Research in Accounting Regulation 27 (2015) 119128

Contents lists available at ScienceDirect

Research in Accounting Regulation

j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / r a c r e g

Regular Paper

Contingency liabilities: The effect of three alternative

reporting styles
Bruce Lagrange a,*, Chantal Viger b, Asokan Anandarajan c
a
Universit du Qubec Rimouski, Rimouski, Qubec, Canada
b Universit du Qubec Montral, Montral, Qubec, Canada
c New Jersey Institute of Technology, USA

A R T I C L E I N F O A B S T R A C T

Article history: The International Accounting Standards Boards (IASB) new international nancial report-
Available online 9 October 2015 ing standards (IFRS) relating to contingencies became effective on January 1, 2011, ocially
replacing the CICAs (Canadian Institute of Chartered Accountants) contingent liability ac-
Keywords: counting standards for publicly accountable enterprises. Although both sets of standards
Contingent liabilities (IFRS and CICA) are based on fundamentally similar conceptual frameworks, they differ sig-
IAS 37
nicantly in certain respects. This study examines the changes now required in contingency
exposure draft IAS 37
reporting and their implications for regulators. Rules for contingency reporting were pre-
IFRS
International Accounting Standards Board viously dictated by Canadian GAAP (CGAAP), as formulated by the Canadian Institute of
Chartered Accountants, but are now subject to the IASBs IAS 37. However, to enhance clarity
and ease of understanding for nancial statement users, the IASB has proposed a new version
of contingent liability accounting standards under IFRS, titled exposure draft IAS 37. The
message conveyed by the three different types of reporting is investigated, with ndings
that have implications for other similar rules adopted by IASB. Results indicate variations
in four types of judgments by the Canadian loan ocers in the experiment. Although their
loan granting decisions were not inuenced by the change to IASBs IAS 37, the ocers
charged signicantly different interest premiums according to the type of nancial state-
ment received, i.e. based on former Canadian requirements, the original IAS 37 or the
proposed IAS 37 exposure draft. Loan ocers judgments are therefore inuenced by the
way contingent liabilities are presented, a nding that has implications for regulators, mainly
in view of the fact that the proposed IAS 37 reporting style could facilitate clarity and un-
derstanding of these liabilities.
2015 Elsevier Ltd. All rights reserved.

1 Objective If judgments are found to be affected, the results could be

This study examines whether the signals conveyed by
the application of different formatting requirements, espe-
cially for contingent liabilities, affect the judgments and
decisions of loan ocers. A comparison is made of reac-
tions to the requirements of the post International Financial
Reporting Standard (IFRS) regime and its predecessor, Ca-
nadian Generally Accepted Accounting Principles (CGAAP).
* Corresponding author. Tel.: +1 418 723 1986, ext. 1726.
E-mail address: bruce_lagrange@uqar.ca (B. Lagrange).
extrapolated to similar situations in which regulators con-
sider alternative forms of disclosure. Regarding contingent
liabilities, the former Canadian Institute of Chartered Ac-
countants (CICA) requirements (CICA section 3290) did not
stipulate separate recording in the balance sheet or disclo-
sure of the probability of occurrence. The potential liability
could therefore be camouaged by including it with other
liabilities. Under IFRS, both IAS 37 and the newly updated
IAS 37 exposure draft require that the events probability
(in this particular study, a lawsuit) be disclosed in the note
and the amount of the potential contingent liability pro-
vided separately in the balance sheet. The fundamental

http://dx.doi.org/10.1016/j.racreg.2015.09.003
1052-0457/ 2015 Elsevier Ltd. All rights reserved.
120 B. Lagrange et al. / Research in Accounting Regulation 27 (2015) 119128
difference between the IFRS standards stems from the tech-
nique for recognizing the loss. The current IAS 37 requires
that the amount recorded be based on a reliable estimate
of the gure most likely to be paid for settling the obliga-
tion at the closing date, whereas the proposed IAS 37
exposure draft stipulates that the liability should be re-
corded at the amount that would rationally be paid at the
end of the reporting period. As a result of the new method,
compared to the value reported under the original IAS 37
requirements (as well as CICA section 3290), the revised IAS
37 entails a change in the contingent liability and its at-
tendant loss value (and, under some circumstances, results
in an even lower value). Reported earnings could thus be
higher under the revised exposure draft.
The objective of this study is to explain how differ-
ences in contingent liability reporting inuence the
judgments of sophisticated users. This knowledge and the
information it provides on the effect of various reporting
formats have implications for regulators, especially IFRS ad-
ministrators. Although contingent liabilities are the focus
of the current investigation, the results can most likely be
extrapolated to other areas of nancial accounting for which
regulators are considering alternative forms of reporting.
This study uses an experimental design in which all vari-
ables are held constant and are subject to manipulation of
the reporting and disclosure formats for contingent liabili-
ties. The results obtained are mixed overall but nonetheless
show signicant variations in four types of judgments, es-
pecially the interest rate premium charged by the three loan
ocer groups (for which nancial statements were pro-
vided based on 1) the IAS 37 exposure draft; 2) current IAS
37; or 3) Canadian accounting standards). Specically, the
groups judgments and decisions are signicantly differ-
ent under the IAS 37 exposure draft, an observation
discussed in section 3. In the next section some back-
ground on the International Accounting Standards Board
(IASB) is provided to explain the current research objec-
tives and rationale.
2 Background and rationale
The International Accounting Standards Board (IASB)
works to develop a single set of high quality and compa-
rable reporting standards that will offer an improved basis
for decision making for business and investors (Canadian
Institute of Chartered Accountants, 2009, p. 1). It has also
been suggested that the IASB provides the best informa-
tion possible to support strategic and tactical decision
making (Canadian Institute of Chartered Accountants, 2009,
p. 1), with the result that it fosters better allocation of capital.
In addition, having common standards across nations is ex-
pected to contribute to a global profession for accountants
and reduce the burden on regulators.
The IASB attempts to establish a balance between rule-
based and principle-based standards. However, the road map
to convergence does not provide that different nations should
have the exact same standard, but rather that some differ-
ences are possible despite the standards being structured
more or less around the same principles. IFRS and CICA stan-
dards are both based on fundamentally similar conceptual
frameworks (with similar style and form), but differ sig-
nicantly in some ways.
One major difference is in the area of contingencies, since
January 1, 2011, when IFRS ocially replaced the CICA han-
dbooks accounting standards for publicly accountable
enterprises. As shown in Table 1, under Section 3290 of the
CICA Handbook, a contingent liability must be recorded if
(1) the event causing the loss is likely to occur (high prob-
ability of occurrence), and (2) the amount of the contingent
loss can be appraised with a degree of certainty. The amount
to be recorded should be based on a reasonable estimate of
the most likely amount to be paid. This recorded liability
is presented with other liabilities (i.e., not separately) in the
balance sheet. In contrast, under current IAS 37, contin-
gent losses are recorded as liabilities when (1) there is a
current obligation, and (2) the outow of resources em-
bodying economic benets is probable (i.e. more probable
than improbable); this occurs when the likelihood of such
losses exceeds 50%, which is a probability threshold lower
than that recommended by CICAs section 3290), and (3)
there is a reliable estimate. The amount most likely to bring
about the elimination of the obligation at the closing date
must be recorded (Chlala, Lavigne, & Vendette, 2009). The
amount under IAS 37 is, in most instances, similar to the
amount recorded under the previous CICAs section, but
the liability under IAS 37 is presented as a separate item
in the balance sheet, a departure from Canadian GAAP
(CGAAP). These requirements are in a state of ux, however,
as the IASB has issued proposals in an exposure draft
(Canadian Institute of Chartered Accountants, 2009) that
eliminate a provisions recognition criterion with refer-
ence to the probability of occurrence, and include the
likelihood of disbursement in the measurement of the li-
ability. Thus, as indicated in Table 1, the measurement would
be the amount that the entity would rationally pay at the
measurement date to relieve itself of the liability (i.e. the
amount equivalent to the expected value). This recorded li-
ability is presented in the same way as under IAS 37, i.e. as
a separate liability even if the recorded amount under IAS
37 is likely to be different.1 In the current investigation, tests
were run to check whether loan ocers differ in their judg-
ments or decisions as a result of the inuence of specic
changes required by IFRS (both current and proposed IAS
37) compared to the types of decisions made when infor-
mation is presented according to CICA accounting standards,
especially as regards contingent liabilities.
The research question at issue is whether nancial state-
ments based on IFRS (specically, contingent liabilities using
the current and proposed IAS 37) are easier to read and
convey a better understanding than those based on require-
ments of Canadian accounting standards. The study was
conducted as an experiment using Canadian loan ocers
assigned randomly into three groups according to the type
of nancial statements they received and the method used
1
Even if net income is different under the IAS 37 exposure draft versus
the current IAS37, the information provided in the note according to the
IAS 37 exposure draft would allow reconciliation of net income dis-
closed according to IAS 37. Thus, the information is available but located
in a different place.
 B. Lagrange et al. / Research in Accounting Regulation 27 (2015) 119128
Table 1
Comparison of the accounting treatment related to contingencies.
Section 3290
Recording criteria Section 3290 indicates that a liability
must be recorded when:
1. It is probable that a future event
will conrm that a liability exists
at the date of the nancial
statement.
2. There is a reasonable estimate.
Amount to be recorded A reasonable estimate of the amount
should be based on: most likely to be paid.
Disclosure In the balance sheet, the recorded
liability is presented with other
liabilities.
If a liability has been recognized, there
is no disclosure under the
contingencies note unless there is a
risk that loss would be higher than the
recorded amount. It may, however, be
relevant to mention the existence and
nature of potential losses.
for reporting and disclosing the contingency. The rst group
received statements that followed the Canadian standards
by recording the liability based on a reasonable estimate of
the amount most likely to be paid, but without presenting
it separately from other liabilities or disclosing any prob-
ability of the contingency occurring. The other two groups
were given information that followed current and pro-
posed IAS 37 by disclosing the contingency events
probability and presenting the contingent liability sepa-
rately. The difference between current IAS 37 and the
exposure draft is in the recording method. In the former,
as mentioned previously, the amount to be recorded should
be the most likely amount necessary for settling the obli-
gation at the closing date, while the latter is based on the
amount that the entity would rationally pay (i.e., the ex-
pected value) at the measurement date to be relieved of the
liability. Hence, any differences between the three types of
disclosures should be due to the reporting and disclosure
methods including reporting the probability of the
121
IAS 37 Exposure draft (ED) to IAS 37
IAS 37 states that a liability must be ED indicates that a liability must be
recorded when the following three recorded when the following three
criteria are met: criteria are met:
1. The obligation must be current. 1. The obligation must be current.
2. It must be probable (more 2. If the element meets the denition
probable than improbable) that an of a liability, irrespective of
outow of resources embodying probability, it must be recorded as
economic benets would be a provision.
required to settle the obligation. 3. There is a reliable estimate.
This means that there is more than
50% likelihood that the debt will be
carried out.
3. A reliable estimate can be made of
the amount of the liability.
A reliable estimate of the amount most The amount that should be rationally
likely to be paid. paid at the end of the reporting period
In other words, it is the best estimate for relief from the present liability.
of the expenditure necessary for the This amount should be the lowest of
extinguishment of the liability at the the following three items:
closing date (which is based upon The present value of the resources
professional judgment, expert opinion, required to fulll the obligation
and other factors). (this is the expected present value,
corresponding at the probability-
weighted average of the present
values of the outows for the
possible outcomes);
The amount that the entity would
have to pay to cancel the
obligation;
The amount that the entity would
have to pay to transfer the
obligation to a third party.
In the balance sheet, the recorded liability is presented as a separate item. Also, for
each class of provision, an entity shall disclose:
A brief description of the nature of the obligation and the expected timing
of any resulting outows of economic benets;
A reconciliation between amounts at the beginning and end of the period;
An indication of the uncertainties about the amount or timing of those
outows;
Where necessary, the major assumptions made concerning future events;
The amount of any expected reimbursement, stating the amount of any
asset that has been recognized for the expected reimbursement.
contingency and highlighting the liability as a separate line
item in the balance sheet.
This investigation was deemed important because it is
thought to be the rst experimental study in Canada to
examine whether reporting under IFRS (using contingen-
cy reporting as an example) conveys a stronger signal than
CICA reporting. Although IFRS applies to many topics, the
current study focused only on contingent liabilities, par-
ticularly in regard to the aforementioned issue. It is thus
hoped that insights can be provided into the potential con-
sequences and benets of IAS 37 (current and proposed) for
nancial statement users. These ndings should be useful
to regulators of IFRS as they consider alternative presen-
tation styles for similar types of regulations.
From a theoretical perspective, the results allow the re-
searchers to check whether previously documented effects
of uncertainty reporting on users (Anandarajan, Viger, &
Curatola, 2002; Bamber & Stratton, 1997; Ben-Amar & Viger,
2000; Gul, 1987; and Viger, Anandarajan, Curatola, &
122 B. Lagrange et al. / Research in Accounting Regulation 27 (2015) 119128
Ben-Amar, 2004, among others) apply to the new IFRS
regime.
3 Literature review
The investigators examined whether differences in the
presentation of contingent liabilities, i.e. disclosure versus
recognition, inuence the message conveyed to users. They
focused on variations in the reporting styles required under
the three different contingent liability reporting regimes
(CICA, IAS and IAS revised), given that the issue of contin-
gent liabilities has not received attention in the literature.
The literature review is broken down into categories re-
ecting the methodologies used in the relevant studies.
3.1 Experimental studies
A number of experimental studies examined disclo-
sure versus recognition (Anandarajan et al., 2002; Bamber
& Stratton, 1997; Belzile, Fortin, & Viger, 2006; Ben-Amar
& Viger, 2000; Elliott, 2006; Gul, 1987; Hirst & Hopkins,
1998; Hirst, Hopkins, & Wahlen, 2004; Hopkins, 1996;
Hopkins, Houston, & Peters, 2000; Maines & McDaniel, 2000;
Sami & Schwartz, 1992; Viger et al., 2004; Viger, Belzile, &
Anandarajan, 2008). They found that users judgments and
valuations are impacted by the method in which the infor-
mation is presented. Some studies demonstrated that
information placement and format inuence users judg-
ments (Anandarajan et al., 2002; Bamber & Stratton, 1997;
Belzile et al., 2006; Ben-Amar & Viger, 2000; Elliott, 2006;
Frederickson & Miller, 2004; Gul, 1987; Hirst & Hopkins,
1998; Hirst et al., 2004; Hopkins, 1996; Hopkins et al., 2000;
Maines & McDaniel, 2000; Sami & Schwartz, 1992; Viger
et al., 2004, 2008). Disclosure in the form of a note was found
not to be a substitute for recognition in the income state-
ments, a conclusion that applies to all types of disclosure.
If this nding holds true, then the current results are ex-
pected to indicate that information under revised IAS 37
would trigger new perceptions compared to other presen-
tation methods.
3.2 Empirical studies
In a study of rms credit ratings and ratings of the quality
of their annual report disclosures, Hein, Shaw, and Wild
(2011) suggested that commitment to annual report dis-
closure is related to lower cost of capital. Regulators may
thus wish to take note that this decrease is related to greater
disclosure costs. Alaraini, Healy, and Stephens (2003) ex-
amined the effect of cash ow information in SFAS 95 on
securities valuation and found that SFAS 95 requiring greater
disclosure and improving the quality and content of cash
ow information had an impact on the market. Shaw (2008)
investigated the relationship between revised pension rules
requiring recognition versus disclosure of pension infor-
mation and the cost of capital. Bond investors were found
to use both recognized and disclosed pension information
in their pricing decisions, suggesting little potential impact
of SFAS 158 on the cost of debt. This study appears to show
that investors understand the message conveyed whether
it is in the form of disclosure or recognition.
Conversely, other studies indicate that recognition sends
a stronger signal than disclosure. Legoria and Sellers (2005)
report that new requirements under SFAS 109 requiring sep-
arate recognition of deferred tax assets, deferred tax
liabilities and valuation allowances enhance clarity and ease
of understanding better than disclosure. In a study of the
SFAS 158 requirement to recognize previously disclosed net
pension and postretirement benet obligations on the
balance sheet, Fried (2012) also found that recognition of
post retirement obligations conveys greater clarity and ease
of understanding than disclosure. These ndings are im-
portant to the recognition versus disclosure debate because
they document that management reacts differently to dis-
closure in the nancial statement footnotes versus disclosure
in the balance sheet, especially in regard to information on
net pension and post retirement obligations. The lesson for
regulators appears to be that disclosure and recognition are
not equivalent. The current investigation further adds to the
debate because none of the above studies examined dis-
closure or recognition of contingent liabilities from the
perspective of the issue under study.
4 Hypotheses development
This study examines whether loan ocers perceptions
and decisions are affected by the loss contingency report-
ing format of nancial statements provided in loan
applications, specically by comparing their reactions to the
format under IFRSs current and proposed IAS 37 versus Ca-
nadian accounting standards. When the ocers focus only
on reported earnings numbers, they are expected to (1) base
their credit judgments and decisions on these gures irre-
spective of the contingency information disclosed in the note,
and (2) associate reported earnings in the income state-
ment with different levels of ability to repay loans (even
though the rms cash ow remained unchanged across ex-
perimental groups). It can be postulated that most of the
loan ocers judgments and decisions would be inu-
enced by the more favorable net income numbers (reported
in the income statement) when the nancial statements are
based on the proposed IAS 37 exposure draft (loan ocer
group referred to as G3 in this experiment) compared to
when they are based on current IAS 37 (loan ocer group
G2) or on Canadian accounting standards (loan ocer group
G1). As discussed in the next section, the ocers were ran-
domly assigned to one of three experimental groups and
given a package of material containing the same case study,
with the only manipulated variable being contingent lia-
bility reporting and disclosure. Those randomly assigned to
the rst group (G1) received nancial statements based on
the accounting standards in the CICA handbook, while the
second (G2) and third (G3) groups both received nancial
statements based on IFRS but differing in the method for
recording the liability. G2 had the same liability amount as
G1, but for G2 only, the liability was presented separately
in the balance sheet, consistent with current IAS 37. G3 had
a liability based on IAS 37 exposure draft requirements,
which lowered the reported liability more than for G1 and
G2, and was presented with the liability separately, similar
to G2. Despite these differences, cash ows from operating
 B. Lagrange et al. / Research in Accounting Regulation 27 (2015) 119128
activities and total cash ows were identical for the three
groups.
The rst two hypotheses stated in the alternate form are
the following:
H1. Loan ocers assign a lower (i.e. less risky) overall risk
rating (ORR)2 when nancial statements follow the pro-
posed IFRS exposure draft (G3) rather than current IAS 37
(G2) or Canadian accounting standards (G1). Alternatively
stated, ORRG3 < ORRG2 = ORRG1.
H2. Loan ocers assign a more optimistic (lower) overall
trend rating (OTR)3 when nancial statements follow the
proposed IFRS exposure draft (G3) rather than current IAS
37 (G2) or Canadian accounting standards (G1). Alterna-
tively stated, OTRG3 < OTRG2 = OTRG1.
Based on the rst two expected results concerning ORR
and OTR, nancial statements based on the proposed IFRS
exposure draft (G3) are assumed to lead to a higher per-
centage of loan approval (A) and a lower premium over the
prime rate (P); hence the following two hypotheses stated
in the alternate form:
H3. Loan ocers are more likely to approve the loan (A)
when nancial statements are based on the proposed IFRS
exposure draft (G3) than when they are based on current
IAS 37 (G2) or Canadian accounting standards (G1). Alter-
natively stated, AG3 > AG2 = AG1.
H4. Loan ocers charge a lower premium over the prime
rate (P) when nancial statements are based on proposed
IFRS exposure draft (G3) than when they are based on
current IAS 37 (G2) or Canadian accounting standards (G1).
Alternatively stated, PG3 < PG2 = PG1.
5 Methodology
The sample was composed of bank loan ocers4 repre-
senting thirty corporate nancial centers of the largest
nancial institution5 in the Province of Quebec, Canada. Case
studies were sent to 359 respondents, 176 of whom com-
pleted the questionnaire, resulting in an overall response
rate of 48.75 percent. This percentage compares favorably
with the average rate of other studies involving loan o-
cers. Almost seventy percent (120/176) of those who
2 Loan ocers were asked to circle the number corresponding to their
assessment of ABC Inc.s overall risk rating (ORR) (on a six-point Likert scale
where 1 means low risk and 6 high risk).
3 Loan ocers were asked to circle the number corresponding to their
assessment of the overall trend rating [positive (1), stable (2) or negative
(3)].
4
Loan ocers are important because of their signicant role in allo-
cating capital to rms, particularly small- and medium-sized enterprises,
and more importantly because they tend to be homogeneous and make
similar decisions. The results of this study should therefore have exter-
nal validity in extrapolation to the loan ocer population at large and the
sophisticated investor population in general.
5 With more than Cdn$190 billion in assets, Le Mouvement Desjardins
is the largest nancial institution in the Province of Quebec and the sixth
largest in Canada.
123
participated in the study had at least an undergraduate
degree, while 10 percent (18/176) and 22 percent
(38/176) held a masters or a doctorate degree respective-
ly. Thirty-three percent (58/176) had 15 years of experience
or more; 19 percent (34/176) had 1014 years; 18 percent
(31/176), 59 years; and 30 percent (53/176) less than ve
years. Sixty-three percent (108/170) of the respondents had
personal lending authority under $150,0006; 18 percent
(29/170) had between $150,000 and $225,000; and 19
percent (33/170), over $225,000. Eighty-nine percent
(156/175) had professional certication, 69 percent
(120/175) were not conned to one area of specialty, and
57 percent (100/175) were male.
All the loan ocers in the sample received an experi-
mental package that included descriptive information on a
hypothetical company (ABC Inc.), a standard auditors report,
a complete set of nancial statements for two scal periods,
and accompanying notes and information on the loan ap-
plication. The sample was divided into three groups, as
shown in Table 2.
6 Discussion of results
6.1 Preliminary checks
Three preliminary checks were performed: rst, manip-
ulation checks were conducted to ascertain that the loan
ocers understood the information presented to them. They
were asked to indicate if the experimental material con-
tained (1) the probability that the provision would be paid,
and (2) whether contingent liabilities were disclosed sep-
arately. The participants were also asked to indicate the
extent to which the companys net income varied between
2009 and 2010.
The rst manipulation check indicated that the sample
generally understood the information presented to them.
A minority failed this manipulation check and was elimi-
nated from the sample. Overall, 87.3 percent (48 out of 55
who responded) of the loan ocers assigned to G1 cor-
rectly acknowledged the fact that there was no probability
of occurrence disclosed, while 76.8 percent (43/56) and 80
percent (48/60) of those assigned to G2 and G3 correctly
acknowledged the probability of occurrence within the range
of 6080 percent.
About 70.2 percent (40 out of 57 who responded) of the
loan ocers assigned to G1 correctly acknowledged the fact
that there was no separate provision in the current liabili-
ties to reect the fact that the company was being sued (as
disclosed in footnote), while 96.4 percent (54/56) and 96.7
percent (59/61) of those assigned to G2 and G3 correctly
acknowledged a separate provision in the current liabili-
ties reecting the lawsuit.
6 The loan ocer reviews the documents, performs a nancial analy-
sis, and makes a decision under a designated lending authority. When the
loan request is above the loan ocers lending authority, the loan
application analysis is performed by the loan ocer and his (her) recom-
mendation is conveyed for approval and decision to a senior ocer with
greater lending authority. The loan ocers surveyed were therefore ac-
customed to analyzing loan applications of the size ($600,000) requested
in the case material.
124 B. Lagrange et al. / Research in Accounting Regulation 27 (2015) 119128

Table 2
Experimental design.

Experimental groups

Group 1 (G1) Group 2 (G2) Group 3 (G3)

Experimental material
Financial statement
presentation format,
especially in regard to
contingent liabilities
Disclosure of the probability of
occurrence
Separate liability regarding the
suit
Each participant received an experimental package that included descriptive information on ABC Inc.,
a standard auditors report, and a complete set of nancial statements for two scal periods (including all
nancial statements and accompanying notes and main nancial ratios).
CICA Handbook IFRS in effect since January 1, 2011 (IAS IASBs IAS 37 exposure draft
37)
A contingent liability must be Contingent losses are recorded as Contingent losses are recorded
recorded if 1) it is probable liabilities when the likelihood of such as liabilities when the
that an event causing a loss losses is greater than fty percent, elements meet the denition
will occur, and 2) the amount among other conditions. The most of a liability, irrespective of the
of the contingent loss can be likely amount necessary for settling probability.
appraised with a degree of the obligation at the closing date must The measurement would be
certainty. The amount to be be recorded. the amount that the entity
recorded should be based on a would rationally pay at the
reasonable estimate of the measurement date to be
amount most likely to be paid. relieved of the liability.
No Yes. A 70% probability of occurrence is disclosed both in G2 and G3.
No Yes. The information available concerning the measurement of the expense
for the provision for dispute is the same for experimental groups 2 and 3.

Recent results and main ratios 2010 2009 2010 2009

Net income $104,747 $149,448 $200,355 $149,448

Increase in revenue (%) 18.11 % 14.57 % 18.11 % 14.57 %
Working capital ratio 1.93 2.53 2.34 2.53
Gross prot margin 2.18 % 3.68 % 4.17 % 3.68 %
Return on shareholders 14.18 % 23.46 % 23.97 % 23.46 %
equity
Earnings per share $0.70 $0.99 $1.34 $0.99
Capitalization ratio 48.81 % 56.34 % 55.95 % 56.34 %
(equity/asset)
Dividends/share $0 $0.17 0% $0.17
Cash ows Cash ows from operating activities and total cash ows are identical for the three experimental groups.

Finally, 96.6 percent (56/58) and 91.1% (51/58) of G1 and early and late respondents for all demographic variables
G2 respondents respectively, correctly noted a decrease in (using both full and reduced samples).
2010 net income while 93.4 of G3 respondents (57/61) ac-
knowledged this decrease. As a result of the three
6.2 Test results for judgment and decision hypotheses
manipulations checks, 14 respondents (6 in G1, 6 in G2 and
(H1H2 and H3H4)
2 in G3) were eliminated from the sample, leaving a nal
usable sample of 162 loan ocers.
The two judgments under investigation were (1) overall
Second, an analysis of the demographic characteristics
risk rating (ORR) and (2) overall trend rating (OTR), and the
of the subjects assigned to each group (from the usable
decision outcomes related to (1) were whether the loan
sample) indicated that the randomization resulted in an ap-
ocer would grant the loan and (2) what premium would
proximately equal distribution of demographic characteristics
be charged.
among the groups (see Table 3), leading to the conclusion
Regarding the primary judgment of the ORR (H 1 :
that the study results were not driven by variations in these
ORRG1 = ORRG2 > ORRG3), the F-test and the results of a
characteristics and that none of the demographic vari-
KruskalWallis one-way ANOVA (Panel A of Table 4) did not
ables (educational background, level of experience, personal
reveal signicant differences (0.792 and 0.749 respective-
lending authority, professional title, area of specialization
ly) among the three experimental groups, indicating a lack
and gender) was signicant (given respective p-values of
of evidence to support H1. In contrast, as shown in Panel B
0.135, 0.139, 0.991, 0.949, 0.491 and 0.748).7
of Table 4, the results for the overall trend rating (H2:
The third and last manipulation check was to test for non-
OTRG1 = OTRG2 > OTRG3) reveal signicant results from both
response bias. No signicant difference was found between
the F-test and the KruskalWallis one-way ANOVA (0.002
and 0.001 respectively). Results from the Tukey HSD tests
7 None of the demographic variables (educational background, level of
indicate that the differences between G1 and G3 and
experience, personal lending authority, professional title, area of special-
between G2 and G3 are statistically signicant (0.025 and
ization and gender) was signicant in the full sample either (given respective 0.000 respectively), while the difference between G1 and
p-values of 0.243, 0.101, 0.984, 0.849, 0.491 and 0.869). G2 is not signicant (p = 0.22). This is consistent with H2.
 B. Lagrange et al. / Research in Accounting Regulation 27 (2015) 119128 125

Table 3
Statistics of demographic variables: comparison by experimental group.

Response Group Pearsons 2 p-value

G1 G2 G3

Educational background
What is your highest education level?
1. High school 1 3 1 0 12.375 0.135
2. General and vocational college 2 6 3 2
3. Bachelors degree 3 27 34 37
4. Masters degree 4 8 2 6
5. Doctorate degree 5 8 10 15
6. Other Total 52 50 60
Levene statistic for homogeneity of variance = 0.465 (p-value = 0.629)
Level of experience
How many years of experience do you have in commercial bank lending?
1. Less than 5 years 1 21 11 15 9.686 0.139
2. 59 years 2 4 12 14
3. 1014 years 3 12 8 12
4. 15 years or more 4 15 19 19
Total 52 50 60
Levene statistic for homogeneity of variance = 1.185 (p-value = 0.308)
Personal lending authority
What is your personal lending authority?
1. Under $125,000 1 22 19 22 1.606 0.991
2. Under $150,000 2 10 13 16
3. Under $200,000 3 6 5 6
4. Under $225,000 4 3 2 4
5. $225,000 or more 5 9 10 10
Total 50 49 58
Levene statistic for homogeneity of variance = 0.097 (p-value = 0.908)

Group N Yes No

Professional title
Do you have a professional title? 1 52 6 46 0.105 0.949
Yes/no 2 50 5 45
If yes, please specify 3 59 7 52
___________________ Total 161 18 143
Area of specialization
Are you specialized in any particular 1 51 14 37 1.422 0.491
industry or industries? 2 50 19 31
Yes/no 3 60 18 42
If yes, please specify Total 161 51 110
___________________

Group N Male Female

Gender
Male/female 1 51 30 21 0.581 0.748
2 50 30 20
3 60 32 28
Total 161 92 69

Group 1: Canadian GAAP; Group 2: newly adopted IFRS; Group 3: proposed IFRS exposure draft.

A Pearson chi-square test for the decision to grant the
loan (H3: AG1 = AG2 < AG3) indicates that the loan ocers will-
ingness to grant the loan did not signicantly differ
(p < 0.315) among the three experimental groups (Panel C
of Table 4). A second set of chi-square tests shows a mar-
ginal difference between G1 and G3 (p < 0.066, one-tailed)
but not between G2 and G3 (p = 0.272, one-tailed) or
between G1 and G2 (p = 0.195). H3 is therefore not sup-
ported. ANOVA results (Panel D of Table 4) for the premium
charged (H4: PG1 = PG2 > PG3) show signicant outcomes from
the F-test and the KruskalWallis one-way ANOVA (0.084
and 0.048 respectively) for the three groups, indicating that
the loan ocers were inuenced by the different formats.
The Tukey HSD Multiple Comparison test (Panel D of Table 4)
indicates marginally signicant differences between G1 and
G3 (p = 0.055, one-tailed) and between G2 and G3 (p = 0.087,
one-tailed), but no signicant difference between G1 and
G2 (p = 0.487). This is consistent with H4.
Continuing the analysis, three secondary judgments
(CONDITION, GROWTH and DEBT) were examined to check
whether they were inuenced by presentation format. As
shown in Table 5, the results of the F-test and the Kruskal
Wallis one-way ANOVA of the three secondary judgments
reveal signicant differences among the three experimen-
tal groups (with p-values of 0.001 and 0.000 for both
CONDITION and DEBT respectively and p-values of 0.96 and
0.037 for GROWTH). The results from the Tukey HSD tests
indicate that the differences between G1 and G3 are sta-
tistically signicant for both CONDITION and DEBT (0.001
and 0.000 respectively) but not signicant (at 0.118) for
126 B. Lagrange et al. / Research in Accounting Regulation 27 (2015) 119128

Table 4
Tests on primary judgments and decisions.

Panel A: Overall risk rating (ORR) Group N Mean Std. Dev. F (p-value) KruskalWallis one-way
ANOVA (p-value)

On this scale, circle the number corresponding to your assessment 1 52 3.06 0.42 0.233 (0.792) 0.579 (0.749)
of ABC Inc.s overall risk rating (on a 6-point Likert scale where 1 2 50 3.04 0.40
indicates low risk and 6 indicates high risk) 3 60 3.10 0.57
Levene statistic = 4.583 (p-value = 0.012) Total 162 3.07 0.47

H1: Comparisons among experimental groups (Tukeys HSD)a: p-value

Group 1 = Group 2 0.490

Group 1 > Group 3 0.443
Group 2 > Group 3 0.394

Panel B: Overall trend rating (OTR) Group N Mean Std. Dev. F (p-value) KruskalWallis one way
ANOVA (p-value)

On this scale, circle the number corresponding to your evaluation 1 52 2.23 0.76 6.743 (0.002)*** 13.342 (0.001)***
of ABC Inc.s overall trend rating [positive (1), stable (2) or 2 50 2.42 0.81
negative (3)]. 3 58 1.88 0.77
Levene statistic = 0.659 (p-value = 0.519) Total 160 2.16 0.81

H2: Comparisons among experimental groups (Tukeys HSD)a: p-value

Group 1 = Group 2 0.22

Group 1 > Group 3 0.025**
Group 2 > Group 3 0.000***

Panel C: Decision to grant the loan (A) Group N Yes No Pearsons 2 p-value

Notwithstanding your business relations, would you be willing to 1 51 38 13 2.310 0.315

grant the loan requested by ABC Inc.? Yes/No 2 49 40 9
3 57 49 8
Total 157 127 30

H3: Comparisons among experimental groups: Pearsons chi-squarea (p-value)

Group 1 = Group 2 0.739 (p-value = 0.195)

Group 1 < Group 3 2.255 (p-value = 0.066*)
Group 2 < Group 3 0.367 (p-value = 0.272)

Panel D: Interest rate premium (P) Group N Mean Std. Dev. F (p-value) KruskalWallis one-way
ANOVA (p-value)

Please indicate the appropriate rate you would ask for the loan 1 50 7.15 3.06 2.515 (0.084)* 6.080 (0.048)**
requested, considering that the repayment of principal is a xed 2 49 7.01 2.67
monthly installment and the interest is variable and calculated 3 57 5.94 3.40
on the basis of the premium added to the preferential rate in
effect at your institution (whether or not you recommend the
loan).
Levene statistic = 1.477 (p-value = 0.231) Total 156 6.66 3.11

H4: Comparisons among experimental groups (Tukeys HSD)a: p-value

Group 1 = Group 2 0.487

Group 1 > Group 3 0.055*
Group 2 > Group 3 0.087*
a
Test performed when controlling with an experiment-wide error rate of 0.10.
*** Signicant at 0.01.
** Signicant at 0.05.
* Signicant at 0.10.

GROWTH. The differences between G2 and G3 are statisti-
cally signicant for both CONDITION and DEBT (0.004 and
0.081 respectively) but not signicant for GROWTH
(p = 0.426). These results provide additional support for the
hypotheses regarding varying judgment impacts arising from
differences in presentation (proposed IAS 37 exposure draft
(G3) compared to current IAS 37 (G2) and Canadian ac-
counting standards (G1)). An unexpected signicant
difference (with a p-value of 0.007) between G1 and G2 was
found in the underlying judgment of DEBT (and a margin-
al p-value of 0.049 for GROWTH). It would thus appear that
the loan ocers perceptions were inuenced by the move
away from Canadian GAAP to IFRS even though the eco-
nomic scenario for the two formats was identical. This result
may have been due to the fact that a liability related to con-
tingencies was presented separately in accordance with IFRS,
and not in the manner stipulated in Canadian GAAP.
7 Conclusion
As of January 1, 2011, the IFRS has ocially replaced the
CICA handbooks accounting standards for publicly
 B. Lagrange et al. / Research in Accounting Regulation 27 (2015) 119128 127

Table 5
Tests on secondary judgments.

Panel A: Overall nancial condition (CONDITION) Group N Mean Std. Dev. F (p-value) KruskalWallis one-way
ANOVA (p-value)

I believe ABCs overall nancial condition is _____________. (on a 1 52 6.73 0.95 7.364 (0.001)*** 18.923 (0.000)***
10-point Likert scale where 1 indicates very poor and 10 2 50 6.82 0.92
indicates very good) 3 60 7.35 0.92
Total 162 6.99 0.97
Levene statistic for homogeneity of variance = 0.294 (p-value = 0.746)

Comparisons among experimental groups (Tukeys HSD)b p-value

Group 1 = Group 2 0.439

Group 1 < Group 3 0.001***
Group 2 < Group 3 0.004***

Panel B: Ability to sustain growth (GROWTH) Group N Mean Std. Dev. F (p-value) KruskalWallis one-way
ANOVA (p-value)

I believe ABCs growth potential is__________. (on a 10-point Likert 1 52 6.44 0.96 2.374 (0.96) 6.612 (0.037)**
scale where 1 indicates very poor and 10 indicates very good) 2 50 6.92 1.08
3 60 6.80 1.36
Total 162 6.72 1.17
Levene statistic for homogeneity of variance = 1.601 (p-value = 0.205)

Comparisons among experimental groups (Tukeys HSD)b p-value

Group 1 = Group 2 0.049

Group 1 < Group 3 0.118
Group 2 < Group 3 0.426

Panel C: Ability to pay its debts (DEBT) Group N Mean Std. Dev. F (p-value) KruskalWallis one-way
ANOVA (p-value)

I believe ABCs ability to pay its debts as they become due is 1 52 6.17 1.83 11.614 (0.000)*** 20.276 (0.000)***
________. (on a 10 point-Likert scale where 1 indicates very poor 2 50 7.06 1.63
and 10 indicates very good) 3 60 7.62 1.29
Total 162 6.98 1.69
Levene statistic for homogeneity of variance = 5.253 (p-value = 0.006)

Comparisons among experimental groups (Tukeys HSD)b p-value

Group 1 = Group 2 0.007***

Group 1 < Group 3 0.000***,a
Group 2 < Group 3 0.081*

Panel D: Pearson correlation coecients (p-value)

CONDITION GROWTH DEBT

CONDITION 0.398 (p < 0.001) 0.514 (p < 0.001)

GROWTH 0.398 (p < 0.001) 0.409 (p < 0.001)
DEBT 0.514 (p < 0.001) 0.409 (p < 0.001)
a One-tailed p-value.
b
Test performed when controlling with an experiment-wide error rate of 0.10.
*** Signicant at 0.01.
** Signicant at 0.05.
* Signicant at 0.10.

accountable enterprises. Although both sets of standards are
based on fundamentally similar conceptual frameworks, they
differ signicantly in some ways, for example in regard to
contingencies. This difference provides an opportunity to test
the impact of IFRS on sophisticated nancial statement users
when contingent liabilities vary in location and value. A
study of a sample of loan ocers revealed that their deci-
sions varied signicantly with respect to three underlying
judgments and the main judgment of the overall risk trend
(OTR). Their loan granting decisions were unaffected by the
change to IFRS, but loan ocers (G3) who worked with -
nancial statements based on the proposed IAS 37 exposure
draft charged a different premium than those who ana-
lyzed statements based on current IAS 37 (G2) or Canadian
accounting standards (G1). These ndings indicate that the
CICA is partially justied in its support of IFRS with respect
to contingent liabilities. Future research should examine
other areas to determine whether the use of IFRS provides
clearer information on the areas it covers.
In addition, the ndings also show that the type of stan-
dard followed signicantly impacts perceptions and some
decisions. The CICA therefore has a strong argument for con-
vincing future adopters of the need for international
convergence, especially in regard to the exposure draft on
contingency reporting. More studies on other key varia-
tions in CICA requirements and other local GAAPs would be
needed to draw conclusions on the effectiveness of the new
requirements in other signicant areas.
128 B. Lagrange et al. / Research in Accounting Regulation 27 (2015) 119128

References Gul, F. (1987). The effects of uncertainty reporting on lending ocers

perception of risk and additional information required. Abacus, 23(2),
172179.
Alaraini, S. A., Healy, J. P., & Stephens, R. G. (2003). SFAS 95 cash ow
Hein, F., Shaw, K. W., & Wild, J. J. (2011). Credit ratings and Disclosure
information and securities valuation. Research in Accounting Regulation,
Channels. Research in Accounting Regulation, 23, 2033.
16, 243255.
Hirst, E., & Hopkins, P. (1998). Comprehensive income reporting and
Anandarajan, A., Viger, C., & Curatola, A. P. (2002). An experimental
analysts valuation judgments. Journal of Accounting Research, 36, 4775.
investigation of alternative going concern reporting formats: A
Hirst, E., Hopkins, P., & Wahlen, J. (2004). Fair values, income measurement,
canadian experience. Canadian Accounting Perspectives, 1(2), 141
and bank analysts risk and valuation judgments. The Accounting Review,
162.
79(2), 453472.
Bamber, E. M., & Stratton, R. A. (1997). The Information content of the
Hopkins, P. (1996). The effect of nancial statement classication of hybrid
uncertainty modied audit report: Evidence from bank loan ocers.
nancial instruments on nancial analysts stock price judgments.
Accounting Horizons, 11(2), 111.
Journal of Accounting Research, 34(Suppl.), 3350.
Belzile, R., Fortin, A., & Viger, C. (2006). Recognition versus disclosure of
Hopkins, P., Houston, R. W., & Peters, M. F. (2000). Purchase, pooling and
stock option compensation: An analysis of judgments and decisions
equity: Analysts valuation judgments. The Accounting Review, 75(3),
of nonprofessional investors. Canadian Accounting Perspectives, 5(2),
257281.
133.
Legoria, J., & Sellers, K. F. (2005). The analysis of SFAS No. 109s usefulness
Ben-Amar, W., & Viger, C. (2000). Limpact du rapport du vricateur en
in predicting future cash ows from a conceptual framework
prsence de doutes quant la continuit dexploitation sur les
perspective. Research in Accounting Regulation, 18, 143161.
perceptions et dcisions des investisseurs: Une comparaison entre le
Maines, L. A., & McDaniel, L. (2000). Effects of comprehensive-income
Canada et les tats-Unis. Comptabilit, Contrle, Audit, 1(March),
characteristics on nonprofessional investors judgments: The role of
101118.
nancial-statement presentation format. The Accounting Review, 75(2),
Canadian Institute of Chartered Accountants. (2009). The CICAs Guide to
179207.
IFRS in Canada, 104 pages.
Sami, H., & Schwartz, B. N. (1992). Alternative pension liability disclosure
Chlala, N., Lavigne, A., & Vendette, M. (2009). Linformation nancire IFRS.
and the effect on credit evaluation: An experiment. Behavioral Research
CICA.
in Accounting, 4, 4962.
Elliott, W. B. (2006). Are investors inuenced by pro forma emphasis and
Shaw, K. W. (2008). Revised pension rules and the cost of debt. Research
reconciliations in earnings announcements? Accounting Review, 81(1),
in Accounting Regulation, 20, 325.
113133.
Viger, C., Anandarajan, A., Curatola, A. P., & Ben-Amar, W. (2004). Behavioral
Frederickson, J. R., & Miller, J. S. (2004). The effects of pro forma earnings
implications of alternative going concern reporting formats. Advances
disclosures on analysts and nonprofessional investors equity valuation
in Accounting Behavioral Research, 7, 5171.
judgments. The Accounting Review, 79(3), 667686.
Viger, C., Belzile, R., & Anandarajan, A. (2008). Disclosure versus recognition
Fried, A. H. (2012). Disclosure versus recognition: Evidence from lobbying
of stock compensation: Effect on the credit decisions of loan ocers.
behavior in response to SFAS 158. Research in Accounting Regulation,
Behavioral Research in Accounting, 20(1), 93114.
24, 2532.