 2015 Elsevier B.V. All rights reserved.

active market, or by referring to the price of a recent comparable

transaction between knowledgeable and willing parties (similar
transaction or 2nd level). In the event that such information is
unavailable, the fair value can be estimated using nancial and statistical models of common acceptance (mark-to-model or 3rd
level). This approach may require the use of assumptions not
entirely disclosed to investors, and leads to greater information
asymmetry between the rm and nancial markets.
By dening liquidity risk as the inability to sell assets on the
market for a foreseeable price (asset liquidity risk) (Brunnermeir
and Pedersen, 2009), Level 1 of the Fair Value Hierarchy appears
to be the most liquid asset class, as it is composed of assets/liabilities with values based on their market prices. In contrast, the
value of securities included in level 3 of the Fair Value Hierarchy
is estimated through nancial models, resulting in assets/liabilities
with higher liquidity risk. The higher the level 3 amounts are, the
greater the assets liquidity risk will be and, therefore, the higher
the total liquidity risk of the company (Brunnermeir and
Pedersen, 2009). If the relationship between liquidity risk and fair
value disclosure exists, then during a liquidity contraction in the
nancial markets, investors will tend to penalize illiquid companies (i.e., those with high percentages of level 3 assets) (Lev and
Zhou, 2009). This increased uncertainty should generate a negative
impact on company value. In an efcient nancial market, this

328

O. Roggi, A. Giannozzi / Journal of Banking & Finance 58 (2015) 327342

would be reected in stock prices. Therefore, a high level of liquidity risk may lead to a decline in corporate stock prices. This could
explain the price decrease that occurred during the nancial crisis
for companies where assets and liabilities became illiquid and
investors decided to no longer include the stocks of illiquid companies in their portfolios.
The aim of this paper is to investigate the impact of company
liquidity risk on the stock prices of nancial and non-nancial
companies by analyzing investors reactions to key crisis events
during the period from 2008 to 2010. The rm liquidity risk is measured by the three levels of fair value information (level 1-mark to
market, 2-market observable input and 3-mark to model) disclosed
on the balance sheet. The presence of level 3 illiquid assets and liabilities represents a greater asset liquidity risk and therefore a
higher company liquidity risk. The effect of liquidity risk information on investors decision-making is explored by analyzing their
reactions to 106 events with liquidity-constraining and
liquidity-expanding effects for European nancial and
non-nancial companies. This analysis will indicate whether rm
liquidity risk as conveyed by the fair value hierarchy affected
investment decisions during the crisis. Building on the analysis of
Lev and Zhou (2009), this paper investigates investors reactions
to crisis events during the period between February 17th, 2008
and June 22nd, 2010. This study uses data collected on the fair
value disclosures of 313 European nancial and non-nancial companies under IAS 39 and IFRS 7 (for the list of companies, see
Table A1 in the Appendix).
In the analysis, 106 liquidity crisis events were classied as follows: 14 Distress/Liquidity-Constraining Events, 26 Bank Bailout
Events and 66 Capital Injection Events. In this classication, the
Capital Injection group consists of events that signal expanded liquidity across the economy, while Distress Events signal constrained
liquidity. The Bank Bailouts1 group consists of events in which nancial institutions have been nancially helped or bailed out by governments or other nancial institutions with the aim of stabilizing
the nancial system. The second group, Distress Events, consists of
episodes related to the failure or nancial trouble of nancial rms,
increases in interest rates by Central Banks (such as the ECB rate
increases on July 3, 2008) and the expiration of guarantee programs
for monetary market funds (such as the expiration of US guarantee
program on September 18, 2009). The third group, Capital Injection
Events, includes announcements of liquidity injections into the
nancial markets and interest rate cuts by major Central Banks
(e.g., ECB, FED, Bank of England), and government policies and legislative actions by major governments intended to address the crisis.
We applied cross-sectional regressions of the event group
returns2 on the levels of fair value assets and liabilities,3 along with
control variables, using all 106 events registered in the investigated
period. Using xed effects models and a non-parametric methodology (PLS), this paper demonstrates a complex reaction to the liquidity risk information conveyed by the three fair value levels. The
liquidity risk information captured by the three fair value levels
affects investors reactions to the crisis events.
We improve upon the existing literature in several ways. First,
differently from Lev and Zhou (2009), these ndings demonstrate
a negative reaction by investors to bank bailout events, especially
for banks with higher degrees of liquidity risk and higher leverage.
Second, this analysis is based on a large database of crisis events
over a longer period of time, allowing us to investigate investors
reactions for both nancial and non-nancial rms in the
European nancial markets. We also investigate the effect of
1
As previously asserted by Lev and Zhou (2009), Bank Bailouts indirectly signal
expanded liquidity across the economy.
2
We run separate regressions for each group of events.
3
For the list of variables, see Table A2 in the Appendix.

recording liabilities at fair value, a topic that has received little

attention in the mainstream research. Finally, the ndings are
based on partial least squares regressions in addition to the traditional xed effect models.
The paper is structured as follows. Section 2 surveys the most
relevant literature. In Section 3, the hypotheses are stated.
Section 4 describes the data. Sections 5 and 6 are devoted to measuring investors reactions to crisis events. In particular, the groups
of crisis events and the cumulative abnormal returns (CARs) are
introduced. Section 7 provides the results of the empirical analysis,
and Section 8 presents the conclusions.

2. Literature review
This paper is mainly related to three broad strands of the literature on nancial rms: event studies investigating the impact of
policy interventions on stock prices, value-relevance studies of fair
value information and market discipline.
The rst strand focuses on the stock market reaction to regulatory events, especially during the nancial crisis. Several papers
(Bomn, 2003; Veronesi and Zingales, 2010; Elyasiani et al.,
2011; Rangel, 2011; Fiordelisi et al., 2014; Pennathur et al., 2014)
have investigated the effects of policy interventions on companies
stock prices. For instance, Veronesi and Zingales (2010) examined
the costs and benets for investors of the U.S. government plan
that injected equity into nine U.S. commercial banks. These authors
conducted an event study of bonds and common equity value
around the date of the planned infusion, concluding that the government intervention reduced the enterprise value by 2.5%. In contrast, Elyasiani et al. (2011) examined the market reactions to TARP
capital infusions and found that investors reacted positively to the
news of injections but negatively to the news of capital injections
through non-TARP programs.
Pennathur et al. (2014) investigated the effects of policy interventions during the nancial crisis on several types of institutions
(banks, insurance companies, savings and loan associations and
REITs). They demonstrated that the interventions were
wealth-decreasing and risk-increasing events for nancial institutions. In particular, they found that seven of the nine interventions
events elicited highly signicant negative market reactions.
Semaan and Drake (2011) noted a decrease in the market risk for
rms following deregulation, expected in the case of insurance
brokers.
Fiordelisi et al. (2014) analyzed the effect of monetary policy
interventions on the stock returns of 27 Globally-Systemically
Important Financial institutions (G-SIFIs) throughout the period
from June 1st, 2007 to June 30th, 2012. Using cumulative abnormal
returns around the announcement of each monetary policy intervention, they found that non-conventional monetary policy measures (liquidity support or monetary easing decisions) have been
more effective than traditional measures (interest rate cuts) in
restoring the stability of the banking system. In addition, they
observed positive CARs around expansionary monetary policy
actions, while policy inaction and restrictive measures had negative effects on bank stock prices.
The second strand of the literature relevant to this paper is
value-relevance studies, which examine the effects of the use of
market prices for asset and liability pricing and the
value-relevance of fair value information. Starting in the early
nineties, most researchers focused primarily on the
value-relevance of fair value (Barth, 1994; Barth et al., 1995,
2001; Eccher et al., 1996; Holthausen and Watts, 2001). Barth
(1994) supported the existence of a strong relationship between
fair value information and stock prices and believed that banks
stock price changes can be measured using the securities at market

O. Roggi, A. Giannozzi / Journal of Banking & Finance 58 (2015) 327342

value. This important result was conrmed by Barth et al. (1995).

They demonstrated that banks net prots measured using fair
value accounting are more volatile than those calculated with historical cost asset valuations.
Unlike previous studies, Barth et al. (1996) noted for the rst
time that the additional explanation required by SFAS 107 signicantly affected bank stock prices. In contrast, in the same year,
Eccher et al. (1996) demonstrated how the historical cost has a
greater explanatory power over bank stock prices, both in absolute
and relative terms. Previously, Beaver and Landsman (1983),
Beaver and Ryan (1985) and Bernard and Ruland (1987) found that
the fair value items calculated for certain types of assets according
to SFAS 33 have no incremental explanatory power over bank stock
prices. In addition, other studies such as Murdoch (1986) and
Cornett et al. (1996) demonstrated how the announcements of
new fair value standards are followed by negative abnormal
returns on bank stock prices. Barlev and Haddad (2003) argued
that Historical Cost Accounting captures an incorrect picture of
the current situation and induces managers to create hidden
reserves of capital to cover administrative errors. For Freixas and
Tsomocos (2004), historical cost accounting allows for the distribution of a greater amount of bank dividends over time. According to
these scholars, when moral hazard and information asymmetry are
very high, the fair value accounting could increase market discipline. Freixas and Tsomocos (2004) concluded that Fair Value
Accounting leads to lower indirect bankruptcy costs because it promotes the bankruptcy of distressed companies, guaranteeing the
redistribution of resources amongst healthy companies and avoiding risk taking incentives for bank managers.
The third strand of the literature relevant to this paper is market
discipline studies of nancial rms (Cubillas et al., 2012;
Demirg-Kunt and Huizinga, 2004; Flannery and Nikolova,
2004; Honohan and Klingebiel, 2003; Martinez Peria and
Schmukler, 2001).
Demirg-Kunt and Huizinga (2004) showed that the presence
and the generosity of explicit deposit insurance weakens market
discipline. They used an international database of banks in 51
countries to show that explicit deposit insurance makes depositors
less likely to monitor banks, weakening the degree of market
discipline.
For Cubillas et al. (2012) market discipline is enhanced with
accounting disclosure and institutional quality but weakens with
the extent of government safety nets, restrictions on bank activities,
and ofcial supervision. Cubillas et al. (2012) analyzed a panel data
set of banks from 66 countries around 79 banking crises and they
concluded that the adoption of an explicit guarantee, government
recapitalization, and nationalization programs have a weakening
effect on market discipline. The interventions to increase safety nets
and depositor protection that governments usually adopt to resolve
banking crises make depositors less concerned about future banking
crises and reduce their incentives to exert discipline.
Honohan and Klingebiel (2003) suggested that following bank
interventions and failures, depositors may become more aware of
the risk of losing deposits, so they may start exercising stricter
market discipline. Then again, when governments usually respond
to banking crises with resolution policies that strengthen bank
safety nets, depositors may be more relaxed if a new banking crisis
occurs and have fewer incentives to exercise discipline. On the
contrary, Martinez Peria and Schmukler (2001) found in
Argentina, Chile, and Mexico during the 1980s and 1990s that
the relative importance of market discipline increases after crises
and that deposit insurance does not appear to diminish the extent
of market discipline.
The potential contagion effects of the use of fair value accounting during a liquidity crisis were investigated by Allen and Carletti
(2007), who concluded that Fair Value Accounting can lead to a

329

contagion effect between banks and insurers. Allen and Carletti

(2008) demonstrated that when economic crises occur and asset
valuations are not adjusted for illiquidity, the only way to mitigate
a contagion effect is to temporarily suspend the application of Fair
Value Accounting. Plantin et al. (2008) found that the damage
caused by the market value reaches its maximum when this
method is used to assess long-term illiquid and senior instruments,
which constitute a large portion of banks assets. The IMF Global
Financial Stability Report (2008) stated that standard setters are
following the correct direction with the implementation of Fair
Value Accounting, despite the measurement difculties and associated volatility. This is explained by the better assessment of the
companys nancial situation (especially for nancial rms), which
is obtained through the use of market prices. Laux and Leuz (2009)
analyzed the role of FVA for U.S. banks in the nancial crisis and
they concluded that it is unlikely that FVA contributed to U.S.
banks problems in the nancial crisis.
After the 20072008 crisis, several scientic papers (Song et al.,
2009; Kolev, 2008; Goh et al., 2009) analyzed the usefulness of
Fair Value information for investors by regressing the market values
(or stock returns) of companies on book values, earnings and Fair
Value disclosure. While this conventional approach is essentially a
valuation analysis, examining whether investors price the disclosed
information, this study aims to identify the impact of differential liquidity risks, reected by the three fair value levels, on investors
decisions. We differ from the value-relevance studies because we
evaluate the value-relevance of the risk information embedded in
the three levels of fair value. Lev and Zhou (2009) contributed to this
eld of literature by analyzing the separation of nancial assets and
liabilities into three liquidity levels. In particular, they investigated
investors reactions to crisis events during the nal months of
2008 (the peak of the nancial crisis) conditioned on the fair value
disclosures of 3,929 U.S. nancial and non-nancial companies
under SFAS 157. Using ve groups of events (Distress, Rescue,
Policy, FED liquidity injections/interest rate cuts and Capital infusions), they documented that the fair value level separation is useful
to investors. Specically, for nancial rms, they demonstrated positive/negative investor reactions to liquidity-expanding/liquidi
ty-constraining events. Lev and Zhou (2009) were the rst to investigate the role of liquidity risk information (fair value hierarchy) in
investors decision making by analyzing their reactions to
liquidity-constraining or liquidity-expanding events during the
nancial crisis for both U.S. nancial and non-nancial companies.
Our paper brings together the strands of literature discussed
above, analyzing the impact of policy and monetary actions and
key crisis events on nancial companies stock prices, according
to their differential liquidity risk as reected by the three fair value
levels. We also investigate investors reactions to crisis events conditioned on the fair value hierarchy of non-nancial companies, a
topic little explored in the mainstream research. In addition, we
believe that the liquidity risk information conveyed by the fair
value hierarchy should increase market discipline. Therefore, when
a liquidity constraining events occurs, investors should penalize
riskier banks (with higher asset liquidity risk) by selling stocks,
generating a negative abnormal return.

3. The hypotheses
To analyze the impact of company liquidity risk as conveyed by
the fair value hierarchy on the stock prices of nancial and
non-nancial companies during crisis events, the relationship
between the market price and the three fair value levels are investigated in order of increasing liquidity.
The hypotheses tested in this study are based on the denition
of 3 groups of crisis events: Distress Group (liquidity-constraining

330

O. Roggi, A. Giannozzi / Journal of Banking & Finance 58 (2015) 327342

events), Capital Injections (liquidity-expanding events) and Bank

Bailouts. For a detailed description of each group, see Section 5.
We expect that investors react differently to crisis events where
nancial and non-nancial companies are considered. In fact, as
asserted by Lev and Zhou (2009), during crisis events, liquidity risk
has a complex effect on the value of assets and liabilities, and this
effect differs between nancial and non-nancial companies due to
the different business models of the two sectors. In addition, previous studies investigating the stock market reaction to regulatory
events (Pennathur et al., 2014; Semaan and Drake, 2011;
Elyasiani et al., 2011) analyzed nancial rms separately. For this,
the entire sample is split into two subsets of nancial and
non-nancial rms. As a result, the following hypotheses are provided separately for non-nancial and nancial companies.

default risk). However, a liquidity-constraining event will decrease

rms ability to rollover their short-term liabilities (renancing risk),
thereby damaging stockholders. Given the interplay of the
above-mentioned opposing forces, investors overall reaction to
liquidity-constraining events on the liabilities of non-nancial rms
primarily depends on the maturity of those liabilities (Lev and Zhou,
2009). In addition, the nancial leverage should exacerbate both rollover and credit risk concerns. If the rms fair-valued liabilities are
primarily short-term, then investors will react negatively (rollover
concerns), whereas if most liabilities are long-term and of Levels 1,
investors will react positively (credit risk or default concerns). Due
to this ambiguity, we make no a priori hypothesis about the impact
of fair-value liabilities of non-nancial rms on stock prices and we
decide to treat the question as an empirical issue.

3.1. Non-nancial companies and fair value

3.2. Financial companies and fair value

The fair value assets of non-nancial companies do not pertain

generally to their core businesses, and they should be retained on
balance sheets exclusively kept for liquidity and hedging purposes
(Lev and Zhou, 2009). In general, these assets are highly liquid and
mainly classied in levels 1 and 2 of the fair value hierarchy.4
Moreover, non-nancial companies often do not report any fair value
assets or liabilities since they do not invest in such kind of instruments. When liquidity-contracting events occur, investors tend to
require liquid securities, leading to an increase in the value of the
assets held by non-nancial companies, a phenomenon better
known as ight to quality. The value of such liquid assets increases
when liquidity-constraining events occur because demand for liquid
assets increases. We expect that this effect will be greater for highly
leveraged companies.
This phenomenon can be translated into the following
proposition:

Level 3 assets are more illiquid and present a higher degree of

information asymmetry than Level 1 and Level 2 assets because
they are measured with unobservable inputs. As asserted by
Lev and Zhou (2009), liquidity constraining events should therefore decrease the value of these assets, causing a negative reaction by investors. We expect investors to react more negatively
to the higher liquidity risk Level 3 assets than to the lower liquidity risk Level 1 and 2 assets. The more illiquid the assets of
nancial rms are, the more negative the investors reaction is
expected to be. In addition, we expect that this effect will be
greater for highly leveraged nancial companies. The following
hypothesis is stated:

Proposition 1. when Distress/Liquidity-constraining events occur,

investors will react positively to Levels 1 fair value assets and will not
react (or will react only mildly and negatively) to Level 3 fair value
assets.
According to this statement, the stock market prices of
non-nancial companies should behave in the opposite way when
liquidity-expanding events occur, such as a capital injection to
banks or an interest rate cut. In these type of events, we believe
that the value of liquid assets (Level 1) will decrease, damaging
shareholders. In contrast, investors reactions to illiquid Level 3
assets should be positive, or due to the low percentages of Level
3 Assets, not signicant. Consistent with Lev and Zhou (2009),
we formulate the following proposition:
Proposition 1-bis. when liquidity-expanding events occur, investors
will react negatively to Levels 1 fair value assets and will not react (or
only mildly and positively) to Level 3 assets.
Concerning the relationship between the fair value liabilities of
non-nancial companies and stock returns, consistently with Lev
and Zhou (2009), we expect that fair value liabilities will mainly
consist
of
relatively
illiquid
instruments
such
as
mortgage-backed securities, whose values tend to decrease during
liquidity-constraining events. We do believe that for
liquidity-constraining events, the value of liabilities will decrease
due to a wealth transfer from lenders to shareholders5 (higher
4
Non-nancial companies often do not any fair value assets or liabilities since they
do not invest in such kind of instruments. In this study, 80 non-nancial companies
do not report any fair value assets or liabilities.
5
As asserted by Lev and Zhou (2009), in liquidity-constraining events, the
company default risk will increase and the value of liabilities will decrease.

Proposition 2. when Distress/Liquidity-constraining events occur,

investors react more negatively to level 3 assets than to levels 1 and
more negatively to nancial companies with higher degree of nancial
leverage.
In contrast, in a liquidity-expanding event we expect investors
to react more positively towards nancial rms with greater
amounts of illiquid Level 3 assets. Higher liquidity risk nancial
companies should benet more from capital injections.
Proposition 2-bis. when liquidity-expanding events occur, investors
react more positively to level 3 assets than to levels 1.
Our descriptive statistics suggest (see Table A4 in the
Appendix) that nancial companies liabilities generally consist
of liquid items, mainly classied in Levels 1 and 2 of the fair
value hierarchy. Furthermore, there is a large amount of government insurance backing customers deposits. Due to these two
considerations, consistently with Lev and Zhou (2009), we expect
that investors do not react to liquid Level 1 liabilities because
they are largely unaffected by crisis events. In contrast, Level 3
liabilities often consist of illiquid non-insured items, and therefore we expect investors to react negatively to Level 3 liabilities
during liquidity-constraining events and to react positively to
L3L during liquidity-expanding events. The following statement
can be tested:
Proposition 3. when liquidity-expanding/liquidity-constraining
events occur, investors do not react to the level 1 fair value liabilities
and will react in a positive/negative way to level 3 liabilities.
For bank bailout events, we observe in the preliminary results
of ANOVA a negative mean reaction (Pennathur et al., 2014;
Veronesi and Zingales, 2010). Relying on ANOVA preliminary ndings, and differently from Lev and Zhous study based on U.S. data
that documented a signicantly positive market reaction (Mean
CAR of 2.58%) to bank bailout events, we believe investors

O. Roggi, A. Giannozzi / Journal of Banking & Finance 58 (2015) 327342

reactions to be negative as a consequence of the risk of nationalization. We believe that this negative reaction should be stronger
on the company level, especially for highly leveraged nancial
institutions with large amounts of illiquid Level 3 assets. In other
terms, nancial rms with higher liquidity risk (higher amount
of L3A) should be perceived as having a greater risk of nationalization, as well as higher exposure to default risk. The following
hypothesis is tested:
Proposition 4. when a bank bailout event occurs in the market,
investors react more negatively to Level 3 fair value assets than to
Level 1, and more negatively to nancial companies with higher
degree of nancial leverage.

4. Data sample
To test our hypotheses, we collected data on fair value
assets/liabilities and control variables for European nancial and
non-nancial companies. The sample selected for this analysis consisted of all 313 companies (59 nancial companies and 254
non-nancials), under the IAS 39 and IFRS 7, listed in the
Eurostoxx index6 (Table A1 in the Appendix provides the list of companies). This index is widely used in nancial practice as an underlying instrument for securities such as ETFs, futures, options and
structured products, and as a benchmark for measuring risk and corporate performance. The Eurostoxx index consists of a variable number of small, medium and largely capitalized companies located
throughout 12 Eurozone countries (Austria, Belgium, Finland,
France, Germany, Greece, Ireland, Italy, Luxembourg, Netherlands,
Portugal and Spain). The observation period is the 2007, 2008 and
2009 scal years.
IAS 32 and IAS 39 in 2005 introduced disclosures related to the
methods and signicant assumptions used to determine fair value
for different classes of nancial assets and nancial liabilities. The
required disclosures include:
- Whether the fair value is based on quoted prices or valuation
techniques.
- Whether the fair value is based on a valuation technique that
includes assumptions not supported by market prices or rates,
and if so, the amount of change in fair value recognized in
prot or loss that arises from the use of the valuation
technique.

IFRS 7. These data were then computed with the same valuation
criteria of the 2009 amendments to IFRS. This allows us to conduct
the analysis on homogeneous data.
As we believe that investors have different reactions to crisis
events between nancial and non-nancial companies, we study
the two sectors separately. We built two subsamples (nancial
companies and non-nancial ones) and ran our analysis separately
for the two subsamples. 80 companies did not report any fair value
assets and liabilities (Non-fair value sample) and were excluded
from the regression analysis.7
The Table A3 shows each level of the fair hierarchy as a percentage of total assets/total liabilities.8 Table A4 provides the mean percentage of fair value assets and liabilities in both nancial and
non-nancial companies. Financial rms show greater L1A and L2A
than those of non-nancial companies in all years analyzed.
Financial rms achieved a lower ROA and OPACT than
non-nancial companies.

5. Crisis events
During the period from February 2, 2008 to June 22, 2010, we
identied 106 crisis events.9 To select the crisis events we used
the Federal Reserve ofcial publication The Financial Crisis: a
Timeline of Events and Policy Actions, the 79th Annual Report
2008/2009 compiled by BIS, and we augmented it with press
releases from the major news agencies and newspapers (i.e., Time
on Line, Wall Street Journal, IlSole24Ore). In addition to European
crisis events, we consider U.S. and UK events in this analysis, as
we assume that European companies are also impacted by policy
interventions in other relevant economies.
Partially consistent with Lev and Zhou (2009),10 we classied
the crisis events into three groups according to the nature of the
events:
- Bank Bailout Events;
- Distress/liquidity-constraining events;
- Capital Injection Events (liquidity-expanding events).
The rst group, Bank Bailout Events, consists of 26 events in
which nancial institutions have been nancially helped or bailed
out to stabilize the nancial system, mergers between major banks
have been approved, and investment banks have been granted
7

IFRS 7, effective for nancial statements issued for the scal

years beginning from January 1st 2007, retains the IAS 32 disclosure requirements mentioned above. Later, in 2009, the IASB issued
amendments to IFRS 7, requiring enhanced disclosures about fair
value measurements and liquidity risk. The amendments to IFRS
7 required companies to classify fair value measurements for
nancial instruments using a three-level fair value hierarchy, with
reference to the observability and signicance of the inputs used in
making the valuation.
For the nancial statement issued before the 2009 amendments
to IFRS, we manually derive the three levels of fair value data from
the notes to the balance sheets of each company in the sample
where the amounts of assets and liabilities for each type of valuation technique (quoted prices, valuation techniques supported by
market prices, valuation techniques not supported by unobservable inputs) were indirectly disclosed according to IAS 39 and
6
At the time of the analysis, the Eurostoxx index included 313 European
companies. Our analysis is based on the entire Eurostoxx database. We also clarify
that insurance companies listed in the Eurostoxx index were considered nancial
rms.

331

These companies with no fair value data are used only for the ANOVA.
The ratio between fair value assets and total assets strongly decreases from to
2007 to 2008 as a consequence of decision made by the IASB in mid-October 2008
which permitted to record nancial assets/liabilities at their historical cost. We
believe that a decrease in the percentages of fair value assets/liabilities does not affect
our conclusions since we demonstrate the existence of signicant relationships
between fair value items and investors reactions, using different methodologies.
9
For the list and description of the events, please contact the corresponding author.
We did not voluntarily consider events related to the Euro sovereign debt crisis
(mainly taking place in years 2011 and 2012) as we believe that these types of events
could bias our analysis. The rm-specic investors reactions to such events should be
affected by the crisis in their own country. We could expect an Italian bank to be
highly affected by the Italian government crisis (since it owns large amounts of Italian
Bonds), whereas it could be unaffected by the sovereign crisis in Portugal. In this case,
investors should react according to the bank exposure to the countrys default risk,
and liquidity risk information reected by fair value hierarchy could become
irrelevant (especially if we consider that government bonds are classied as L1A).
For this type of event, a portion of assets assumed to be liquid items (L1A) becomes
highly illiquid due to the sovereign crisis, completely biasing our assumptions and
ndings. We believe that a specic study with different objectives should be devoted
to analyzing the impact of the Euro sovereign crisis on banks stock prices.
10
The authors classied the events into ve groups: Distress, Rescue (government
actions taken to rescue US nancial and non-nancial rms, Policy events (US
government policies and legislative action to reverse the crisis), Capital infusions
(such as the TARP program) and FED events (interest rate cuts implemented by the
FED).
8

332

O. Roggi, A. Giannozzi / Journal of Banking & Finance 58 (2015) 327342

permission to convert into commercial banks (such as Goldman

Sachs).
The second group, Distress/liquidity-constraining Events, consists
of 14 episodes related to the investment fraud in the capital markets (e.g., Madoff fraud), the bankruptcy or distress of nancial
rms (such as the Lehman bankruptcy). These events all signal
constrained liquidity across the economy.
The third group, Capital Injection Events, includes 66 announcements of liquidity injections and interest rate cuts instituted by
major Central Banks (ECB, FED, Bank of England, etc.), government
policies and legislative actions intended to reverse the crisis by
governments in European countries and the U.S., and bank capital
purchases (such as under the TARP program).
In this classication, Capital Injections group contains events
that signal expanded liquidity across the economy, while Distress
events signal constrained liquidity. For bank bailout events, we separately analyze investors reactions to these different types of
events using specic hypotheses as previously illustrated.
As numerous events take place within a relatively short period
of time, we use a two-day event window for stock return computation. In particular, the effects of each individual event are
assumed to terminate within a two-day period (the day of the
event and the rst trading day following the event).11 For the
date of each event and the observation window, see Table A5 in
the Appendix. The mean cumulative abnormal returns (CARs) are
estimated to assess the stock market reactions to the 106 crisis
events. This analysis revealed that while most events directly
affected nancial rms, investors in non-nancial companies also
reacted to the events. In general, investors had negative reactions
for both nancial and non-nancial rms in response to alarming
events such as Lehmans implosion. Nevertheless, certain events
were apparently perceived as being benecial to nancial rms
but harmful to non-nancial rms, or vice versa. This supports
our decision to separately analyze nancial and non-nancial
companies.
The sample of 313 companies includes rms with and without
fair value information. In particular, 233 companies reported fair
value data (Fair Value sample), while 80 companies did not report
any fair value assets or liabilities (Non-Fair Value sample). In order
to check the signicance of crisis events, we applied the ANalysis
Of Variance (ANOVA) to these sub-samples (fair value vs.
non-fair value for the entire sample, fair value vs. non-fair value
sample for non-nancial companies, Fair value sample-Financial
companies vs. Fair value sample non-nancial companies, Fair
value
sample-Financial
companies
vs.
Non-Fair
value
sample-Non-Financial companies). As previously discussed, we
expect a signicant difference in mean reactions between nancial
vs. non-nancial companies and between fair value sample vs.
non-fair value sample. We used t-tests to test the differences in
mean, and , , and to denote signicance at the 10, 5, and 1
percent levels, respectively. Table 1 presents the results of
ANOVA. To avoid the introduction of a selection bias, we decided
to run our regression analysis on all 106 crisis events.
6. Measuring investors reactions to crisis events
Investors reactions to fair value hierarchy were measured
through the concept of mean cumulative abnormal returns
(CARs). In this case, the abnormal returns were calculated as
follows:

11
If an event occurred on either Saturday or Sunday, we use the following Monday
(or the next trading day) as the event day. For events on consecutive trading days, we
use a one-day event window.

PT
CARsij

t1 Rit

 REt

where
CARsij = average cumulative abnormal returns for the relevant
j-th event for the i-th rm
T = number of days of measuring abnormal returns = 2 days
Rit = rate of return on day t of the i-th rm
REt = rate of return on the Eurostoxx index on day t
The hypotheses were tested with 36 cross-sectional regressions
and 12 Partial Least Square regressions12 (PLS), with separate regressions for the group of nancial companies and the non-nancial
companies. Partial Least Squares Regression is based on the NIPALS
algorithm. It relates two matrices of variables, the independent X
and the dependent Y. From matrix X, the algorithm extracts PLS
components that simultaneously summarize the variable X and
explain the dependent variables Y. Compared to other methods,
the principal advantages of the PLS regression are:
- PLS solves the problem of missing data: the algorithm can be
run with missing values.
- The observations matrix JX of independent variables and matrix
JY made up of dependent variables can be correlated;13
- Some, or even all, of the JX independent variables and the JY
dependent variables can be qualitative or nominal in type.
Specically, we cross-sectionally regress the event group
returns on the levels of fair value assets and liabilities, using all
106 events registered in the investigated period. To avoid multicollinearity problems (especially with traditional data panel
methodologies) related to the use of all levels of the fair value hierarchy in the same regression model, we decided to consider the
Level 1 and Level 3 of fair valued assets and liabilities. In doing
so, L2A and L2L were excluded from the regression analysis.14 In
addition to PLS methodology, three regressions were run for each
group of events: an OLS regression, a regression with the xed
effects correction, and nally, a regression with the ability to record
nancial assets/liabilities at their historical cost as permitted by the
IASB in mid-October 2008 (labeled OLS with HC). In this latter regression model, we introduce two additional variables: the rst is a
dummy variable, called HC RECLASS (1 = reclassication of specic
assets at historical cost as permitted by the IASB on October 2008;
0 = no reclassication of assets at historical cost), the second variable
(HCA RECLASS) represents the amount of assets reclassied at historical value as a percentage of total assets.
For the list of independent variables, see Table A2 in the
Appendix.

7. The relevance of fair value hierarchy to European companies

stock returns during the nancial crisis
Sections 7.1 and 7.2 present our ndings based on all 106 crisis
events, separately for nancial and non-nancial companies. The
estimates relate to each of the three groups of events considered
in our study. The results of the regression analysis for nancial
companies are shown in Tables 57. Tables 24 report the results
of the analysis for non-nancial companies.
12
For a discussion of how to identify signicant variables using PLS regression, see
Jun et al. (2009).
13
PLS regression allows up solving the problems of multicollinearirty among
independent variables. For details about PLS regression see Tenenhaus (1998).
14
Regarding the omitted variables L2A/L2L, we believe that this decision did not
affect our results since we tried to include L2A/L2L in the model resulting in
insignicant coefcients for almost all regressions runs.

Table 1
Results of ANOVAs.
Fair value
sample
(nancial)

Vs

Non-fair value
sample (Nonnancial)

Fair value sample

(Non-nancial)

Vs

Non-fair value
sample (Nonnancial)

Fair value
sample (Full
sample)

Vs

Non-fair value sample

(Full sample)

Fair value
sample (Financial
companies)

Vs

Fair value sample

(Non-nancial companies)

02/17/08
03/16/08
06/03/2008
07/04/2008
07/11/08
07/13/08
09/07/08
09/15/08
09/16/08
09/18/08
09/19/08
09/22/08
09/25/08
09/29/08
09/30/08
10/03/08
10/06/08
10/07/08
10/08/08
10/09/08
10/10/08
10/13/08
10/14/08
10/15/08
10/16/08
10/17/08
10/19/08
10/21/08
10/24/08
10/28/08
10/29/08
10/30/08
11/03/08
11/05/08
11/09/08
11/10/08
11/12/08
11/14/08
11/20/08
11/21/08
11/24/08
11/25/08
11/26/08
12/02/08
12/03/08
12/05/08
12/09/08
12/12/08
12/15/08
12/16/08
12/19/08
12/23/08
12/29/08

0.01598
0.01084
0.02512
0.00572
0.0315
0.04728
0.018718
0.02491
0.05147
0.0154
0.082776
0.069023
0.025935
0.04864
0.02982
0.028446
0.02005
0.06232
0.04371
0.03788
0.017713
0.00021
0.079893
0.01756
0.07693
0.03706
0.00065
0.032043
0.04171
0.07967
0.012865
0.082053
0.02458
0.04917
0.023919
0.00923
0.06646
0.00017
0.0632
0.04404
0.04488
0.07592
0.008548
0.01963
0.01985
0.01371
0.07206
0.02683
0.03023
0.00624
0.007006
0.01744
0.000833

0.01101
0.00764
0.01683
0.00623
0.02431
0.03513
0.01024
0.01625
0.03546
0.01552
0.05634
0.05085
0.00911
0.04867
0.03059
0.01528
0.02705
0.05844
0.04709
0.02917
0.004122
0.004917
0.053551
0.01736
0.05115
0.02444
0.00497
0.00839
0.03304
0.0301
0.035999
0.066492
0.027462
0.025139
0.013728
0.009942
0.03868
0.00494
0.0347
0.01964
0.026201
0.034237
0.005837
0.01929
0.0012
0.01668
0.029867
0.01378
0.01256
1,4E-05
0.00902
0.005516
0.006515

0.02568
0.00915
0.00087
0.00074
0.03335
0.03414
0.00342
0.00527
0.01846
0.02653
0.02151
0.02258
0.00517
0.04045
0.01602
0.00708
0.04049
0.04963
0.03031
0.02732
0.035686
0.03586
0.075618
0.02577
0.06337
0.00726
0.04316
0.020086
0.02612
0.01106
0.051839
0.058458
0.029935
0.012279
0.013236
0.022018
0.03502
0.009936
0.03934
0.02344
0.04712
0.06429
0.005336
0.01909
0.013946
0.02182
0.04721
0.01988
0.0134
0.006578
0.00499
0.00896
0.004915

0.01845
0.00305
0.00041
0.0052
0.0246
0.02685
0.001114
0.00614
0.01888
0.01795
0.02611
0.02523
0.00199
0.03733
0.01839
0.00668
0.03531
0.04824
0.0333
0.0294
0.023993
0.021803
0.065694
0.01703
0.05496
0.01496
0.017656
0.017844
0.02988
0.01662
0.043049
0.050462
0.024952
0.015582
0.016841
0.01614
0.0317
0.001751
0.03737
0.02542
0.028792
0.043353
0.005247
0.01457
0.011492
0.01975
0.032761
0.01251
0.00804
0.00584
0.00152
0.00487
0.005934

0.02127
0.01032
0.00021
0.00111
0.03379
0.04181
0.00707
0.01490
0.03503
0.02224
0.05178
0.04578
0.01534
0.04599
0.02328
0.00969
0.03229661
0.05770158
0.03795372
0.03348311
0.028486606
0.019978578
0.080694613
0.02285852
0.07240952
0.02185586
0.023858496
0.026628764
0.03464830
0.04443323
0.035204584
0.072122241
0.028561831
0.030477496
0.01889375
0.016760441
0.0515146
0.005482149
0.0523676
0.03426518
0.047930471
0.07244843
0.007091072
0.02011302
0.017341502
0.01880024
0.061029843
0.02401036
0.02201030
0.006678794
0.000568957
0.01339088
0.003152484

0.01233
0.00386
0.00024
0.00098
0.024548
0.028321
0.002731
0.007930
0.021816
0.017518
0.031470
0.029775
0.003256
0.039338
0.020555
0.002788
0.033848
0.050045
0.035747
0.029361
0.020469
0.018808
0.063541
0.017085
0.054283
0.016639
0.013644
0.016167
0.030443
0.019008
0.041798
0.053304
0.025396
0.017276
0.016288
0.015041
0.032936
0.000563
0.036894
0.024391
0.028332
0.041736
0.005351
0.015404
0.009241
0.019084
0.032247
0.012734
0.008838
0.004801
0.002858
0.003031
0.006036

0.02801245
0.01182757
0.00098
0.00079
0.03436229
0.05157333
0.020419946
0.02717222
0.05614678
0.01680415
0.090300755
0.075297657
0.028292596
0.05305901
0.03253144
0.031032394
0.02187355
0.06798070
0.04767880
0.04132551
0.019323687
0.00023421
0.087155713
0.01915647
0.08391991
0.04043101
0.00070744
0.034955547
0.04550222
0.08691319
0.014034131
0.089512817
0.026814442
0.053639576
0.026093976
0.010069256
0.07250258
0.00018577
0.06894211
0.04804314
0.048962376
0.082831726
0.009324877
0.02141543
0.021663011
0.01495453
0.078618301
0.02926399
0.03297395
0.006807664
0.007643428
0.01902546
0.000908897

0.015984975
0.009145902
0.00065
0.00661
0.033346309
0.034139326
0.003417904
0.005271214
0.018455609
0.0265281
0.021517771
0.022588908
0.005175259
0.040451073
0.01601907
0.007077602
0.040486158
0.049625127
0.030312589
0.027321238
0.035686042
0.035860061
0.075618035
0.025767282
0.063365648
0.007261104
0.043160311
0.020086292
0.026120234
0.011056124
0.051838511
0.058458217
0.029934779
0.012278718
0.013236428
0.022017801
0.03502415
0.009935522
0.039344943
0.023439633
0.047119689
0.064290125
0.00533594
0.019089696
0.01394603
0.021821882
0.047210339
0.019882508
0.013396018
0.006577538
0.004989556
0.008963722
0.004915302

333

(continued on next page)

O. Roggi, A. Giannozzi / Journal of Banking & Finance 58 (2015) 327342

Date

334

Table 1 (continued)
Fair value
sample
(nancial)

Vs

Non-fair value
sample (Nonnancial)

Fair value sample

(Non-nancial)

Vs

Non-fair value
sample (Nonnancial)

Fair value
sample (Full
sample)

Vs

Non-fair value sample

(Full sample)

Fair value
sample (Financial
companies)

Vs

Fair value sample

(Non-nancial companies)

12/31/08
01/08/09
01/15/09
01/16/09
01/19/09
01/26/09
02/02/09
02/05/09
02/10/09
02/18/09
02/25/09
02/26/09
03/02/09
03/03/09
03/05/09
03/07/09
03/09/09
03/18/09
03/20/09
03/25/09
03/29/09
04/01/09
04/02/09
04/06/09
04/09/09
04/14/09
04/17/09
04/21/09
05/06/09
05/14/09
05/20/09
05/21/09
05/28/09
06/02/09
06/10/09
06/17/09
06/26/09
06/29/09
07/02/09
07/08/09
07/20/09
08/06/09
09/18/09
09/23/09
10/05/09
12/17/09
04/11/10
05/02/10
05/03/10
05/10/10
05/18/10
06/07/10
06/22/10

0.013101
0.00425
0.03702
0.00834
0.02111
0.019944
0.01667
0.021729
0.00027
0.02713
0.00389
0.026555
0.04317
0.04259
0.0219
0.03094
0.02225
0.013681
0.016426
0.005316
0.0147
0.032214
0.047772
0.002747
0.03539
0.02397
0.021573
0.02754
0.019444
0.03236
0.016005
0.0055
0.00824
0.014256
0.01232
0.02593
0.000859
0.008856
0.0073
0.01265
0.00751
0.011293
0.01541
0.005325
0.0096
0.00086
0.018785
0.001541
0.00049
0.025555
0.006155
0.02528
0.00409

0.007702
0.00078
0.04231
0.02386
0.05028
0.034453
0.01459
0.023362
0.001482
0.01765
0.01442
0.010538
0.03564
0.03561
0.03476
0.011205
0.005374
0.005305
0.034306
0.005929
0.00878
0.040311
0.054598
0.015717
0.022633
0.018186
0.020634
0.02315
0.025756
0.02939
0.03418
0.00419
0.002347
0.021639
0.00827
0.03475
0.00545
0.010318
0.00587
0.01601
0.008104
0.011906
0.01233
0.010205
0.01161
0.00018
0.021305
0.001127
0.0003
0.024627
0.002689
0.02816
0.002726

0.013227
0.00876
0.03046
0.0028
0.00091
0.007232
0.01178
0.019414
0.017509
0.01864
0.00486
0.008028
0.02055
0.02255
0.003844
0.0111
0.00822
0.00446
0.00627
0.004969
0.01369
0.020597
0.032206
0.001806
0.019834
0.005738
0.020262
0.01904
0.010033
0.01371
0.015351
0.0033
6,03E-05
0.018615
0.007362
0.01501
0.00059
0.009718
0.00084
0.0098
0.010311
0.001275
0.01474
0.004789
0.00536
9,13E-05
0.013526
0.0019
0.00116
0.020946
0.004209
0.01599
0.002358

0.00947
0.01074
0.03185
0.0066
0.00989
0.014197
0.01238
0.019995
0.010129
0.01924
0.00656
0.009695
0.03025
0.02974
0.00582
0.01129
0.00902
0.0004
0.010217
0.007693
0.0136
0.026605
0.03898
0.001165
0.019663
0.009519
0.016924
0.01878
0.011845
0.01454
0.016952
0.0033
0.00123
0.016227
0.006146
0.0158
0.001
0.008264
0.00115
0.01118
0.008448
0.002732
0.01632
0.004546
0.00636
0.00043
0.015512
8,92E-05
0.001781
0.020874
0.002779
0.0179
0.002158

0.013695604
0.00683387
0.03232029
0.00448331
0.0067365
0.010776485
0.01332145
0.020856597
0.012393635
0.02093192
0.00483937
0.013156088
0.02701740
0.02828505
0.00331338
0.01718408
0.01248511
0.000596784
0.009168199
0.00526166
0.01398675
0.023907694
0.036749061
0.001929982
0.024155824
0.010907395
0.020714932
0.02131187
0.012529106
0.0190147
0.015515832
0.0039884
0.0024056
0.017322991
0.008733786
0.01819665
0.0002154
0.009459283
0.0015584
0.0105797
0.009518973
0.004033059
0.00911
0.00500483
0.00657138
0.00010156
0.014709752
0.00112843
0.000788316
0.021983308
0.004646952
0.01807859
0.000907333

0.009156
0.009572
0.025903
0.002830
0.006601
0.010929
0.012522
0.020834
0.013486
0.020789
0.007289
0.006484
0.025998
0.028613
0.004492
0.01194
0.009647
0.003569
0.010607
0.006046
0.010076
0.024285
0.035969
0.004695
0.018975
0.009161
0.018905
0.018563
0.010733
0.011488
0.018953
0.000471
0.001434
0.016485
0.005643
0.016364
0.001008
0.007586
0.000696516
0.010347059
0.007751
0.002339
0.00085
0.003648
0.005647
0.000959
0.016245
0.000220
0.001517
0.021349
0.002767
0.019197405
0.002229

0.014291828
0.00140212
0.03696669
0.00868822
0.02130537
0.019638861
0.01717681
0.022357574
9,969E-05
0.02665157
0.00478304
0.02597582
0.04318876
0.04262645
0.02120592
0.03238828
0.02314949
0.01323635
0.016413535
0.005994193
0.01472242
0.032183484
0.048107563
0.002240861
0.034961411
0.023831812
0.021846794
0.02698518
0.018769222
0.03227234
0.015928921
0.00570006
0.0085704
0.014092065
0.012164397
0.0261735
0.000718959
0.008811949
0.00755561
0.01252212
0.007539424
0.010929372
0.00221
0.005544558
0.0095891
0.00058375
0.018784528
0.001540727
0.0004928
0.025555354
0.006154615
0.0252777
0.0040850

0.013227142
0.00900657
0.03046173
0.002801348
0.000909026
0.007231535
0.011779303
0.020256206
0.017168173
0.018644066
0.004861914
0.008028195
0.020548861
0.022548494
0.003843623
0.011102403
0.008219354
0.004459043
0.006270065
0.004968646
0.013692482
0.020597378
0.03220566
0.00180563
0.019833589
0.005737629
0.020262187
0.019042553
0.010033059
0.013711779
0.015350597
0.003303774
6,02753E-05
0.018615361
0.007361541
0.015005909
0.000589161
0.009718216
0.000840427
0.009802817
0.010310792
0.001274534
0.00326
0.004788939
0.005364265
9,13184E-05
0.013526018
0.001903829
0.001160493
0.020945619
0.004208972
0.015987219
0.002357628

In this table, events marked with

reported a statistically signicant coefcient at the 90% level, while those with

and

were found to be signicant at the 95% and 99%, respectively.

O. Roggi, A. Giannozzi / Journal of Banking & Finance 58 (2015) 327342

Date

335

O. Roggi, A. Giannozzi / Journal of Banking & Finance 58 (2015) 327342

7.1. The results for non-nancial companies

7.1.1. Distress group

For non-nancial companies, we expected investors to react
positively to Level 1 fair-valued assets and to not react (or to react
only mildly and negatively) to level 3 fair value assets (Proposition
1). The results of the analysis partially conrm our hypothesis. In
fact, Level 1 Assets exert a positive and signicant effect

For the sub-sample of non-nancial companies, we ran 18

regressions using cross-sectional OLS/xed effects model and 6
PLS regressions. The results of the empirical analysis for
non-nancial companies are presented in the following tables.

Table 2
OLS with control variables (106 crisis events) Non-nancial companies.
Variables

Non-nancial sample
Bank bailouts

INTERCEPT
L1A
L3A
TOTNFA
L1L
L3L
TOTNFVL
ROA
OPACT
LEV
HC RECLASS
HCA RECLASS
F stat
Adj R2
Number of companies
Number of events

Capital Injections

Distress

OLS

OLS with HC

OLS + F.E.

OLS

OLS with HC

OLS + F.E.

OLS

OLS with HC

OLS + F.E.

0.0134
0.0016
0.0022
0.00009
0.0016
0.0315
0.0001
0.0645
0.0026
0.0681

0.0194
0.0711
174
26

0.0191
0.0017
0.0024
0.0001
0.0019
0.0321
0.0002
0.0712
0.0021
0.0874
0.0001
0.0002
0.0365
0.0698

0.0311
0.0013
0.0027
0.00008
0.0021
0.0254
0.00009
0.0815
0.0029
0.0755

0.0401
0.0899

0.0581
0.0045
0.0401
0.0011
0.072
0.0012
0.0028
0.0596
0.0193
0.0728

0.0981
0.1781

0.0624
0.0049
0.0378
0.0018
0.0801
0.0025
0.0026
0.0693
0.0225
0.0745
0.0001
0.0009
0.0832
0.1863

0.0653
0.0047
0.0454
0.0021
0.0896
0.0054
0.0047
0.0783
0.00196
0.0803

0.0863
0.1899

0.0568
0.0075
0.0396
0.0009
0.0006
0.0455
0.00001
0.0015
0.0210
0.0135

0.0542
0.0632

0.0526
0.0042
0.0387
0.0016
0.0007
0.0699
0.00001
0.0014
0.0196
0.0149
0.0003
0.0001
0.6530
0.0412

0.0541
0.0336
0.0393
0.0001
0.0008
0.0663
0.0001
0.0023
0.0145
0.0201

0.0649
0.0463

OLS with HC

OLS + F.E.

66

14

Table 3
OLS without control variables (106 events) Non-nancial companies.
Variables

Non-nancial sample
Bank bailouts
OLS

INTERCEPT
L1A
L3A
TOTNFA
L1L
L3L
TOTNFVL
HC RECLASS
HCA RECLASS
F stat
Adj R2
Number of companies
Number of events

Capital Injections
OLS with HC

0.0215
0.0020
0.0041
0.00251
0.00162
0.0226
0.0021

0.0489
0.0910
174
26

0.0174
0.0055
0.00357
0.0078
0.0069
0.0125
0.0033
0.0009
0.00224
0.0559
0.1168

OLS + F.E.
0.0711
0.0022
0.0016
0.0255
0.0021
0.0366
0.0039

0.0741
0.1036

OLS

Distress

OLS with HC

0.0199
0.0062
0.0550
0.0019
0.0199
0.0011
0.0035

0.2251
0.1226

0.0198
0.0059
0.0573
0.0028
0.0274
0.0150
0.0041
0.0001
0.0005
0.2241
0.1632

OLS + F.E.

0.0314
0.0061
0.0511
0.0026
0.0241
0.0021
0.0058

0.2069
0.2031

66

OLS

0.0209
0.0133
0.0201
0.0010
0.0012
0.0399
0.0001

0.1114
0.0163

0.0245
0.0117
0.0326
0.0015
0.0052
0.0481
0.0003
0.0009
0.0001
0.1221
0.0257

0.0299
0.0138
0.0299
0.0024
0.0047
0.0423
0.0004

0.1411
0.0399

14

Table 4
PLS results (106 events) Non-nancial companies.
Variables

Non-nancial sample
Bank bailouts

INTERCEPT
L1A
L3A
TOTNFVA
L1L
L3L
TOTNFVL
Number of companies
Number of events
***

Capital Injections

Distress

PLS

PLS cross-val

PLS

PLS cross-val

PLS

PLS cross-val

0.0998
0.0062
0.0369
0.0085
0.0052
0.0542
0.0031
174
26

0.0785
0.0058
0.0411
0.0082
0.0044
0.0623
0.0022

0.0425
0.00778
0.0498
0.0019
0.0447
0.0060
0.0016

0.0399
0.00699
0.0509
0.0015
0.0512
0.0057
0.0018

0.0528
0.0333
0.0625
0.0011
0.0057
0.0396
0.0025

0.0579
0.0315
0.0633
0.0013
0.0063
0.0411
0.0020

VIP > 1.5, **1 < VIP > 1.5 *0.80 < VIP > 1.

66

14

336

O. Roggi, A. Giannozzi / Journal of Banking & Finance 58 (2015) 327342

(0.0336) on investors reactions, while the coefcient of Level 3

Assets is not signicant. Both OLS and xed effect models with or
without control variables exhibit this relationship. When a
Distress Event occurred, non-nancial companies with greater percentages of liquid assets had higher cumulative abnormal returns.
This result is also conrmed by PLS regression, showing a positive
and signicant coefcient for Level 1 Assets (0.0315).
Concerning the impact of liabilities on stock prices, regression
with xed effects demonstrated a negative and signicant coefcient for Level 3 liabilities (0.0663). PLS regressions conrmed
that Level 3 liabilities exert a negative and signicant effect on
investors reactions (0.0411). Liquidity-constraining events
decrease rms ability to rollover their short-term liabilities due
to the renancing risk, thus harming investors. Consistent with
the ndings of Lev and Zhou (2009) based on US data, our results
demonstrate that investors reactions were driven by rollover concerns, whereas credit risk concerns do not seem to be relevant for
equity investors. When liquidity-constraining events occur, companies with greater illiquid liabilities showed worse cumulative
abnormal returns.
Of the control variables, only nancial leverage exerts a negative and signicant effect (0.0201) on investors reactions to
Distress events: non-nancial companies with greater leverage
had lower abnormal returns.
7.1.2. Capital injections group
We expected investors to react negatively to Level 1 assets and
to not react (or to only mildly and positively react) to level 3 fair
value assets (Proposition 1-bis). The results of the analysis where
the stock prices were regressed against the levels of fair valued
assets and liabilities conrm our hypothesis. Both OLS and xed
effects models demonstrate that Level 1 Assets has a negative
and signicant effect (0.0047) on investors reactions. In addition, Level 3 Assets show a positive and signicant coefcient
(0.0454) in all regression runs.
PLS regressions conrmed these ndings with greater signicance: Level 1 Assets exert a negative and signicant effect
(0.00699) on investors reactions whereas Level 3 Assets show
a positive and signicant coefcient (0.0509).
When liquidity-expanding events occur, non-nancial companies with greater percentages of liquid assets (L1A) on their balance sheets showed lower positive CARs. In contrast, companies
with greater percentages of illiquid assets (L3A) benet from
higher positive CARs.

Fair-valued liabilities did not affect investors reactions.

Similarly, to Lev and Zhou (2009), this result suggests that investors did not perceive a signicant risk of renancing resulting from
liquidity-expanding events. Investors do not react to Level 3 fair
value liabilities.
Of the control variables, only the return on assets (ROA) has a
positive and highly signicant effect (0.0783) on stock prices.
When liquidity-expanding events occur, non-nancial companies
with a higher ROA had higher abnormal returns. Financial leverage
does not exert signicant effect on investors reactions.
7.1.3. Bank bailouts group
As expected, bank bailouts do not have a signicant impact on
the stock returns of non-nancial rms. In fact, there were no signicant coefcients except in the case of the level 1 Liabilities
(L1L), which has a negative and signicant coefcient (0.0021)
in the xed effects model. Of the control variables, leverage (LEV)
shows a signicant and negative coefcient.
7.2. The results for nancial rms
For the nancial sample, we ran 18 regressions using the
cross-sectional OLS/xed effect models and 6 PLS regressions.
The results of the empirical analysis for the nancial companies
are presented in the following tables.
7.2.1. Distress group
The xed effects model (with and without control variables),
investigating the relationship between investors reactions and fair
value asset and liabilities reveals that L3A exerts a negative and
signicant effect (0.0022) on investors reaction. PLS regressions
showed negative and signicant coefcients both for Level 3 Assets
(0.0234) and for Level 1 Assets (0.0029). In addition, the coefcient for L3A is greater than the coefcient for L1A.
The investors reactions were more negative for those banks
with greater percentages of illiquid assets (L3A). High liquidity risk
nancial companies had worse negative abnormal returns than
those with a low degree of liquidity risk. Based on these results,
it is possible to conrm proposition 2: when liquidity-contracting
events occur, investors react more negatively to level 3 assets than
to level 1 assets. Financial leverage did not exert a signicant effect
on investors reactions.
On the liabilities side of nancial companies, the results show
that the Level 1 and 3 fair-valued liabilities are not signicant.

Table 5
OLS with control variables (106 crisis events) Financial companies.
Variables

Financial sample
Bank bailouts

INTERCEPT
L1A
L3A
TOTNFA
L1L
L3L
TOTNFVL
ROA
OPACT
LEV
HC RECLASS
HCA RECLASS
F stat
Adj R2
Number of companies
Number of events

Capital Injections

Distress

OLS

OLS with HC

OLS + F.E.

OLS

OLS with HC

OLS + F.E.

OLS

OLS with HC

OLS + F.E.

0.0087
0.00142
0.0366
0.0003
0.0132
0.0796
0.0041
0.0936
0.0359
0.079

0.4151
0.1596
59
26

0.00781
0.00215
0.0374
0.0004
0.01235
0.0777
0.0039
0.0841
0.0478
0.081
0.0012
0.0063
0.4012
0.2033

0.0093
0.0028
0.0399
0.0005
0.0163
0.0685
0.0055
0.0811
0.0632
0.0799

0.3966
0.2266

0.0510
0.0069
0.0741
0.0039
0.0068
0.0497
0.0001
0.0098
0.0009
0.0079

0.3467
0.2568

0.0741
0.0081
0.0562
0.0029
0.00674
0.0579
0.0002
0.0111
0.0008
0.0066
0.0063
0.0014
0.4156
0.2965

0.0766
0.0082
0.0763
0.0031
0.0073
0.0631
0.0003
0.0123
0.0006
0.0071

0.5124
0.3126

0.0029
0.0017
0.0016
0.0001
0.0007
0.0028
0.00001
0.0007
0.0055
0.0074

0.0551
0.0496

0.0031
0.0025
0.0024
0.0005
0.0006
0.0033
0.0003
0.0002
0.0061
0.0061
0.0002
0.0008
0.0631
0.0632

0.0030
0.0027
0.0022
0.0006
0.0009
0.0035
0.0006
0.0005
0.0065
0.0063

0.0515
0.0621

66

14

337

O. Roggi, A. Giannozzi / Journal of Banking & Finance 58 (2015) 327342

Table 6
OLS without control variables (106 events) Financial companies.
Variables

Financial sample
Bank bailouts

INTERCEPT
L1A
L3A
TOTNFA
L1L
L3L
TOTNFVL
HC RECLASS
HCA RECLASS
F stat
Adj R2
Number of companies
Number of events

Capital Injections

Distress

OLS

OLS with HC

OLS + F.E.

OLS

OLS with HC

OLS + F.E.

OLS

OLS with HC

OLS + F.E.

0.0211
0.0022
0.0298
0.0006
0.0121
0.0966
0.0519

0.4010
0.1863
59
26

0.0252
0.0015
0.0278
0.0005
0.0020
0.0874
0.0599
0.0009
0.0005
0.3963
0.2036

0.0186
0.0024
0.0226
0.0008
0.0155
0.0813
0.00482

0.4682
0.2190

0.0599
0.0139
0.0942
0.0002
0.0413
0.0603
0.0009

0.3487
0.1958

0.0475
0.0187
0.0866
0.0002
0.0493
0.0582
0.0008
0.0005
0.0007
0.5236
0.2741

0.0519
0.0193
0.0790
0.0001
0.0519
0.0653
0.0010

0.4139
0.1963

0.0198
0.0041
0.0173
0.0002
0.0009
0.0266
0.0005

0.0761
0.0526

0.01753
0.0019
0.0196
0.0001
0.0010
0.0043
0.0006
0.0002
0.0001
0.0298
0.00552

0.0215
0.0026
0.0210
0.0003
0.0006
0.0058
0.0002

0.0396
0.0541

66

14

Table 7
PLS results (106 events) Financial companies.
Variables

Financial sample
Bank bailouts

INTERCEPT
L1A
L3A
TOTNFVA
L1L
L3L
TOTNFVL
Number of companies
Number of events
***

**

Capital Injections

Distress

PLS

PLS cross-val

PLS

PLS cross-val

PLS

PLS cross-val

0.0621
0.0029
0.0511
0.0002
0.0191
0.0698
0.0058
59
26

0.0650
0.0026
0.0558
0.0003
0.0026
0.0741
0.0061

0.0079
0.0099
0.0598
0.0049
0.0079
0.0690
0.0002

0.075
0.0091
0.0612
0.0042
0.0077
0.0631
0.0001

0.0298
0.0027
0.0211
0.0002
0.0066
0.0091
0.0001

0.0317
0.0029
0.0234
0.0001
0.0052
0.0093
0.0001

66

14

VIP > 1.5, 1 < VIP > 1.5 0.80 < VIP > 1.

We expected that investors would not react to the liquid level 1

fair value liabilities (as they mainly consist of insured items) and
that they would react in a negative way to illiquid Level 3
Liabilities in liquidity-constraining events (proposition 3). We are
not able to conrm proposition #3. Level 3 liabilities did not exert
signicant effects on investors reactions to liquidity-constraining
events.

7.2.2. Capital Injections group

For liquidity-expanding events, we expected investors to react
more positively to Level 3 fair value assets than to Level 1 assets
(Proposition 2-bis). On the liabilities side, we expected investors
not to react to Level 1 liabilities and to react in a positive way to
level 3 liabilities (Proposition 3). The results of the analysis where
the stock price was regressed against the levels of fair valued assets
and liabilities conrm our hypothesis. The xed effects model
(with and without control variables) shows a positive coefcient
of Level 3 Assets (0.0763) signicant and larger than the coefcients of Level 1 (0.0082). This result is also conrmed by PLS
regression and appears stable in all regression runs.
Level 1 liabilities is not signicant, whereas the coefcient for
level 3 Liabilities is positive and signicant in all regressions.
Of the control variables, only the return on assets (ROA) is signicant, with a negative coefcient: the cumulative abnormal
return was higher for less protable banks.
The investors reactions are conditional to the fair value hierarchy. Financial companies with greater amounts of Level 3 Assets
and Level 3 liabilities on their balance sheets had higher CARs.

High liquidity risk nancial rms benet more from capital

injections.

7.2.3. Bank Bailouts group

Despite the liquidity-expanding nature of the Bank Bailouts
group, and similarly to Pennathur et al. (2014) and Veronesi and
Zingales (2010), studies based on US data, we documented an average negative reaction to these types of events. As a consequence, in
contrast with the ndings of Lev and Zhou (2009), we expected
investors to react more negatively to Level 3 fair value assets than
to Level 1 and more negatively towards nancial companies with
higher leverage (Proposition 4). The results of the analysis where
the stock price was regressed against the levels of fair valued assets
and liabilities conrm our hypothesis.
The cross-sectional regression and xed effects model show a
negative coefcient of the Level 3 Asset (0.0399),15 signicant
at 95%. Moreover, the coefcient of Level 3 Assets is larger than
the coefcient of Level 1 Assets in all regression runs. Investors
reacted more negatively to banks with a high degree of liquidity risk
(high amounts of L3A).
On the liabilities side of nancial rms, illiquid Level 3 liabilities
(L3L) show a negative and signicant coefcient (0.0685).16 We
found the investors reactions to be more negative for those nancial
companies that record greater percentages of illiquid and uninsured
liabilities. The Level 3 Assets and Level 3 liabilities are signicant in
all regression runs.
15
16

The coefcient pertains to the xed effects model with control variables.
The coefcient pertains to the xed effects model with control variables.

338

O. Roggi, A. Giannozzi / Journal of Banking & Finance 58 (2015) 327342

Of the control variables, protability (ROA) exerts a positive and

signicant effect on investors reactions, whereas nancial leverage
(LEV) shows a signicant negative coefcient. The investors
reacted less negatively to bank bailouts for nancial rms with
higher ROA and more negatively for companies with greater
leverage.
PLS regressions conrmed that L1A (0.0558) and L3A
(0.0741) exert a negative and signicant effect on investors
reactions.
This leads us to conclude and conrm proposition 4: when bank
bailouts occur, investors react more negatively to Level 3 fair value
assets than to Level 1 assets, and more negatively towards nancial
companies with greater leverage. In addition, differently from Lev
and Zhou (2009), we found that Level 3 liabilities exert a negative
effect on investors reaction to bank bailouts, whereas ROA exhibits
a positive effect.
These results could be explained by the negative perception of
the risk of nationalization, which is strongly felt for nancial
companies with greater amounts of illiquid Level 3 assets, greater
leverage, and greater amounts of Level 3 liabilities and lower ROA.
The risk of being the next to be saved and then undergo governmental interference seems to be negative news for equity
investors.
Our ndings can be summarized as follows: for non-nancial
companies, when liquidity-constraining events occur, investors
reaction to L1A is positive, whereas L3A did not exert any effect
on investors reactions. In addition, investors reactions were
affected by the rollover risk related to the level 3 Liabilities (L3L)
during liquidity-constraining events. When liquidity-expanding
events occur, investors react positively to L3A and negatively to
L1A.
For nancial companies, when liquidity-expanding events
occur, investors react more positively to L3A than to L1A. When
liquidity-constraining events occur, investors react more negatively to L3A than to L1A. When bank bailouts occur, investors
react more negatively to L3A than to L1A, negatively to L3L, and
more negatively to nancial companies with greater leverage.

8. Conclusions and further considerations

This study investigated investors reactions to crisis events during the period from February 17, 2008 to June 22, 2010 based on
the fair value disclosures of 313 European nancial and
non-nancial rms. We rst identied 106 crisis events and classied them into three groups of different natures. We then examined
whether the three level fair value information disclosed by
non-nancial and nancial rms affects investors reaction to the
crisis events by regressing the event group returns on the levels
of fair value assets and liabilities, along with control variables.
Using xed effects models and a non-parametric methodology
(PLS), we demonstrated a complex reaction to the liquidity risk
information conveyed by the fair value hierarchy. The fair value
separation of three liquidity levels affects investors reactions to
crisis events.
For nancial companies, high liquidity risk (high percentage of
level 3 assets), level 3 assets exert a negative effect on investors
reactions to liquidity-constraining events and a positive effect on
liquidity-expanding events. Investors react more negatively to

the most illiquid banks in response to Distress/liquidity-constraini

ng Events and more positively in response to Capital Injections.
For non-nancial companies, having low liquidity risk (low percentage of level 3 assets), Level 1 assets exert a positive effect on
investors reactions to liquidity-constraining events and a negative
effect on reactions to liquidity-increasing events. In addition, level
3 assets exert a positive effect on investors reactions to Capital
Injections. Non-nancial companies with higher illiquid liabilities
had worse cumulative abnormal returns in response to Distress
Events due to the rollover risk perceived by investors. These results
for European companies are consistent with the study by Lev and
Zhou (2009) based on US companies, in explaining investors reactions to Distress/Liquidity-constraining Events and to Capital
Injections. Our ndings concerning bank bailout events differ from
the previous literature: we demonstrated that investors react more
negatively for the most illiquid nancial rms (companies with
greater percentages of level 3 assets on their balance sheets)
affected by high leverage and low protability (ROA).
These ndings have several implications for both regulators and
equity investors in reference to asset allocation. First, as was previously observed in US markets under SFAS 157, we conrm that fair
value hierarchy under IFRS 7 is useful for investors, providing
information on the liquidity risk of European nancial and
non-nancial companies. This result contributes to the on-going
debate about Fair Value Accounting vs. Historical Cost
Accounting, documenting an important attribute of fair value valuations in European markets. In addition, we documented that
excluding potential contagion effects not considered in this study,
bank rescues engineered by major governments were not benecial for equity investors over the short term.
Second, the relationships found in this study among fair value
assets/liabilities and stock prices could be useful for equity investors or investment funds, enabling them to create a specic
short-term asset allocation strategy in the event of a crisis. For
example, according to our results, when a bank bailout is
announced investors should expect a negative short run reaction
for higher liquidity risk banks affected by a high degree of nancial
leverage, and they should avoid investing in these types of companies. When an interest rate cut is announced, investors could benet from higher positive abnormal returns by investing in higher
liquidity risk banks (banks with higher amounts of level 3 assets).
Our ndings provide investors with useful asset allocation information to respond to policy/monetary interventions, bailouts or
company distress in the European nancial markets.
Our ndings could be affected by several constraints, such as
the number of observations, the process of aggregation into groups
of events and the approach used to estimate abnormal returns.
Despite these potential limitations, this work answers the initial
research question, demonstrating that rms liquidity risk as conveyed by the fair value hierarchy affected investment decisions
during the crisis. We believe that the limitations described above
are not likely to affect the results, as the scope of our empirical
analysis is the identication of the relationship (and not the creation of a forecast model) between stock returns and the levels
of fair value disclosure. Additional studies based on larger databases will support or refute our ndings.
Appendix A
See Tables A1A5

339

O. Roggi, A. Giannozzi / Journal of Banking & Finance 58 (2015) 327342

Table A1
List of companies in the data sample.
Company

Sector

Company

Sector

Company

Sector

Company

Sector

AEGON GROUP
AIR LIQUID
ALCATEL LUCENT
ALLIANZ
ARCELORMITTAL
ASS. GENERALI
AXA
BANCO SANTANDER
BASF
BAYER
BBVA
BNP PARIBAS
CARREFOUR
CREDIT AGRICOLE
DAIMLER
DEUTSCHE BANK
DEUTSCHE BORSE
DEUTSCHE TELEKOM
E ON
ENEL
ENI
FRANCE TELECOM
DANONE
SOCIETE GENEREALE
IBERDROLA
ING GROUP
INTESA SANPAOLO
LOREAL
LVMH
MUNICH RE
NOKIA
PHILIPS
RENAULT
REPSOL YPF
RWE
SAINT GOBAIN
SANOFI AVENTIS
SAP
SCHNEIDER ELECTRIC
SIEMENS
GDF SUEZ
TELECOM ITALIA
TELEFONICA
TOTAL
UNICREDIT GROUP
UNILEVER
VINCI
VIVENDI
VOLKSWAGEN
UNIBAIL-RODAMCO
CRH
ALSTOM
ANHEUSER BUSH
A2A
ABERTIS INFR.
ACCOR
ACERINOX
ACKERM. & VAN HAAREN
ACS
ADIDAS
ADP
AGEAS
AHOLD
AIR FRANCE-KLM
AIXTRON
AKZO NOBEL
ALLIED IRISH BANKS
ALPHA BANK
ANDRITZ
ARKEMA
ASML HLDG
ATLANTIA
ATOS ORIGIN
ATRIUM EUR.RE

F
NF
NF
F
NF
F
F
F
NF
NF
F
F
NF
F
NF
F
F
NF
NF
NF
NF
NF
NF
F
NF
F
F
NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
F
NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
F
NF
NF
NF
F
NF
NF
NF
NF
F
F
NF
NF
NF
NF
NF
NF

BCA CARIGE
BCA MPS
BCA POP. DI MILANO
BCA POP. DI SONDRIO
BCA POP. EMIL. ROMAGNA
BCO COMERCIAL PORTUG.
BCO DE VALENCIA
BCO ESPIRITO SANTO
BCO POPOLARE
BCO POPULAR ESPANOL
BCO SABADELL
BEIERSDORF
BEKAERT
BELGACOM
BIC
BILFINGER BERGER
BIOMERIEUX
BMW
BOLSAS Y MERCADOS ESP.
BOSKALIS WESTMINSTER
BOURBON
BOUYGUES
BRISA
BUREAU VERITAS
BWIN INT. ENT.
C&C GRP
CHRISTIAN DIOR
CIMPOR
CNP ASSURANCES
COCA-COLA HBC
COFINIMMO
COMMERZBANK
COMP NAT. A PORTEFEU
CONTINENTAL
CORIO
CAP GEMINI
CASINO GUICHARD
CATTOLICA ASS.
CELESIO
CGG VERITAS
CREDITO VALTELLINES
CRITERIA CAIXACORP
CRUCELL
CSM
DASSAULT SYSTEMS
DCC
DELHAIZE GRP
DELTA LLOYD
DEUTSCHE POST
DEUTSCHE POSTBANK
DEXIA
DRAGON OIL PLC
EADS
EBRO FOODS
EDENRED
EDF
EDP
EDP RENOVAVEIS
EFG EUROBANK ERGASIAS
EIFFAGE
ELAN CORPORATION
ELISA CORPORATION
ENAGAS
ENDESA
ERAMET
ERSTE GROUP BANK
ESSILOR INTERNATIONAL
ETS COLRUYT
EURAZEO
EUTELSAT COMMUNIC.
FERROVIAL
FIAT
FINMECCANICA
FOM. DE CONSTR. CONTRA

F
F
F
F
F
F
F
F
F
F
F
NF
NF
NF
NF
NF
NF
NF
F
NF
NF
NF
NF
NF
NF
NF
NF
NF
F
NF
NF
F
NF
NF
NF
NF
NF
F
NF
NF
F
F
NF
NF
NF
NF
NF
F
NF
F
F
NF
NF
NF
NF
NF
NF
NF
F
NF
NF
NF
NF
NF
NF
F
NF
NF
F
NF
NF
NF
NF
NF

FRESENIUS PREF
FUGRO
GALP ENERGIA
GAMESA
GAS NATURAL SDG
GRP BRUX. LAMBERT
GEA GRP
GECINA
GEMALTO
GESTEVISION TELECINCO
GRIFOLS
GRP EUROTUNNEL
GRUPO ACCIONA
HANNOVER RUECK
HEIDELBERGCEMENT
HEINEKEN
HEINEKEN HLDG
HENKEL PREF
HERMES INTERNATIONAL
HOCHTIEF
IBERDROLA REN.
IBERIA
ICADE
ILIAD
IMERYS
IMMOFINANZ
IMTECH
INDITEX
INDRA SISTEMAS
INFINEON TECHNOLOGIES
JCDECAUX
JERONIMO MARTINS
K+S
KBC GRP
KEMIRA
KERRY GRP
KESKO
KLEPIERRE
KLOECKNER & CO
KONE B
KONECRANES
KONINKLIJKE DSM
KPN
LAFARGE
LAGARDERE
LANXESS
LEGRAND
LINDE
LUFTHANSA
LUXOTTICA
M6 METROPOLE TV
MAN
MAPFRE
MARFIN INV. GRP
MAUREL ET PROM
MEDIASET
MEDIOBANCA
MERCK
METRO
METSO
MICHELIN
MOBISTAR
MTU AERO ENGINES HLDG
NAT. BANK OF GREECE
NATIXIS
NEOPOST
NESTE OIL
NEXANS
NOKIAN RENKAAT
NUTRECO
OMV
OPAP
ORION B
OTE

NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
F
NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
F
NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
F
F
NF
NF
F
NF
NF
NF
NF
NF
NF
F
F
NF
NF
NF
NF
NF
NF
NF
NF
NF

PERNOD RICARD
PEUGEOT
PIRELLI & C.
POHJOLA BANK
PORSCHE PREF
PORTUGAL TELECOM
PPR
PRYSMIAN
PUBLIC POWER CORP.
PUBLICIS GRP
PUMA
QIAGEN
RAIFFEISEN INT. BANK
RANDSTAD
RAUTARUUKKI K
RED ELECTRICA CORP.
REED ELSEVIER NV
RHEINMETALL
RHODIA
RHOEN KLINIKUM
RYANAIR
SAFRAN
SAIPEM
SALZGITTER
SAMPO
SANOMA
SBM OFFSHORE
SCOR
SEB
SES
SGL CARBON
SNAM RETE GAS
SODEXO
SOFINA
SOFTWARE
SOLARWORLD
SOLVAY
STADA ARZNEIMITTEL
STMICROELECTRONICS
STORA ENSO R
SUEDZUCKER
SUEZ ENVIRONNEMENT
SYMRISE
TECHNIP
TECNICAS REUNIDAS
TELEKOM AUSTRIA
TELENET GRP HLDG
TELEPERFORMANCE
TENARIS
TERNA
TF1
THALES
THYSSENKRUPP
TITAN CEMENT
TNT
TOGNUM
UBI BCA
UCB
UMICORE
UNITED INTERNET
UPM KYMMENE
VALEO
VALLOUREC
VEOLIA ENV.
VERBUND
VIENNA INSURANCE
VOESTALPINE
VOPAK
WACKER CHEMIE
WARTSILA
WENDEL
WERELDHAVE
WIENERBERGER
WINCOR NIXDORF

NF
NF
NF
F
NF
NF
NF
NF
NF
NF
NF
NF
F
NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
F
NF
NF
F
NF
NF
NF
NF
NF
F
NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
NF
F
NF
NF
NF
NF
NF
NF
NF
NF
F
NF
NF
NF
NF
F
NF
NF
NF

(continued on next page)

340

O. Roggi, A. Giannozzi / Journal of Banking & Finance 58 (2015) 327342

Table A1 (continued)
Company

Sector

Company

Sector

Company

Sector

Company

Sector

AURUBIS
AZIMUT HLDG
BANK OF IRELAND
BANKINTER

NF
F
F
F

FONCIERE DES REGIONS

FORTUM
FRAPORT
FRESENIUS MEDIC. CARE

NF
NF
NF
NF

OUTOKUMPU
OUTOTEC
PAGESJAUNES
PARMALAT

NF
NF
NF
NF

WOLTERS KLUWER
YIT
ZARDOYA OTIS
ZODIAC AEROSPACE

NF
NF
NF
NF

Table A2
List of variables.
Label

Variable

L1A
L2A
L3A
TOTFVA
TOTNFVA
L1L
L2L
L3L
TOTFVL
TOTNFVL
ROA
OPACT
LEV
HC RECLASS
HCA RECLASS

Level 1 assets/Total fair value assets

Level 2 assets/Total fair value assets
Level 3 assets/Total fair value assets
Total fair value assets/Total assets
Total non-fair value assets /Total assets
Level 1 liabilities/Total fair value liabilities
Level 2 liabilities/Total fair value liabilities
Level 3 liabilities/Total fair value liabilities
Total fair value liabilities/Total liabilities
Total non-fair value liabilities/Total liabilities
Return on Assets
Cash ows generated by operating activities/total assets
Leverage; (Long-term debt + debt in current liabilities)/Total assets
Dummy (1 = reclassication of specic derivatives at historical cost as permitted by IASB on October 2008; 0 = no reclassication at historical cost)
Amount of assets reclassied at historical value as permitted by IASB/Total assets

Table A3
Fair value assets/liabilities as a percentage of Total Assets/Liabilities for nancial companies vs. non-nancial companies (Full sample).
Financial companies

Level 1 Assets/Total assets

Level 2 Assets/Total assets
Level 3 Assets/Total assets
TOT Fair value assets/Total assets
TOT Non-fair value assets/Total assets
Level 1 liabilities/Total liabilities
Level 2 liabilities/Total liabilities
Level 3 liabilities/Total liabilities
TOT Fair value liabilities/Total liabilities
TOT Non-fair value liabilities/Total liabilities
ROA
OPACT
LEV
Number of companies

Non-nancial companies

2007

2008

2009

2007

2008

2009

0.344
0.193
0.014
0.551
0.449
0.109
0.162
0.011
0.289
0.711
0.02
0.025
0.95
59

0.014
0.015
0.002
0.031
0.969
0.005
0.013
0.002
0.021
0.979
0.01
0.04
0.89

0.010
0.012
0.001
0.023
0.977
0.002
0.011
0.001
0.014
0.986
0.01
0.007
0.85

0.063
0.070
0.039
0.172
0.828
0.073
0.166
0.068
0.306
0.694
0.09
0.105
0.62
254

0.104
0.072
0.015
0.191
0.809
0.006
0.008
0.003
0.018
0.982
0.08
0.061
0.67

0.028
0.021
0.002
0.051
0.949
0.006
0.021
0.004
0.031
0.969
0.04
0.112
0.62

Table A4
Fair value hierarchy in the nancial and non-nancial companies (Fair value sample).
Financial companies
(Mean)

L1A
L2A
L3A
L1L
L2L
L3L
Number ofcompanies

Non-nancial companies
(Mean)

2007

2008

2009

2007

2008

2009

0.6244
0.3508
0.0248
0.3872
0.5723
0.0404
59

0.4594
0.4774
0.0632
0.2632
0.649
0.0878

0.4326
0.5225
0.045
0.1484
0.8016
0.0499

0.935
0.05
0.015
0.35
0.45
0.2
174

0.871
0.072
0.057
0.21
0.47
0.32

0.941
0.021
0.038
0.2
0.42
0.38

341

O. Roggi, A. Giannozzi / Journal of Banking & Finance 58 (2015) 327342

Table A5
Summary of the events, dates and observation windows.
ID

Date of the event

Observation window

ID

Date of the event

Observation window

ID

Date of the event

Observation window

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37

02/17/08
03/16/08
06/03/08
07/04/08
07/11/08
07/13/08
09/07/08
09/15/08
09/16/08
09/18/08
09/19/08
09/22/08
09/25/08
09/29/08
09/30/08
10/03/08
10/06/08
10/07/08
10/08/08
10/09/08
10/10/08
10/13/08
10/14/08
10/15/08
10/16/08
10/17/08
10/19/08
10/21/08
10/24/08
10/28/08
10/29/08
10/30/08
11/03/08
11/05/08
11/09/08
11/10/08
11/12/08

02/18
03/17
06/03-04
07/04-07
07/11-14
07/14
09/08-09
09/15-16
09/16-17
09/18-19
09/19-22
09/22-23
09/25-26
09/29-30
09/30-10/01
10/03-06
10/06-07
10/07-08
10/08-09
10/09-10
10/10-13
10/13-14
10/14-15
10/15-16
10/16-17
10/17-20
10/20
10/21-22
10/24-27
10/28-29
10/29-30
10/30-31
11/03-04
11/05-06
11/10
11/10-11
11/12-13

38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74

11/14/08
11/20/08
11/21/08
11/24/08
11/25/08
11/26/08
12/02/08
12/03/08
12/05/08
12/09/08
12/12/08
12/15/08
12/16/08
12/19/08
12/23/08
12/29/08
12/31/08
01/08/09
01/15/09
01/16/09
01/19/09
01/26/09
02/02/09
02/05/09
02/10/09
02/18/09
02/25/09
02/26/09
03/02/09
03/03/09
03/05/09
03/07/09
03/09/09
03/18/09
03/20/09
03/25/09
03/29/09

11/14-17
11/20-21
11/21-24
11/24-25
11/25-26
11/26-27
12/02-03
12/03-04
12/05-08
12/09-10
12/12-15
12/15-16
12/16-17
12/19-22
12/23-24
12/29-30
12/31-01/01
01/08-09
01/15-16
01/16-19
01/19-20
01/26-27
02/02-03
02/05-06
01/10-11
02/18-19
02/25-26
02/26-27
03/02-03
03/03-04
03/05-06
03/09
03/09-10
03/18-19
03/20-23
03/25-26
03/30

75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106

04/01/09
04/02/09
04/06/09
04/09/09
04/14/09
04/17/09
04/21/09
05/06/09
05/14/09
05/20/09
05/21/09
05/28/09
06/02/09
06/10/09
06/17/09
06/26/09
06/29/09
07/02/09
07/08/09
07/20/09
08/06/09
09/18/09
09/23/09
10/05/09
12/17/09
04/11/10
05/02/10
05/03/10
05/10/10
05/18/10
06/07/10
06/22/10

04/01-02
04/02-03
04/06-07
04/09-10
04/14-15
04/17-20
04/21-22
05/06-07
05/14-15
05/20-21
05/21-22
05/28-29
06/02-03
06/10-11
0617-18
06/26-29
06/29-30
07/02-03
07/08-09
07/20-21
08/06-07
09/18-21
09/23-24
10/05-06
12/17-18
04/12
05/02-03
05/03-04
05/10-11
05/18-19
06/07-08
06/22-23

References
Allen, F., Carletti, E., 2007. Mark-to-market accounting and liquidity pricing. Journal
of Accounting and Economics 45 (1), 358378.
Allen, F., Carletti, E., 2008. Should nancial institutions mark to market? Banque de
France Financial Stability Review, October.
Bank for International Settlements 2009. 79th Annual Report. Available on www.
bis.org.
Barlev, B., Haddad, J.R., 2003. Fair value accounting and the management of the rm.
Critical Perspectives on Accounting 14 (1), 383415.
Barth, M.E., 1994. Fair value accounting: evidence from investment securities and
the market valuation of banks. The Accounting Review 69, 125, January.
Barth, M.E., Landsman, W.R., Wahlen, J.M., 1995. Fair value accounting: effects on
banks earnings volatility, regulatory capital, and value of contractual cash
ows. Journal of Banking and Finance 19, 577605.
Barth, M.E., Beaver, W.H., Landsman, W.R., 1996. Value-relevance of banks fair
value disclosures under SFAS No. 107. Accounting Review 1, 513537.
Barth, M.E., Beaver, W.H., Landsman, W.R., 2001. The relevance of the value
relevance literature for nancial accounting standard setting: another view.
Journal of Accounting and Economics 31, 77104.
Beaver, W.H., Landsman, W.R., 1983. Incremental information content of Statement
33 disclosures. FASB: Stamford, Connecticut.
Beaver, W.H., Ryan, S., 1985. How well do Statement No. 33 earnings explain stock
returns? Journal of Financial Analysts 41, 6671.
Bernard, V.L., Ruland, R., 1987. The incremental information content of historical
cost and current cost income numbers: time series analyses for 1962-1980. The
Accounting Review 62, 707722.
Bomn, A.N., 2003. Pre-announcement effects, news effects, and volatility:
monetary policy and the stock market. Journal of Banking & Finance 27, 133
151.
Brunnermeir, M.K., Pedersen, L.H., 2009. Market liquidity and funding liquidity. The
Review of Financial Studies 22 (6), 22012238.
Cornett, M.M., Rezaee, Z., Tehranian, H., 1996. An investigation of capital market
reactions to pronouncements on fair value accounting. Journal of Accounting
and Economics 22, 119154.
Cubillas, E., Fonseca, A.R., Gonzlez, F., 2012. Banking crises and market discipline:
International evidence. Journal of Banking & Finance 36, 22852298.

Demirg-Kunt, A., Huizinga, H., 2004. Market discipline and deposit insurance.
Journal of Monetary Economics 51, 375399.
Eccher, E.A., Ramesh, K., Thiagarajan, S.R., 1996. Fair value disclosures by bank
holding companies. Journal of Accounting and Economics 22, 79117.
Elyasiani, E., Mester, L., Pagano, M., 2011. Large capital infusions, investor reactions,
and the return and risk performance of nancial institutions over the business
cycle and recent nancial crisis. Research Department, Federal Reserve Bank of
Philadelphia, working paper.
Fiordelisi, F., Galloppo, G., Ricci, O., 2014. The effect of monetary policy
interventions on stock markets and G-SIFIs during the crisis. Journal of
Financial Stability 11, 4961.
Flannery, M.J., Nikolova, S., 2004. Market discipline of U.S. nancial rms: Recent
evidence and research issues. In: Hunter, W.C., Kaufman, G.G., Borio, C.,
Tsatsaronis, K. (Eds.), Market Discipline across Countries and Industries. MIT
Press.
Freixas, X., Tsomocos, D., 2004. Book vs. fair value accounting in banking, and
intertemporal smoothing. Oxford Financial Research Centre, working paper.
Goh, B.W., Ng, J., Yong, K.O., 2009. Voluntary Fair Value Disclosures Beyond SFAS
157s Three-level Estimates. doi: 10.2139/ssrn.1335848.
Holthausen, R.W., Watts, R.L., 2001. The relevance of the value-relevance literature
for nancial accounting standard setting. Journal of Accounting and Economics
18, 375.
Honohan, P., Klingebiel, D., 2003. The scal cost implications of an accommodating
approach to banking crises. Journal of Banking and Finance 27, 15391560.
International Monetary Fund, 2008. Global nancial stability report. Chapter 3: Fair
value accounting and procyclicality, October.
Jun, C.H., Lee, S.H., Park, H.S., Lee, J.H., 2009. Use of partial least squares regression
for variable selection and quality prediction. International Conference On
Computers & Industrial Engineering, 13021307.
Kolev, K.S., 2008. Do investors perceive marking-to-model as marking-to-myth?
Early evidence from FAS 157. Disclosure. http://dx.doi.org/10.2139/
ssrn.1336368.
Laux, C., Leuz, C., 2009. The crisis of fair value accounting: making sense of the
recent debate. Accounting, Organizations and Society 34, 826834.
Lev, B., Zhou, N., 2009. Unintended Consequence: Fair Value Accounting Informs on
Liquidity Risk. http://dx.doi.org/10.2139/ssrn.1466009.
Martinez Peria, M.S., Schmukler, S.L., 2001. Do depositors punish banks for bad
behavior? Journal of Finance 56, 10291051.

342

O. Roggi, A. Giannozzi / Journal of Banking & Finance 58 (2015) 327342

Murdoch, B., 1986. The information content of FAS 33 returns on equity. The
Accounting Review 61, 273287. April.
Pennathur, A., Smith, D., Subrahmanyam, V., 2014. The stock market impact of
government interventions on nancial services industry groups: Evidence from
the 20072009 crisis. Journal of Economics and Business 71, 2244.
Plantin, G., Sapra, H., Shin, H.S., 2008. Marking-to-market: panacea or Pandoras
box? Journal of Accounting Research 46 (21), 435460.
Rangel, J.G., 2011. Macroeconomic news, announcements, and stock market jump
intensity dynamics. Journal of Banking & Finance 35, 12631276.

Semaan, E., Drake, P.P., 2011. Deregulation and risk. Financial Management 40, 295
329.
Song, C.J., Thomas, W.B., Yi, H., 2009. Value Relevance of FAS 157 fair value
hierarchy information and the impact of corporate governance mechanisms.
Accounting Review 85 (4).
Tenenhaus, M., 1998. La regression PLS: theorie et pratique. Editions TECHNIP.
Veronesi, P., Zingales, L., 2010. Paulsons gift. Journal of Financial Economics 97,
339368.