The SEC’s Elimination of 20-F Reconciliation and the Cost of Bank Loans Saiying Deng * , Southern Illinois University Lucy Huajing Chen, Villanova University Parveen P. Gupta, Lehigh University Heibatollah Sami, Lehigh University (Abstract) We examine the impact of SEC’s elimination of 20-F reconciliation on cost of bank loans. To the extent that information production is costly, the elimination of 20-F reconciliation is associated with increased information collection costs. Consistent with such view, we find that banks charge higher interest spread to IFRS firms after the elimination rule, to compensate for the information loss. Furthermore, we document such effect is more pronounced in IFRS firms borrowing from distant lenders such as foreign banks, and in IFRS firms with more opaque financial reporting and less proprietary information. Our main results are robust to several sensitivity checks. First Draft: April 27, 2014 Preliminary draft, please do not quote. JEL Classification: G31 and M48 Key Words: 20-F reconciliation; cost of bank loans; information loss; loan spread * Corresponding author. Department of Finance, College of Business, Southern Illinois University, Carbondale, IL 62901. Email: [email protected], phone: 618-453-1418.
39
Embed
The SEC’s Elimination of 20-F Reconciliation and the Cost ...
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
The SEC’s Elimination of 20-F Reconciliation and the Cost of Bank Loans
Saiying Deng*, Southern Illinois University
Lucy Huajing Chen, Villanova University
Parveen P. Gupta, Lehigh University
Heibatollah Sami, Lehigh University
(Abstract)
We examine the impact of SEC’s elimination of 20-F reconciliation on cost of bank
loans. To the extent that information production is costly, the elimination of 20-F reconciliation
is associated with increased information collection costs. Consistent with such view, we find that
banks charge higher interest spread to IFRS firms after the elimination rule, to compensate for
the information loss. Furthermore, we document such effect is more pronounced in IFRS firms
borrowing from distant lenders such as foreign banks, and in IFRS firms with more opaque
financial reporting and less proprietary information. Our main results are robust to several
sensitivity checks.
First Draft: April 27, 2014
Preliminary draft, please do not quote.
JEL Classification: G31 and M48
Key Words: 20-F reconciliation; cost of bank loans; information loss; loan spread
* Corresponding author. Department of Finance, College of Business, Southern Illinois
University, Carbondale, IL 62901. Email: [email protected], phone: 618-453-1418.
To respond to the global accounting convergence and especially the adoption of
International Financial Reporting Standards (IFRS) by the European Union, the U.S. Securities
and Exchange Commission (SEC) adopted a rule to accept financial statements from foreign
private issuers prepared in accordance with IFRS as issued by the International Accounting
Standards Board (IASB) without reconciliation to the U.S. Generally Accepted Accounting
Principles (GAAP) (SEC 2007) on December 21, 2007. This new rule became effective on
March 4, 2008, for all foreign private issuers with fiscal years ending on and after November 15,
2007. Prior to the adoption of this rule, foreign private issuers which prepared their financial
statements using accounting standards other than U.S. GAAP were required to file Form 20-F
reconciliation with the SEC.1
Prior studies document that reconciliation provides useful information to shareholders in
the U.S. equity market to assess the financial condition of the foreign private issuers and suggest
that the elimination of the reconciliation is “premature” (e.g., Harris and Muller 1999; Chen and
Sami 2008, 2013; Henry, Lin, and Yang 2009). On the other hand, more recent studies find
there are no significant economic consequences in the equity market associated with the
elimination of 20-F reconciliation (e.g., Kim, Li, and Li 2012; Jiang, Petroni, and Wang 2010).
However, there is scant empirical evidence on how the 20-F reconciliation information is used
by private lenders such as banks and the economic consequence of the elimination rule in private
debt market. In this paper, we attempt to fill this void by examining whether the elimination of
1 Form 20-F is an annual report filed by foreign private issuers with the SEC. Foreign firms were required to file
form 20-Fs within six months after fiscal year end. On Form 20-Fs, foreign firms reported the quantitative and
qualitative differences between their primary accounting standards, either home country GAAP or IFRS, and the
U.S. GAAP under item 17 or item 18.
2
20-F reconciliation is associated with cost of bank loans in the U.S. for foreign private issuers
that prepared their financial statements under IFRS as issued by the IASB.2
We focus on bank loans because they become the primary source of corporate debt
financing in the past decade (Sufi 2009). We argue that the impact of elimination of 20-F
reconciliation on the private debt market may differ from that of equity market for the following
reasons: First, banks act as delegated monitor (Diamond 1984) and are specialized in information
collection and processing. Banks collect borrower-specific information over time and across
multiple products, perform due diligence and act as delegated monitors to borrowers. This
information advantage makes lending relationship banks special (Fama 1985; James 1987;
Lummer and McConnell 1989; Gande and Saunders 2012;), compared to arm’s-length investors
such as stockholders.3 To the extent that information production is costly, the elimination of 20-F
reconciliation may result in higher information collection costs for banks.
Second, banks, if possible, collect both soft and hard information about borrowers to
help assess their credit quality and reduce information uncertainty (Petersen 2004). Soft
information comes from onsite visits, and face-to-face interactions, etc., while hard information
primarily comes from financial statements, including 20-F reconciliation. Although the U.S.
GAAP continues to converge with IFRS, understanding the differences between these two
accounting standards is informative when establishing debt covenants, especially in cross-border
transactions (PWC 2013). From this perspective, 20-F reconciliation provides valuable
supplemental information for lenders to better interpret the financial statements and confirm the
2 The standards developed and approved by the IASB since 2001 are IFRS. Before that, they were International
Accounting Standards (IAS). For convenience, we refer to IFRS as both IFRS and its predecessor IAS. 3 The information advantage may make banks less susceptible to the information loss associated with the
elimination of 20-F reconciliation, even if reconciliation can be value relevant to arm’s length investors as
documented in Harris et al. (2009), and Chen and Sami (2013).
3
accounting numbers. Unlike shareholders, banks have asymmetric payoff structure: they bear
substantial downside risk without the potential of reaping upside benefits (Watts 2003).
Accordingly, banks could be more sensitive to information risk engendered by the elimination of
20-F reconciliation.
Lastly, many cross-listed firms get syndicated loans in international loan market, of
which most of the lead arrangers are foreign banks. Lenders may find the cost of collecting soft
information about and monitoring borrowers at a great distance (i.e., foreign borrowers)
prohibitive, and may rely more on hard information to make loan decision. Therefore, we expect
the information loss associated with elimination of 20F reconciliation is more pronounced for
lenders that are at great distance from their borrowers. To sum up, the information loss
associated with elimination of 20-F reconciliation, coupled with greater information collection
and monitoring costs associated with geographic distance, may lead the banks to charge higher
interest rate to foreign private issuers after the elimination of 20-F reconciliation.
Based on the arguments presented above, we predict that banks charge higher interest
rates to IFRS firms after the elimination rule. We also predict that such effect is more
pronounced in IFRS firms that borrow from distant banks such as foreign banks, where the
borrower’s headquarters is located in a different country from the country of syndication of the
syndicated loan.
It is notable that there are alternative information sources such as corporate voluntary
disclosures that may aid a lender to evaluate its borrowers’ creditworthiness. The impact of
elimination of 20-F on information loss crucially depend on the usefulness of the alternative
information sources in assessing borrowers’ credit quality.
4
We also examine whether a firm’s financial reporting opacity affects the association
between the elimination of 20-F and cost of bank loans. According to Franco et al. (2012),
“because banks rely primarily on periodic financial accounting reports and disclosure, along with
private information, to assess the creditworthiness (Spiro 2007), if a firm’s financial reporting is
more opaque, then the demand for private information and other supplemental and
complementary information that help interpret financial statements will be greater.” We thus
expect that 20-F reconciliation provides more useful information for lenders to interpret financial
statements in more opaque reporting firms. Consequently, lenders may charge higher interest
spread in firms with more opaque financial reporting after the elimination of 20-F reconciliation.
We further investigate whether a firm’s proprietary information affect the cost of bank
loans after the elimination rule. Banks need to collect more soft information in firms with more
proprietary information to make prudential lending decision. On the contrary, banks rely more on
hard information in firms with less proprietary information to make wise lending decision.
Therefore, we postulate that 20-F reconciliation provides more useful information for firms with
less proprietary information. As a result lenders charge higher interest spread in firms with less
proprietary information after the elimination of 20-F reconciliation.
We test our hypotheses using foreign private issuers that borrow at least one bank loan
before and after the elimination rule during 2005─2012 sample period. Our test sample consists
of foreign firms that prepared financial statements under IFRS as issued by the IASB and did not
reconcile to U.S. GAAP after the elimination of 20-F reconciliation. Our control firms include
foreign firms that prepared financial statements under their home country GAAP or U.S. GAAP;
that is, firms that do not change their reconciling behavior after the no-reconciliation rule. We
5
use a difference-in-differences approach and cluster standard errors at firm level with industry
fixed effects.
Our findings are generally consistent with our predictions. We find that loan spread
increases significantly in the post-elimination period for the test firms vis-a-vis the control firms.
Our result is consistent with our conjecture that there is significant information loss associated
with the elimination rule to private lenders, and in turn banks charge higher loan spread to IFRS
firms. We also find that IFRS firms borrowing from foreign banks experience a significantly
higher loan spread after the elimination of 20-F reconciliation compared to the control firms,
while there is no association between loan spread and the elimination of 20-F reconciliation for
IFRS firms borrowing from domestic banks. The evidence is consistent with our predication that
the elimination of 20-F reconciliation is a significant information loss event for private lenders
especially foreign banks that are far away from borrowers. It also suggests that the increasing
geographic distance between the borrower and banks leads to a more pronounced information
loss due to the elimination of 20-F reconciliation, as a result lenders charge a higher loan spread.
Third, We further document foreign IFRS firms with greater financial reporting opacity and less
proprietary information experience higher loan spread after the elimination rule. The result is
consistent with our predication that 20-F reconciliation provides more useful information for
firms with more opaque financial reporting/less proprietary information; as a result lenders
charge higher interest spread to compensate for the greater information loss.
Our paper contributes to the literature in the following ways: first, to the best of our
knowledge, this is the first paper to examine the economic consequence of elimination of 20-F
reconciliation in private debt market. Banks are special, and have information collection and
monitoring advantages over dispersed arm’s-length investors (Rajan, 1992). The elimination of
6
20-F reconciliation may impose a significant information loss for lenders, especially
geographically distant ones, which rely more on hard information provided by financial
statements in their lending decisions. We find that the elimination of 20-F reconciliation is
associated with significant increase in loan spread, suggesting that there is a significant
information loss associated with the elimination of 20-F reconciliation in private debt market.
Second, our study is related to Kim et al. (2012) who find no evidence that the
elimination of 20-F reconciliation has a significant impact on market liquidity, probability of
informed trading, cost of equity, analyst forecasts, institutional ownership, and stock price
efficiency and synchronicity. Our results contrast with theirs as we document that banks charge
higher loan spread to IFRS borrowers after the elimination rule. Furthermore, we find a more
pronounced information loss effect for IFRS firms borrowing from distant banks such as foreign
banks, since great distance makes information collection and monitoring prohibitive and erodes
lenders’ information advantage. We do not find information loss when the IFRS firms borrow
from domestic lenders. We argue that banks are better at collecting and processing information,
compared to arm’s-length investors in the equity market. 20-F reconciliation provides important
supplemental information for banks to interpret the financial statements, while shareholders may
not rely on the 20-F reconciliation to assess the credit worthiness of borrowers, and therefore
there are no significant economic consequences associated with the elimination rule in the equity
market. 4
Third, our study complements existing studies on why banks are special (James 1987;
Lummer and McConnell 1989; Gande and Saunders 2012; etc.). We show that banks are
4 In addition, cross-listed firms possess much thinner trading volume than their underlying local markets
(Chakravarty, Chiyachantana, and Jiang 2011), which may make it harder to detect any price impact in these
markets using equity data.
7
different from shareholders or bondholders in the public market and experience significant
information loss due to the elimination of 20-F reconciliation, especially when the lender-
borrower distance increases.
Fourth, our study relates to an emerging stream of literature on how geography affects
important financial outcomes and how geographic proximity is related to information advantage
of being local. For example, Chen, Hong, Huang and Kubik (2004), Coval and Moskowitz
(1999 and 2001) document that mutual fund managers investing in geographically proximate
companies enjoy superior returns due to the informational advantage from selecting nearby
stocks. Francis, Hasan, and Waisman (2007) find that rural firms experience a higher cost of
debt than urban firms because of the difficultly associated with monitoring their activities.
Similarly, Butler (2008) documents that local investment banks sell municipal bonds at lower
yields and charge lower fees than their non-local counterparts do. Such pricing advantage arises
from local investment banks’ soft information advantage. Using the elimination of 20-F
reconciliation as a unique setting, our study complements this stream of research by documenting
that distant lenders lose information advantage and charge a higher loan spread to foreign IFRS
firms after the elimination rule.
Finally, our paper contributes to the ongoing debate that challenges whether the SEC was
judicious in eliminating the 20-F reconciliation requirement. We document that eliminating the
20-F reconciliation requirement is associated with higher cost of bank debt for foreign IFRS
issuers. From this perspective, our result is more in line with Henry et al. (2009) and Chen and
Sami (2013) that reconciliation may provide some useful information to a subset of investors.
Overall, our findings shed some light on the economic consequences in the private debt market
8
associated with SEC’s decision to eliminate the 20-F reconciliation requirement for foreign
private issuers.
The remainder of the paper proceeds as follows: Section II presents related literature and
hypotheses development. Section III describes sample selection and research design. Section IV
presents the empirical results. Section V concludes.
II. Related literature and hypotheses development
2.1 Existing studies on 20-F reconciliation
There are split views about the economic consequences of elimination of 20-F
reconciliation in equity market. Hopkins et al. (2008), representing American Accounting
Association’s Financial Accounting and Reporting Section of the Financial Reporting Policy
Committee, argues that reconciliation provides useful information to assess the financial
condition of the foreign private issuers and suggests that the elimination of the reconciliation is
“premature” (Hopkins et al. 2008). Consistent with this view, prior literature provides empirical
evidence that reconciliation to U.S. GAAP contains some information content beyond IFRS.5 For
example, Harris and Muller (1999) find that earnings reconciliation from International
Accounting Standards (IAS) (forerunner to the current IFRS) to U.S. GAAP is value relevant in
the market valuation model, although earnings reconciliation per share is not associated with
price per share. In addition, Chen and Sami (2008) find that investors in the U.S. capital markets
trade on the earnings reconciliation from IAS to U.S. GAAP during the 1995-2004 period.
Henry, Lin, and Yang (2009) document that both net income reconciliation and book value
reconciliation from IFRS to U.S. GAAP are value relevant in the market valuation model but the
5 The standards developed and approved by the IASB since 2001 are IFRS. Before that, they were International
Accounting Standards (IAS). For convenience, we refer to IFRS as both IFRS and its predecessor IAS.
9
net income reconciliation is not value relevant in the long-term return model during the 2005-
2006 period. Also, Gordon, Jorgensen, and Linthicum (2008) show that accrual reconciliation
from IFRS to U.S. GAAP provides incremental information content beyond IFRS accruals from
2004 to 2006. Chen and Sami (2013) report that the relation between abnormal trading volume
and earnings reconciliation from IFRS to U.S. GAAP during the 2005-2006 period is driven by
first-time IFRS users with low institutional ownership.
On the other hand, Jamal et al. (2008), representing American Accounting Association’s
Financial Accounting Standards Committee, argue that the cost of preparing and auditing 20-F
reconciliation is substantial and “allowing foreign companies to use IFRS without costly
reconciliations to U.S. GAAP is likely to make U.S. stock exchanges more competitive and
provide useful feedback to U.S. accounting standard-setters about the efficacy of their standards”
(Jamal et al. 2008). There are also empirical studies echoing such view. For example, Kim et al.
(2012) compare foreign IFRS firms in the year before with the year after the no-reconciliation
rule. They find no evidence that removing the reconciliation has a negative impact on these
firms’ market liquidity, probability of informed trading, cost of equity, analyst forecasts,
institutional ownership, and stock price efficiency and synchronicity. Overall, their results do not
support information loss for equity investors. Similarly, Jiang et al. (2010) document no
abnormal trading volume, abnormal return volatility, or bid-ask spread during the short term
window around the release of 20-F in the year after the rule. In addition, Jiang et al. (2010) find
an accelerated filing of 20-F in the year after the rule compared to the year before, which they
interpret as an indication of the cost savings for these foreign IFRS firms.
2.2 Information, geographic distance, and bank lending
10
Banks collect borrower-specific information over time and across multiple products,
perform due diligence and act as delegated monitor to borrowers (Diamond 1984). This
information advantage makes lending relationship banks special (Fama 1985; Lummer and
McConnell 1989; Gande and Saunders 2012;). Banks can collect two types of borrower-specific
information: (i) Soft information, which is information loan officers collect through face-to-face
interaction with a potential borrower, onsite visits, and local contacts, etc., including
entrepreneur’s competence, honesty, management style, employee morale, etc.). Soft information
is hard to verify and credibly transfer within an organization hierarchy. (ii) Hard information,
which is information that is verifiable and can be credibly transferred (i.e. faxed or emailed)
within an organization hierarchy. The major sources of hard information are financial statements
and credit scoring, etc. 20-F reconciliation provides useful supplemental information about the
borrowers, which may help lenders especially lenders at a great distance from borrowers confirm
the accounting numbers. More specifically, reconciling to U.S. GAAP provides banks an
opportunity to confirm the IFRS-based accounting numbers, which potentially improves the
precision of information, provides additional assurance on the financial reporting quality, and
reduces information risk. Therefore the elimination of 20F reconciliation creates information loss
for lenders that rely on hard information to make loan decisions, for which lenders may charge
higher loan spread. Based on the above arguments, our first hypothesis is as follows:
H1: IFRS firms experience higher interest rates on bank loans after the elimination of 20-
F reconciliation.
However, there are alternative information sources such as corporate voluntary
disclosures that aid a lender to evaluate the borrowers’ creditworthiness. Those could become
countervailing forces against this prediction. IFRS firms may adopt voluntary disclosures to
11
make up the information loss associated with the elimination rule. Kim et al. (2012), Jiang et al.
(2010) find that none of their sample firms provide voluntary reconciliation after the rule. In
contrary, Yu (2011) find that IFRS firms significantly increase their overall voluntary disclosures
in annual financial reports and earnings announcement press releases after the elimination of 20-
F reconciliation. However, according to Cheng, Liao, and Zhang (2013), voluntary disclosure
does not serve as a credible commitment mechanism as mandatory disclosure.
Existing studies document that distance is related to information collection and
monitoring. Agarwal and Hauswald (2010) find that borrower proximity facilitates lenders to
collect soft information, thereby enhancing banks’ private information quality, and thus their
information advantage. Petersen and Rajan (2002) and Berger et al. (2005) document as the
physical distance between a lender and its borrowers increases, soft information collection
becomes more difficult and costly, eroding banks’ information advantage. Mian (2006)
document that foreign banks are willing to make transaction loans based on hard information, but
they are at a comparative disadvantage when making relational loans based on soft information.
Similarly, Hauswald and Marquez (2006) argue that lenders that are geographically
proximate to borrowers act as more effective monitors, and the quality of banks’ information
generation process is a decreasing function of lender-borrower distance. Along a similar vein,
Almazan (2002), and Sussman and Zeira (1995) indicate that a bank’s monitoring expertise is a
decreasing function of lender-borrower distance, and lenders’ monitoring costs are an increasing
function of distance. Hollander and Verriest (2011) find that distance erodes lenders’ ability to
collect and process borrower-specific information, as a result, lenders rely more on covenants to
resolve contracting frictions. Ayers, Ramalingegowda and Yeung (2011) posits that institutional
investor’s cost of acquiring monitoring information increases with the distance between a firm
12
and institutional investor, and find that local institutional investors lead to less financial reporting
discretion by the corporate managers.
Based on the above arguments, banks located closer to their borrowers enjoy significant
information advantage and more efficient monitoring, resulting in lower cost of information
collection and monitoring which the bank may pass on to borrowers in the form of lower interest
spread. On the other hand, banks may charge higher interest rates to distant borrowers due to the
greater information and monitoring costs. Therefore we propose the following hypothesis:
H2: IFRS firms that borrow from distant banks experience higher interest rates on bank
loans after the elimination of 20-F reconciliation.
2.3. Financial reporting opacity
We examine whether a firm’s financial reporting opacity may exacerbate the information
loss associated with the elimination rule. According to Franco et al. (2012), in a firm with more
opaque financial reporting, the demand for private information and other supplemental and
complementary information that may aid interpreting financial statements becomes greater. We
expect 20-F reconciliation provides more useful information in firms with more opaque financial
reporting. Therefore lenders may charge higher loan spread in firms with opaque financial
reporting after the elimination of 20-F, accordingly we propose the following hypothesis:
H3: IFRS firms that have more opaque financial reporting experience higher interest
rates on bank loans after the elimination of 20-F reconciliation.
2.4. Proprietary information
We investigate whether 20-F reconciliation provides more useful information in firms
with less proprietary information. In firms with more proprietary information, the demand for
banks collecting soft information is greater. In contrary, lenders rely more on hard information
13
(i.e., 20-F reconciliation) when making loans to firms with less proprietary information.
Therefore elimination of 20-F reconciliation leads to a greater information loss for banks that
lend to firms with less proprietary information. As a result, we expect cost of bank loan increase
more in firms with less proprietary information after the elimination of 20-F. Accordingly we
propose the following hypothesis:
H4: IFRS firms that have less proprietary information costs experience higher interest
rates on bank loans after the elimination of 20-F reconciliation.
III. Data and empirical methodology
3.1. Sample selection
We obtain syndicated loan data from the Dealscan database provided by the Loan Pricing
Corporation of Thomson Reuters, which contains detailed information on bank loans worldwide,
such as borrower and lender identities, loan amounts, LIBOR spread, issuing and maturity dates,
etc. We link the Dealscan and the Compustat database using a link file compiled by Chava and
Roberts (2008).6 Loan-specific, borrower-specific, and stock return information are obtained,
respectively, from Dealscan, Compustat, and CRSP databases. We hand collect the accounting
standard data from the SEC website. Since syndicate structure and loan contract terms (e.g., loan
amount, maturity, pricing) differ across different facilities in a deal, our analysis is conducted at
the loan facility level. Our sample spans from January 1, 2005 to December 31, 2012. We start
on January 1, 2005 as the European Union (EU) mandated all EU-listed companies adopt IFRS
beginning on January 1, 2005. As the elimination rule is effective for all foreign private issuers
with fiscal years ending on and after November 15, 2007, we classify the pre-elimination loan as
6 We would like to thank Michael R. Roberts for generously providing us the Dealscan-Compustat link file.
14
loan originated from January 1, 2005 to November 14, 2007 and the post-elimination loan as
loan originated from November 15, 2007 to December 31, 2012.
The sample selection process is as follows: We first hand collect accounting standard
information for 992 cross listed firms between January 1, 2005 and December 31, 2012. We
delete 147 firms which switch accounting standard before and after the elimination rule and
hence obtain 845 cross listed firms which use the same accounting standard during the entire
sample period. From these 845 firms, we extract Compustat data to construct firm-specific
variables and obtain 4,123 firm-year observations with 781 unique firms after deleting missing
values. We then merge these 4,123 observations with Dealscan dataset through Dealscan-
Compustat link file compiled by Chava and Roberts (2008). After deleting observations with
missing values, awe get 808 loan facilities for 179 unique cross listed firms over 2005~2012
sample period. There are 56 IFRS firms with 292 facilities, and 123 non-IFRS control firms with
516 facilities.
3.2. Research design
To investigate the effect of elimination of 20-F reconciliation on bank loan pricing, we
follow Graham, Li, and Qiu (2008) by controlling for firm-specific, loan-specific, and
macroeconomic factors that may affect loan pricing. We include industry fixed effects in all
models. We estimate the models with standard errors clustered at firm level.
This table reports the Pearson correlation matrix for major variables. All variables are defined in Appendix A. Correlations significant at the 1%–10% level are
This table presents the OLS regression results using a difference-in-differences approach to check the robustness of our main results. Column (1) uses U.S.
GAAP firms as the control firms; Column (2) uses home country GAAP firms as the control firms; Column (3) presents the results after deleting 2011 and 2012
observations; Column (4) presents the results after winsorizing all continuous variables at 1% and 99% level. Column (5) reports the results by accounting for
the recent 2007~2009 financial crisis effect. The dependent variable is LogSpread. Crisis is an indicator for the recent 2007~2009 financial crisis, which equals
one for 2007Q3 and 2009Q2, and zero otherwise. All variables are as defined in Appendix A. ***, **, and * denote statistical significance at the 1%, 5%, and
10% levels, respectively. Standard errors are clustered at firm level.
(1) (2) (3) (4) (5)
Variable US GAAP Home country GAAP Deleting 2011 and 2012 Excluding extreme values Crisis effect