THE JOURNAL OF THE AMERICAN TAXATION ASSOCIATION American Accounting Association

Vol. 46, No. 1 DOI: 10.2308/JATA-2022-035

Spring 2024
pp. 113–135

Tax Indemnification and Unrecognized Tax Benefits in

Mergers and Acquisitions
Patrick L. Hopkins
Texas Christian University

ABSTRACT: Tax indemnification transfers the risk of potential tax settlements from a taxpayer to an external party.
Current accounting standards require firms to reflect unrecognized tax benefits (UTBs) even when the underlying tax
position has indemnification. Practitioners suggest that financial statement disclosures regarding tax indemnification
are necessary to allow users to assess tax risk adequately. However, acquirers rarely disclose tax indemnification
arrangements from merger and acquisition (M&A) contracts. Evidence suggests that tax indemnification attenuates
the positive association between UTBs and future tax cash outflows. Moreover, results are consistent with
indemnification agreements resulting in minimal settlements. Because of the effect that removing accounts
associated with lapsed indemnification agreements has on effective tax rates and the attenuating effect that
indemnification has on UTB’s association with future tax cash outflows, the findings could be of interest to the
Financial Accounting Standards Board regarding their income tax disclosure project.
JEL Classifications: M48; M41; G34.
Keywords: income tax reporting; business combinations; unrecognized tax benefits; tax indemnification.

I. INTRODUCTION

T
ax indemnification transfers the risk of potential tax settlements from a taxpayer to an external party and has
traditionally been used to facilitate mergers and acquisitions (M&As). Despite this risk transfer, accounting
standards may still require a firm to report indemnified positions as unrecognized tax benefits (UTBs), resulting
in the reporting of a liability that is contractually not the firm’s responsibility.1 UTBs are contingent liabilities represent-
ing potential cash outflows associated with uncertain tax positions that tax authorities could successfully challenge.
Although public firms are required to provide detailed disclosures regarding UTBs, they are not required to provide dis-
closures regarding tax indemnification. As a result, firms have wide latitude in these disclosures, possibly hindering a
financial statement user’s ability to assess the potential consequences of indemnified positions within UTBs
(PricewaterhouseCoopers (PwC) 2020). However, there is little empirical evidence on whether and how indemnified tax
positions are disclosed, their impact on various tax accounts, and their effect on a financial statement user’s ability to
assess future tax cash flows.

I thank Ryan J. Wilson and two anonymous reviewers for their helpful comments and suggestions. Also, I am forever grateful for the invaluable guid-
ance given to me by my dissertation committee—T. J. Atwood (Chair), Cory Cassell, and Jonathan Shipman. Additionally, I am grateful for discus-
sions with participants of the 2020 26th Journal of the American Taxation Association Conference, the Arkansas Alumni Virtual Accounting Research
Seminar, and workshops at the University of Arkansas, University of Connecticut, Louisiana State University, and Texas Christian University, as well
as for feedback from Audra Boone, William Grieser, Stephen Lusch, Todd Kravet, Trent Krupa, Frank Murphy, James Myers, Linda Myers, Karen
Nelson, Elizabeth Plummer, John Robinson (discussant), Roy Schmardebeck, Hannah Smith, Mary Stanford, and Jake Thornock.
Patrick L. Hopkins, Texas Christian University, Neeley School of Business, Accounting Department, Fort Worth, TX, USA.
Editor’s note: Accepted by Senior Editor Ryan J. Wilson.

Submitted: November 2022

Accepted: May 2023
Early Access: August 2023

1
For example, Accounting Standards Codification (ASC) 805 coupled with ASC 740 requires that an acquirer continue to reflect acquired tax posi-
tions in UTBs after a completed M&A transaction, regardless of whether the seller indemnifies the positions. Although tax indemnification encom-
passes all types of taxes (e.g., income, payroll, property, sales, excise, etc.), the issues documented in this study are unique to income taxes because
UTBs only reflect uncertain income tax positions, accounting standards require separate disclosure of income taxes paid, and federal regulations pre-
vent reversing indemnification assets against income tax contingencies.

113
114 Hopkins

In this study, I investigate whether tax indemnification affects the association between UTBs and future tax cash
outflows for firms completing M&As. M&As are economically significant transactions and are an ideal setting to study
tax indemnification for two reasons. First, public acquirers often file M&A contracts with the Securities and Exchange
Commission (SEC), which usually include information regarding whether indemnification exists. Publicly available
information regarding indemnification arrangements is not widely available outside this setting. Second, public acquirers
file annual reports with the SEC that may include details regarding indemnification.
I begin my study by obtaining a sample of 495 M&A contracts from 2008 to 2016 via a manual search of the SEC’s
Electronic Data Gathering, Analysis, and Retrieval (EDGAR) system.2 I review these contracts to identify those includ-
ing tax indemnification clauses. I find that tax indemnification clauses are present in approximately 77 percent of this
sample but disclosed in approximately 3 percent of the associated 10-Ks, where investors can see and incorporate the
information regardless of whether the contract is publicly available.
With this sample, I document a positive association between current UTBs and future tax cash outflows for firms
without tax indemnification but an insignificant association for firms with tax indemnification. These findings suggest
that including indemnified positions within UTBs attenuates the association between UTBs and future tax cash out-
flows. In my sample, this effect reduces the explanatory power of UTBs to predict future tax cash outflows by an eco-
nomically meaningful 71.2 percent.
The attenuating effect of indemnification may be due to firms netting indemnification settlements against tax pay-
ments for disclosure purposes or indemnified positions not resulting in settlements/payments. To disentangle these alter-
native explanations, I analyze effective tax rates (ETRs) for firms with and without indemnification. When reversals of
indemnification positions occur, current accounting standards prevent reversing the initial indemnification entries and
require entries that reduce income tax expense to remove the associated UTBs and raise pre-tax book income to remove
the associated indemnification assets. Although these entries have no net effect on after-tax book income, they deflate
generally accepted accounting principles (GAAP) ETRs and inflate cash ETRs.3 ETR analyses are consistent with rever-
sals of indemnified positions happening more often than settlements. Specifically, I document lower GAAP and higher
cash ETRs for firms with indemnification compared with firms without indemnification, suggesting sellers may offer
indemnification when they perceive the risk of payout to be low.
Because unmodeled characteristics of the M&A transaction could affect future tax cash outflows and the choice to
obtain indemnification, omitted variable bias may be an issue in this study. To mitigate this concern, I conduct tests uti-
lizing a two-stage, endogenous-switching regression model that accounts for the treatment choice in the first stage. I pro-
pose a treatment model where the choice to obtain indemnification depends on operational uncertainties regarding the
target and other characteristics of the acquirer. First-stage results support this notion by documenting that acquirers are
likelier to obtain indemnification when acquiring a target from a different industry. Second-stage regressions include a
hazard variable produced in the first stage and reflect similar inferences to the primary analyses.
Limited disclosure of indemnification in non-M&A settings makes it challenging to directly assess the external valid-
ity of the M&A analyses. However, I document that my sample can be used to document similar conclusions to that of
a study using a general population of firms without considering indemnification. Furthermore, I also show that the
industry composition of my sample is similar in several respects to a more general population of public firms. These pat-
terns increase confidence that the results could generalize to non-M&A settings with indemnification.
To my knowledge, this is the first academic study to examine the disclosure and accounting issues arising from tax
indemnification. PwC (2020) warns that “it may often be necessary to provide additional disclosure in regard to the
terms of any [tax] indemnification arrangements so that financial statement readers can appropriately assess the net eco-
nomic exposure to the entity.” I find, however, that the norm is not to disclose these arrangements in the associated
financial statements. Given the frequency of tax indemnification within the M&A setting and the paucity of associated
10-K disclosure, my results suggest that investors may be unable to assess tax risk accurately in other settings with tax
indemnification.4 Therefore, firms with indemnified tax positions within their UTBs may want to consider voluntarily
disclosing these positions to facilitate investors’ assessments of their impact on future tax cash outflows.5

2
Although my sample runs from 2008 through 2016, I require data through 2021 to calculate my dependent variable.
3
See Appendix C for an example of the accounting and potential issues associated with tax indemnification.
4
In my setting, tax risk refers to the potential payment of additional taxes, interest, and penalties due to settlements with income tax authorities.
5
Separating UTBs into with and without indemnification groups would likely be considered a higher-quality form of disclosure by investors.
However, Robinson and Schmidt (2013) find that investors may penalize firms for providing higher-quality disclosures regarding UTBs. Therefore,
managers may be reluctant to disclose this information voluntarily. Furthermore, firms (and their indemnifiers) may not disclose indemnification
because they see it as alerting external parties (e.g., tax authorities) that risky tax positions exist (due to buyers requiring indemnification). This rea-
soning is consistent with earlier arguments by firms opposing the uniform reporting of uncertain tax positions under ASC 740.

The Journal of the American Taxation Association

Volume 46, Number 1, 2024
Tax Indemnification and Unrecognized Tax Benefits in Mergers and Acquisitions 115

The use of tax indemnification is growing and expanding beyond the M&A setting, mainly through tax indemni-
fication insurance.6 As a result, financial statement users need to understand how tax indemnification affects future
tax cash outflows and their associated disclosures. Under current guidance, firms can choose to net or not net
indemnification settlements against tax payments for disclosures regarding income taxes paid on the cash flow state-
ment, which could hinder the comparability of income taxes paid disclosures among firms. Given that minimal dis-
closure regarding tax indemnification exists in non-M&A settings, the issues documented in my study may make it
difficult for financial statement users to assess firms’ tax disclosures. As such, my findings are of interest to the
Financial Accounting Standards Board (FASB) for their ongoing “Targeted Improvements to Income Tax
Disclosures” project, which focuses on disclosures surrounding income taxes paid.7 Furthermore, I provide insights
into characteristics of deals obtaining indemnification, which could be helpful to future studies proxying for indemni-
fication when contracts are not readily available.
This study also contributes to the literature on UTBs. Consistent with prior research, I find that UTBs are associ-
ated with future tax cash outflows but only when indemnification is not present. Although past research generally con-
cludes that ASC 740 resulted in more consistent disclosures of uncertain tax positions, it also questions the
informativeness of these disclosures (Gleason, Mills, and Nessa 2018; Financial Accounting Foundation (FAF) 2012)
because of managers’ efforts to obfuscate tax strategies (Robinson and Schmidt 2013) and provisions within ASC 740
that generally result in overestimations (Robinson, Stomberg, and Towery 2016). My evidence suggests that tax indem-
nification could also impact the association between UTBs and future tax cash outflows.
This study is subject to several caveats. First, I can only obtain contracts for M&As that the reporting firm files
with the SEC. Analyzing the full population of M&A contracts may result in different conclusions regarding the
prevalence of indemnification and its effect on future tax cash flows. Second, firms often do not file supporting
schedules for M&A contracts, which may disclose specific tax positions or limits to the indemnification, thus signifi-
cantly impacting indemnification coverage. Third, M&A parties may obtain or provide third-party insurance instead
of offering indemnification by contract. By providing insurance, some contracts may not indicate that there is
indemnification although the acquirer is protected through insurance. These misidentifications could bias coefficients
documented in this study.

II. BACKGROUND AND HYPOTHESIS DEVELOPMENT

Tax Indemnification Accounting under ASC 805

M&As contain many risks (Karlinsky and Burton 2016), and sellers often offer indemnification to protect acquirers
from these risks, including tax-related risks (Logue 2005).8 ASC 805 requires mirror accounting for indemnified liabili-
ties, including indemnified tax positions. Specifically, ASC 805-20-25-27 states, “the acquirer shall recognize an indemni-
fication asset at the same time that it recognizes the indemnified item.”9 This accounting results in recognizing an asset
equal to a liability, resulting in no effect on the acquirer’s equity or overall income.
Although removing most types of indemnified liabilities and their corresponding assets is often unproblematic, the
same cannot be said about the underlying positions of UTBs that are indemnified. When these indemnified positions set-
tle or their statutes of limitations expire, their corresponding assets and UTB liabilities must be eliminated. When settle-
ments occur, the asset and related liability decrease without affecting the acquirer’s equity or overall income. However,
when adjustments occur (e.g., there is never a settlement or the settlement does not equal the UTB recorded), the adjust-
ments affect both pre-tax income and tax expense. Specifically, indemnification assets must adjust through pre-tax
accounts, and UTBs must adjust through tax expense.10 This inability to reverse the original mirror accounting entries
creates a permanent book-tax difference and distorts ETRs. Specifically, GAAP ETRs decrease, and cash ETRs
increase. Unless the book-tax difference is greater than 5 percent of total taxes calculated using the firm’s statutory rate,
it is not required to be separately disclosed in the ETR reconciliation or elsewhere in the tax disclosure.11

6
Aon Transaction Solutions reported an increase of 500 percent in their tax indemnification coverage over a span of three years, increasing from $2.1
billion in 2013 to $12.6 billion in 2016 (Rosen and Blitz 2017).
7
See [Link]
8
When M&A contracts provide tax indemnification, they include clauses that protect the acquirer from potential future tax liabilities involving the
target’s past. Refer to Appendix B for a discussion regarding contract language and specific examples of contract language.
9
For example, a UTB for an indemnified tax position is reflected by a corresponding (mirror) tax indemnification asset under mirror accounting.
10
ASC 740-10-25-8 allows benefit recognition when the tax position’s statutory period expires, whereas SEC Code of Federal Regulations §210.5-03
allows only “taxes based on income tax” to be reported under income tax expense.
11
See SEC Regulation S-X, Rule 4-08(h).

The Journal of the American Taxation Association

Volume 46, Number 1, 2024
116 Hopkins

Potential Issues with Indemnification Disclosure

Although the accounting for indemnification settlements results in no net effect for the indemnified acquirer, the
indemnified positions in the acquirer’s UTBs may reduce UTBs’ association with future tax cash outflows. Specifically,
ASC 740 requires the separate disclosure of UTBs, but neither ASC 740 nor ASC 805 requires the disclosure or discus-
sion of indemnification, potentially resulting in an undisclosed asset. Therefore, financial statement users may be unable
to differentiate indemnified from nonindemnified liabilities.12
Furthermore, neither ASC 740 nor ASC 805 addresses how to handle settlements when disclosing taxes paid on the
cash flow statement, resulting in two disclosure scenarios. In scenario one, the indemnified party nets out the settlement
against taxes paid, reflecting that they did not pay the tax. In scenario two, the indemnified party does not net out the
settlement against taxes paid, reflecting payment of the tax. Both scenarios are problematic. In scenario one, not reflect-
ing the paid tax reduces the association of UTBs with future tax cash outflows. In scenario two, reflecting the taxes paid
by an external party overstates taxes paid by the acquirer. Thus, current disclosure requirements (or lack thereof) may
hinder ASC 740’s original purpose of UTBs providing information about future tax cash outflows.

Prior Research on Tax Reserves

The FASB issued Interpretation No. 48 (FIN 48) in 2007, now codified under ASC 740, to facilitate the comparabil-
ity of financial statements and standardize tax reserve reporting. Before FIN 48, tax reserves were evaluated as contin-
gent losses and were often inadequate for assessing tax uncertainties (Gleason and Mills 2002; Dhaliwal, Gleason, and
Mills 2004; Gupta, Laux, and Lynch 2016). Under FIN 48, firms assess the likelihood of successfully defending authori-
tative challenges of uncertain tax positions. In determining this likelihood, they assume that tax authorities will chal-
lenge the positions and will have access to all private information regarding the positions. If firms determine that they
are “more-likely-than-not” to defend the positions successfully, they assess the associated benefits using a cumulative
probability approach. Any benefits associated with positions not meeting the “more-likely-than-not” threshold are
recorded as UTBs.
Since the issuance of FIN 48, several academic studies have explored the informativeness of UTBs (Guenther,
Matsunaga, and Williams 2017; Robinson et al. 2016; Hanlon, Maydew, and Saavedra 2017; Ciconte, Donohoe,
Lisowsky, and Mayberry 2016; Nesbitt 2014). For example, Guenther et al. (2017) find that estimated UTBs are not
associated with the volatility of future cash ETRs. In addition, Robinson et al. (2016) find that ASC 740 reduced the
ability of tax accounts to predict future tax cash outflows. In contrast, Hanlon et al. (2017) and Ciconte et al. (2016) find
that UTBs are positively associated with future tax cash outflows, whereas Nesbitt (2014) finds that nondiscretionary
UTBs are positively associated with future tax cash outflows but discretionary UTBs have no association with future tax
cash outflows.
Prior research suggests several reasons for substantial variation in the informativeness of UTBs, including (1) the
requirements of FIN 48 (De Simone, Robinson, and Stomberg 2014; Robinson and Schmidt 2013; Robinson et al. 2016;
Cazier, Rego, Tian, and Wilson 2015), (2) variations in the behavior and characteristics of firms (Lisowsky, Robinson,
and Schmidt 2013; Gleason et al. 2018; FAF 2012; Bozanic, Hoopes, Thornock, and Williams 2017; Abernathy, Beyer,
Gross, and Rapley 2017), and (3) complexities in tax law (FAF 2012). I contribute to this line of research by examining
the previously overlooked role of indemnified tax positions in reducing UTBs’ association with future tax cash outflows.

Development of Hypothesis
UTBs appear on a firm’s financial statements as reserves for potential settlements associated with uncertain tax posi-
tions. Additionally, Bozanic et al. (2017) suggest that tax authorities access firms’ financial statements and associated
UTB disclosures through the SEC’s EDGAR system. If a tax authority notices a firm reporting higher UTBs than its
peers, it may trigger a tax audit of that firm, focusing on the positions underlying the UTBs. Settlements of these audits
will likely result in additional taxes and interest. Due to UTBs’ underlying construct and tax authorities’ interest in
them, UTBs should be associated with future tax cash outflows. Consistent with this conclusion, Robinson et al. (2016)
document that a portion of these reserves results in cash outflows. In addition, Hanlon et al. (2017), Ciconte et al.
(2016), and Nesbitt (2014) document that UTBs are positively associated with future tax cash outflows. Thus, I expect
UTBs without indemnification to be positively associated with future tax cash outflows. However, indemnified tax posi-
tions within UTBs may attenuate this association. Specifically, a successfully challenged tax position with

12
The disclosure of indemnified tax positions is allowable; therefore, a firm may disclose this information voluntarily. Per my review of 10-K disclo-
sures, indemnified firms disclose the mirror asset and indemnified liability in 3.68 and 1.84 percent of associated 10-Ks, respectively.

The Journal of the American Taxation Association

Volume 46, Number 1, 2024
Tax Indemnification and Unrecognized Tax Benefits in Mergers and Acquisitions 117

indemnification would result in a cash outflow for the indemnifier, but the indemnified party would experience minimal
cash outflows.
Because sellers often use tax indemnification to facilitate M&As, the acquirer is often not financially responsible for
the target’s uncertain tax positions. However, the acquirer is required under ASC 805-740-25-2 to report uncertain tax
positions associated with the target (i.e., acquired UTBs) and the transaction (i.e., generated UTBs), regardless of
indemnification. Consistent with this standard, Lisowsky et al. (2013) document a positive association between the pres-
ence of M&As and the level of UTBs. Furthermore, Towery (2017) analyzes the composition of uncertain tax positions
reported by U.S. firms on Schedule UTP and documents that a significant component of these positions relates to
M&As. Therefore, a significant portion of a firm’s UTBs may relate to an M&A, but this portion may not be the finan-
cial responsibility of the reporting firm due to indemnification.
I predict that indemnified tax positions within UTBs will attenuate the association between the acquirer’s UTBs and
future tax cash outflows. This moderation could occur for two reasons. First, acquirers may net indemnification settle-
ments against tax payments for disclosure purposes. Second, sellers may offer indemnification when they perceive the
risk of payout as low, and the indemnified tax positions never result in tax authority settlements. If either or both of
these scenarios exist on average, I expect firms that report indemnified positions to have a lower association between
UTBs and future tax cash outflows than firms that do not. Accordingly, I formally state my first hypothesis in alterna-
tive form as follows:
H1: For acquirers in an M&A transaction, the association between UTBs and future tax cash outflows is less
positive for firms with indemnified uncertain tax positions than those without.
At least three scenarios exist where indemnification would not affect the association between UTBs and future tax
cash outflows. First, a firm may not net indemnification settlements against tax payments for disclosure purposes.
Second, a firm may have an immaterial amount of indemnified tax positions compared with all uncertain tax positions.
Third, a firm may be unable to enforce an indemnification clause.13 If any of these scenarios exist on average, I expect a
similar association between UTBs and future tax cash outflows for firms that report indemnified tax positions and those
that do not.

III. EMPIRICAL METHOD AND SAMPLE SELECTION

Methodology
To test whether the association between current UTBs and future tax cash outflows varies with the inclusion of
indemnified tax positions (H1), I interact the presence of indemnification clauses, INDEMN, with variables expected to
explain future tax cash outflows, FUT_TXPD, as follows14:

FUT TXPDt ¼ b0 þ b1 INDEMN t þ b2 UTBt þ b3 INDEMN t  UTBt þ bk CONTROLk

þ bj INDEMN t  CONTROLk þ et (1)

FUT_TXPDt is the log of the ratio of cumulative tax cash outflows over the years t+1 through t+5 to ending total
assets in year t and represents my measure of future tax cash outflows.15 I utilize a five-year window for FUT_TXPD to
account for the time lag between taking a tax position and losing a challenge to that position. Specifically, the statute of
limitations for U.S. federal income tax returns is generally three years, beginning on the date the return is filed.
Therefore, a position taken in year t is typically subject to scrutiny for up to four years, the year of filing, year t+1,
through the expiration of the statute, year t+4. A five-year window captures tax payments in year t+5 on successful tax
authority challenges in year t+4.
M&A contracts often indicate the existence of indemnification but rarely disclose the size of the indemnified item.
Therefore, indemnification is proxied as an indicator variable, INDEMN, rather than a continuous variable.
Indemnification may (or may not) affect future tax cash outflows depending on whether the acquirer obtained (or did
not obtain) any indemnified tax positions. Therefore, I offer no prediction on the main effect of INDEMN. The presence
of indemnification may mitigate associations with future tax cash outflows for UTBs, net operating loss (NOL)

13
10-K disclosures often warn that indemnification by third parties may not be enforceable or significant enough to mitigate all risks.
14
Firm subscripts are omitted from models for ease of exposition. Appendix A provides the definitions of all variables.
15
Hanlon et al. (2017) use a logged version of the ratio of future tax outflows to ending total assets, whereas Ciconte et al. (2016) use an unlogged ver-
sion. I use a logged transformation of the dependent variable to reduce concerns regarding the skewness of the dependent variable. As a robustness
in Table 5, column (2), I repeat the primary analysis with the unlogged ratio, and inferences remain the same.

The Journal of the American Taxation Association

Volume 46, Number 1, 2024
118 Hopkins

carryforwards, deferred tax assets, and deferred tax liabilities. Therefore, I include interaction terms between each
explanatory variable and INDEMN to allow associations to vary across observations with and without indemnification.
The main variable of interest is UTB. Consistent with prior research, the coefficient on UTB is expected to be positive
(b2 > 0) and reflects the association between UTBs and future tax cash outflows for the subsample of firms not obtaining
indemnification from the M&A contract. The joint effect of interest (b2 + b3) is expected to be lower than the association
documented on UTB (b2) and reflects the equivalent association for the subsample of firms obtaining indemnification.
The interaction between INDEMN and UTB is expected to be negative (b3 < 0) and tests whether the association of
interest varies between the two subsamples.
CONTROL is a vector of control variables similar to those in Hanlon et al. (2017). TXPD represents
current taxes paid, whereas D_TXPD represents changes in taxes paid. Consistent with prior research, I
expect a positive coefficient on TXPD and a negative coefficient on D_TXPD. D_PTBI represents changes in
pre-tax book income. Because of the stickiness of earnings, increases in taxable income should result in per-
sistent increases to future tax cash outflows. Therefore, I predict a positive coefficient on D_PTBI. NOL car-
ryforwards, NOLCF, can offset future taxable income and reduce future tax cash outflows. Thus, NOLCF is
predicted to have a negative coefficient. DEFTAX represents net deferred tax assets and liabilities adjusted
for UTBs relating to temporary book-tax differences. These assets (liabilities) should indicate reductions
(additions) to future tax liabilities and be negatively associated with future tax cash outflows. Consideration
paid during an M&A, CONS, can provide substantial tax benefits to the acquirer through cost recovery
deductions. As consideration paid becomes a larger component of ending assets, acquirers should experience
more considerable immediate reductions in future tax cash outflows. However, acquisitions may provide sub-
stantial income in subsequent years, suggesting that acquirers will experience increases in future tax cash out-
flows. Therefore, I refrain from predicting an association between CONSt and future tax cash outflows. All
models in this study use robust errors and include fixed effects for unmodeled industry and year
characteristics.

Sample Selection
I begin the sample formation by obtaining a pool of firm-year observations whose M&A contracts could be
available on the EDGAR database. Specifically, I merge all firm-years reporting more than one million dollars in
total assets from the Compustat North America database from the years 2008 to 2016 with data from the Thomson
SDC Platinum Mergers and Acquisitions database, resulting in 68,835 firm-year observations from 12,183 public
firms.16 Next, I remove observations that are real estate investment trusts (REITs) or foreign filers and observations
without all necessary data, which results in 11,928 firm-year observations from 2,344 firms. I then remove observa-
tions that did not complete an M&A during the firm-year, which results in 3,999 firm-year observations from 1,232
firms.
M&As are complex transactions, which makes exploring aspects of M&As difficult. These difficulties are amplified
if the acquirer has multiple M&As with varying terms during the same fiscal year.17 To reduce these confounding effects,
I remove observations with more than one completed M&A in a given firm-year, which results in 1,996 firm-year obser-
vations from 1,034 firms. Lastly, I remove observations reporting less than 100 percent ownership of the target after the
M&A completion. The final pool of observations for which M&A contracts may be available consists of 1,857 firm-year
observations from 995 firms.
SEC Regulation S-K requires that firms file all material contracts that come into effect during an accounting period
as exhibits to their associated 10-Q or 10-K. Accordingly, I review all 10-Ks associated with the final M&A sample pool
for acquisition disclosures and M&A contracts and remove any M&A observations that were not disclosed or discussed
in detail within the 10-K. With the remaining observations, I attempt to locate M&A contracts by manually searching
14 As, 14 D-1s, S-4s, 8-Ks, 10-Qs, and 10-Ks filed by the acquiring and selling firms (if available) during the two-month
windows surrounding the M&A announcement and effective dates. This search resulted in the final contract sample of
495 firm-year observations (contracts) from 385 firms (Table 1).

16
Blouin, Gleason, Mills, and Sikes (2010) document preemptive removal and reestablishment of tax reserves around the implementation of FIN 48,
effectively contaminating UTBs in 2006 and 2007. Therefore, I begin my study in 2008 to avoid contamination from this documented behavior.
17
This is highly problematic when examining the annually updated financial characteristics of a firm (e.g., UTBs). Moreover, it is doubtful that a firm
with multiple M&As in a given year would file all associated M&A contracts. Considering these issues, I concentrate on firms with only one M&A
in a given year.

The Journal of the American Taxation Association

Volume 46, Number 1, 2024
Tax Indemnification and Unrecognized Tax Benefits in Mergers and Acquisitions 119

TABLE 1
Sample Selection

Firms Observations
Total public company observations with more than $1,000,000 in total assets from 2008 through 2016 12,183 68,835
Less observations from REITs and foreign filers (4,130) (22,991)
Less observations without UTB data (3,295) (19,544)
Less observations without data for remaining variables (2,414) (14,372)

Full sample with model variables 2,344 11,928

Less observations without current-year M&As (1,112) (7,929)
Less observations with multiple current-year M&As (198) (2,003)
Less observations without 100% ownership after M&A (39) (139)

Potential contract pool 995 1,857

Less observations without M&A detail or contract (610) (1,362)

Final contract sample 385 495

This table presents the sample selection criteria.

IV. MAIN RESULTS

Descriptive Statistics
Table 2 presents descriptive statistics on indemnification clauses within the 495 hand-collected contracts. Panel A
reveals that indemnification is often provided to the acquirer through the indemnification of tax representations.18
Specifically, in my contract sample, 76.77 percent of acquirers are indemnified from losses associated with tax misrepre-
sentations of the target, suggesting that indemnification is common among firms completing M&As. Furthermore, 231
of those 380 contracts require escrow accounts to secure indemnification claims. Although indemnification is often pre-
sent in M&As, firms rarely disclose the indemnification asset and indemnified liability within the associated 10-K. In my
sample, 13 observations disclose the indemnification assets, whereas only six disclose the indemnified tax positions.
Table 2, Panel B indicates that 50.82 percent of merger contracts, 89.40 percent of share purchase contacts, 94.87
percent of asset purchase contracts, and 93.18 percent of other contracts provide tax indemnification.19 Moreover, Panel
C shows that as M&A deals grow in size, it is less likely for tax indemnification to be present. Specifically, 91.12 percent
of M&A deals valued less than $100 million have tax indemnification, compared with 34.38 percent of M&A deals val-
ued at over $1 billion. Lastly, Table 2, Panel D suggests that approximately 93 percent of M&A deals in which the target
is a subsidiary or a private company have tax indemnification. However, tax indemnification is present in roughly 1 per-
cent of M&As between public companies, likely due to difficulties in convincing large pools of shareholders to provide
personal guarantees or escrow funds to service indemnification claims.
Table 2, Panel E reports that 42 industries of the 48 Fama-French industries are represented in my sample. Approximately
33 percent of my sample is from the business service, electronic equipment, and pharmaceutical industries, whereas less than 1
percent is from the shipbuilding, tobacco, and utilities industries. The business service, electronic equipment, and pharmaceuti-
cal industries have indemnification frequencies of 72.97, 69.64, and 51.61 percent, respectively. Focusing on industries with
more than ten contracts in my sample, I find that the electrical equipment industry has the highest frequency of tax indemnifica-
tion clauses (100.00 percent). In contrast, the pharmaceutical industry has the lowest frequency (51.61 percent).
Table 3, Panel A presents descriptive statistics for the full contract sample. The sample consists primarily of large
firms with average year-end assets of $3.670 billion and is split relatively evenly between multinational (54.10 percent)
and domestic-only firms (45.90 percent). On average, tax cash outflows are approximately 10.40 percent of year-end
assets over the subsequent five years, and UTBs are 1.23 percent of year-end assets. These reserves are economically sig-
nificant and comparable with Hanlon et al.’s (2017) UTB percentage of 1.00 percent.
Table 3 also provides descriptive statistics for the subsamples of observations without (Panel B) and with indemnification
(Panel C). On average, firms that do not obtain indemnification are significantly larger (mean of $7.730 billion versus $2.442

18
Tax representations include a plethora of topics. Important to this study, these representations often state that all returns have been filed and all
taxes have been paid. See Appendix B for a discussion on contract language.
19
Other contracts consist primarily of partnership and equity interest purchase contracts.

The Journal of the American Taxation Association

Volume 46, Number 1, 2024
120 Hopkins

TABLE 2
Contracts and Indemnification Characteristics

Panel A: General
Yes No Total Contracts % Yes
Indemnification 380 115 495 76.77
Escrow 231 149 380 60.79
Discloses escrow 85 146 231 36.80
Discloses asset 13 367 380 3.42
Discloses liability 6 374 380 1.58
Panel B: Type
Indemnification No Indemnification Total Contracts % with Indemnification
Merger 93 90 183 50.82
Share purchase 135 16 151 89.40
Asset purchase 111 6 117 94.87
Other 41 3 44 93.18

380 115 495 76.77

Panel C: Consideration (in Millions)

Indemnification No Indemnification Total Contracts % with Indemnification
0 < $ < 100 195 19 214 91.12
100  $ < 250 90 17 107 84.11
250  $ < 500 44 18 62 70.97
500  $ < 1,000 29 19 48 60.42
1,000  $ 22 42 64 34.38

380 115 495 76.77

Panel D: Target Status

Indemnification No Indemnification Total Contracts % with Indemnification
Publicly held 1 87 88 1.14
Subsidiary 172 17 189 91.01
Privately held 199 10 209 95.22
Joint venture 8 1 9 88.89

380 115 495 76.77

(continued on next page)

billion), are more highly levered (mean of 28.30 percent versus 19.60 percent), and pay more for M&As as a percentage of
year-end assets (mean of 27.40 percent versus 15.20 percent) than firms that do obtain indemnification. These differences may
be attributable to the higher concentration of public-to-public M&As within the subsample of firms without indemnification.
Table 4 presents Pearson (below diagonal) and Spearman (above diagonal) correlations for all model variables.
Highly significant correlations (p-value < 0.01) are in bold. Correlations are consistent with expectations and, therefore,
are not discussed for brevity.

Future Tax Cash Outflows

Table 5, column (1) presents model (1) results using the full sample of M&A contracts. As predicted, future tax cash
outflows are positively associated with UTBs for firms without tax indemnification (b2 ¼ 22.018; p-value < 0.01).
Moreover, the equivalent association for firms with tax indemnification is not significant (b2 + b3 ¼ 0.016; p-value ¼
0.99) and is significantly lower than those without indemnification (b3 ¼ 22.002; p-value < 0.01). These results are con-
sistent with H1 and suggest that including indemnified tax positions in UTBs mutes the association between UTBs and
future tax cash outflows. Column (2) presents the standardized coefficients from an unlogged version of model (1) to

The Journal of the American Taxation Association

Volume 46, Number 1, 2024
Tax Indemnification and Unrecognized Tax Benefits in Mergers and Acquisitions 121

TABLE 2 (continued)
Panel E: Industries
Indemnification No Indemnification Total Contracts % with Indemnification
Business Services 54 20 74 72.97
Electronic Equipment 39 17 56 69.64
Pharmaceutical Products 16 15 31 51.61
Computers 17 5 22 77.27
Machinery 18 4 22 81.82
Medical Equipment 16 5 21 76.19
Wholesale 19 1 20 95.00
Communication 15 3 18 83.33
Measuring and Control Equipment 14 1 15 93.33
Automobiles and Trucks 12 2 14 85.71
Retail 9 5 14 64.29
Chemicals 7 6 13 53.85
Food Products 11 1 12 91.67
Steel Works 10 2 12 83.33
Trading 11 1 12 91.67
Business Supplies 8 3 11 72.73
Construction Materials 10 1 11 90.91
Electrical Equipment 11 0 11 100.00
Healthcare 8 3 11 72.73
Petroleum and Natural Gas 8 3 11 72.73
Consumer Goods 8 1 9 88.89
Apparel 7 1 8 87.50
Entertainment 8 0 8 100.00
Construction 7 0 7 100.00
Fabrication and Other 3 3 6 50.00
Rubber and Plastic Products 5 1 6 83.33
Transportation 5 1 6 83.33
Personal Services 3 2 5 60.00
Restaurants, Hotels, Motels 5 0 5 100.00
Insurance 1 3 4 25.00
Aircraft 2 1 3 66.67
Banking 2 1 3 66.67
Printing and Containers 2 1 3 66.67
Recreation 3 0 3 100.00
Precious Metals and Mining 2 0 2 100.00
Shipbuilding, Railroad Equipment 2 0 2 100.00
Textiles 2 0 2 100.00
Tobacco Products and Utilities 0 2 2 0.00

380 115 495 76.77

This table presents the indemnification characteristics of hand-collected M&A contracts.

allow the comparison of these results with those derived from a framework similar to Ciconte et al. (2016). They report
that “UTB explains as much as 9.6 percent of future income tax cash payments as do contemporaneous income tax cash
payments” (Ciconte et al. 2016, 4). My results suggest that among firms completing M&As, UTBs without indemnifica-
tion explain up to 22.3 percent of future tax cash outflows, whereas UTBs with indemnification explain only 6.4 percent,
a reduction of 71.2 percent.20 This evidence is consistent with indemnified tax positions resulting in minimal future tax

20
Although this reduction seems large, the ratio of indemnified UTBs to total UTBs can be quite large. For example, a subsample of eight firms in my
contract sample disclosed the size of their indemnification assets and how these assets were eventually removed. Among these firms, indemnified
UTBs accounted for approximately 67 percent of total UTBs, comparable to the 71 percent reduction in explanatory power.

The Journal of the American Taxation Association

Volume 46, Number 1, 2024
122 Hopkins

TABLE 3
Descriptive Statistics

Panel A: Full Contract Sample

Variables n Mean Std. Dev. p(25) Median p(75)
Total assets (in billions) 495 3.670 8.050 0.351 0.909 2.735
Pre-tax return on assets 495 0.037 0.119 0.001 0.050 0.101
Leverage ratio 495 0.216 0.186 0.052 0.198 0.332
Market to book ratio 495 2.946 3.971 1.350 2.062 3.155
Multinational 495 0.541 0.499 0.000 1.000 1.000
FUT_TXPDt (unlogged) 495 0.104 0.106 0.025 0.074 0.145
UTBt 495 0.012 0.020 0.002 0.006 0.014
INDEMNt 495 0.768 0.423 1.000 1.000 1.000
NOLCFt 495 0.238 0.630 0.000 0.030 0.186
TXPDt 495 0.018 0.021 0.003 0.011 0.030
DEFTAXt 495 0.005 0.062 0.039 0.003 0.019
CONSt 495 0.180 0.149 0.076 0.133 0.250
RDt 495 0.047 0.077 0.000 0.010 0.066
INTANGt 495 0.305 0.208 0.129 0.279 0.458
FOREIGNt 495 0.013 0.035 0.000 0.002 0.021
Panel B: No Indemnification Subsample (INDEMNt 5 0)
Variables n Mean Std. Dev. p(25) Median p(75)
Total assets (in billions) 115 7.730 11.968 0.795 2.530 9.322
Pre-tax return on assets 115 0.022 0.098 0.010 0.026 0.073
Leverage ratio 115 0.283 0.193 0.141 0.277 0.431
Market to book ratio 115 3.783 5.646 1.306 2.117 3.558
Multinational 115 0.530 0.501 0.000 1.000 1.000
FUT_TXPDt (unlogged) 115 0.088 0.096 0.024 0.056 0.123
UTBt 115 0.016 0.022 0.003 0.008 0.018
NOLCFt 115 0.356 0.894 0.000 0.065 0.324
TXPDt 115 0.015 0.018 0.002 0.009 0.022
DEFTAXt 115 20.022 0.062 0.070 20.016 0.005
CONSt 115 0.274 0.187 0.117 0.247 0.406
RDt 115 0.051 0.079 0.000 0.013 0.084
INTANGt 115 0.363 0.224 0.186 0.368 0.550
FOREIGNt 115 0.013 0.035 0.000 0.003 0.024
(continued on next page)

cash outflows. After weighting these results between firms with and without indemnification, my results suggest that
UTBs explain approximately 10.1 percent of future tax cash outflows as do “contemporaneous income tax cash
payments,” which is comparable with Ciconte et al.’s (2016) 9.6 percent. In later analyses, I explore whether the docu-
mented moderation is due to the reversal of indemnified positions or the netting of settlements for disclosure purposes.
The remaining results are consistent with expectations and are not discussed for brevity.

Biases in Estimates Concerns

Entropy balancing. Data collection procedures yielded a sample of 495 contracts, consisting of 380 treatment obser-
vations and 115 control observations. When an imbalance in representation exists between treatment and control
groups, concerns arise that significant covariate differences can result in biased coefficient estimates. Specific to this
study, inherent differences between treatment and control covariates may explain the differences documented in the
associations between UTBs and future tax cash outflows for firms with and without indemnification. To address this
concern, I perform first- and second-moment entropy balancing.21 Table 6, Panel A shows that this process results in the

21
Hainmueller (2012) suggests that entropy balancing “effectively adjusts for systematic and random inequalities in representation.”

The Journal of the American Taxation Association

Volume 46, Number 1, 2024
Tax Indemnification and Unrecognized Tax Benefits in Mergers and Acquisitions 123

TABLE 3 (continued)
Panel C: Indemnification Subsample (INDEMNt 5 1)
Variables n Mean Std. Dev. p(25) Median p(75)
Total assets (in billions) 380 2.442 5.906 0.293 0.721 2.036
Pre-tax return on assets 380 0.042 0.124 0.006 0.058 0.107
Leverage ratio 380 0.196 0.179 0.029 0.164 0.304
Market to book ratio 380 2.692 3.269 1.353 2.009 3.066
Multinational 380 0.545 0.499 0.000 1.000 1.000
FUT_TXPDt (unlogged) 380 0.108 0.108 0.025 0.080 0.151
UTBt 380 0.011 0.020 0.001 0.005 0.012
NOLCFt 380 0.202 0.520 0.000 0.023 0.138
TXPDt 380 0.019 0.022 0.003 0.012 0.030
DEFTAXt 380 20.000 0.061 0.032 0.000 0.021
CONSt 380 0.152 0.123 0.068 0.122 0.190
RDt 380 0.046 0.077 0.000 0.010 0.063
INTANGt 380 0.287 0.200 0.118 0.268 0.425
FOREIGNt 380 0.013 0.036 0.000 0.002 0.021
This table presents various descriptive statistics on the full contract sample, the subsample of contracts without indemnification, and the subsample of contracts
with indemnification in Panels A, B, and C, respectively. Significant differences in means and medians at the 1 percent level between Panels B and C are in
bold. All continuous variables are Winsorized at the 1 and 99 percent levels.
All variables are defined in Appendix A.

TABLE 4
Correlations

Variables (1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
(1) FUT_TXPDt 0.073 0.014 0.671 20.355 0.059 0.002 20.122 0.095 0.293
(2) INDEMNt 0.079 20.144 0.081 20.121 0.161 20.302 0.032 20.140 0.026
(3) UTBt 20.125 0.097 0.045 0.211 0.077 0.020 0.401 0.000 0.131
(4) TXPDt 0.666 0.087 20.154 20.399 0.015 0.077 20.131 0.085 0.260
(5) NOLCFt 20.226 0.103 0.440 20.215 0.126 0.108 0.380 0.070 0.029
(6) DEFTAXt 0.001 0.145 0.002 0.060 0.039 20.156 0.226 20.246 0.108
(7) CONSt 0.061 20.346 0.045 20.162 0.046 20.130 0.085 0.414 0.040
(8) RDt 20.130 0.031 0.491 20.153 0.441 0.094 0.074 0.047 0.079
(9) INTANGt 0.050 20.153 0.109 0.006 0.067 20.209 0.429 0.078 0.022
(10) FOREIGNt 0.230 0.011 0.078 0.232 0.105 0.048 0.067 20.207 0.017
This table presents Pearson’s (below diagonal) and Spearman’s (above diagonal) correlation coefficients for the sample (n ¼ 495). Significant cor-
relations at the 1 percent level are in bold. All continuous variables are Winsorized at the 1 and 99 percent levels.
All variables are defined in Appendix A.

treatment group having similar means and variances as the control group across all explanatory variables. In Table 6,
Panel B, I use this reweighted sample to repeat the analyses from Table 5 and find similar inferences.
Outlier bias. With any small sample, concerns exist regarding whether a few observations bias the results.
Therefore, all continuous variables are Winsorized at the 1 and 99 percentiles to reduce outlier bias. Additionally, I per-
form DFBETA analyses described by Belsley, Kuh, and Welsch (2005) as a robustness check to determine the influence
of each observation and remove any observations with significant influence on the overall regression. This procedure
results in the removal of 21 observations. With this smaller sample, I repeat the primary analyses (untabulated) and con-
tinue to find that UTBs of firms without indemnification have a positive association with future tax cash outflows
(p-value < 0.01), and UTBs of those without have no association with future tax cash outflows (p-value ¼ 0.82).

The Journal of the American Taxation Association

Volume 46, Number 1, 2024
124 Hopkins

TABLE 5
Interaction Model of UTBs’ Association with Future Tax Cash Outflows

(1) (2)
Variables (+,2) FUT_TXPDt Unlogged FUT_TXPDt
INDEMNt (b1) ? 0.011 0.010
(0.033) (0.125)
UTBt (b2) + 22.018 0.160
(3.221) (2.388)
INDEMNt  UTBt (b3) - 222.002 20.114
(22.730) (21.758)
TXPDt (b4) + 40.368 0.638
(5.652) (4.482)
INDEMNt  TXPDt (b5) ? 1.616 0.061
(0.206) (0.392)
D_TXPDt (b6)  13.315 0.210
(1.355) (2.041)
INDEMNt  D_TXPDt (b7) ? 0.041 0.010
(0.003) (0.090)
D_PTBIt (b8) + 2.830 0.120
(1.788) (1.540)
INDEMNt  D_PTBIt (b9) ? 2.288 0.066
(1.318) (0.827)
NOLCFt (b10)  1.175 0.225
(7.717) (4.226)
INDEMNt  NOLCFt (b11) ? 0.868 0.094
(3.801) (2.262)
DEFTAXt (b12)  6.986 0.133
(3.283) (1.917)
INDEMNt  DEFTAXt (b13) ? 8.089 0.183
(3.321) (2.735)
CONSt (b14) ? 0.145 0.036
(0.271) (0.551)
INDEMNt  CONSt (b15) ? 0.580 0.029
(0.743) (0.436)
Constantt ? 2.830 0.165
(1.788) (1.845)
b2 + b3 5 0 0.016 0.046
[0.00] [0.33]
Observations 495 495
Adjusted R2 0.407 0.513
Year and Industry FE Yes Yes
,, Indicate significance at the 0.10, 0.05, and 0.01 levels, respectively (based on one-tailed tests when a direction is predicted, two-tailed
otherwise).
This table presents in column (1) the results of model (1) using the full sample, whereas column (2) presents the standardized coefficients of an
unlogged model (1). All continuous variables are Winsorized at the 1 and 99 percent levels. Robust t-statistics are presented in parentheses, and
F-statistics are presented in brackets.
All variables are defined in Appendix A.

The Journal of the American Taxation Association

Volume 46, Number 1, 2024
Tax Indemnification and Unrecognized Tax Benefits in Mergers and Acquisitions 125

TABLE 6
Entropy Balancing

Panel A: First Two Moments before and after Weighting

Before Weighting After Weighting

Variables n Mean Variance Mean Variance

Control Group
UTBt 115 0.011 0.000 0.011 0.000
TXPDt 115 0.019 0.001 0.019 0.001
D_TXPDt 115 0.003 0.000 0.003 0.000
D_PTBIt 115 0.001 0.012 0.001 0.012
NOLCFt 115 0.202 0.271 0.202 0.271
DEFTAXt 115 0.000 0.004 0.000 0.004
CONSt 115 0.152 0.015 0.152 0.015
Treatment Group
UTBt 380 0.016 0.000 0.011 0.000
TXPDt 380 0.015 0.000 0.019 0.001
D_TXPDt 380 0.000 0.000 0.003 0.000
D_PTBIt 380 0.018 0.006 0.001 0.012
NOLCFt 380 0.356 0.799 0.205 0.274
DEFTAXt 380 0.022 0.004 0.000 0.004
CONSt 380 0.274 0.035 0.153 0.015

Panel B: Interaction Model of UTBs’ Association with Future Tax Cash Outflows with Weighted Sample
Variables (+,2) FUT_TXPDt
INDEMNt (b1) ? 0.321
(0.940)
UTBt (b2) + 22.286
(3.601)
INDEMNt * UTBt (b3) 2 226.188
(23.406)
Observations 495
R2 0.626
Year and Industry FE Yes
Controls and Constant Yes
 Indicates significance at the 0.01 level (based on two-tailed tests).
This table presents the results of performing model (1) with an entropy balanced sample. Panel A presents the means and variances of variables
used in model (1) before and after the weighting procedure to reduce differences in the first and second moments between the control and treat-
ment group. Panel B presents the results of model (1) using the weighted sample. All continuous variables are Winsorized at the 1 and 99 percent
levels. Robust t-statistics are presented in parentheses.
All variables are defined in Appendix A.

Endogeneity Concerns regarding Treatment

Obtaining indemnification is likely not a random event but a conscious decision made by the acquirer and target.
As a result, coefficient bias could be a concern in model (1) due to self-selection for treatment.22 Specific to this study,
unmodeled characteristics of the M&A deal may explain future taxes paid and the nonrandom assignment of firms to
the indemnified (treatment) and not indemnified (control) groups. To mitigate this concern, I utilize a two-stage,

22
I follow Larcker and Rusticus (2010) and estimate the first stage of my treatment model to obtain a predicted value of INDEMN to incorporate,
along with the observed value of INDEMN, into a second-stage ordinary least squares regression. The predicted value of INDEMN is insignificant
at conventional levels in this regression, which fails to support the endogeneity concerns.

The Journal of the American Taxation Association

Volume 46, Number 1, 2024
126 Hopkins

endogenous-switching model described by Maddala (1983).23 The first-stage Probit model estimates the treatment effect
and a hazard variable for each observation. The first-stage model is as follows:

INDEMN t ¼ k0 þ k1 MAC INDt þ kk CONTROLk þ et (2)

MAC_IND is an indicator variable equal to 1 if the target operates in a different macroindustry than the acquirer,
and 0 otherwise. MAC_IND represents a scenario where acquirers (sellers) may have less (more) of an understanding
regarding the targets’ risks and, therefore, demand (provide) some assurances. Therefore, I expect a positive association
between MAC_IND and INDEMN. However, I do not expect acquiring a target from a different industry to influence
future taxes paid differently than acquiring a target from the same industry.24 Therefore, I expect no association between
MAC_IND and FUT_TXPD, suggesting that MAC_IND may be an appropriate instrument (exclusion restriction) in
this model.25 CONTROL represents all of the explanatory variables (without their interactions) from model (1). The sec-
ond stage incorporates the hazard variable from model (2)’s estimation as an additional control into model (1).
Table 7, column (1) presents the results of estimating model (2), the first stage of the endogenous-switching model. I
find that acquirers are more likely to obtain indemnification from sellers when acquiring targets from a different industry
(k1 ¼ 0.313; p-value < 0.05), suggesting that when acquirers are less familiar with the complexities of operating in the
target’s industry, they are likelier to demand indemnification. Furthermore, I find that as the consideration paid
increases, it is less likely for the acquirer to obtain indemnification; this finding is consistent with inferences drawn from
Table 2, Panel C. Table 7, column (2) presents the results of introducing the hazard variable from the first-stage Probit
model into model (1). The second-stage results are consistent with earlier results and provide some comfort that earlier
conclusions were not solely attributable to unmodeled characteristics of the M&A deal that explain future taxes paid
and the nonrandom assignment of treatment.

V. ADDITIONAL ANALYSES

Competing Explanations
Table 5 suggests that firms with indemnification experience lower associations between UTBs and future tax cash
outflows than firms without indemnification. Two scenarios may explain this result. First, sellers offer indemnification
on positions they believe are low risk, and the positions eventually reverse. Second, firms offset tax payments with
indemnification settlements for disclosure purposes.
In Section II and Appendix C, I discuss how the accounting for indemnification reversals could result in deflated
and inflated ETRs. Based on these discussions and to disentangle the alternative explanations, I compare the future five-
year GAAP and five-year cash ETRs of firms without and with indemnification for the entire sample. Untabulated uni-
variate tests reveal that future five-year GAAP ETRs are not statistically different between the two groups (p-value ¼
0.79). However, future five-year cash ETRs are statically higher for indemnified firms compared with not indemnified
firms (p-value < 0.10). This result is consistent with reversals of indemnification assets inflating cash ETRs.
However, numerous firm characteristics could explain firms’ ETRs. To ensure that other firm characteristics cannot
explain the ETR univariate results, I follow Drake, Hamilton, and Lusch (2020) and control for those expected to
explain ETRs. The ETR multivariate model is as follows:

ETR ¼ b0 þ b1 INDEMN t þ bk ETR CONTROLk þ et (3)

ETR represents future five-year GAAP and future five-year cash ETRs. INDEMN is as previously defined. If rever-
sals of indemnification assets happen more often than settlements, I expect a negative (positive) association between
GAAP_ETR (Cash_ETR) and INDEMN. ETR_CONTROL represents several variables that are expected to explain

23
A two-stage, endogenous switching regression model is similar to the Heckman (1976) self-selection approach. However, a key difference is that the
endogenous switching regression allows the treatment effect to interact with explanatory variables, thus, allowing associations between explanatory
variables and the dependent variable to vary between the treatment and control groups. Furthermore, in my setting, the dependent variable (future
taxes paid) is consistently observed whether or not treatment exists, a condition that is not appropriate for using the Heckman self-selection
approach.
24
In untabulated analyses, I include MAC_IND and its interaction with INDEMN in model (1). Results show that neither variable has associations
with FUT_TXPD that are significantly different from 0. Moreover, the overall inferences remain unchanged with the inclusion of these variables.
Specifically, UTBs of firms without indemnification have a positive association with future tax cash outflows (p-value < 0.01), whereas UTBs of
firms with indemnification have no association (p-value ¼ 0.99).
25
To further validate the appropriateness of MAC_IND as an instrument, I follow Larcker and Rusticus (2010) and check the correlation between the
error terms of the second-stage regression with MAC_IND. This analysis fails to reveal a correlation between the error terms and instrument vari-
able at conventional levels (p-value ¼ 0.48), suggesting the instrument is appropriate.

The Journal of the American Taxation Association

Volume 46, Number 1, 2024
Tax Indemnification and Unrecognized Tax Benefits in Mergers and Acquisitions 127

TABLE 7
Two-Stage, Endogenous Treatment Switching Model

(1) (2)
Stage 1 Stage 2

INDEMNt 5 0 INDEMNt 5 1

Variables (+,2) INDEMNt FUT_TXPDt

MAC_INDt (k1) + 0.313
(2.040)
TXPDt (k2) 0.636 39.706 38.654
(0.170) (5.060) (11.440)
D_TXPDt (k3) 0.813 13.448 13.733
(0.170) (1.430) (3.250)
D_PTBIt (k4) 0.523 2.644 0.398
(0.780) (1.750) (0.660)
NOLCFt (k5) 0.122 1.116 0.261
(1.100) (6.060) (1.950)
DEFTAXt (k6) 2.665 8.285 0.382
(2.410) (3.580) (0.300)
CONSt (k7) 2.867 1.158 1.370
(6.530) (0.790) (1.250)
UTBt (k8)/(b1) 23.422 23.069 1.403
(20.990) (3.290) (0.370)
Diff. in (b1) 21.666
[8.390]
Observations 495 495
Year and Industry FE No Yes
Hazard No Yes
Constant Yes Yes
,, Indicate significance at the 0.10, 0.05, and 0.01 levels, respectively (based on two-tailed tests).
This table presents the results of performing a two-stage, endogenous-switching model. Column (1) presents the results of model (2), the first stage,
whereas column (2) presents the results of including the hazard variable in model (1), the second stage. All continuous variables are Winsorized at
the 1 and 99 percent levels. Z-statistics are presented in parentheses, and v2-statistics are presented in brackets.
All variables are defined in Appendix A.

future ETRs. Controls include the observation’s size, market-to-book ratio, intangible intensity, tangible intensity, earn-
ings in equity, leverage, return on assets, research and design (R&D) intensity, advertising intensity, free cash flows, and
NOL carryforwards. In Table 8, columns (1) and (2) present the GAAP_ETR and Cash_ETR results, respectively. In
column (1), INDEMN has a negative association with future GAAP_ETR (b1 ¼ 0.058; p-value < 0.10), which is con-
sistent with reversals of indemnification assets happening more often than settlements. However, I fail to document an
association between INDEMN and Cash_ETR. The lack of significance may be explained by some indemnified positions
resulting in settlements, suggesting that the documented attenuation could be a product of both reversals and settlements
of indemnified positions.26

Sample Generalizability
Utilizing a sample of firms engaging in only one M&A introduces concerns regarding the external validity of con-
clusions drawn on that sample. However, minimal indemnification disclosures within 10-Ks make it challenging to

26
To further explore this possibility, I reviewed subsequent 10-Ks of all observations that disclosed indemnification assets to determine how these
assets unraveled over time. However, only 8 of 13 potential observations disclosed how their assets eventually unraveled. Specifically, six of these
eight observations eventually reversed more than half of their indemnification assets, consistent with reversals happening more often than
settlements.

The Journal of the American Taxation Association

Volume 46, Number 1, 2024
128 Hopkins

TABLE 8
ETR Multivariate Analysis

(1) (2)
Variables GAAP_ETRt Cash_ETRt
INDEMNt (b1) 20.058 20.005
(21.959) (20.223)
SIZEt1 (b2) 0.005 0.002
(0.539) (0.256)
MTBt (b3) 0.002 0.002
(0.581) (0.670)
INTANGt (b4) 0.133 0.111
(1.884) (1.779)
PPEt (b5) 0.110 0.066
(2.409) (1.541)
EQINCt (b6) 3.270 1.542
(0.842) (0.347)
LEVt (b7) 0.043 0.009
(0.604) (0.144)
ROAt (b8) 0.190 0.030
(1.399) (0.197)
RDt (b9) 0.383 0.381
(1.730) (1.740)
ADVt (b10) 0.938 0.761
(1.975) (1.176)
FCFt (b11) 0.392 0.591
(1.987) (3.016)
NOLCFt (b12) 0.012 0.021
(0.416) (1.305)
Constant 0.444 0.241
(4.694) (2.706)
Observations 495 495
Adjusted R2 0.066 0.131
Year and Industry FE Yes Yes
,, Indicate significance at the 0.10, 0.05, and 0.01 levels, respectively (based on two-tailed tests).
This table presents the results of model (3) using the full sample. The dependent variables in columns (1) and (2) are the cumulative future
five-year GAAP and future five-year cash ETRs, respectively. All continuous variables are Winsorized at the 1 and 99 percent levels. Robust
t-statistics are presented in parentheses.
All variables are defined in Appendix A.

establish indemnification’s presence in a broader sample of firms. With this limitation in mind, I attempt to address
this concern in two ways. First, the choice to provide the standardized, unlogged results in Table 5, column (2) was
partially based on generating results that could be compared with those of Ciconte et al. (2016), which use a larger
population of firms in a non-M&A setting. As stated earlier, I find that UTBs explain 10.1 percent of future tax cash
outflows as do contemporaneous income tax payments, which is comparable with Ciconte et al.’s finding of 9.6
percent.
Additionally, I compare the industry composition of the contract sample with that of the full sample with model
variables (see Table 1). Based on an observation count, the two samples share eight of the top ten industries, which com-
pose 53.33 and 41.99 percent of the contract and full samples, respectively. Furthermore, the six industries not included
in the contract sample compose only 1.84 percent of the full sample. Thus, the industry composition is similar to a larger
population of firms. The similarities with Ciconte et al.’s results and the full sample’s industry composition suggest that
my conclusions may generalize to a non-M&A setting.

The Journal of the American Taxation Association

Volume 46, Number 1, 2024
Tax Indemnification and Unrecognized Tax Benefits in Mergers and Acquisitions 129

Additional Robustness Checks (Untabulated)

Exclusion of Asset and Partnership Contracts
The main results include contracts involving the purchase of assets and partnership interests. However, unless sub-
stantially all of a target’s assets are acquired, acquisition by an asset purchase contract may not result in a requirement
to record the tax positions of the target. Moreover, because of the flow-through taxation nature of partnerships,
acquirers of partnerships would likely not have tax positions to record regarding past tax uncertainties. Accordingly, I
remove these contracts, repeat primary analyses, and continue to find that UTBs of firms without indemnification have
a positive association with future tax cash outflows (p-value < 0.01), and UTBs of those without have no association
with future tax cash outflows (p-value ¼ 0.53).

Inclusion of Additional Control Variables

Numerous firm characteristics likely explain future tax cash outflows, such as R&D intensity, foreign operations
intensity, and intangible investment intensity. Model (1) does not include variables for research intensity, foreign opera-
tions intensity, and intangible investment intensity because the inclusion of contemporaneous taxes paid should account
for other business aspects that could explain future taxes paid (Ciconte et al. 2016). For robustness, I include controls
for these firm characteristics and continue to find that UTBs of firms without indemnification have a positive association
with future tax cash outflows (p-value < 0.01), and UTBs of those without have no association with future tax cash out-
flows (p-value ¼ 0.80).

Exclusion of Public-to-Public Mergers

Because of the complex nature of M&As, the main results are based on a sample of firms engaging in only one
M&A during the current year. This design choice resulted in a control group consisting of several large firms engaging
in public-to-public M&As. As robustness, I remove observations engaging in public-to-public M&As and continue to
find that UTBs of firms without indemnification have a positive association with future tax cash outflows (p-value <
0.01), and UTBs of those without have no association with future tax cash outflows (p-value ¼ 0.80).

Exclusion of Years Affected by the Tax Cuts and Jobs Act of 2017
The Tax Cuts and Jobs Act of 2017 requires firms to remit a transition tax on post-1986 untaxed foreign earnings.
Because of the significant burden created by this tax law change, Congress allowed this tax to be paid over time starting
in 2017, possibly creating spikes in cash taxes paid. Therefore, to ensure the results are not the product of these spikes in
taxes paid, I exclude observations that require 2017–2021 data (years t ¼ 2012–2016) and find that UTBs of firms with-
out indemnification have a positive association with future tax cash outflows (p-value < 0.05), and UTBs of those with-
out have no association with future tax cash outflows (p-value ¼ 0.60).

Exclusion of Observations with Subsequent M&As

M&As occurring in subsequent periods may explain any documented differences in future tax cash flows for firms with
and without indemnification. Therefore, I remove observations with M&As in t+1, t+1 through t+2, and t+1 through t+3
and reperform analyses from Table 5.27 These sample cuts result in sample sizes of 305, 200, and 141, respectively. In the 305
and 200 observation samples, I continue to document an attenuating effect from indemnification (p-values < 0.01 and < 0.05,
respectively). However, the attenuation effect documented with the smallest sample of observations is insignificant (b3 ¼
18.130, p-value ¼ 0.16). This lack of significance is likely due to the loss of power from the sample attrition of 72 percent.

VI. CONCLUSION
Tax indemnification is common in M&A contracts to minimize the acquirer’s exposure to the target’s tax uncertain-
ties, but it is rarely disclosed in the associated financial statements. The lack of disclosure and the accounting for tax
indemnification can be problematic for investors. Specifically, UTBs should be informative of tax risks due to their asso-
ciation with future tax cash outflows. However, this study finds that tax indemnification can completely attenuate this
association between UTBs and future tax cash outflows. Therefore, if investors rely on UTB disclosures to assess the tax
cash outflows of a potential investment, then they could incorrectly assess the potential investment.

27
Removing observations with no M&A in years t+1 through t+4 and t+1 through t+5 results in sample attritions of 82 and 89 percent, respectively.
Due to the low power of these regressions, I exclude discussing them in this study.

The Journal of the American Taxation Association

Volume 46, Number 1, 2024
130 Hopkins

Moreover, I find that practitioners’ concerns about accounting and disclosure for tax indemnification are valid
(PwC 2020). My results suggest that researchers, auditors, and investors should consider the potential impact of tax
indemnification when developing tax risk and expectation models. This study opens a new avenue of research by provid-
ing evidence that tax indemnification is common, rarely disclosed, and consequential.

REFERENCES
Abernathy, J., B. Beyer, A. Gross, and E. Rapley. 2017. Income statement reporting discretion allowed by FIN 48: Interest and
penalty expense classification. Journal of the American Taxation Association 39 (1): 45–66. [Link]
Belsley, D., E. Kuh, and R. Welsch. 2005. Regression Diagnostics: Identifying Influential Data and Sources of Collinearity. New
York, NY: John Wiley & Sons.
Blouin, J., C. Gleason, L. Mills, and S. Sikes. 2010. Pre-empting disclosure? Firms’ decisions prior to FIN no. 48. The Accounting
Review 85 (3): 791–815. [Link]
Bozanic, Z., J. Hoopes, J. Thornock, and B. Williams. 2017. IRS attention. Journal of Accounting Research 55 (1): 79–114. https://
[Link]/10.1111/1475-679X.12154
Cazier, R., S. Rego, X. Tian, and R. Wilson. 2015. The impact of increased disclosure requirements and the standardization of
accounting practices on earnings management through the reserve for income taxes. Review of Accounting Studies 20 (1):
436–469. [Link]
Ciconte, W., M. Donohoe, P. Lisowsky, and M. Mayberry. 2016. Predictable uncertainty: The relation between unrecognized tax
benefits and future income tax cash outflows. (Working paper). [Link]
De Simone, L., J. Robinson, and B. Stomberg. 2014. Distilling the reserve for uncertain tax positions: The revealing case of black
liquor. Review of Accounting Studies 19 (1): 456–472. [Link]
Dhaliwal, D., C. Gleason, and L. Mills. 2004. Last-chance earnings management: Using the tax expense to meet analysts’ fore-
casts. Contemporary Accounting Research 21 (2): 431–459. [Link]
Drake, K. D., R. Hamilton, and S. J. Lusch. 2020. Are declining effective tax rates indicative of tax avoidance? insight from effec-
tive tax rate reconciliations. Journal of Accounting and Economics 70 (1): 101317. [Link]
Financial Accounting Foundation (FAF). 2012. Post-Implementation Review Report on FASB Interpretation No. 48,
Accounting for Uncertainty in Income Taxes. Norwalk, CT: FAF.
Gleason, C., and L. Mills. 2002. Materiality and contingent tax liability reporting. The Accounting Review 77 (2): 317–342. https://
[Link]/10.2308/accr.2002.77.2.317
Gleason, C., L. Mills, and M. Nessa. 2018. Does FIN 48 improve firms’ estimates of tax reserves? Contemporary Accounting
Research 35 (3): 1395–1429. [Link]
Guenther, D., S. Matsunaga, and B. Williams. 2017. Is tax avoidance related to firm risk? The Accounting Review 92 (1): 115–136.
[Link]
Gupta, S., R. Laux, and D. Lynch. 2016. Do firms use tax reserves to meet analysts’ forecasts? Evidence from the pre- and post-
FIN 48 periods. Contemporary Accounting Research 33 (3): 1044–1074. [Link]
Hainmueller, J. 2012. Entropy balancing for causal effects: A multivariate reweighting method to produce balanced samples in
observational studies. Political Analysis 20 (1): 25–46. [Link]
Hanlon, M., E. Maydew, and D. Saavedra. 2017. The taxman cometh: Does tax uncertainty affect corporate cash holdings?
Review of Accounting Studies 22 (3): 1198–1228. [Link]
Heckman, J. 1976. The common structure of statistical models of truncation, sample selection and limited dependent variables
and a simple estimator for such models. Annals of Economic and Social Measurement 5 (Fall): 475–492.
Karlinsky, S., and H. Burton. 2016. Tax professionals’ perception of large and mid-size business tax law complexity. Electronic
Journal of Tax Research 14: 61. [Link]
Large_and_Mid-Size_Business_Tax_Law_Complexity#:~:text=The%20current%20study%20concentrates%20on%20issues%
20related%20to,of%20tax%20complexity%20across%2040%20different%20tax%20issues
Larcker, D., and T. Rusticus. 2010. On the use of instrumental variables in accounting research. Journal of Accounting and
Economics 49 (3): 186–205. [Link]
Lisowsky, P., L. Robinson, and A. Schmidt. 2013. Do publicly disclosed tax reserves tell us about privately disclosed tax shelter
activity? Journal of Accounting Research 51 (3): 583–629. [Link]
Logue, K. D. 2005. Tax law uncertainty and the role of tax insurance. Virginia Tax Review 25 (2): 339–414. [Link]
[Link]/cgi/[Link]?article=2741&context=articles#:~:text=TAX%20LAW%20UNCERTAINTY%20AND
%20THE%20ROLE%20OF%20TAX,and%20public%20insurance%20can%20play%20in%20managing%20it
Maddala, G. S. 1983. Limited-Dependent and Qualitative Variables in Econometrics. Cambridge, U.K.: Cambridge University Press.
Nesbitt, W. 2014. Unrecognized tax benefits: Disentangling the effects of tax aggressiveness and financial reporting discretion.
Doctoral dissertation, University of Georgia.

The Journal of the American Taxation Association

Volume 46, Number 1, 2024
Tax Indemnification and Unrecognized Tax Benefits in Mergers and Acquisitions 131

PricewaterhouseCoopers (PwC). 2020. Income taxes. [Link]

[Link]
Robinson, L., and A. Schmidt. 2013. Firm and investor responses to uncertain benefit disclosure requirements. Journal of the
American Taxation Association 35 (2): 85–120. [Link]
Robinson, L., B. Stomberg, and E. Towery. 2016. One size does not fit all: How the uniform rules of FIN 48 affect the relevance
of income tax accounting. The Accounting Review 91 (4): 1195–1217. [Link]
Rosen, P., and G. Blitz. 2017. Trends and features of transactional liability insurance and its effects on the M&A marketplace.
Proceedings of the American College of Coverage and Extracontractual Counsel 2017 University of Michigan Law School
Symposium, Ann Arbor, MI, October 20. [Link]
[Link]
Towery, E. 2017. Unintended consequences of linking tax return disclosures to financial reporting for income taxes: Evidence
from schedule UTP. The Accounting Review 92 (5): 201–226. [Link]

APPENDIX A
Variable Descriptions

Variables Definition
ADV The ratio of advertising expense (XAD) to ending total assets in year t (AT). If XAD is missing, then XAD is set
to 0.
Cash_ETR The ratio of the summation of future tax cash outflows (TXPD) over the next five years to the summation of
future pre-tax book income (PI) over the next five years (i.e., t+1 through t+5).
CONS The ratio of merger consideration (hand-collected from 10-K disclosures) to ending total assets in year t (AT).
DEFTAX The ratio of net deferred taxes (TXNDB) adjusted for UTBs relating to temporary book-tax differences
(TXTUBEND – TXTUBTXTR) to ending total assets in year t (AT).
EQINC The ratio of earnings in equity (ESUB) to ending total assets in year t (AT). If ESUB is missing, then ESUB is set
to 0.
FCF The ratio of net cashflows from operations (OANCF  CAPX) to ending total assets in year t (AT).
FOREIGN The ratio of foreign pre-tax income (PIFO) to total pre-tax income in year t (PI). If PIFO is missing, then PIFO is
set to 0.
FUT_TXPD The log of the summation of future tax cash outflows (TXPD) over the next five years (i.e., t+1 through t+5) to
ending total assets in year t (AT).
GAAP_ETR The ratio of the summation of future tax expense (TXT) over the next five years to the summation of future pre-
tax book income (PI) over the next five years (i.e., t+1 through t+5).
INDEMN Indicator variable equal to 1 if the observation obtained indemnification through M&A contract, 0 otherwise.
INTANG The ratio of total intangible assets (INTAN) to ending total assets in year t (AT). If INTAN is missing, then
INTAN is set to 0.
LEV The ratio of debt (DLTT + DLC) to ending total assets in year t (AT). If DLTT is missing, then DLTT is set to 0.
MAC_IND Indicator variable equal to 1 if the acquirer and target are in different macro industries, 0 otherwise (according to
the Thomson Reuters SDC Mergers and Acquisitions database).
MTB The ratio of market value (PRCC_F  CSHO) to book value of equity in year t (CEQ).
NOLCF The ratio of net operating loss carryforwards (TLCF) to beginning total assets (AT) in year t.
PPE The ratio of gross plant, property, and equipment (PPEGT) to ending total assets in year t (AT). If PPEGT is
missing, then PPEGT is set to net plant, property, and equipment (PPENT). If PPEGT and PPENT are
missing, then PPEGT is set to 0.
D_PTBI The ratio of the difference in pre-tax book income in year t and pre-tax book income in year t1 to ending total
assets in year t (AT).
ROA The ratio of pre-tax book income (PI) to ending total assets in year t (AT).
RD The ratio of total research and design expenses (XRD) to ending total assets in year t (AT). If XRD is missing,
then XRD is set to 0.
SIZE The log of ending total assets in year t (AT).
UTB The ratio of reported unrecognized tax benefits (TXTUBEND) to ending total assets (AT) in year t.
TXPD The ratio of taxes paid (TXPD) to ending total assets in year t (AT).
D_TXPD The ratio of the difference between taxes paid in year t and taxes paid in year t1 to ending total assets in year t (AT).

The Journal of the American Taxation Association

Volume 46, Number 1, 2024
132 Hopkins

APPENDIX B
M&A Contracts and Excerpts from M&A Contracts on Indemnification

Indemnification Language
I review the indemnification and tax matters sections of M&A contracts to identify tax indemnification clauses.
These contracts describe tax indemnification in various ways. For instance, the language in the stock purchase agree-
ment between Forward Air Corporation, TQI Holdings, Inc., and various sellers discloses within the tax matters section
that “Each Entity has filed…all…material Tax Returns…All such Tax Returns were correct and complete in all mate-
rial respects. All Taxes owed by the Entities…have been paid.” In other words, the seller represents that all taxes associ-
ated with the target were appropriately done and paid. Later in the tax matters section, the sellers agree that “the Sellers
shall…indemnify the Entities and Buyer…against, and protect, save and hold harmless each Indemnified Taxpayer
from, any and all [Losses] resulting from: (1) except to the extent reflected in the calculation of Closing Date Working
Capital, any Taxes of any Entity allocable to any period ending on or prior to the Closing Date…(4) any misrepresenta-
tion or breach of any representation, warranty or obligation set forth in this [section].” Simply put, if the sellers misrep-
resented something to the buyer and it results in a loss, then the sellers would make the buyer whole again.
Contracts often include separate discussions of indemnification that specifically address taxes. However, the indem-
nification of tax representations normally protects the buyer against all tax losses associated with past positions, includ-
ing complex issues such as transfer pricing. Therefore, I concentrate on the indemnification of tax representations
instead of specific tax indemnification clauses.
Contracts that do not provide indemnification are not always obvious. In a straightforward case, the contract speci-
fies that any representations about the target will not survive the closing. For instance, refer to the agreement and plan
of merger below between Visa and Cybersource. Cybersource states within the “General Provisions” section that “The
representations and warranties of the Company, Parent and Merger Sub contained in this Agreement…shall terminate
at the Effective Time [of the M&A].” In other words, any misrepresentation of the target that results in a loss for Visa
will not be reimbursed by Cybersource, unless Cybersource was fraudulent in their representations or a separate agree-
ment exists within the contract, which bypasses this language. When this language is not present, an external party
would have to review a significant portion of the contract to determine whether indemnification is present. See below for
contract excerpts.

Example of Indemnification Language

Stock Purchase Agreement between Forward Air Corporation, TQI Holdings, Inc., and the “Sellers” Named Herein
(Dated March 4, 2013)
Tax Matters:
8.1 Representations and Obligations Regarding Taxes. The Company hereby represents and warrants to and agree
with Buyer that, except as set forth on Schedule 8.1 hereto: (a) Each Entity has filed all federal income and all
other material Tax Returns that it has been required to file. All such Tax Returns were correct and complete in all
material respects. All Taxes owed by the Entities (whether or not shown on any Tax Return and whether or not
any Tax Return was required) have been paid. Since March 7, 2008, no claim has been made by a taxing author-
ity in a jurisdiction where any Entity does not file Tax Returns that such Entity is subject to taxation by that
jurisdiction.
8.2 Indemnification for Taxes. (a) From and after the Closing, the Sellers shall severally (in accordance with each
Seller’s Proportionate Share) indemnify the Entities and Buyer (each herein sometimes referred to as an
“Indemnified Taxpayer”) against, and protect, save and hold harmless each Indemnified Taxpayer from, any and
all out-of-pocket damages, deficiencies, losses and reasonable expenses, including, without limitation, reason-
able attorneys’, accountants’ and experts’ fees and disbursements (all herein referred to as “Losses”) resulting
from: (1) except to the extent reflected in the calculation of Closing Date Working Capital, any Taxes of any
Entity allocable to any period ending on or prior to the Closing Date or, as provided in Section 8.3(c) hereof, allo-
cable to any period that begins before and ends after the Closing Date; (2) any Tax of any Person other than an
Entity for periods ending on or before the Closing Date imposed upon any Entity as a result of the Entity being
included prior to the Closing Date in a combined, consolidated or unitary Tax group under Treasury
Regulation Section 1.1502-6 (or any similar provision of the applicable law of any Governmental Authority) or,

(continued on next page)

The Journal of the American Taxation Association

Volume 46, Number 1, 2024
Tax Indemnification and Unrecognized Tax Benefits in Mergers and Acquisitions 133

APPENDIX B (continued)

as a transferee or successor, by contract or otherwise, except for a Contract the principal purpose of which is not
to indemnify or pay the Taxes of another person; (3) the failure to pay state income Taxes or state franchise
Taxes in any state other than the State of Michigan prior to the Closing Date but not thereafter (the “Potential
State Tax Matters”); (4) any misrepresentation or breach of any representation, warranty or obligation set forth in
this Article VIII; or (5) any real property transfer, recordation or similar tax imposed by the State of Michigan
with reference to real property owned by QSX.

Example of No Indemnification Language

Agreement and Plan of Merger between Visa, Inc. and Cybersource, Corp. (Dated April 20, 2010)
Representations and Warranties of the Company:
2.6 Taxes.
(b) Taxes, Tax Returns and Audits.
(i) The Company and each of its Subsidiaries have (1) duly and timely filed or caused to be filed all Tax Returns
and such Tax Returns are true, correct, and complete in all material respects, (2) duly and timely paid or with-
held (and timely paid over any withheld amounts to the appropriate Governmental Entity) all Taxes required
to be paid or withheld whether or not shown as due on any Tax Return, and (3) established reserves in accor-
dance with GAAP that are adequate for the payment of all Taxes not yet due and payable with respect to the
assets and operations of the Company and each Subsidiary through the date of this Agreement.
General Provisions:
8.1 Non-Survival of Representations and Warranties. The representations and warranties of the Company, Parent
and Merger Sub contained in this Agreement, or any instrument delivered pursuant to this Agreement, shall ter-
minate at the Effective Time, and only the covenants that by their terms survive the Effective Time and this
Article VIII shall survive the Effective Time.

APPENDIX C
Illustrative Example of Accounting under ASC 805 and Potential ETR Issues
ASC 805 addresses various aspects of business combinations, including recording indemnified tax positions. Within
the context of business combinations, indemnified tax positions are acquired contingent tax liabilities in which the seller
has agreed to hold the acquirer harmless. ASC 805 requires mirror accounting for indemnified liabilities; specifically,
ASC 805 states, “the acquirer shall recognize an indemnification asset at the same time that it recognizes the indemnified
item.” For an illustrative example, consider the accounting for a $100,000,000 indemnified tax position under ASC 805.
Firm A acquires Subsidiary T from Firm S. Prior to the acquisition, Subsidiary T maintains a $100,000,000 UTB for
uncertain tax positions that are currently susceptible to authoritative challenges. To sweeten the deal, Firm S has agreed
to indemnify Firm A for all taxes, interest, and penalties associated with successful challenges to Subsidiary T’s uncer-
tain tax positions. Although Firm S is contractually responsible for the uncertain tax positions, Firm A is still required,
under ASC 805, to reflect Subsidiary T’s uncertain tax positions in its consolidated financial statements. Therefore, Firm
A will account for the M&A as follows:

(1) Other M&A Assets $###,###,###

Tax Indemnification Asset $100,000,000
UTB $100,000,000
Other M&A Liabilities $###,###,###
Consideration Paid for M&A $###,###,###

(continued on next page)

The Journal of the American Taxation Association

Volume 46, Number 1, 2024
134 Hopkins

APPENDIX C (continued)

Although recognizing these indemnified tax positions increases Firm A’s assets and liabilities, Firm A will not experience
any net effect on equity or income.
When the tax position is successfully challenged and a tax payment is due or when the statute of limitations expires
for the tax position, the tax indemnification asset and UTB associated with that position must be eliminated. Let us con-
tinue the example. Subsidiary T loses a challenge from a tax authority on the indemnified tax position. Firm A pays the
settlement on behalf of Subsidiary T and submits a reimbursement request per the terms of the indemnification agree-
ment with Firm S. Firm S approves the reimbursement request and submits the payment to Firm A. The entries are as
follows:

(2) UTB $100,000,000

Cash $100,000,000
(3) Cash $100,000,000
Tax Indemnification Asset $100,000,000

These entries result in removing the asset and liability representing the indemnified tax position without affecting Firm
A’s equity or income.
After accounting for the reimbursement, Firm A needs to prepare its statement of cash flows. ASC 230 provides
standards for preparing the statement of cash flows and requires that “income taxes paid during the period shall be dis-
closed.” In this example, Firm A has technically paid $100,000,000 to the tax authority but effectively paid nothing due
to the reimbursement. ASC 230 does not address this situation. Therefore, Firm A may reflect or net the $100,000,000 in
their income tax disclosure requirement. The former is misleading since they did not effectively pay the tax, and the latter
makes UTBs less informative.
It is also possible that the tax indemnification accounts may need to be adjusted without any actual cash flows due
to the expiration of statutes of limitations associated with indemnified positions or due to specific provisions within the
M&A contract. Unlike other indemnified liabilities, adjusting indemnified tax liabilities results in changes to both pre-
tax income and tax expense. When tax statutes expire, ASC 740 standards coupled with SEC regulations result in the
recognition of a tax benefit without the ability to offset the recognized tax benefit with the reversal of the tax indemnifi-
cation asset. Therefore, the tax indemnification asset must be adjusted through a pre-tax account, whereas the UTB
must be adjusted through tax expense. Now, return to the earlier example and assume the position was never challenged.
After three years, the statute of limitations expires on the uncertain tax position taken by Subsidiary T. In response,
Firm A removes the tax indemnification asset and associated UTB from their financial records. Their entries are as
follows:

(4) UTB $100,000,000

Tax Benefit $100,000,000
(5) Other Expense $100,000,000
Tax Indemnification Asset $100,000,000

These entries result in removing the asset and liability representing the indemnified tax position without affecting Firm
A’s equity or overall income.
However, the above accounting does influence pre-tax income and tax expense, creating a permanent book-tax dif-
ference and distorting GAAP and cash ETRs.28 Assume that prior to removing the UTBs associated with the indemni-
fied tax position, Firm A’s consolidated financial statements reflected pre-tax book income of $2,500 million, GAAP tax
expense of $525 million, and cash taxes paid of $525 million, which results in GAAP and cash ETRs of 21.00 percent.

(continued on next page)

28
A similar difference is created if the firm has a settlement that is more or less than the booked UTB reserve. If the firm settled for more than the
reserve and it was fully indemnified, it would record increases to the UTB reserve and a tax indemnification asset in the period in which it becomes
evident that the reserve is not adequate per ASC 740-10-25-15. Per SEC regulations, mirror accounting is not allowed to adjust the reserve.
Therefore, firms would record an increase in other income (to increase the indemnification asset) and tax expense (to record the higher settlement).

The Journal of the American Taxation Association

Volume 46, Number 1, 2024
Tax Indemnification and Unrecognized Tax Benefits in Mergers and Acquisitions 135

APPENDIX C (continued)

By removing the $100 million UTB reserve and the $100 million tax indemnification asset, Firm A’s consolidated finan-
cial statements reflect GAAP tax expense of $425 million (decreased), cash outflows related to income tax of $525 mil-
lion (unchanged), and pre-tax book income of $2,400 (decreased), which results in a GAAP ETR of 17.71 percent and a
cash ETR of 21.88 percent. Therefore, the GAAP ETR is deflated, and the cash ETR is inflated.

The Journal of the American Taxation Association

Volume 46, Number 1, 2024
Copyright of Journal of the American Taxation Association is the property of American
Accounting Association and its content may not be copied or emailed to multiple sites or
posted to a listserv without the copyright holder's express written permission. However, users
may print, download, or email articles for individual use.