Private Equity Fund Reporting Quality and External Monitors, Third-party Service

Providers, and Fund Attributes

Peter Easton
Mendoza College of Business, University of Notre Dame
(574) 631-6096
peaston@nd.edu

Stephannie Larocque
Mendoza College of Business, University of Notre Dame
(574) 631-6136
larocque.1@nd.edu

Paul Mason
Hankamer School of Business, Baylor University
(254) 710-6129
p_mason@baylor.edu

Steven Utke
School of Business, University of Connecticut
(860) 486-2374
sutke@uconn.edu

March 2, 2023

Keywords: Private Equity, Monitoring, Fair Value, Net Asset Values, Financial Reporting,
Valuation Specialists

JEL codes: G1, G14, G30, M4, M41

Data Availability: Data used in this study are available from public sources listed in the paper.

_______________
We gratefully acknowledge the support of our respective universities. We thank several current and
former private equity fund employees and consultants for helpful discussions. We appreciate helpful
comments from Brian Baik, Maria Borysoff, Mary Ellen Carter, John Donovan, John Heater, Rob Raney,
Sophie Shive, Jennifer Stevens, and workshop participants at the Institute for Private Capital’s 2022
Spring Research Symposium, Lancaster University, University of Bristol, University of Connecticut, and
University of Notre Dame.
 Private Equity Fund Reporting Quality and External Monitors, Third-party Service
Providers, and Fund Attributes

Abstract
We provide descriptive evidence about variation in the reporting quality of private equity funds,
in particular buyout and venture capital funds, relative to external monitoring, third-party service
providers, and fund attributes. We find limited evidence that investor monitoring is associated
with private equity fund reporting quality, measured as the accuracy and bias of funds’ reported
net asset values (NAVs). However, we observe that fund NAV accuracy and bias vary with
auditor involvement and the use of third-party service providers (i.e., valuation specialists,
marketers, and administrators). Further, these relations differ across fund types and often differ
from evidence from public markets. We provide important evidence for investors and regulators
especially as private equity supersedes public markets as the main vehicle to raise capital.
 I. Introduction

We provide descriptive evidence regarding the factors associated with private equity (PE)

fund reporting quality. While drivers of reporting quality, such as monitoring by investors and

auditors, have been studied in public firms, PE funds are fundamentally different. These

differences include PE’s sophisticated investor base, long-term contracts with investors, distinct

agency issues faced by PE fund managers, and PE’s difficulty in valuing their assets, which

qualify as “Level 3” category assets under fair value accounting. Thus, it is an empirical question

as to what attributes affect PE fund reporting quality and the direction of any relations. The

evidence of bias and inaccuracy in PE fund net asset values (NAVs) in Jenkinson, Landsman,

Rountree, and Soonawalla (2020) forms an important foundation for future accounting research

in the emerging and poorly understood PE arena (e.g., Borysoff, Mason, and Utke 2022). We

extend Jenkinson et al. (2020) with evidence of how PE fund reporting quality, measured using

the bias and accuracy of PE fund NAVs, varies relative to external monitoring, the use of third-

party service providers, and fund attributes. Our evidence is largely descriptive, in that we seek

to better understand private equity reporting, and refrain from making causal inferences (Gow,

Larcker, and Reiss 2016).

PE funds – buyout (BO) and venture capital (VC) funds for our study – comprise an

increasingly important sector of the overall economy. The amount of capital raised by PE funds

has grown by 12 percent per year since 2010 (Zimmerman 2016). Global PE assets under

management were $3.9 trillion in 2019 (McKinsey 2020) and are expected to exceed $9 trillion

by 2025 (Joyce 2020). Further, newly raised private capital exceeded new public capital for the

first time in 2017, allowing firms to forgo initial public offerings (e.g., Witte and Brown 2019;

Ewens and Farre-Mensa 2020). PE growth has been fueled, in part, by the promise of high rates

1
of return for investors. This growth correspondingly increases the importance of understanding

the quality of financial information provided by PE funds, as indicated by recent Securities and

Exchange Commission (SEC) attention in the area (e.g., Gensler 2022).

Over a PE fund’s life (prior to liquidation), fund returns are based on the fund’s cash

distributions and the reported fair value of the PE fund’s investments, known as the fund’s NAV.

NAVs provide decision-relevant information to PE investors, both because existing investors

incorporate the reported fair value of their PE fund investments in their own financial statements

(Easton, Larocque, and Stevens 2021) and because existing and potential investors look to the

current fund’s performance in deciding whether to invest in the PE fund manager’s next fund

(Barber and Yasuda 2017). Further, PE NAVs provide important information for investors’ asset

allocation decisions and understanding of the risk and return of their investments (see Longstaff

2001, Gârleanu 2009, and Ang et al. 2014). 1 As such, accurate and unbiased NAVs are important

given that PE funds are private, otherwise opaque entities with limited disclosure and regulation.

An advantage of using the PE setting to identify factors associated with financial

reporting quality is that we can observe ex post cash flows over the life of the fund. We use fund-

level NAV and cash flow data from Preqin to calculate quarterly NAV bias (i.e., signed error)

and accuracy (i.e., the unsigned error) based on the difference between each fund’s reported

NAV (i.e., management’s estimated value) and its DCF (i.e., based on actual net cash flows),

following Jenkinson et al. (2020). 2 We supplement the NAV and DCF data with fund-level

information collected from Form ADVs, filed with the SEC, for 4,502 fund-quarters from 2009

1
Pensions funds and university endowments, for example, increased their asset allocation to private assets, or
‘alternatives’ such as PE, from about 7% to over 25% over the last 20 years (Thinking Ahead Institute 2021).
Discussions with these ‘asset allocators’ indicate that inaccurate or biased NAVs will distort allocations for
alternatives and, correspondingly, tilt the balance of their portfolio in the offsetting direction.
2
Following Jenkinson et al. (2020), we use the term DCF rather than NPV throughout our study.

2
to 2017. Form ADV information includes investor ownership, use of a Big 4 or non-Big 4

auditor, and use of a valuation specialist, a marketer, or a fund administrator. We divide the

attributes that may affect the bias and accuracy of PE fund NAVs into three categories – external

monitoring, the use of third-party service providers, and fund attributes. We separately examine

BO and VC funds (e.g., Jenkinson et al. 2020) because these funds differ in several ways, which

could lead to variation in the relation between each attribute and NAV accuracy and bias. For

example, VC funds tend to take smaller stakes in newer, riskier underlying investments whereas

BO funds tend to own controlling interests in more mature firms (e.g., Gompers and Lerner

2001; Kaplan and Strömberg 2009). Moreover, VC funds tend to report NAVs that are more

optimistically biased whereas BO funds tend to report more pessimistic NAVs relative to their ex

post discounted cash flows (Jenkinson et al. 2020).

While the external monitors we examine (i.e., investors and auditors) have been studied

in public firms, PE funds are fundamentally different along multiple dimensions. For example,

PE funds: 1) are primarily equity-financed by sophisticated “accredited” investors, who may be

skilled at monitoring, but face little or no external monitoring from creditors; 3 2) lock investors

into long-term contracts, preventing governance through “exit” (Edmans 2014); 3) are entirely

controlled by fund managers with few, if any, voting rights for external investors regardless of

ownership levels, limiting governance through “voice” (Black 1992); and, 4) face distinct agency

issues between the funds and external investors due to, for example, funds’ unique compensation

structures, which may provide explicit and implicit incentives to bias NAVs (e.g., Barber and

Yasuda 2017; Brown, Gredil, and Kaplan 2019; Easton et al. 2021; see Section 2). 4 Further,

3
PE investments are often debt-financed, but those creditors are not involved with the PE fund, only the underlying
investment. For example, Toys “R” Us, not its PE owners, went bankrupt; the PE funds did not carry the debt.
4
Implicit incentives refer to funds’ incentives to report strong performance to aid in raising capital for future,
follow-on funds. Explicit incentives refer to pay-based incentives, such as carried interest and management fees.

3
because PE funds exist for finite periods of time (Kaplan and Strömberg 2009), some external

investors may focus on cash distributions from liquidated investments rather than interim NAVs.

External investors may also focus less on NAVs due to the private information channels that

exist between investors and PE fund managers (i.e., Reg FD does not apply). Moreover, insiders’

influence on NAVs may vary depending on their incentives. For instance, if insider ownership is

higher, insiders may see less need to manipulate NAVs since fewer investors may be influenced

by the manipulation. Conversely, funds with high insider ownership may be more willing and

able to manipulate NAVs to aid in future fundraising efforts given lower external monitoring.

Monitoring by auditors may also differ for PE funds compared to public firms given that

funds are private, with lower reputation/litigation costs and greater incentive to cater to PE

investors, and have difficult-to-audit/monitor fair values (e.g., Ayres et al. 2019; Carcello, Neal,

Reid, and Shipman 2020). Importantly, the unique PE setting generally provides ambiguous

predictions regarding investor and auditor monitoring, unlike public firm settings where, for

example, it is well known that auditors and certain institutional owners mitigate misreporting.

In examining the association between external monitors and PE funds’ financial

reporting, we find only limited evidence of associations between external investors and NAV

bias or accuracy for both BO and VC funds. These results stand in striking contrast to the

evidence that external oversight affects reporting in public firms (e.g., Cornett, Marcus, and

Tehranian 2008), as well as evidence that investor oversight affects PE fund financial reporting

choices (Gaver, Mason, and Utke 2020). 5 With respect to auditors’ oversight role, we observe

greater NAV accuracy in large BO funds with Big 4 auditors and in VC funds with non-Big 4

auditors. In addition, Big 4 auditors are associated with higher bias (i.e., less pessimism) in large

5
One concern with our inability to document monitoring effects is low power arising from splitting the sample into
BO and VC funds. To reduce this concern, we re-run our analysis in the full sample of funds. Results are similar.

4
and higher-performing BO funds, while non-Big 4 auditors are associated with lower bias (i.e.,

less optimism) in VC funds. Our evidence is consistent with matching between BO funds and

Big 4 auditors and smaller VC funds and non-Big 4 auditors (cf. Mason and Utke 2020). This

adds to the mixed evidence on Big N effects for private entities (e.g., Chaney, Jeter, and

Shivakumar 2004 versus Che, Hope, and Langli 2020) and differs from the well-known Big N

effect for public firms (e.g., Defond, Erkens, and Zhang 2017). Our findings are especially

important as regulators debate whether PE funds should be required to obtain an audit, as in

recent SEC proposals (e.g., Gensler 2022), and for PE funds selecting auditors.

We next explore whether the use of third-party service providers influences NAV

reporting. Importantly, we can observe third-party providers that are generally unobservable in

other settings. 6 Directly related to NAVs, funds can engage valuation specialists to prepare fair

value estimates. Funds can also employ administrators to handle the routine, back-office

assembly of financial statements. Finally, funds can engage marketers to help attract investors in

fundraising for new funds, which could provide an incentive to upwardly bias NAVs for existing

funds. Section 3 discusses possible ways each third-party service provider can affect managers’

incentives and/or NAV reporting.

We observe a positive association between third-party valuation specialists and NAV

accuracy in BO funds, but little evidence for VC funds. We also find limited evidence of

valuation specialists affecting NAV bias. These results contrast with evidence from the insurance

industry (Hanley et al. 2018), where valuation specialists constrain managers’ inflation of fair

values. This could be due to PE’s fundamental differences from insurance; for example, because

PE fund managers have deep knowledge of the assets (i.e., businesses) they control, external

6
See Hanley, Jagolinzer, and Nikolova (2018) and Koo, Sivaramakrishnan, and Zhao (2020) for exceptions specific
to insurance companies.

5
specialists may add little value. We further find that the use of third-party administrators is

positively (negatively) associated with NAV accuracy in VC (BO) funds. We also find that the

use of third-party marketers is positively related to bias in BO funds and top-performing VC

funds. This evidence suggests that a fund may increase NAVs as part of its marketing strategy.

These findings are important to investors selecting among PE funds. 7

Finally, we examine funds’ propensity to adjust their reported NAVs. We examine how

external monitoring, the use of third-party service providers, and fund attributes are associated

with either positive or large (i.e., larger than 10 percent) valuation adjustments as compared to

NAVs reported in the previous quarter. For both BO and VC funds, we find evidence that large

NAV adjustments are more likely in audited quarters (by both Big 4 and non-Big 4 auditors).

This suggests that auditors induce adjustments, even though we find mixed evidence as to

whether and how these adjustments affect accuracy or bias. Positive adjustments are also more

likely for BO funds using valuation specialists, suggesting specialists may cater to clients or

funds may shop for favorable reports.

Our study makes several contributions to the literature. First, we extend the literature

with descriptive evidence about the external monitoring of PE funds, where both managers and

investors have vastly different rights and incentives compared to heavily regulated and

informationally transparent public firms. We find limited evidence that investors – typical

external monitors in the public equity setting – improve PE fund reporting on average. We also

7
In Section 6.1, we discuss our supplemental examination of the determinants of funds’ use of third-party valuation
specialists, administrators, and marketers. While only a small percentage of fund assets are valued by third-party
specialists (2.2 [1.6] percent on average for BO [VC]), specialists are increasingly important for BO funds; the
percentage increases from 2.1 percent in 2009 to 4.1 percent in 2017, reaching as high as 7.1 percent in 2016.
However, the percentage of assets valued by specialists has declined over time for VC funds. Meanwhile, the use of
administrators and marketers increased over our sample period for VC funds but fluctuated for BO funds. With a
few exceptions, we find little evidence that external monitors and fund attributes affect the decision to use any of
these third-party service providers.

6
extend audit research to PE funds, where nearly all assets are reported as difficult-to-audit Level

3 fair values. We observe differences in the impact of Big 4 / non-Big 4 auditors on the accuracy

and bias of BO versus VC funds. Finally, we add to the sparse literature on how third-party

service providers affect fair value estimates, also finding variation in effects for bias and

accuracy by fund type and service provider.

Second, we add to the broad literature on fair value accounting. This is particularly

important because accounting research related to new valuation issues, including how fair values

are estimated, is incomplete (Kaplan 2011). Fair value accounting is complicated, difficult to

audit, and subject to discretion (e.g., Griffith, Hammersley, and Kadous 2015; Ayres et al. 2019).

PE funds follow fair value accounting rules, allowing us to identify several attributes that affect

the accuracy and bias of PE fund fair value NAVs, and therefore the quality of financial

information provided to PE investors.

Finally, our study is increasingly important as PE funds’ economic significance grows

and as private markets open to more investors, supplanting public markets. PE fund investments

are typically complex, illiquid, and difficult to value (Phalippou 2009), raising regulatory

concerns regarding NAV reporting (Lattman 2012; Gensler 2022). By identifying attributes

associated with greater or less bias and inaccuracy in NAVs, we provide PE fund managers and

investors, as well as regulators, with important new information to guide their decision-making.

II. Setting

PE funds operate as financial intermediaries. Fund managers, known as general partners

(GPs), raise capital in the form of commitments from external investors, known as limited

partners (LPs). GPs or their affiliates (i.e., insiders) can also commit capital to each fund. Once

the final total amount of capital is committed, funds “close” and then “call” the committed

7
capital from investors when they are ready to deploy it. 8 The called capital is typically invested

by GPs purchasing interests in other entities, known as portfolio companies. BO funds typically

buy controlling interests while VC funds purchase smaller stakes. 9

Importantly, we conduct our study at the fund-level, the level at which GPs raise capital

from LPs and where NAV reporting is relevant to external fund investors (i.e., LPs), and not at

the portfolio company-level, where the PE fund invests the raised capital into portfolio

companies. 10 In order for investors to evaluate and account for their fund-level investments on an

interim basis, PE funds provide investors with financial information in the form of fund NAVs.

In addition to external investors, numerous other parties are interested (as well as involved) in

the preparation, review, and receipt of NAVs, including fund managers, auditors, third-party

valuation specialists, and third-party fund administrators.

As with other financial intermediaries, PE funds face agency conflicts between external

investors and fund managers at the fund-level due to asymmetric information between these

parties. 11 Because PE funds, unlike other financial intermediaries, are relatively opaque, external

investors generally must rely heavily on the fund’s reported NAV in evaluating fund

performance and value, with few other sources of information. 12 Under ASC 820, funds record

8
Closed funds do not accept new capital commitments or LPs. It is difficult, but not impossible, for LPs to exit (e.g.,
Mason and Utke 2022). Separately, capital calls generally occur over time and cannot exceed the commitment.
9
Borysoff et al. (2022) provide a detailed overview of PE fund structures and operations.
10
Using mutual funds as a poor analogy to PE – because mutual funds are public and regulated – the PE fund
mirrors the mutual fund, whereas portfolio companies mirror the mutual fund’s holdings. All investor-to-PE fund
interaction occurs at the fund-level, similar to investor-to-mutual fund interaction. Even heavily regulated mutual
funds provide only limited data about holdings to investors (i.e., shares and value, not financial statements); PE
funds may provide no portfolio company-level information, making the fund-level the appropriate unit of analysis.
11
There is also agency conflict at the portfolio company-level (Phalippou 2009; Borysoff et al. 2022), which PE
funds mitigate through active roles in the operating, accounting, financial, and governance activities of portfolio
companies (e.g., Jensen 1986, 1989; Gompers and Lerner 2001; Hellman and Puri 2002; Kaplan and Strömberg
2004, 2009; Katz 2009; Badertscher, Katz, and Rego 2013; Zimmerman 2016; Gilson and Gordon 2019; Inderst and
Vladimirov 2019). We focus on the fund-level agency conflict, where NAV reporting is pertinent.
12
While recent regulations (e.g., the 2010 Dodd-Frank Act) intend to increase PE fund transparency (Gaver et al.
2020), this limited oversight does not necessarily improve NAV reporting. For example, recall that investors may
have limited ability or incentive to influence fund behavior and already have direct access to management.

8
their underlying portfolio companies at fair value rather than consolidating them for financial

reporting purposes (regardless of ownership level). Funds’ fair value accounting typically

involves Level 3 assets, where the fund must use valuation models to determine valuations since

there is no established market. Further, (non-individual) LPs typically include the reported fair

value of their PE fund investments in their own financial statements under the ASU 2009-12

amendment to ASC 820, commonly known as the “practical expedient” (Easton et al. 2021). In

sum, interim NAV estimation can be difficult, leading to inaccuracy and/or bias, yet NAVs

provide investors with critical financial data regarding fund performance and other specific

information required to calculate their performance and complete their own financial statements.

Inaccurate or biased NAVs are also important to LPs in their asset allocation decisions. That is,

misreported NAVs may lead LPs to inaccurately believe they are under- or over-allocated to

their PE fund asset class, which will affect their future investment decisions.

Fund managers face various incentives when estimating NAVs, and these incentives

fundamentally differ from public firms’ incentives to manage earnings. For instance, fund

managers have implicit incentives to report favorable past performance to raise capital in

subsequent follow-on funds. Relatedly, managers have explicit incentives to upwardly bias

NAVs due to the typical compensation structure of PE funds. Funds often have a “2 and 20”

compensation structure where manager compensation consists of a fixed component equal to 2

percent of the capital committed to the fund and a variable, or incentive component, equal to 20

percent of the profits of the fund above a set benchmark and known as ‘carried interest’ (Kaplan

and Strömberg 2009). 13 As a result of this compensation structure, managers have an incentive to

inflate (i.e., upwardly bias) NAV values, which can improve their ability to raise future funds so

13
While the specific details of the compensation can vary slightly across funds, the general structures – and more
importantly, the relevant incentives – are similar across funds. See Borysoff et al. (2022) for additional details.

9
as to increase managers’ compensation (e.g., Barber and Yasuda 2017; Brown et al. 2019; Baik

2022). 14 At the same time, overstated NAVs may lead LPs – as allocators – to believe they are

over-allocated to the PE asset class, requiring them to decrease future capital commitments. We

highlight that PE fund managers’ incentives typically relate to raising future funds, unlike public

firm managers who receive compensation and stock price benefits at their current firm. As such,

PE fund (public firm) misreporting incentives potentially harm future (current) investors.

III. Prior Literature and Hypothesis Development

We consider three general categories that may affect NAV accuracy and bias: external

monitoring, the use of third-party providers, and fund attributes.

3.1 External Monitoring

We begin by focusing on the external monitoring of PE funds. External monitors play a

key role in improving reporting quality for public firms (e.g., Rajgopal and Venkatachalam 1997;

Becker, Defond, Jiambalvo, and Subramanyam 1998; Francis, Maydew, and Sparks 1999; Gaver

and Paterson 2001; Ramalingegowda and Yu 2012) and potentially play a similar role for PE

NAVs. 15 We examine two external monitors that can influence the bias and accuracy of interim

NAVs: external (and conversely, insider) owners and Big 4 and non-Big 4 auditors.

Funds’ ownership structure could affect the accuracy and bias of NAVs. First, as insider

ownership by GPs increases, the incentives of investors and managers could be more closely

aligned (e.g., Denis, Denis, and Sarin 1997), potentially reducing bias in NAV reporting or

increasing managers’ efforts to estimate accurate NAVs. Further, the larger the insider

14
While most research finds bias in NAVs, some evidence suggests that NAVs closely align with accounting
fundamentals and market prices and are conservative on average (Ferreira et al. 2019)
15
Tangentially related research examines closed-end fund discounts in the secondary market (Grullon and Wang
2001; Lawrence, Siriviriyakul, and Sloan 2016). However, this research does not focus on accuracy and bias. Our
study is also somewhat related to management forecast research (a large body of work beginning with Foster 1973)
in that NAVs are similar to management’s expectation of fair value while DCF represents the actual, ex post, value.
Notably, our setting lacks the equivalent of an analyst’s forecast.

10
ownership, the fewer external investors that can be affected by NAV bias, reducing managers’

incentives to bias NAVs. However, because insider ownership in funds is often relatively small,

the incentive alignment effect may be small. To the extent that fund managers’ ownership is

relatively low and their income primarily stems from successful fundraising and management

fees, insider ownership may not affect NAV reporting. Finally, higher insider ownership may

encourage managers to exploit the (lower) external ownership, or otherwise reduce monitoring

and insiders’ care in developing accurate, unbiased NAVs.

External investors can also affect the accuracy and bias of reported NAVs. Agency costs

increase as the proportion of external ownership increases (Jensen and Meckling 1976) and, as

such, external investors demand financial information from PE funds (Gaver et al. 2020). To

ensure insiders do not extract private benefits at their expense, LPs (e.g., pensions, endowments)

likely monitor fund managers (e.g., Cumming and Walz 2010), potentially reducing GPs’ ability

to bias NAVs. If external investors effectively monitor NAVs, accuracy and bias in NAVs

should improve in the level of LP ownership. However, unlike public firms where investor

monitoring is well established, monitoring of PE funds may be too costly due to large

information asymmetries (Johan and Zhang 2020) and the lack of regulatory oversight. 16

Importantly, PE investor voting rights are minimal or non-existent and there is essentially no

“exit” option for PE fund investors. While investors can exit in infrequent cases (typically only

with the permission of the GP) they must find a replacement investor. Further, exits do not affect

the “market price” of the fund because there is no public market, and exit has no direct effect on

16
PE funds are often solely equity financed, and thus lack monitoring from creditors. Further, numerous studies find
that IRS monitoring improves financial reporting quality (e.g., Mason and Williams 2020; Zhao 2021), but PE funds
face little or no IRS monitoring (e.g., Drucker and Hakim 2021)

11
managers’ wealth. 17 Finally, LPs may prefer overstated NAVs because these overstatements can

positively influence the LPs’ reported investment performance and financial statements.

The type of external investor may affect the success of monitoring efforts. For instance,

sophisticated investors, specifically fund-of-funds, are likely better able to process financial

information, monitor fund managers, and access inside information from managers (e.g.,

Rajgopal and Venkatachalam 1997; Ramalingegowda and Yu 2012; Gao, Haight, and Yin 2020;

Gaver et al. 2020). 18 As a result, higher ownership by fund-of-funds may limit any inaccuracy

and bias in NAVs. On the other hand, fund-of-funds may have strong implicit incentives to

report favorable current fund performance to raise capital for their own follow-on funds.

Therefore, they might pressure, or potentially collaborate with, fund managers to inflate NAVs.

Further, fund-of-funds remain subject to the previously discussed challenges in monitoring PE

(e.g., lack of the option to exit).

In addition to sophisticated ownership, ownership by foreign investors, who provide a

significant amount of capital to PE funds, may increase oversight of managers. Foreign investors,

however, face limitations due to their distance from fund management and other cultural and

political roadblocks (Ayers, Ramalingegowda, and Yeung 2011; Gao, Wong, Xia, and Yu 2013).

Given these impediments, fund managers may be more likely to bias NAVs as the level of

foreign ownership increases. Alternatively, as the level of foreign ownership increases, fund

managers may provide stronger financial information to alleviate the additional costs foreign

investors face (e.g., Gaver et al. 2020).

17
For evidence that voting and exit are important monitoring mechanisms in non-PE settings, see, for example:
Brochet, Ferri, and Miller 2021; Fan, Radhakrishnan, and Zhang 2021; Li, Maug, and Schwartz-Ziv 2021.
18
Fund-of-funds are entities that invest in other funds (see Harris, Jenkinson, Kaplan, and Stucke 2018).

12
 Under Dodd-Frank, PE funds are generally required to obtain an annual audit (see Gaver

et al. 2020) and the auditor may act as a monitor (e.g., Becker et al. 1998; Francis et al. 1999;

Gaver and Paterson 2001). 19 However, auditors may have little effect on the interim (i.e., non-

year-end) financial reporting of NAVs since quarterly reports are typically not audited (e.g.,

Jenkinson et al. 2020). 20 Further, it is well known that auditors have difficulty auditing fair value

estimates (e.g., Ayres et al. 2019; Carcello et al. 2020), so they may struggle to constrain

managers’ NAV bias or improve accuracy.

We examine whether auditors affect NAV reporting, following existing literature which

finds that Big N auditors (Big 4 in our time period) provide higher quality financial reporting to

public firms (e.g., Defond et al. 2017). We note, however, existing literature based on private

entities provides mixed evidence of a Big N effect (e.g., Chaney et al. 2004; Hardies

Vandenhaute, and Breesch 2018; Che et al. 2020).

In sum, we state our hypothesis related to the external monitors discussed above

(investors and auditors) in the null form as follows:

H1: Private equity fund NAV reporting accuracy and bias are not related to the fund’s
external monitoring.

3.2 Third-party Service Providers

We consider the effect of a fund’s use of the following third-party service providers on

NAV accuracy and bias: valuation specialists; administrators; and, marketers. 21

3.2.1 Valuation Specialists

19
Exempt investment advisers are not required to obtain annual audits. The exemption is based on the type of funds
the adviser manages (advisers managing only VC funds are exempt) and adviser size (advisers with total assets
under management below $150 million are exempt). Most exempt advisers obtain audits (Gaver et al. 2020). Exempt
advisers are still required to disclose certain information about the PE funds they manage. All funds in our sample
obtain audits (see Section 4 for sample details).
20
Discussion with practitioners indicates that PE quarterly reports are also rarely reviewed, unlike in public firms.
However, auditors may “check-up” on more complex investments at interim periods in preparation for the audit.
21
In supplemental analyses, we explore determinants of funds’ use of third-party service providers.

13
 Valuation specialists can play an important role in NAV estimation. In a sample of

insurers, Hanley et al. (2018) find that 60 percent of insurers do not use valuation specialists for

Level 3 valuations, and that these self-estimated cases report higher valuations. Koo et al. (2020)

find that valuation specialist reports can also be opportunistically managed by insurers. In

contrast to insurers, Altamuro and Zhang (2013) find that banks use their information advantage

to generate more accurate fair values for their assets through Level 3 (i.e., more subjective)

estimates. Extending these predictions to PE, funds using valuation specialists may have more or

less accurate or biased NAVs. Of note, PE funds generally own highly specialized assets (e.g., a

portfolio firm) in which the PE fund has deep knowledge because they generally control the

operating, financial, and accounting decisions of the asset (e.g., Borysoff et al. 2022). Thus,

valuation specialists may not improve NAV bias and accuracy for PE funds. If managers

opportunistically manage specialist reports (Koo et al. 2020), valuation specialist use may

increase NAV bias by acting as a certification mechanism for management’s preferred valuation.

3.2.2 Administrators

Third-party fund administrators can be used to outsource fund financial reporting.

Administrators provide services related to regulatory compliance, investor reporting, capital

calls/distributions, fee and expense allocation, and rate of return calculations for the assets under

management (Private Funds CFO 2020; SS&C Technologies 2020). Third-party administrators

may also serve as a signal to external investors that fund managers desire to meet investor

demands for transparency and additional oversight (NES Financial 2020). Further, the presence

of a third-party administrator could make managers more wary of misreporting. If these aspects

are descriptive, PE funds using an administrator are less likely to bias their NAVs consistent with

recent research, which finds that entities allocating more resources to their administrative

14
reporting functions have higher reporting quality (Abramova, Gillette, and Weber 2021; Lee and

Yu 2021). On the other hand, because administrators are not independent and receive

information directly from the PE fund, administrators may not play a significant oversight role

regarding NAV bias or accuracy. Finally, and more broadly, funds outsourcing tasks to

administrators no longer have to deal with these routine tasks, allowing them to focus on more

important items such as managing and valuing their assets. 22 This could lead to more or less bias

or accuracy in NAVs.

3.2.3 Marketers

Funds can hire marketers to assist with fundraising. Marketers could be associated with

NAV reporting accuracy and bias because reported NAVs affect fundraising (e.g., Barber and

Yasuda 2017). Hiring a marketer could indicate that the fund has less need to use NAVs to

attract investors due to the marketer’s services, or could indicate that the fund is especially

concerned with raising funds and may manipulate NAVs to assist with fundraising. For IPOs,

Gustafson, Henry, Kim, and Pisciotta (2021) find that marketing increases IPO underpricing (i.e.,

generates larger day-one returns). Marketing also attracts capital to mutual funds (Roussanov,

Ruan, and Wei 2021).

3.2.4 Summary

Regarding the three third-party service providers (i.e., valuation specialists,

administrators, and marketers), we present our hypothesis in the null form as follows:

H2: Private equity fund NAV reporting accuracy and bias are not related to PE funds’
use of third-party service providers.

3.3 Fund Attributes

22
The relation between administrator use and NAV accuracy and bias may vary with fund size. For example, small
funds may easily be able to handle their administration, or may need the most help. Similarly, large funds may be
able to fully internally staff their back-office functions, or may need extra help. We explore this further in Section 5.

15
 We underscore that our discussion above focuses on PE funds overall. However,

important differences across fund types may result in different relations between NAV reporting

and each type of external monitoring and the use of third-party providers. We separately analyze

BO and VC funds throughout our study as the associations we study are likely to vary across

fund type for multiple reasons. We highlight three important differences between BO and VC

funds that can affect our results. First, BO and VC funds differ fundamentally in their strategies

and the type of assets they own, with BO funds typically investing in more mature businesses

and VCs focusing on higher risk startups. These differences in risk-taking preferences drive

differences in performance, which likely affect NAV reporting incentives. Second, VC funds

typically overstate their NAVs, compared to BO funds, which tend to understate their NAVs,

relative to ex post discounted cash flows (Jenkinson et al. 2020). Third, fund size varies widely

within and by fund type, with BO fund size generally exceeding VC fund size. Larger funds may

be more sophisticated, but also more complex and unwieldy. As a result of these differences, we

perform cross-sectional analyses splitting our sample according to fund size and performance.

In addition to fund size and performance, we also consider fund managers’ experience

and their current fundraising activities. Managers that have significant experience managing PE

funds may have less incentive to manage NAVs due to their reputation. That said, more

experienced managers may have a greater ability to manage NAVs. Finally, we expect managers

that are currently fundraising for new funds to have a greater incentive to upwardly bias current

fund NAVs (Barber and Yasuda 2017).

IV. Data, Research Design, and Summary Statistics

4.1 Data

We obtain our data from two sources. First, we use Preqin’s Private Equity Cash Flow

16
data and Performance Analysis Database. Preqin is a pre-eminent source of PE fund performance

data, used in a multitude of studies (e.g., Harris, Jenkinson, and Kaplan 2014; Barber and Yasuda

2017; Easton et al. 2021). From Preqin we obtain quarterly fund NAVs, cash flow information

(i.e., contributions and distributions), performance, fund vintage year, fundraising information,

and fund manager experience. We require complete cash flow information for funds in our

sample to compute the DCF, meaning that we only include fully liquidated funds. Our sample

ends in 2017 so that we have a sufficient number of observations for funds that are fully

liquidated at the time we downloaded Preqin data (i.e., as of August 2021). We begin our sample

in 2009, after the implementation of ASC 820, to ensure consistent NAV accounting (see Easton

et al. 2021 for details). This period is also after the Global Financial Crisis, which may have

affected the estimation and measurement of fair values.

Second, we use annual SEC filings required for PE funds under the 2010 Dodd-Frank Act

(see Gaver et al. 2020 and Borysoff et al. 2022 for details). PE GPs must disclose information on

Form ADV regarding each fund they advise. The information reported includes annual fund-

level information on auditors and other third-party service providers of the fund, as well as fund

ownership data for several categories of investors, minimum investment required, and fund type.

Following prior literature, we limit our sample to U.S. private equity funds identified as BO and

VC fund types. We exclude fund-of-funds – which own other PE funds – because, as noted

earlier, these funds rely on other funds’ NAV reporting.

We manually match fund-level (quarterly) Preqin data to (annual) Form ADV data using

fund names. Because fund-level Form ADV filings begin in 2012, following the passage of

Dodd-Frank, the Form ADV funds matched to Preqin include only fund-quarters after 2011.

However, because we study closed-end, finite life PE funds that do not allow for divestment or

17
new investment after fund formation (Kaplan and Strömberg 2004, 2009), some fund attributes

(e.g., fund ownership) typically do not change through the fund’s life. Therefore, we use

observable fund attributes disclosed after the passage of Dodd-Frank to backfill fund attributes

for earlier fund-quarters. We acknowledge that this potentially adds noise to some of our

measures. Our sample consists of 4,502 fund-quarters, based on 1,240 fund-years for 242 funds.

Table 1, Panel A, presents our sample selection. Table 1, Panel B, presents the fund-quarter

sample by year and fund type. We discuss additional descriptive statistics below.

INSERT TABLE 1 HERE

4.2 Research Design

We investigate whether NAV accuracy and bias vary with external monitoring and third-

party providers using an ordinary least squares (OLS) estimation of the following model,

combining attributes of Easton et al. (2021), Gaver et al. (2020), and Jenkinson et al. (2020):

Accuracyiq/Biasiq = α0 + α1*Ln_Ownersit + α2*%Insiderit +

α3*%Own_Fund_of_Fundit + α4*%Own_Fgnit + α5*Ln_Min_Invit + α6*Big4, Q1-Q3it +
α7*Big4, Q4it + α8*Non-Big4, Q4it + α9*Perc_Assets_Valuedit + α10*Adminit +
α11*Marketerit + α12*Ln_Ttl_Committed_Capitali + α13*Experienceit + α14*Fundraisingiq
+ α15*Quartileit + α16*Exemptit + α17*SP500_Retq + α18*SP500_Retq-1 + αt*ΣFund_Ageit
+ eiq (1)

The dependent variable, either Accuracy or Bias, measures the accuracy or bias of the NAV of

fund i in quarter q in year t. Following Jenkinson et al. (2020), we calculate these variables as the

natural logarithm of the difference between NAV and DCF (formed using an 11 percent discount

rate) and scaled by contributions, where accuracy is unsigned and bias is signed. 23 Both variables

are constructed so that they are increasing in accuracy or bias. We follow Jenkinson et al. (2020)

23
An 11 percent discount rate is the average internal rate of return for PE funds documented in Kaplan and Schoar
(2005) and used in Jenkinson et al. (2020). Using this rate assumes no fund-specific performance information. In
additional analyses, we use each of a 10 percent and 12 percent discount rate and observe similar inferences..

18
and estimate the model separately for BO and VC funds. We cluster standard errors by fund.

Appendix A presents a complete list of variable definitions.

As discussed, we group factors into three groups when assessing their relation to NAV

accuracy and bias. First, we examine attributes of external monitoring. We consider the number

of owners (Ln_Owners), as well as the percentage of the fund owned by either insiders

(%Insider), sophisticated fund-of-fund investors (%Own_Fund_of_Fund), or foreign investors

(%Own_Fgn). We include a related variable for the minimum investment that the fund will

accept (Ln_Min_Inv), because funds requiring larger investments likely have more sophisticated

investors. Related, we examine oversight from auditors, another component of external

monitoring. We include indicator variables for funds that employ Big 4 auditors for both fiscal

quarters one through three (Big4, Q1-Q3) and fiscal quarter four (Big4, Q4), which corresponds

to the period the fund audit occurs. We similarly include an indicator variable for funds that

employ non-Big 4 auditors in fiscal quarter four (Non-Big4, Q4). The benchmark or holdout

group represents fund-quarter observations for fiscal quarters one through three for funds that

employ non-Big 4 auditors. We also conduct tests of the significance of differences across the

Big 4 coefficients, comparing audited to unaudited quarters for funds with Big 4 auditors.

Relative to hypothesis 1, our interest is in the estimates of the coefficients on these external

monitoring variables related to investor and auditor oversight.

Our second group of fund attributes relates to funds’ use of third-party service providers.

We include a measure of the portion of the fund’s assets valued by a valuation specialist

(Perc_Assets_Valued) as well as indicator variables for a fund’s use of an administrator (Admin)

or a marketer (Marketer). Relative to hypothesis 2, we are interested in the estimates of the

coefficients on these third-party service provider variables.

19
 Third, we examine fund attributes that may affect accuracy and bias. We include fund

size represented by the total capital commitment (Ln_Ttl_Committed_Capital) because larger

funds may commit more resources in preparing NAVs, or may have more complex or diverse

assets that are harder to value. We note that Jenkinson et al. (2020) find mixed results for fund

size. 24 We include the number of funds previously managed by the fund’s GP (Experience)

because more experienced managers may have less need (e.g., due to reputation), or conversely

be better able, to manage the accuracy and bias of NAVs. We expect managers to positively bias

NAVs in periods around fundraising (Fundraising) as suggested by Barber and Yasuda (2017)

and Pham, Turner, and Zein (2021); we make no prediction regarding accuracy. We include the

fund’s quartile performance ranking (Quartile), as measured by Preqin, which represents how the

fund’s prior year’s interim IRR and fund multiples compare to a portfolio of benchmark funds. 25

We code this variable so that higher rankings of Quartile represent better performance. This

measure captures a manager’s NAV reporting incentives that may be based on fund performance,

such as the incentive to improve recent poor performance or to continue to show good

performance. This measure also captures a simplistic measure of systematic or market risk for

the fund since each fund’s IRR, or performance, is compared to a benchmark return in the similar

way a market model may use the S&P 500 return as a benchmark return (Kaplan and Schoar

2005). We include an indicator for the fund’s regulatory status (Exempt) reflecting exemption

from certain SEC reporting regulations. This measure captures the potential effects of another

external monitor, the SEC, and whether its oversight, or the lack thereof, affects NAV reporting.

24
We cannot control for time-varying size of the fund because the time-varying size of the fund is its NAV.
25
The benchmark group of funds is based on funds’ vintage year, investment strategy (i.e., growth, early stage,
venture, etc.), and geographic focus (i.e., all regions, U.S., etc.). Within each benchmark, funds are assigned an
overall quartile ranking taking into account the fund’s net IRR and fund multiple where a fund’s multiple captures
the distribution and future unrealized value of the fund.

20
 Finally, we control for market performance using the current and prior quarter’s S&P 500

returns (SP500_Ret and SP500_Ret_Lag). We also include fund age fixed effects. That is, we

include an indicator set equal to one for funds in their first year, and zero otherwise; another

indicator set to one for funds in their second year, and zero otherwise; and so on. This captures

the possibility that either NAV reporting incentives or the cash flows from the funds change

through the life of the fund. 26 For example, managers may have a greater incentive to bias NAVs

upward as the fund nears liquidation and seeks external capital for a new, follow-on fund.

4.3 Summary Statistics

Table 2, Panel A presents descriptive statistics for BO and VC funds separately. Our

Accuracy and Bias measures are in line with Jenkinson et al. (2020) (with our measures

transformed so that higher values reflect increases and, therefore, our signs are opposite theirs).

The average BO fund NAV differs from the DCF by 13.4 percent of contributions and the

average bias is negative 4.4 percent of contributions, consistent with BO funds conservatively

(i.e., pessimistically) reporting NAVs. 27 We observe a similar accuracy for VC funds of 13.0

percent; however, we observe a positive 3.5 percent average bias for VC funds, suggesting more

optimistic reporting by VC funds. For both types of funds, the magnitude of the difference for

Accuracy exceeds that of Bias because Accuracy is unsigned. 28 Bias has a wide range despite the

median fund being relatively unbiased. For ease of interpretation, we present the unlogged values

of the other logged variables we use in our regression, as well as the logged regression variables

(Ln_Owners, Ln_Min_Inv, Ln_Size), in Panel A. The median BO (VC) fund has 59 (48) owners,

26
Jenkinson et al. (2020) include groupings related to variations in fund life (e.g., an early-life indicator).
27
We transform the Accuracy and Bias variables from logarithms into percentages for ease of interpretation. E.g.,
Accuracy = e^(-0.126*-1) – 1 (multiplying by negative 1 to reverse our transformation); Bias = e^(-0.045) – 1.
28
Accuracy is unsigned, but we express it as a negative variable so that higher measures are more accurate. That is, a
value of zero represents a perfectly accurate NAV while more negative numbers are less accurate.

21
2 (1) percent insider ownership (which is consistent with the known structuring of PE funds;

Borysoff et al. 2022), 11 (16) percent ownership by fund-of-funds, 7 (8) percent ownership by

foreign investors, and a minimum investment of $5 ($0.15) million. About 79 (70) percent of

fund-quarters relate to BO (VC) funds with Big 4 auditors, consistent with Gaver et al. (2020),

with 17.2 (14.7) percent of the BO (VC) sample fund-quarters (roughly 21% of the total Big 4

fund-quarters for both BO and VC funds, untabulated) being fourth quarter observations when a

Big 4 auditor is used. Only about 2.2 (1.6) percent of BO (VC) fund assets are valued by a third-

party, while 18.1 (5.2) percent of BO (VC) fund-quarters use an administrator. 29 In addition,

only 6.8 (1.9) percent of BO (VC) fund-quarters use a marketer. The median BO (VC) fund has

committed capital of $530 ($225) million and the median BO (VC) GP operates two (one)

fund(s). In sum, a) there are numerous differences between fund types, supporting our separate

analyses; and b) BO funds tend to be larger than VC funds.

INSERT TABLE 2 HERE

Table 2, Panel B (C) presents descriptive statistics for several of the hypothesized factors

related to accuracy and bias by year for BO (VC) funds. We find that the number of owners and

insider ownership percentage remain relatively constant over time, except that VC insider

ownership increases. Fund-of-fund ownership increases over time for BO funds whereas it

remains relatively flat for VC funds. Foreign ownership increases somewhat over time for both

VC and BO funds. Big 4 auditor usage decreases over time, consistent with Gaver et al. (2020).

Most notably, valuation specialists value an increasing percentage of assets over time for BO

funds. Administrator and marketer use increase over time for VC funds (ignoring the most recent

years with few available observations) but fluctuates for BO funds.

While perhaps surprising, the low use of valuation specialists in PE is consistent with Allee, Erickson, Esplin, and
29

Yohn’s (2020) survey, which notes most valuations relate to litigation, taxes, or purchase and sale of businesses.

22
 Table 2, Panel D (E) presents univariate statistics for Accuracy and Bias partitioned

across several of our indicator variables for BO (VC) funds. This univariate analysis presents

some initial insight regarding our hypotheses. We find that BO (VC) funds with Big 4 auditors

have higher (lower) accuracy and lower (higher) bias, or more pessimistic (optimistic) NAVs. In

contrast, BO funds using administrators have lower accuracy and higher (i.e., less pessimistic)

bias, while VC funds exhibit higher (i.e., more optimistic) bias when using administrators. BO

funds that use a valuation specialist have more accurate and less pessimistically biased NAVs,

whereas VC fund bias or accuracy does not differ based on valuation specialist usage. BO funds

using marketers have significantly lower accuracy. VC funds, on the other hand, have higher

accuracy and less optimistically biased NAVs when using a marketer. While this provides some

initial evidence, we caution that these results do not control for other fund characteristics.

Therefore, we draw our main conclusions from our regression results in Section 5.

Table 3, Panel A (B) presents correlations among the variables of interest for BO (VC)

funds. We find significant correlations between many variables and NAV accuracy and bias.

Specifically, we find various ownership attributes, auditors, and the use of third-party service

providers are significantly correlated with NAV accuracy and bias. However, most correlations

are relatively low in magnitude, with the exception of the correlations between the number of

owners and fund size in the BO sample (both above 0.62). This suggests multicollinearity is not

likely a concern in our tests. To ensure that size does not cause multicollinearity issues or

otherwise mask or affect inferences, we also conduct cross-sectional splits based on size.

INSERT TABLE 3 HERE

V. Multiple Regression Analyses of NAV Accuracy and Bias

23
 For our main analysis of how PE fund reporting quality varies across external monitoring,

use of third-party service providers, and fund attributes, we estimate equation (1). Following

Jenkinson et al. (2020), we separately examine BO and VC funds because the differences,

discussed in Section 3.3, across these two fund types could lead to variation in the association of

each attribute on NAVs. Tables 4 and 5 present the results from our regression analyses for BO

and VC funds, respectively. In each table, Panel A (B) reports estimates in which NAV accuracy

(bias) is the dependent variable. In each panel, column 1 presents results for all BO (Table 4) or

VC (Table 5) funds in our sample. Columns 2 and 3 present results for large and small funds,

respectively, based on terciles of committed capital within fund type. Columns 4 and 5 presents

results for top and bottom performing funds, respectively, based on Preqin’s performance

quartile designations. We anticipate funds’ incentives to manipulate NAV reporting are affected

by fund size (e.g., differences in resources or complexity) and performance so that particular

attributes may affect NAV reporting differently depending on the fund’s performance or size.

INSERT TABLE 4 HERE

5.1 Buyout (BO) Funds

We find little evidence that investor ownership or monitoring is associated with NAV

reporting accuracy in BO funds in Table 4, Panel A. None of the estimates of the coefficients on

our investor monitoring variables are statistically significant in the overall sample (column 1).

However, we observe a positive association between accuracy and both insider ownership and

foreign ownership for small BO funds (column 3). We find lower accuracy as fund-of-fund

ownership increases for top performing funds (column 4). This may suggest sophisticated

investors of high-performing funds put forth less effort to monitor high performing funds given

the high returns. In contrast, we find higher accuracy as foreign ownership increases for high

24
performing BO funds (column 4). This suggest foreign investors are more interested in ensuring

accurate NAV reporting in high performing funds. Finally, we observe NAV accuracy is lower

for bottom-performing funds (column 5) as the minimum required investment increases.

We also find little evidence that investor ownership or monitoring is related to NAV

reporting bias in BO funds in Table 4, Panel B. For the full sample in column 1, we observe a

weakly positive coefficient on insider ownership (0.0031, p<0.10) suggesting less pessimistic

bias as insider ownership increases. This is consistent with managers having additional discretion

with higher levels of ownership. However, this effect is small, indicating a roughly 0.3 percent

increase (i.e., exp[0.0031] – 1) in bias for each additional percent of insider ownership.

Moreover, the effect is limited to the small BO funds in our sample (column 3), which suggests,

together with evidence in Panel A of Table 4, that insiders focus on NAV bias and accuracy

when they own more of small funds. In column 4 of Table 4, Panel B, we find lower bias as

fund-of-fund ownership increases but higher bias as foreign ownership increases for high

performing BO funds. Thus, we find some evidence rejecting the null hypothesis 1 in statistical

terms, but the economic significance is generally small.

We find some evidence of a Big 4 effect for BO funds for accuracy. Across all columns

of Panel A of Table 4, which focuses on NAV accuracy, the coefficient on the Big 4 audited

quarters are statistically larger than on Big 4 unaudited quarters (p<0.05). 30 In Panel B of Table

4, which focuses on NAV bias, the Big 4 audited quarter variable is statistically higher than Big

4 unaudited quarters (p<0.10) but only for large or top-performing funds. Taken together, these

results for BO funds provide some support for rejecting hypothesis 1 related to the effect of Big

4 auditors on NAVs. In both Panels A and B, none of the non-Big 4 audit coefficients are

Recall that the main effect coefficients on Big 4 auditor quarters are relative to non-Big 4 unaudited quarters.
30

While informative, we are more interested in the effect of Big 4 audited versus Big 4 unaudited quarters.

25
significant, suggesting that non-Big 4 auditors are not associated with reporting quality for BO

funds. Given that BO funds are usually large, this could represent a mismatch between non-Big 4

auditors and large funds (cf. Mason and Utke 2020).

For hypothesis 2, we find some evidence that third parties affect NAV reporting for BO

funds. We first examine accuracy in Table 4, Panel A. We observe higher accuracy as the

percentage of assets valued by a third-party valuation specialist increases (column 1) but this

effect is concentrated in small BO funds (column 3). 31 Turning to administrator use, we observe

a negative association with accuracy in our full BO sample (column 1), concentrated in large

(column 2) and bottom-performing (column 5) funds. This suggests administrator use may signal

fund manager time constraints, but does not reduce constraints enough to allow managers to

focus on NAV reporting. We also observer higher accuracy for large BO funds using a marketer

(column 2).

Turning to BO fund NAV bias in Panel B, Table 4, we find significantly less pessimistic

bias for funds that use a third-party marketer. The effect is large, with bias 16.8 percent (i.e.,

exp[0.1551] – 1) higher for funds with marketers versus without, based on column 1.

Conversely, we see no effect in the large or small funds sample, in columns 2 and 3, suggesting

that the effect is concentrated in midsize BO funds or that lack of power impacts our partitioned

results. The association between third-party marketers and NAV bias is concentrated in high-

performing funds (column 4), suggesting a greater incentive to increase NAVs when using a

marketer in high performing funds. We also find evidence suggesting the use of administrators in

small BO funds decreases pessimistic NAV bias (column 3). Collectively, these results suggest

the use of administrators plays a role in the bias of BO fund NAVs. Thus, we reject the null for

31
Top and bottom performing BO funds do not use valuation specialists in our sample.

26
hypothesis 2 suggesting third-party service providers are associated with NAV reporting.

In examining fund attributes, we do not find evidence that BO fund attributes (e.g., fund

size, manager experience, performance) affect NAV accuracy in the overall sample presented in

column 1 of Table 4, Panel A. However, we do observe that higher performing funds and funds

with more experienced managers have lower bias in column 1 of Table 4, Panel B. These

findings likely suggest managers have a lower incentive to bias NAVs when fundraising

difficulty is lower, as is the case when funds have performed well and when the manager is

experienced. We also find that bottom performing funds with more experience have lower bias.

This suggests that even though a fund exhibits poor performance, experienced managers have a

lower incentive to bias NAVs. Further, bottom performing funds that are currently fundraising

have lower bias.

INSERT TABLE 5 HERE

5.2 Venture Capital (VC) Funds

With respect to VC funds, we again find little evidence that investor ownership or

monitoring is associated with NAV accuracy in Table 5, Panel A. In columns 1 through 3, we

find investor oversight has limited relation to accuracy overall and regardless of fund size.

However, we find that the association of external monitors with fund NAV accuracy depends on

fund performance (columns 4 and 5). More specifically, we find some evidence that bottom

performing funds have greater (lower) accuracy as insider (sophisticated) ownership increases

(column 5). We also document that top (bottom) performing funds have a negative (positive)

relation between accuracy and foreign ownership in column 4 (5). This suggests that foreign

owners of VC funds are more interested in monitoring the reporting of low performing funds as

compared to top performing funds.

27
 For NAV bias (Table 5, Panel B), we observe a significant reduction in optimistic bias as

the number of external investors increases (column 2, -0.057, p-value<0.10) suggesting a

disciplining role of LPs in VC funds. However, we also find significantly lower bias as the

percentage of insider ownership increases. We do not observe a negative relation between the

number of owners and bias for large or small VC funds, suggesting the result is concentrated in

midsize VC funds or that lack of power impacts our tests. Further, we find that insiders mitigate

bias for both large (column 2) and small (column 3) funds. Interestingly, we document that

foreign ownership reduces bias for both top (column 4) and bottom (column 5) performing

funds, suggesting foreign owners have incentives to monitor such funds.

Turning to the effect of auditors in VC reporting, we find only limited evidence

suggesting Big 4 auditors are associated with more accurate NAVs in VC funds (Table 5, Panel

A). Specifically, the coefficient on Big 4 auditor quarters is significantly higher than Big 4

unaudited quarters (p<0.02), but only for small and poorly performing VC funds (columns 3 and

5). We find a similar improvement in bias (i.e., lower optimism) only for bottom performing

funds (Panel B, column 5, p<0.03). Alternatively, we observe a significant increase (decrease) in

accuracy (bias) for fourth quarter NAVs of VC funds when using a non-Big 4 auditor (Panel A,

column 1, 0.0134, p-value<0.01; Panel B, column 1, -0.0132, p-value<0.05). This result is

consistent with Jenkinson et al. (2020) who find fourth quarter NAVs are more accurate, but

expands on their findings to show the identity of the auditor (i.e., non-Big 4) is important.

That said, the positive association between non-Big 4 auditors and VC fund accuracy

overall (Table 5, Panel A, column 1) is concentrated in small funds (column 3) that may need the

most assistance from auditors in arriving at reasonable NAVs. Interestingly, we find lower

accuracy in small VC funds for quarter one through three when a Big 4 auditor is used (column

28
3); recall the baseline is unaudited non-Big 4 quarters. Overall, we provide some evidence that

auditors affect NAV reporting in VC funds, contrary to our null hypothesis 1.

We also find some evidence that third-party involvement is significantly associated with

VC fund accuracy (Panel A) and bias (Panel B). We document that the use of valuation

specialists in VC funds is associated with higher (lower) accuracy (bias) for bottom performing

funds in column 5 of Table 5, Panel A (B) suggesting third-party valuation specialists enhance

these funds’ reporting. However, we find valuation specialist usage is associated with higher bias

in top performing funds (column 4), suggesting an inability to constrain reporting in such funds.

Administrators are positively associated with accuracy for VC funds overall (Panel A column 1,

coefficient estimate of 0.0721, p<0.05) and in both top and bottom performing funds (in columns

4 and 5 of Panel A). We find that the use of an administrator is associated with higher bias in

large funds (Panel B, column 2). Marketers are also positively associated with NAV accuracy

and bias for top-performing VC funds (column 4, Panels A and B). Thus, we reject the null

hypothesis 2 related to the impact of third-party service providers for VC funds.

For fund attributes, we find evidence that more experienced VC managers are associated with

increased accuracy (Panel A, column 1). While we fail to find evidence of a relation between

accuracy and exempt status (i.e., the lack of additional SEC oversight) when we combine all VC

funds (Panel A, column 1), we find a positive relation between accuracy and exempt status for

both large and small VC funds (Panel A, columns 2 and 3) and lower performing funds (Panel A,

column 5). We also observe, similar to BO funds, that VC funds with lower fund performance

have lower bias (Panel B, column 1). The association of performance with bias also concentrates

in large VC funds (Panel B, column 2). Among lowest performing VC funds, exempt funds

exhibit lower bias (Panel B, column 5). Coupled with the evidence from Table 5, Panel A that

29
associates NAV accuracy with exempt, low-performing funds, it appears that funds with low

performance avoid manipulating NAVs regardless of SEC oversight. However, we also find a

positive relation between exempt status and bias for large VC funds (Table 5, Panel B, column

2). Therefore, regulators should carefully consider whether certain exempt funds warrant more

SEC attention.

VI. Multiple Regression Analyses of Use of Third-Party Service Providers

6.1 Determinants of Third-Party Service Provider Use

While we find some evidence that the use of third-party service providers is associated

with NAV bias or accuracy, our unique dataset also allows us to gain insight into why firms

engage these providers. 32 To examine this question, we make two modifications to equation (1).

First, we replace the dependent variable (either bias or accuracy previously) with

Perc_Assets_Valued, Valuation (an indicator for using a valuation specialist for any portion of

assets), Admin, or Marketer. As we move through the dependent variables, we control for the

other possible dependent variables (e.g., if Admin is the dependent variable, we control for

Marketer and Valuation). Second, the dependent variables in this analysis come from Form ADV

and are annual measures. Thus, we re-calibrate our model to use fund-year, rather than fund-

quarter, observations. As such, we replace SP500_Ret and SP500_Ret_Lag, which are quarterly

measures, with SP500_Ret_12mo and SP500_Ret12mo_Lag, which are annual measures. We

also compress the Fundraising variable so that it is set equal to one if the PE fund is fundraising

in the next quarter at any point in the year, and zero otherwise.

Table 6 presents results, with Panel A (B) reporting BO (VC) results. We use OLS

models in this section because OLS models, or more accurately linear probability models, tend to

32
Gaver et al. (2020) examine other PE fund reporting choices, such as Big 4 auditor use.

30
be easier to interpret than logit models for binary dependent variables. Column 1 presents results

for Valuation (i.e., the use of external valuations), while column 2 reports results using the

continuous Perc_Assets_Valued measure rather than the Valuation indicator variable. Column 3

presents results for Admin (i.e., the use of an administrator), while column 4 reports results for

Marketer (i.e., the use of a marketer). Results are largely consistent across columns 1 and 2

suggesting external monitoring and fund attributes have little effect on the decision to engage a

third-party valuation specialist. However, we find that VC funds with fewer owners are more

likely to obtain valuations (Panel B, column 1).

INSERT TABLE 6 HERE

Column 3 of both panels in Table 6 present little evidence that any of our fund attributes

explain administrator use, except that BO funds (Panel A) using a valuation specialist are more

likely to use administrators and VC funds (Panel B) with higher foreign ownership are less likely

to use administrators. We also find evidence that VC funds managed by a manager with more

experience are associated with a lower likelihood of using an administrator. This is likely due to

lower need for administrative help as the fund and its support staff gain the experience necessary

to perform administrative tasks. Column 4 of Panel A documents that BO funds with a higher

minimum investment threshold (Ln_Min_Inv) and fund manager experience are more likely to

use a marketer. More experienced funds may be more aware of the value of marketers and

therefore be more likely to engage a marketer. Further, when raising a new fund with a high

minimum investment, managers may need the help of marketers to attract the necessary investor

clientele. In Panel B, we do not find evidence that any of our fund attributes affect the decision to

use a marketer in VC funds. 33

33
Cain, McKeon, and Solomon (2020) examine marketers known as placement agents. Their study differs from ours
because they primarily focus on LP-level data, whereas we focus on fund-level data, and they focus on how

31
 VII. Multiple Regression Analyses of NAV Adjustments

We also examine NAV adjustments in each quarter. While our main focus is on NAV

accuracy and bias, we are also interested in the magnitude and direction of NAV adjustments

because they affect reported NAVs. We examine large NAV adjustments, > 10 Pct Adj (greater

than 10 percent of prior NAV, in absolute value), and positive NAV adjustments, Pos Adj. Table

7 reports results for BO (Panel A) and VC (Panel B) funds.

INSERT TABLE 7 HERE

For BO funds (Panel A), we find that large NAV adjustments are less likely for funds

with higher minimum investment levels (column 1). We also find large NAV adjustments by Big

4 and non-Big 4 auditors in column 1. This is interesting given the mixed evidence of auditor

effects on reporting quality; auditors appear to be driving funds to do something with their

NAVs, but these changes do not necessarily increase accuracy or reduce bias. Positive

adjustments are more likely for BO funds using valuation specialists (column 2), suggesting

specialists may cater to clients or funds may shop for favorable reports. Alternatively, positive

adjustments are less likely for BO funds with more owners, perhaps due to more caution in

increasing NAVs with more external monitors. We also find BO funds using a marketer are less

likely to make positive adjustments (column 2). Experience, committed capital, and performance

are also positively associated with positive adjustments.

For VC funds (Panel B), we find that large NAV adjustments are more likely for funds

with higher minimum investment, or audits by either Big 4 or non-Big 4 auditors (column 1). We

placement agents affect LP returns. They use Preqin to gather LP data, but Preqin is known to have a substantial
amount of missing LP data. For example, Cain et al. (2020) report median fund size of $378 million but having only
four LPs, which is implausible if the LP data were complete. We report median fund commitment of $400 million
(similar to Cain et al. 2020), but median ownership by 48 to 59 LPs. Thus, imposing the requirement for Preqin LP-
level data likely results in a sample that fundamentally differs from ours and appears to miss over 90 percent of LPs.
Importantly, we obtain LP ownership data from Form ADV, which is a mandatory SEC filing.

32
find VC funds using a marketer are less likely to make positive adjustments (column 2), similar

to BO funds. Better performing VC funds and VC funds that fundraise in the near future are

more likely to adjust NAVs upwards (column 2). The fundraising result is consistent with

interim NAVs factoring into funds’ ability to fundraise (Barber and Yasuda 2017).

VIII. Full Sample Analysis

One concern with our inability to document monitoring effects across funds is due to low

power arising from our sample splits. To limit this possibility, we re-run our analysis in the full

sample of funds (untabulated). We continue to find limited evidence of associations between

investors and NAV reporting. We observe that non-Big 4 auditors overall are positively

associated with accuracy. This suggests the limited monitoring effects in PE funds are not a

statistical artifact. We do not find evidence of an effect of valuation specialists, while the use of a

marketer remains positively associated with bias. Further, we identify a negative association

between bias and manager experience and performance, whereas fundraising is associated with

higher accuracy.

IX. Conclusion

In this study, we provide new, descriptive evidence on how external monitors, third-party

service providers, and fund attributes affect PE fund NAV accuracy and bias. In contrast to well-

known monitoring effects in public firms, we find limited evidence that traditional external

monitors (investors and auditors) improve NAV reporting. Similarly, we find limited and mixed

evidence that valuation specialists improve NAV reporting. We also find that certain funds,

which employ a marketer, report more optimistically biased NAVs.

While we avoid making causal inferences, our study highlights that important differences

between PE funds and public firms lead to different relations between various entity attributes

33
and reporting outcomes in the PE versus public space. As the PE market grows, and public

markets shrink, our prior understanding of public firms must be complemented by a new

understanding of private markets, which are now primarily responsible for funding firms and

economic growth. Our study is one of the first to explore these important differences, providing

new insight for regulators, investors, and academics.

34
 APPENDIX A
Variable Definitions (Source: Form ADV filings unless otherwise noted)
Variable Definition
Dependent Variables
Accuracyiq Accuracy of the NAV of fund i in quarter q in year t. This measure is unsigned and is
calculated as the natural logarithm of one plus the absolute value of the difference between
NAV and DCF (calculated using an 11 percent discount rate) scaled by total contributions to
fund i as of quarter q. The measure is then multiplied by -1 so that the measure is increasing in
NAV accuracy. Source: Preqin
Biasiq Bias of the NAV of fund i in quarter q in year t. This measure is signed and calculated as the
natural logarithm of one plus the difference between NAV and DCF (calculated using an 11
percent discount rate) scaled by total contributions to fund i as of quarter q. The measure is
increasing in NAV bias. Source: Preqin
> 10 Pct Adjiq Is an indicator variable that equals one if the absolute value of the change in NAV (excluding
contributions and distributions) from q-1 to q is greater than 10 percent of the q-1 NAV, and
equals zero otherwise. Source: Preqin
Pos Adjiq Is an indicator variable that equals one if the change in NAV (excluding contributions and
distributions) from q-1 to q is greater than 0, and equals zero otherwise. Source: Preqin
Investor Monitoring Variables
Ln_Ownersit The natural logarithm of the raw number of investors (Number of Owners) in fund i in year t.

%Insiderit The percentage of fund i owned by the investment adviser (i.e., GP) or a related party in year
t.
%Own_Fund_of_Fundsit The percentage of fund i owned by other funds (often known as fund-of-funds) in year t.

%Own_Fgnit The percentage of fund i owned by non-U.S. investors in year t.

Ln_Min_Invit The natural logarithm of fund i’s minimum required investment to invest in the fund
(Minimum Investment) in year t.

Auditor Variables
Big4, Q1-3it Is an indicator variable that equals one if the fund engages a Big 4 accounting firm and the
fund-quarter observation is in the first, second, or third quarter, and equals zero otherwise.
Big4, Q4 (audited)it Is an indicator variable that equals one if the fund engages a Big 4 accounting firm and the
fund-quarter observation is in the fourth quarter, which corresponds to the quarter of the audit,
and equals zero otherwise.
Non-Big4, Q4 (audited)it Is an indicator variable that equals one if the fund engages a non-Big 4 accounting firm and
the fund-quarter observation is in the fourth quarter, which corresponds to the quarter of the
audit, and equals zero otherwise.

Third-Party Service Provider Variables (Dependent Variables in Determinants Analyses)

Perc_Assets_Valuedit The percentage of fund i’s assets that were valued by a valuation specialist in year t.

Valuationit Is an indicator variable that equals one if fund i engages a valuation specialist in year t, and
equals zero otherwise.
Adminit Is an indicator variable that equals one if fund i engages a third-party fund administrator in
year t, and equals zero otherwise.

35
 APPENDIX A
Variable Definitions (Source: Form ADV filings unless otherwise noted)
Variable Definition
Marketerit Is an indicator variable that equals one if fund i engages a third-party marketer in year t, and
equals zero otherwise.

Fund Attributes
Ln_Ttl_Committed_Captali The natural logarithm of fund i’s total committed capital. Source: Preqin

Experienceit The number of funds managed by fund i’s manager prior to year t. Source: Preqin

Fundraisingiq Is an indicator variable that equals one if fund i’s manager calls capital for a new fund during
the 12 months following the current quarter-end, and equals zero otherwise. Source: Preqin
Quartileit A rank variable indicating fund i’s quartile in terms of past fund performance, relative to other
funds of similar vintage, strategy, and geographic focus, as defined by Preqin, in year t-1. In
Preqin, “1” (“4”) represents the top (bottom) quartile. We reverse code this so that “4” (“1”)
represents the top (bottom) quartile. Source: Preqin
Exemptit Is an indicator variable that equals one if fund i is exempt from the audit requirement
established under Dodd-Frank, which occurs when the fund’s adviser (i.e., GP) only advises
venture capital (VC) funds or has less than $150 million of assets under management across
all of the GP’s funds, in year t, and equals zero otherwise.
Fund_Ageit Represent separate indicators set equal to one for funds in their first year, and equal to zero
otherwise; another indicator set to one for funds in their second year, and equal to zero
otherwise; and so on. Source: Preqin

Market Attributes
SP500_Retq The return on the S&P 500 index in quarter q. Source: CRSP

SP500_Ret_Lagq-1 The return on the S&P 500 index in quarter q-1. Source: CRSP

SP500_Ret_12mot The return on the S&P 500 index in year t. Source: CRSP

SP500_Ret_12mo_Lagt-1 The return on the S&P 500 index in year t-1. Source: CRSP

36
 References
Abramova, I., J. Gillette, and J. Weber. 2021. Evidence on the importance of the financial
reporting function in municipal governments. Working paper.
Allee, K. D., D. Erickson, A. M. Esplin, and T. L. Yohn. 2020. The characteristics, valuation
methods, and information use of valuation specialists. Accounting Horizons 34 (3): 23-
38.
Altamuro, J., and H. Zhang. 2013. The financial reporting of fair value based on managerial
inputs versus market inputs: Evidence from mortgage servicing rights. Review of
Accounting Studies 18: 833-858.
Ang, A., D. Papanikolaou, and M. M. Westerfield. Portfolio choice with illiquid assets.
Management Science 60 (11): 2737-2761.
Ayers, B. C., S. Ramalingegowda, and P. E. Yeung. 2011. Hometown advantage: The effects of
monitoring institution location on financial reporting discretion. Journal of Accounting
and Economics 52: 41-61.
Ayres, D. R., T. L. Neal, L. C. Reid, and J. E. Shipman. 2019. Auditing goodwill in the post-
amortization era: Challenges for auditors. Contemporary Accounting Research 36 (1):
82-107.
Badertscher, B. A., S. P. Katz, and S. O. Rego. 2013. The separation of ownership and control
and corporate tax avoidance. Journal of Accounting and Economics 56: 228-250.
Bae, G. S., S. U. Choi, and J. E. Lee. 2021. Auditors’ response to auditor business risk: An
analysis using public and private companies. Auditing: A Journal of Practice and Theory
forthcoming.
Baik, B. K. 2022. Private equity fund valuation management during fundraising. Working paper.
Barber, B. M. and A. Yasuda. 2017. Interim fund performance and fundraising in private equity.
Journal of Financial Economics 124: 172-194.
Becker, C. L., M. L. Defond, J. Jiambalvo, and K. R. Subramanyam. 1998. The effect of audit
quality on earnings management. Contemporary Accounting Research 15 (1): 1-24.
Black, B. S. 1992. Agents watching agents: The promise of institutional investor voice. UCLA
Law Review 39 (4): 811-894.
Borysoff, M. N., P. Mason, and S. Utke. 2022. Understanding private equity funds: A guide to
private equity research in accounting. Working paper.
Brochet, F., F. Ferri, and G. S. Miller. 2021. Investors’ perceptions of activism via voting:
Evidence from contentious shareholder meetings. Contemporary Accounting Research
forthcoming.
Brown, G. W., O. R. Gredil, and S. Kaplan. 2019. Do private equity funds manipulate reported
returns? Journal of Financial Economics 132: 267-297.
Cain, M. D., S. B. McKeon, and S. D. Solomon. 2020. Intermediation in private equity: The role
of placement agents. Journal of Financial and Quantitative Analysis 55 (4): 1095-1116.
Carcello, J. V., T. L. Neal, L. C. Reid, and J. E. Shipman. 2020. Auditor independence and fair
value accounting: An examination of nonaudit fees and goodwill impairments.
Contemporary Accounting Research 37 (1): 189-217.
Chaney, P. K., D. C. Jeter, and L. Shivakumar. 2004. Self-selection of auditors and audit pricing
in private firms. The Accounting Review 79 (1): 51-72.
Che, L., O.-K. Hope, and J. C. Langli. 2020. How Big-4 firms improve audit quality.
Management Science forthcoming.
37
Cornett, M. M., A. J. Marcus, and H. Tehranian. 2008. Corporate governance and pay-for-
performance: The impact of earnings management. Journal of Financial Economics 87:
357-373.
Cumming, D., and U. Walz. 2010. Private equity returns and disclosure around the world.
Journal of International Business Studies 41: 727-754.
DeFond, M., D. H. Erkens, and J. Zhang. 2017. Do client characteristics really drive the Big N
audit quality effect? New evidence from propensity score matching. Management Science
63 (11): 3628-3649.
DeFond, M., J. Hu, M. Hung, and S. Li. 2019. The effect of fair value accounting on the
performance evaluation role of earnings. Journal of Accounting and Economics 70:
101341
Denis, D. J., D. K. Denis, and A. Sarin. 1997. Agency problems, equity ownership, and corporate
diversification. Journal of Finance 52 (1): 135-160.
Drucker, J., and D. Hakim. 2021. Private inequity: How a powerful industry conquered the tax
system. New York Times June 12.
Easton, P., S. Larocque, and J. S. Stevens. 2021. Private equity valuation before and after ASC
820. Journal of Investment Management 19:4.
Edmans, A. 2014. Blockholders and corporate governance. Annual Review of Financial
Economics 6: 23-50.
Ewens, M. and J. Farre-Mensa. 2020. The deregulation of the private equity market and the
decline in IPOs. The Review of Financial Studies 33: 5463-5509.
Fan, Z., S. Radhakrishnan, and Y. Zhang. 2021. Corporate governance and earnings
management: Evidence from shareholder proposals. Contemporary Accounting Research
38 (2): 1434-1464.
Ferreira, P. H., R. Kräussal, W. R. Landsman, M. N. Borysoff, and P. F. Pope. 2019. Reliability
and relevance of fair values: Private equity investments and investee fundamentals.
Review of Accounting Studies 24: 1427-1449.
Foster, G. 1973. Stock market reaction to estimates of earnings per share by company officials.
Journal of Accounting Research 11 (1): 25-37.
Francis, J.R., E.L. Maydew, and H.C. Sparks. 1999. The role of Big 6 auditors in the credible
reporting of accruals. Auditing: A Journal of Practice & Theory 18 (2): 17–34.
Gao, C., T. D. Haight, and C. Yin. 2020. Fund selection, style allocation, and active management
abilities: Evidence from funds of hedge funds’ holdings. Financial Management 49: 135-
159.
Gao, X., T.J. Wong, L. Xia, and G. Yu. 2013. Local information advantages and the agency cost
of delegated portfolio management: Evidence from mutual funds investing in China.
Working Paper, The Chinese University of Hong Kong, Shanghai Jiaotong University,
and Harvard Business School.
Gârleanu, N. 2009. Portfolio choice and pricing in illiquid markets. Journal of Economic Theory
144: 532-564.
Gaver, J. J., P. Mason, and S. Utke. 2020. Financial reporting choices of private funds and their
implications for capital formation. Working paper.
Gaver, J., and J. S. Paterson. 2001. The association between external monitoring and earnings
management in the property-casualty insurance industry. Journal of Accounting Research
39 (2): 269-282.

38
Gensler, G. 2022. Statement on private fund advisers proposal. Available at:
https://www.sec.gov/news/statement/gensler-statement-private-fund-advisers-proposal-
020922
Gilson, R. J., and J. N. Gordon. Board 3.0 -- An introduction. The Business Lawyer 74 (Spring).
Gompers, P., and J. Lerner. 2001. The venture capital revolution. Journal of Economic
Perspectives 15 (2): 145-168.
Gow, I. D., D. F. Larcker, and P. C. Reiss. 2016. Causal inference in accounting research.
Journal of Accounting Research 54 (2): 477-523.
Griffith, E. E., J. S. Hammersley, and K. Kadous. 2015. Audits of complex estimates as
verification of management numbers: How institutional pressures shape practice.
Contemporary Accounting Research 32 (3): 833-863.
Grullon, G. and F. Wang. 2001. Closed-end fund discounts with informed ownership differential.
Journal of Financial Intermediation 10: 171-205.
Gustafson, M., J. J. Henry, E. Kim, and K. Pisciotta. 2021. The marketing of initial public
offerings. Working paper.
Hanley, K. W., A. D. Jagolinzer, and A. Nikolova. 2018. Strategic estimation of fair values.
Journal of Accounting and Economics 66: 24-45.
Hardies, K., M.-L. Vandenhaute, and D. Breesch. 2018. An analysis of auditors' going-concern
reporting accuracy in private firms. Accounting Horizons 32 (4): 117-132.
Harris, R. S., T. Jenkinson, and S. N. Kaplan. 2014. Private equity performance: What do we
know? Journal of Finance 69 (5): 1851-1882.
Harris, R. S., T. Jenkinson, S. N. Kaplan, and R. Stucke. 2018. Financial intermediation in
private equity: How well do funds of funds perform? Journal of Financial Economics
129: 287-305.
Hellmann, T., and M. Puri. 2002. Venture capital and the professionalization of start-up firms:
Empirical evidence. The Journal of Finance 57 (1): 169-197.
Inderst, R., and V. Vladimirov. 2019. Growth firms and relationship finance: A capital structure
perspective. Management Science forthcoming.
Jenkinson, T., W. R. Landsman, B. Rountree, and K. Soonawalla. 2020. Private equity net asset
values and future cash flows. The Accounting Review 95 (1): 191-210.
Jensen, M. C. 1986. Agency costs of free cash flow, corporate finance, and takeovers. The
American Economic Review 76(2): 323-329.
Jensen, M. C. 1989. Eclipse of the public corporation. Harvard Business Review 67: 60-70.
Jensen, M. C., and W. H. Meckling. 1976. Theory of the firm: Managerial behavior, agency
costs, and ownership structure. Journal of Financial Economics 3: 305-360.
Johan, S., and M. Zhang. 2020. Information asymmetries in private equity: Reporting frequency,
endowments, and governance. Journal of Business Ethics
Joyce, C. 2020. Future of Alternatives 2025: Private equity AUM will top $9tn in 2025. Preqin
Insights November 4.
Kaplan, R. S. 2011. Accounting scholarship that advances professional knowledge and practice.
The Accounting Review 86 (2): 367-383.
Kaplan, S. N., and A. Schoar. 2005. Private equity returns: Persistence and capital flows. Journal
of Finance 60 (4): 1791-1823.
Kaplan, S. N., and P. Strömberg. 2004. Characteristics, contracts, and actions: Evidence from
venture capitalist analyses. The Journal of Finance 59 (5): 2177-2210

39
Kaplan, S. N., and P. Strömberg. 2009. Leveraged buyouts and private equity. Journal of
Economic Perspectives 23: 121-146.
Katz, S. P. 2009. Earnings quality and ownership structure: The role of private equity sponsors.
The Accounting Review 84(3): 623-658.
Koo, M., K. Sivaramakrishnan, and Y. Zhao. 2020. Fair value opinion shopping or objective
valuation? Working paper
Lattman, P. 2012. Private equity industry attracts S.E.C. scrutiny. The New York Times February
12.
Lawrence, A., S. Siriviriyakul, and R. G. Sloan. 2016. Who’s the fairest of them all? Evidence
from closed-end funds. The Accounting Review 91 (1): 207-227.
Lee, R., and G. Yu. 2021. Investment in human capital and external reporting quality. Working
paper.
Li, S. Z., E. Maug, and M. Schwartz-Ziv. 2021. When shareholders disagree: Trading after
shareholder meetings. The Review of Financial Studies forthcoming.
Longstaff. F. A. 2001. Optimal portfolio choice and the valuation of illiquid securities. The
Review of Financial Studies 14 (2): 407-431.
Makan, A. 2012. Skin in the game is crucial, but how much? Financial Times November 18.
Mason, P., and S. Utke. 2020. Do acquirers attract, retain, or lose customers following mergers
and acquisitions? Evidence from the audit industry. Working paper.
Mason, P., and S. Utke. 2022. Private equity fund pricing in secondary markets: Effects of taxes
and fund structure. Journal of Alternative Investments forthcoming.
Mason, P., and B. Williams. 2020. Does IRS monitoring deter managers from committing
accounting fraud? Journal of Accounting, Auditing and Finance forthcoming.
McKinsey & Company. 2020. A new decade for private markets. McKinsey Global Private
Markets Review February 2020.
Moser, H. 2021. Dressing for the occasion? Audit quality in the presence of competition for new
clients. The Accounting Review forthcoming.
NES Financial. 2020. Outsourcing fund administration in private equity.
Phalippou, L. 2009. Beware of venturing into private equity. Journal of Economic Perspectives
23 (1): 147-166.
Phalippou, L. and O. Gottschalg. 2008. The performance of private equity funds. The Review of
Financial Studies 22 (4): 1747-1776.
Pham, P. K., N. Turner, and J. Zein. 2021. Rush to raise: Does fundraising pressure incentivize
strategic venture capital deal pricing? Working paper.
Private Fund CFO. 2020. 7 reasons US private equity funds are finally outsourcing their fund
administration. January 9.
Rajgopal, S., and M. Venkatachalam. 1997. The role of institutional investors in corporate
governance: An empirical investigation. Working paper, Stanford University.
Ramalingegowda, S., and Y. Yu. 2012. Institutional ownership and conservatism. Journal of
Accounting and Economics 53: 98-114.
Roussanov, N., H. Ruan, and Y. Wei. 2021. Marketing mutual funds. The Review of Financial
Studies 34 (6): 3045-3094.
SS&C Technologies. 2020. Private equity services.
The Clearing House (2016, July). The Custody Services of Banks.
Thinking Ahead Institute. 2021. Global Pension Assets Study.

40
Witte, P., and G. Brown. 2019. A new equilibrium: Private equity's growing role in capital
formation and the critical implications for investors. Ernst & Young and the Institute for
Private Capital.
Zimmerman, J. L. 2016. Private Equity, the Rise of Unicorns, and the Reincarnation of Control-
based Accounting. Journal of Applied Corporate Finance 28: 56-67.
Zhao, L. 2021. The effect of tax authority enforcement on earnings informativeness. European
Accounting Review forthcoming.

41
 Table 1
Sample
Panel A: Sample Selection
# of Fund-Quarters # of Funds
All liquidated U.S. fund-quarters in Preqin with NAV > 0 35,319 937

Less:
Fund-quarters missing Preqin Variables (3,673) (113)
Funds liquidated prior to 2009 or after 2017 (20,475) (285)
Fund-quarters observed more than 15 years after fund inception (1,457) (21)
Funds not able to be manually matched to Form ADV (3,361) (199)
Hedge (and other) funds that are not buyout or venture capital (1,148) (48)
Fund of funds (460) (18)
Fund-quarters not using GAAP (243) (12)

Final Sample of Funds 4,502 241

By Fund Type
Buyout (BO) 3,135 157
Venture Capital (VC) 1,367 85

Final Sample of Funds, by Fund Typea 4,502 242

a: One fund classifies itself as a BO fund in some years a VC fund in other years, leading to 1 more fund
appearing in the "by fund type" breakout than the full sample. That is, the "by fund type" breakout includes this
one fund as both a BO and VC fund.

Panel B: Sample by Year

Total Buyout (BO) % of Venture Capital % of
Year Funds Funds Total (VC) Funds Total

2009 821 528 64.31% 293 35.69%

2010 758 505 66.62% 253 33.38%

2011 756 506 66.93% 250 33.07%

2012 687 476 69.29% 211 30.71%

2013 568 405 71.30% 163 28.70%

2014 416 300 72.12% 116 27.88%

2015 269 212 78.81% 57 21.19%

2016 153 129 84.31% 24 15.69%

2017 74 74 100.00% - 0.00%

Full Sample 4,502 3,135 69.64% 1,367 30.36%

This table presents information regarding sample formation and sample statistics. Panel A provides details
regarding the sample formation whereas Panel B provides sample statistics by year including the total number of
funds as well as the number and proportion of funds by fund type.

42
 Table 2
Panel A: Buyout (BO) versus Venture Capital (VC) Funds
Buyout (BO) Funds Venture Capital (VC) Funds
Variable N Mean StdDev Q1 Median Q3 N Mean StdDev Q1 Median Q3
Accuracy 3,135 -0.126 0.143 -0.172 -0.073 -0.023 1,367 -0.122 0.148 -0.178 -0.056 ** -0.016
Bias 3,135 -0.045 0.326 -0.090 0.000 0.060 1,367 0.035 *** 0.228 -0.014 0.017 *** 0.110
Number of Owners 3,135 81.777 93.295 33.000 59.000 98.000 1,367 50.792 *** 26.043 28.000 48.000 *** 74.000
Ln_Owners 3,135 3.975 0.963 3.497 4.078 4.585 1,367 3.773 *** 0.600 3.332 3.871 *** 4.304
%Insider 3,135 4.667 7.710 1.000 2.000 5.000 1,367 2.884 *** 5.892 1.000 1.000 *** 2.000
%Own_Fund_of_Funds 3,135 14.755 14.622 3.000 11.000 22.000 1,367 21.211 *** 22.375 1.000 16.000 *** 39.000
%Own_Fgn 3,135 10.139 11.364 1.000 7.000 15.000 1,367 14.225 *** 16.028 3.000 8.000 *** 22.000
Minimum Investment (in dollars) 3,135 5,795,641 7,978,094 500,000 5,000,000 10,000,000 1,367 1,113,250 *** 2,124,815 50,000 150,000 *** 1,000,000
Ln_Min_Inv (in dollars) 3,135 13.236 4.893 13.122 15.425 16.118 1,367 11.440 *** 3.990 10.820 11.918 *** 13.816
Big4, Q1-3 3,135 0.617 0.486 0.000 1.000 1.000 1,367 0.550 *** 0.498 0.000 1.000 *** 1.000
Big 4, Q4 (audited) 3,135 0.172 0.377 0.000 0.000 0.000 1,367 0.147 ** 0.354 0.000 0.000 ** 0.000
Non-Big4, Q4 (audited) 3,135 0.045 0.207 0.000 0.000 0.000 1,367 0.072 *** 0.258 0.000 0.000 *** 0.000
Perc_Assets_Valued 3,135 2.207 14.498 0.000 0.000 0.000 1,367 1.615 12.109 0.000 0.000 ** 0.000
Admin 3,135 0.181 0.385 0.000 0.000 0.000 1,367 0.052 *** 0.222 0.000 0.000 *** 0.000
Marketer 3,135 0.068 0.251 0.000 0.000 0.000 1,367 0.019 *** 0.137 0.000 0.000 *** 0.000
Fund Size (in millions) 3,135 1,004.613 1,237.341 300.000 530.000 1,214.000 1,367 280.267 *** 210.393 158.000 225.000 *** 354.000
Ln_Ttl_Committed_Capital 3,135 6.375 1.072 5.704 6.273 7.102 1,367 5.360 *** 0.802 5.063 5.416 *** 5.869
Experience 3,135 2.400 2.061 1.000 2.000 3.000 1,367 1.803 *** 0.966 1.000 1.000 *** 3.000
Fundraising 3,135 0.019 0.136 0.000 0.000 0.000 1,367 0.004 *** 0.060 0.000 0.000 *** 0.000
Quartile 3,135 2.325 1.019 1.000 2.000 3.000 1,367 2.503 *** 1.041 2.000 2.000 *** 3.000
Exempt 3,135 0.038 0.190 0.000 0.000 0.000 1,367 0.950 *** 0.217 1.000 1.000 *** 1.000
SP500_Ret 3,135 0.035 0.079 -0.004 0.047 0.100 1,367 0.036 0.085 -0.004 0.049 0.102
SP500_Ret (lagged) 3,135 0.024 0.095 -0.010 0.047 0.100 1,367 0.023 0.103 -0.010 0.049 0.102

43
 Table 2 (Cont'd)
Descriptive Statistics
Panel B: Selected Descriptives by Year - Buyout (BO) Funds
Year Total Funds Ln_Owners %Insider %Own_Fund_of_Fund %Own_Fgn Big4 Perc_Assts_Valued Admin Marketer
2009 528 4.00 4.31 13.77 9.77 84.09% 2.08 18.75% 7.01%
2010 505 3.98 4.20 13.81 9.55 80.99% 2.18 19.01% 7.13%
2011 506 4.03 4.08 13.82 10.15 81.62% 1.58 19.17% 8.30%
2012 476 4.01 4.61 14.36 9.99 80.67% 2.10 20.38% 8.61%
2013 405 3.97 5.43 15.43 10.73 79.01% 2.07 15.06% 5.93%
2014 300 3.92 5.36 15.94 10.89 71.33% 1.45 14.33% 5.67%
2015 212 3.83 6.11 17.15 10.40 72.64% 2.00 15.09% 3.77%
2016 129 3.81 3.68 15.96 10.35 62.02% 7.13 21.71% 3.88%
2017 74 4.14 5.35 19.69 10.32 74.32% 4.05 18.92% 2.70%
Panel C: Selected Descriptives by Year - Venture Capital (VC) Funds
Year Total Funds Ln_Owners %Insider %Own_Fund_of_Fund %Own_Fgn Big4 Perc_Assts_Valued Admin Marketer
2009 293 3.71 2.33 20.53 12.76 73.04% 2.14 5.12% 1.37%
2010 253 3.74 2.57 19.84 13.21 72.73% 2.03 6.32% 1.58%
2011 250 3.77 2.61 21.42 13.61 74.00% 1.60 4.80% 1.60%
2012 211 3.82 2.82 20.99 14.46 73.93% 1.90 3.79% 1.90%
2013 163 3.81 3.02 22.31 14.26 59.51% 1.23 4.91% 2.45%
2014 116 3.83 3.52 25.56 16.09 56.03% 0.00 4.31% 3.45%
2015 57 3.79 5.26 18.23 19.51 66.67% 1.16 12.28% 3.51%
2016 24 3.86 6.83 22.33 25.25 58.33% 0.00 0.00% 0.00%
2017 - 0.00 0.00 0.00 0.00 0.00% 0.00 0.00% 0.00%

44
 Table 2 (Cont'd)
Descriptive Statistics
Panel D: Partitions by Selected Categories - Buyout (BO) Funds
Var = 1 Var = 0 1-0 Var = 1 Var = 0 1-0
Variable Accuracy Accuracy Difference Bias Bias Difference
Big4 -0.1192 -0.1490 0.0298 *** -0.0558 -0.0058 -0.0500 ***
Valuation -0.0948 -0.1265 0.0318 *** -0.0132 -0.0462 0.0331 *
Admin -0.1560 -0.1188 -0.0373 *** -0.0107 -0.0528 0.0421 ***
Marketer -0.1777 -0.1217 -0.0559 *** 0.0098 -0.0492 0.0590
Panel E: Partitions by Selected Categories - Venture Capital (VC) Funds
Var = 1 Var = 0 1-0 Var = 1 Var = 0 1-0
Variable Accuracy Accuracy Difference Bias Bias Difference
Big4 -0.1332 -0.0976 -0.0356 *** 0.0539 -0.0092 0.0631 ***
Valuation -0.1873 -0.1212 -0.0661 0.0529 0.0345 0.0184
Admin -0.1169 -0.1227 0.0058 0.1140 0.0305 0.0835 ***
Marketer -0.0235 -0.1243 0.1008 *** 0.0191 0.0351 -0.0160 **
This table presents descriptive statistics for our full sample of PE funds by quarter, and broken out between BO and VC
funds. Panel A presents descriptive statistics for our full sample. We include unlogged variables for ease of interpretation.
Panel B presents descriptive statistics broken out by BO and VC funds, and includes tests of differences at the mean and
median. Panel C (D) presents statistics by year for the number of BO (VC) funds in our sample and for the average value
of select variables of interest. Panel E (F) presents univariate analyses comparing Accuracy and Bias, dependent on the
usage of a Big 4 auditor (Big4), valuation specialist (Valuation), fund administrator (Admin), and a third-party marketer
(Marketer) for BO (VC) funds. ***, **, * indicate significant differences in means for each variable across subsamples or
categories at the one percent, five percent, or ten percent level, respectively. Appendix A provides variable definitions.

45
 Table 3
Correlations
Panel A: Buyout (BO) Funds
Variable (1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
(1) Accuracy 1.000 0.263 0.129 0.040 -0.001 0.016 0.023 0.027 0.057 -0.044
(2) Bias 0.077 1.000 -0.104 0.038 -0.008 0.111 0.112 -0.056 0.004 0.031
(3) Ln_Owners 0.110 -0.199 1.000 -0.074 0.123 0.051 -0.006 0.126 0.049 -0.099
(4) %Insider -0.054 0.109 0.129 1.000 -0.195 -0.098 -0.118 0.003 -0.011 -0.001
(5) %Own_Fund_of_Funds 0.024 -0.027 0.147 -0.117 1.000 0.177 -0.098 0.074 0.031 -0.046
(6) %Own_Fgn 0.070 0.110 0.243 0.125 0.229 1.000 -0.067 0.113 0.064 -0.076
(7) Ln_Min_Inv -0.017 0.012 0.096 -0.048 -0.002 0.101 1.000 -0.054 -0.043 0.035
(8) Big4, Q1-3 0.045 -0.059 0.151 -0.004 0.076 0.151 0.063 1.000 -0.578 -0.275
(9) Big 4, Q4 (audited) 0.058 -0.007 0.058 -0.004 0.025 0.086 0.011 -0.578 1.000 -0.099
(10) Non-Big4, Q4 (audited) -0.052 0.033 -0.114 0.008 -0.052 -0.110 -0.053 -0.275 -0.099 1.000
(11) Perc_Assets_Valued 0.039 -0.034 0.000 0.088 -0.021 -0.053 -0.091 -0.048 -0.014 0.032
(12) Admin -0.151 0.026 -0.003 0.056 -0.113 0.053 -0.026 0.021 -0.003 0.002
(13) Marketer -0.041 0.087 -0.001 0.040 -0.131 0.033 0.024 0.035 0.015 -0.009
(14) Ln_Ttl_Committed_Capital 0.133 -0.148 0.692 0.143 0.041 0.303 0.358 0.206 0.075 -0.153
(15) Experience -0.016 -0.245 0.383 0.045 -0.080 0.036 0.035 0.105 0.020 -0.062
(16) Fundraising -0.057 -0.057 -0.013 -0.009 0.015 -0.025 0.021 0.013 0.005 -0.007
(17) Quartile 0.098 -0.290 0.059 -0.064 -0.057 -0.217 0.016 0.107 0.023 -0.073
(18) Exempt -0.111 -0.020 -0.097 0.041 -0.147 -0.161 -0.139 -0.199 -0.072 0.159
(19) SP500_Ret -0.038 -0.023 0.006 -0.002 -0.011 0.004 0.003 -0.123 0.185 0.075
(20) SP500_Ret (lagged) -0.008 -0.006 0.003 -0.006 -0.014 0.006 0.001 -0.050 0.077 0.002

Panel A: Buyout (BO) Funds (Cont’d)

Variable (11) (12) (13) (14) (15) (16) (17) (18) (19) (20)
(1) Accuracy 0.012 -0.100 -0.098 0.163 -0.039 -0.039 0.067 -0.102 -0.011 0.027
(2) Bias 0.023 0.050 0.045 -0.034 -0.225 -0.038 -0.294 -0.014 -0.007 0.009
(3) Ln_Owners -0.021 0.021 0.011 0.623 0.405 -0.010 -0.016 -0.096 0.000 -0.003
(4) %Insider -0.014 0.047 -0.026 -0.025 0.035 -0.019 0.066 0.102 -0.002 0.008
(5) %Own_Fund_of_Funds -0.021 -0.049 -0.121 0.030 -0.174 0.036 -0.065 -0.102 0.000 0.006
(6) %Own_Fgn 0.025 0.035 0.107 0.154 -0.033 -0.022 -0.178 -0.126 -0.007 0.006
(7) Ln_Min_Inv 0.004 0.057 0.097 0.054 -0.103 0.021 -0.062 -0.014 0.003 -0.002
(8) Big4, Q1-3 -0.067 0.021 0.035 0.235 0.096 0.013 0.109 -0.199 -0.141 -0.048
(9) Big 4, Q4 (audited) -0.022 -0.003 0.015 0.086 0.027 0.005 0.024 -0.072 0.187 0.048
(10) Non-Big4, Q4 (audited) 0.044 0.002 -0.009 -0.172 -0.056 -0.007 -0.072 0.159 0.082 0.004
(11) Perc_Assets_Valued 1.000 0.260 -0.041 -0.090 -0.050 0.027 0.037 0.190 0.005 0.000
(12) Admin 0.195 1.000 -0.061 -0.038 0.023 0.008 -0.057 0.073 -0.003 0.000
(13) Marketer -0.048 -0.061 1.000 0.023 0.298 0.000 -0.011 -0.020 0.006 0.004
(14) Ln_Ttl_Committed_Capital -0.033 -0.005 0.034 1.000 0.340 -0.023 0.063 -0.373 0.001 -0.013
(15) Experience -0.020 -0.013 0.133 0.401 1.000 -0.009 0.026 -0.111 0.006 -0.004
(16) Fundraising 0.016 0.008 0.000 -0.032 -0.008 1.000 0.002 0.059 0.004 -0.005
(17) Quartile 0.042 -0.060 -0.016 0.065 0.025 0.004 1.000 -0.117 0.001 -0.002
(18) Exempt 0.149 0.073 -0.020 -0.254 -0.142 0.059 -0.117 1.000 -0.001 0.021
(19) SP500_Ret -0.009 0.002 0.009 0.012 0.001 0.009 0.010 -0.013 1.000 0.105
(20) SP500_Ret (lagged) -0.014 0.007 0.010 0.006 -0.001 0.001 0.010 0.004 0.019 1.000
The sample includes 3,135 fund-quarters. Pearson (Spearman) correlations are above (below) the diagonal. Bolded correlations are significant at
the 0.05 level (two-tailed). Variables are defined in Appendix A.

46
 Table 3 (Cont'd)
Correlations
Panel B: Venture Capital (VC) Funds
Variable (1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
(1) Accuracy 1.000 -0.282 0.000 0.017 0.120 -0.062 0.028 -0.084 -0.026 0.060
(2) Bias -0.359 1.000 -0.081 -0.121 -0.022 0.032 -0.062 0.094 0.032 -0.065
(3) Ln_Owners 0.061 -0.140 1.000 -0.346 0.216 -0.038 -0.132 0.078 0.023 -0.035
(4) %Insider -0.028 -0.083 0.123 1.000 -0.087 0.115 0.008 0.011 0.009 -0.013
(5) %Own_Fund_of_Funds 0.133 -0.012 0.214 -0.107 1.000 0.228 -0.078 0.067 0.007 -0.033
(6) %Own_Fgn -0.157 0.132 0.070 0.301 0.290 1.000 0.190 -0.059 -0.016 0.035
(7) Ln_Min_Inv -0.027 0.077 -0.183 -0.080 0.052 0.153 1.000 -0.069 0.026 0.024
(8) Big4, Q1-3 -0.118 0.132 0.099 0.022 0.055 0.013 -0.111 1.000 -0.459 -0.307
(9) Big 4, Q4 (audited) -0.032 0.032 0.035 0.011 0.020 0.024 -0.011 -0.459 1.000 -0.115
(10) Non-Big4, Q4 (audited) 0.078 -0.077 -0.050 -0.012 -0.031 -0.015 0.056 -0.307 -0.115 1.000
(11) Perc_Assets_Valued -0.025 -0.048 -0.213 -0.022 -0.032 -0.095 0.138 -0.086 -0.041 0.047
(12) Admin 0.063 0.069 -0.154 -0.059 -0.177 -0.062 0.223 -0.166 -0.051 0.088
(13) Marketer 0.101 -0.006 0.071 0.132 0.087 0.101 -0.023 0.061 0.033 -0.039
(14) Ln_Ttl_Committed_Capital 0.010 -0.037 0.436 0.126 0.314 0.269 0.166 0.092 0.011 -0.045
(15) Experience 0.052 0.088 0.238 -0.147 0.228 -0.175 -0.143 0.075 -0.019 -0.025
(16) Fundraising 0.006 0.022 -0.028 -0.013 0.030 -0.008 -0.019 0.006 0.043 -0.017
(17) Quartile 0.134 -0.239 0.078 0.252 -0.061 -0.154 -0.136 -0.002 0.004 0.005
(18) Exempt 0.103 0.094 -0.165 0.041 -0.102 0.024 -0.043 -0.133 -0.010 0.064
(19) SP500_Ret -0.049 0.037 -0.021 -0.022 -0.011 -0.022 -0.002 -0.101 0.146 0.093
(20) SP500_Ret (lagged) -0.038 0.019 -0.004 -0.014 -0.004 -0.009 0.004 -0.044 0.073 0.046

Panel B: Venture Capital (VC) Funds (Cont’d)

Variable (11) (12) (13) (14) (15) (16) (17) (18) (19) (20)
(1) Accuracy -0.026 0.009 0.093 -0.037 0.070 0.023 0.074 0.021 -0.011 0.023
(2) Bias -0.021 0.081 -0.010 -0.067 0.100 0.007 -0.237 0.092 0.012 -0.020
(3) Ln_Owners -0.237 -0.136 0.079 0.433 0.241 -0.014 0.019 -0.182 -0.010 0.013
(4) %Insider -0.039 0.016 0.159 0.009 -0.103 -0.017 0.252 0.045 -0.007 0.004
(5) %Own_Fund_of_Funds -0.071 -0.164 0.048 0.307 0.204 0.014 -0.141 -0.081 -0.002 0.005
(6) %Own_Fgn -0.029 -0.094 0.050 0.158 -0.179 -0.026 -0.146 0.065 -0.005 0.006
(7) Ln_Min_Inv 0.085 0.161 0.010 0.008 -0.235 0.002 -0.188 -0.083 -0.001 0.018
(8) Big4, Q1-3 -0.105 -0.166 0.061 0.118 0.070 0.006 -0.003 -0.133 -0.121 -0.045
(9) Big 4, Q4 (audited) -0.046 -0.051 0.033 0.026 -0.016 0.043 0.004 -0.010 0.156 0.049
(10) Non-Big4, Q4 (audited) 0.057 0.088 -0.039 -0.064 -0.023 -0.017 0.005 0.064 0.108 0.043
(11) Perc_Assets_Valued 1.000 -0.031 -0.019 -0.217 -0.111 -0.008 -0.125 0.031 -0.002 -0.002
(12) Admin -0.032 1.000 -0.033 -0.003 -0.195 -0.014 -0.028 0.054 0.005 -0.001
(13) Marketer -0.019 -0.033 1.000 0.010 0.317 -0.008 0.200 0.032 -0.001 0.005
(14) Ln_Ttl_Committed_Capital -0.188 0.011 0.007 1.000 0.080 -0.096 0.107 -0.254 -0.015 0.017
(15) Experience -0.125 -0.214 0.245 0.186 1.000 0.050 -0.054 -0.179 0.002 0.005
(16) Fundraising -0.008 -0.014 -0.008 -0.074 0.051 1.000 -0.029 0.014 0.013 0.006
(17) Quartile -0.093 -0.028 0.196 0.098 -0.075 -0.031 1.000 -0.051 -0.014 -0.007
(18) Exempt 0.031 0.054 0.032 -0.223 -0.215 0.014 -0.048 1.000 0.014 0.008
(19) SP500_Ret 0.006 0.009 -0.008 -0.022 0.001 0.005 -0.026 0.032 1.000 0.131
(20) SP500_Ret (lagged) 0.009 0.001 -0.003 -0.001 0.011 -0.009 -0.020 0.020 0.055 1.000
The sample includes 1,367 fund-quarters. Pearson (Spearman) correlations are above (below) the diagonal. Bolded correlations are significant at
the 0.05 level (two-tailed). Variables are defined in Appendix A.

47
 Table 4
Multiple Regression Analyses of NAV Accuracy and Bias - Buyout (BO) Funds
Panel A: Accuracy
Accuracy or Biasiq = α0 + α1*Ln_Ownersit + α2*%Insiderit + α3* %Own_Fund_of_Fundit + α4*%Own_Fgnit + α5*Ln_Min_Invit + α6*Big4, Q1-3it + α7*Big4, Q4iq + α8*Non-Big4, Q4it +
α9*Perc_Assets_Valuedit + α10*Adminit + α11*Marketerit + α12*Ln_Ttl_Committed_Capitali + α13*Experienceit + α14*Fundraisingiq + α15*Quartileit + α16*Exemptit + α17*SP500_Retq +
α18*SP500_Retq-1 + αt*ΣFund_Ageit + eiq
(1) (2) (3) (4) (5)
Sample All BO Funds Large BO Funds Small BO Funds Top Perf BO Funds Bottom Perf BO Funds
Dependent Variable Accuracy Accuracy Accuracy Accuracy Accuracy
Variables Coef. (t-stat) Coef. (t-stat) Coef. (t-stat) Coef. (t-stat) Coef. (t-stat)
Investor Monitoring
Ln_Owners 0.0132 (1.086) 0.0072 (0.374) 0.0136 (0.633) -0.0262 (-0.769) 0.0230 (0.730)
%Insider 0.0007 (0.741) 0.0014 (1.470) 0.0048** (2.542) -0.0031 (-1.332) -0.0039 (-1.179)
%Own_Fund_of_Funds -0.0008 (-1.327) -0.0005 (-0.803) -0.0015 (-0.924) -0.0086** (-2.309) 0.0002 (0.190)
%Own_Fgn 0.0003 (0.395) -0.0010 (-0.824) 0.0024** (2.224) 0.0117** (2.339) 0.0013 (0.779)
Ln_Min_inv 0.0006 (0.331) -0.0003 (-0.193) -0.0016 (-0.392) -0.0077 (-1.299) -0.0043** (-2.294)
Auditor Involvement
Big4, Q1-3 0.0093 (0.431) 0.0431* (1.859) -0.0182 (-0.408) -0.0975*** (-3.354) 0.0039 (0.057)
Big 4, Q4 (audited) 0.0271 (1.337) 0.0566** (2.480) 0.0183 (0.436) -0.0690** (-2.337) 0.0332 (0.526)
Non-Big4, Q4 (audited) 0.0087 (1.604) 0.0087 (1.499) 0.0065 (0.850) -0.0347 (-1.151) 0.0215 (1.299)
Third-Party Involvement
Perc_Assets_Valued 0.0008* (1.659) -0.0050 (-0.943) 0.0019** (2.501)
Admin -0.0504* (-1.925) -0.0768*** (-3.546) -0.0324 (-0.427) 0.0189 (0.379) -0.1314* (-1.768)
Marketer -0.0260 (-0.605) 0.1011** (2.027) -0.0218 (-0.505) 0.0876 (1.223) 0.0048 (0.080)
Fund Attributes
Ln_Ttl_Committed_Capital 0.0044 (0.401) -0.0324** (-2.402) 0.0347 (1.190) 0.0651 (1.659) 0.0488* (2.019)
Experience -0.0030 (-0.527) -0.0141** (-2.370) -0.0083 (-0.568) -0.0037 (-0.438) -0.0038 (-0.636)
Fundraising 0.0350 (1.413) 0.0057 (0.120) 0.0689* (1.778) 0.0881 (1.303) 0.0886 (1.563)
Quartile -0.0016 (-0.150) -0.0110 (-1.184) 0.0133 (0.484) n/a n/a n/a n/a
Exempt -0.0282 (-0.535) -0.0469 (-0.650) -0.0481 (-0.540)
Market Attributes
SP500_Ret -0.0166 (-1.048) -0.0217 (-0.965) -0.0499* (-1.869) -0.0496 (-1.585) -0.0210 (-0.733)
SP500_Ret_Lag 0.0072 (0.617) 0.0225 (1.273) -0.0078 (-0.307) -0.0154 (-0.488) 0.0146 (0.493)
Constant -0.2569*** (-3.440) -0.0944 (-0.875) -0.4443** (-2.358) -0.1619 (-0.943) -0.6534*** (-3.539)
Big 4: Audited - Unaudited 0.0178*** 0.0135*** 0.0363** 0.0028** 0.0293**
Observations 3,135 1,045 1,045 459 823
Adjusted R-squared 0.241 0.414 0.198 0.454 0.401
Fund Age FE Yes Yes Yes Yes Yes
Clustered SE By Fund By Fund By Fund By Fund By Fund

48
 Table 4 (Cont'd)
Multiple Regression Analyses of NAV Accuracy and Bias - Buyout (BO) Funds
Panel B: Bias
Accuracy or Biasiq = α0 + α1*Ln_Ownersit + α2*%Insiderit + α3* %Own_Fund_of_Fundit + α4*%Own_Fgnit + α5*Ln_Min_Invit + α6*Big4, Q1-3it + α7*Big4, Q4iq + α8*Non-Big4, Q4it +
α9*Perc_Assets_Valuedit + α10*Adminit + α11*Marketerit + α12*Ln_Ttl_Committed_Capitali + α13*Experienceit + α14*Fundraisingiq + α15*Quartileit + α16*Exemptit + α17*SP500_Retq + α18*SP500_Retq-1 +
αt*ΣFund_Ageit + eiq
(1) (2) (3) (4) (5)
Sample All BO Funds Large BO Funds Small BO Funds Top Perf BO Funds Bottom Perf BO Funds
Dependent Variable Bias Bias Bias Bias Bias
Variables Coef. (t-stat) Coef. (t-stat) Coef. (t-stat) Coef. (t-stat) Coef. (t-stat)
Investor Monitoring
Ln_Owners -0.0182 (-0.657) 0.0397 (0.990) -0.0816* (-1.785) -0.0799 (-0.812) -0.0529 (-0.850)
%Insider 0.0031* (1.923) -0.0040 (-1.661) 0.0113** (2.385) -0.0066 (-1.014) 0.0071 (1.114)
%Own_Fund_of_Funds -0.0011 (-0.897) -0.0021 (-1.269) -0.0009 (-0.397) -0.0215** (-2.095) -0.0021 (-1.034)
%Own_Fgn 0.0011 (0.735) -0.0026 (-0.942) 0.0032 (1.449) 0.0309* (2.028) -0.0002 (-0.072)
Ln_Min_inv 0.0033 (1.040) 0.0010 (0.311) 0.0031 (0.445) -0.0202 (-1.202) 0.0025 (0.615)
Auditor Involvement
Big4, Q1-3 -0.0211 (-0.463) -0.0616 (-0.720) -0.0616 (-0.876) -0.4204 (-1.659) 0.0245 (0.242)
Big 4, Q4 (audited) -0.0054 (-0.121) -0.0298 (-0.353) -0.0467 (-0.672) -0.3378 (-1.317) 0.0023 (0.024)
Non-Big4, Q4 (audited) 0.0138 (0.764) 0.0469 (1.662) -0.0079 (-0.598) -0.0117 (-0.515) 0.0194 (0.561)
Third-Party Involvement
Perc_Assets_Valued 0.0007 (0.923) 0.0144 (1.137) -0.0007 (-0.431) - -
Admin 0.0307 (0.673) -0.0669 (-1.206) 0.2296** (2.255) 0.1039 (0.738) 0.1691 (1.593)
Marketer 0.1551** (2.233) 0.0401 (0.287) -0.0762 (-0.693) 0.3222* (1.875) 0.1137 (1.637)
Fund Attributes
Ln_Ttl_Committed_Capital 0.0228 (0.788) -0.0789*** (-2.825) 0.1932** (2.460) 0.2053* (1.988) -0.0374 (-0.700)
Experience -0.0410*** (-3.304) -0.0364*** (-2.888) -0.0268 (-0.836) -0.0498 (-1.534) -0.0318** (-2.541)
Fundraising -0.0572 (-1.144) -0.0331 (-0.402) -0.0179 (-0.254) 0.4027 (0.936) -0.1452* (-1.696)
Quartile -0.0955*** (-4.060) -0.0891*** (-3.342) -0.1327** (-2.456) n/a n/a n/a n/a
Exempt -0.1155 (-0.711) -0.0673 (-0.276) -0.3513 (-1.364)
Market Attributes
SP500_Ret -0.0347 (-0.838) -0.1016 (-1.608) -0.0077 (-0.150) -0.1577* (-1.777) 0.0154 (0.325)
SP500_Ret_Lag 0.0163 (0.636) 0.0031 (0.093) -0.0021 (-0.034) -0.1095 (-0.819) 0.0877* (1.814)
Constant 0.0705 (0.406) 0.2801 (1.203) -0.2884 (-0.526) -0.2271 (-0.558) 0.7584** (2.669)
Big 4: Audited - Unaudited 0.0157 0.0320* 0.0150 0.0830** -0.0222
Observations 3,135 1,045 1,045 459 823
Adjusted R-squared 0.170 0.242 0.242 0.260 0.291
Fund Age FE Yes Yes Yes Yes Yes
Clustered SE By Fund By Fund By Fund By Fund By Fund
This table presents OLS regression estimation results of NAV accuracy and bias on external monitoring, third-party provider, and fund and market attributes. Accuracy is scaled to be increasing in accuracy
relative to realized fund distributions. Bias is scaled to be increasing in bias relative to realized fund distributions. For ease of presentation, we group attributes into five categories, Investor Monitoring, Auditor
Involvement, Third-Party Involvement, Fund Attributes, and Market Attributes. Panel A (B) presents results for Accuracy (Bias). Column 1 reports the results for the full BO fund sample. Column 2 (3) reports
results for large (small) BO funds. We define a large (small) BO fund by whether the fund is in the top (bottom) tercile of Ln_Ttl_Committed_Capital in the BO sample. Column 4 (5) reports results for top
(bottom) performing BO funds. We define a top (bottom) performing BO fund by whether the fund is in Preqin's top (bottom) quartile. We cluster the standard errors by fund. Appendix A provides variable
definitions. Robust t-statistics are in parentheses with ***, **, and * indicating significance at the one percent, five percent, or ten percent level, respectively.

49
 Table 5
Multiple Regression Analyses of NAV Accuracy and Bias - Venture Capital (VC) Funds
Panel A: Accuracy
Accuracy or Biasiq = α0 + α1*Ln_Ownersit + α2*%Insiderit + α3* %Own_Fund_of_Fundit + α4*%Own_Fgnit + α5*Ln_Min_Invit + α6*Big4, Q1-3it + α7*Big4, Q4iq + α8*Non-Big4, Q4it +
α9*Perc_Assets_Valuedit + α10*Adminit + α11*Marketerit + α12*Ln_Ttl_Committed_Capitali + α13*Experienceit + α14*Fundraisingiq + α15*Quartileit + α16*Exemptit + α17*SP500_Retq +
α18*SP500_Retq-1 + αt*ΣFund_Ageit + eiq
(1) (2) (3) (4) (5)
Sample All VC Funds Large VC Funds Small VC Funds Top Perf VC Funds Bottom Perf VC Funds
Dependent Variable Accuracy Accuracy Accuracy Accuracy Accuracy
Variables Coef. (t-stat) Coef. (t-stat) Coef. (t-stat) Coef. (t-stat) Coef. (t-stat)
Investor Monitoring
Ln_Owners -0.0031 (-0.158) -0.0243 (-0.539) 0.0106 (0.661) 0.0528 (1.037) -0.0105 (-0.444)
%Insider 0.0001 (0.085) 0.0091 (1.265) 0.0009 (0.804) 0.0017 (0.530) 0.0106* (2.008)
%Own_Fund_of_Funds 0.0002 (0.510) -0.0011 (-1.093) 0.0006 (1.216) 0.0002 (0.175) -0.0016* (-2.034)
%Own_Fgn -0.0001 (-0.099) -0.0019 (-1.578) -0.0010 (-0.989) -0.0070*** (-3.092) 0.0066*** (3.145)
Ln_Min_inv 0.0018 (0.716) 0.0048 (1.144) -0.0037 (-1.328) -0.0053** (-2.291) -0.0067 (-1.052)
Auditor Involvement
Big4, Q1-3 -0.0300 (-1.463) 0.0504 (1.156) -0.0665* (-2.013) 0.0231 (0.563) -0.0386 (-1.489)
Big 4, Q4 (audited) -0.0205 (-1.005) 0.0572 (1.268) -0.0479 (-1.489) 0.0340 (0.953) -0.0070 (-0.303)
Non-Big4, Q4 (audited) 0.0134*** (2.897) 0.0021 (1.184) 0.0151** (2.057) 0.0005 (0.132) 0.0231** (2.120)
Third-Party Involvement
Perc_Assets_Valued -0.0005 (-0.516) -0.0000 (-0.046) -0.0005 (-0.782) 0.0044 (1.122) 0.0011* (2.096)
Admin 0.0721** (2.239) -0.0404 (-0.453) 0.0544 (1.024) 0.1468** (2.123) 0.0757* (2.030)
Marketer 0.0500 (1.125) 0.2920*** (3.378)
Fund Attributes
Ln_Ttl_Committed_Capital -0.0239 (-1.294) 0.0971 (1.490) -0.0148 (-0.611) 0.0621* (2.025) 0.0089 (0.195)
Experience 0.0235* (1.680) 0.0067 (0.371) 0.0116 (0.654) -0.0494* (-1.865) -0.0172 (-0.633)
Fundraising 0.0377 (1.172) 0.0785 (1.377)
Quartile -0.0005 (-0.037) 0.0102 (0.531) -0.0178 (-1.000) n/a n/a n/a n/a
Exempt 0.0526 (1.484) 0.1724** (2.411) 0.0759** (2.056) 0.0111 (0.170) 0.0957*** (3.612)
Market Attributes
SP500_Ret 0.0206 (1.338) 0.0244 (1.300) -0.0055 (-0.261) 0.0342 (1.259) 0.0195 (0.718)
SP500_Ret_Lag -0.0335* (-1.792) -0.0450* (-1.853) -0.0045 (-0.127) -0.0662* (-2.066) -0.0172 (-1.197)
Constant -0.4549*** (-3.898) -1.2071*** (-3.097) -0.3271*** (-3.221) -0.5557*** (-2.879) -0.6665** (-2.567)
Big 4: Audited - Unaudited 0.0095 0.0068 0.0190** 0.0109 0.0310**
Observations 1,367 457 460 293 276
Adjusted R-squared 0.468 0.528 0.449 0.584 0.720
Fund Age FE Yes Yes Yes Yes Yes
Clustered SE By Fund By Fund By Fund By Fund By Fund

50
 Table 5 (Cont'd)
Multiple Regression Analyses of NAV Accuracy and Bias - Venture Capital (VC) Funds
Panel B: Bias
Accuracy or Biasiq = α0 + α1*Ln_Ownersit + α2*%Insiderit + α3* %Own_Fund_of_Fundit + α4*%Own_Fgnit + α5*Ln_Min_Invit + α6*Big4, Q1-3it + α7*Big4, Q4iq + α8*Non-Big4, Q4it +
α9*Perc_Assets_Valuedit + α10*Adminit + α11*Marketerit + α12*Ln_Ttl_Committed_Capitali + α13*Experienceit + α14*Fundraisingiq + α15*Quartileit + α16*Exemptit + α17*SP500_Retq +
α18*SP500_Retq-1 + αt*ΣFund_Ageit + eiq
(1) (2) (3) (4) (5)
Sample All VC Funds Large VC Funds Small VC Funds Top Perf VC Funds Bottom Perf VC Funds
Dependent Variable Bias Bias Bias Bias Bias
Variables Coef. (t-stat) Coef. (t-stat) Coef. (t-stat) Coef. (t-stat) Coef. (t-stat)
Investor Monitoring
Ln_Owners -0.0571* (-1.837) -0.0424 (-0.757) -0.0593 (-1.551) -0.0368 (-0.345) -0.0077 (-0.311)
%Insider -0.0044** (-2.566) -0.0190** (-2.101) -0.0054** (-2.361) -0.0046 (-0.687) -0.0103 (-1.662)
%Own_Fund_of_Funds -0.0001 (-0.142) 0.0011 (0.987) -0.0011 (-0.968) -0.0028 (-1.151) 0.0016 (1.710)
%Own_Fgn 0.0009 (0.754) -0.0001 (-0.031) 0.0033** (2.718) -0.0158*** (-2.936) -0.0058** (-2.380)
Ln_Min_inv -0.0037 (-1.029) -0.0063 (-0.859) 0.0035 (0.670) 0.0069 (1.708) 0.0074 (1.107)

Auditor Involvement
Big4, Q1-3 0.0814** (2.071) 0.1584** (2.057) 0.1142** (2.153) -0.1884* (-1.761) 0.0681** (2.427)
Big 4, Q4 (audited) 0.0766* (1.889) 0.1737** (2.123) 0.1092** (2.112) -0.1639 (-1.618) 0.0401 (1.458)
Non-Big4, Q4 (audited) -0.0132** (-2.188) -0.0070 (-0.811) -0.0119 (-1.180) -0.0041 (-0.517) -0.0270** (-2.481)
Third-Party Involvement
Perc_Assets_Valued -0.0002 (-0.157) 0.0003 (0.266) -0.0002 (-0.216) 0.0273*** (3.145) -0.0026*** (-4.271)
Admin 0.0741 (1.244) 0.3751** (2.519) 0.0744 (0.636) 0.0017 (0.011) -0.0399 (-0.868)
Marketer 0.0603 (0.885) 0.5276** (2.662)
Fund Attributes
Ln_Ttl_Committed_Capital 0.0170 (0.690) -0.0067 (-0.068) -0.0562 (-1.293) -0.0351 (-0.654) -0.0090 (-0.178)
Experience 0.0146 (0.731) 0.0355 (1.175) 0.0104 (0.311) -0.0451 (-0.775) 0.0213 (0.638)
Fundraising 0.0238 (0.413) -0.0089 (-0.137)
Quartile -0.0368* (-1.757) -0.0504* (-1.734) 0.0064 (0.280) n/a n/a n/a n/a
Exempt 0.0762 (0.906) 0.2592** (2.475) 0.0287 (0.418) -0.1316 (-1.617) -0.1102*** (-3.668)
Market Attributes
SP500_Ret 0.0016 (0.080) 0.0321 (1.005) -0.0070 (-0.263) 0.0582 (1.132) -0.0399 (-1.430)
SP500_Ret_Lag -0.0060 (-0.178) -0.0198 (-0.501) 0.0300 (0.592) -0.0825 (-0.860) 0.0286 (1.496)
Constant 0.5339*** (2.889) 0.3322 (0.551) 0.4967** (2.223) 0.7008 (1.439) 0.6974** (2.492)
Big 4: Audited - Unaudited -0.0048 0.0153 -0.0050 0.0250 -0.0280**
Observations 1,367 457 460 293 276
Adjusted R-squared 0.264 0.439 0.244 0.438 0.692
Fund Age FE Yes Yes Yes Yes Yes
Clustered SE By Fund By Fund By Fund By Fund By Fund
This table presents OLS regression estimation results of NAV accuracy and bias on external monitoring, third-party provider, and fund and market attributes. Accuracy is scaled to be increasing in accuracy
relative to realized fund distributions. Bias is scaled to be increasing in bias relative to realized fund distributions. For ease of presentation, we group attributes into five categories, Investor Monitoring, Auditor
Involvement, Third-Party Involvement, Fund Attributes, and Market Attributes. Panel A (B) presents results for Accuracy (Bias). Column 1 reports the results for the full VC fund sample. Column 2 (3) reports
results for large (small) VC funds. We define a large (small) VC fund by whether the fund is in the top (bottom) tercile of Ln_Ttl_Committed_Capital in the VC sample. Column 4 (5) reports results for top
(bottom) performing VC funds. We define a top (bottom) performing VC fund by whether the fund is in Preqin's top (bottom) quartile. We cluster the standard errors by fund. Appendix A provides variable
definitions. Robust t-statistics are in parentheses with ***, **, and * indicating significance at the one percent, five percent, or ten percent level, respectively.

51
 Table 6
Multiple Regression Analyses of Use of Third-Party Service Providers
Panel A: Buyout (BO) Funds
DVit = α0 + α1*Ln_Ownersit + α2*%Insiderit + α3* %Own_Fund_of_Fundit + α4*%Own_Fgnit + α5*Ln_Min_Invit + α6*Big4it +
[α7*Adminit] + [α8*Marketerit] + [α9*Valuationit] + α10*Ln_Ttl_Committed_Capitali + α11*Experienceit + α12*Fundraisingiq +
α13*Quartileit + α14*Exemptit + α15*SP500_Ret_12mot + α16*SP500_Ret_12mot-1 + αt*ΣFund_Ageit + eit
(1) (2) (3) (4)
Dependent Variable Valuation Perc_Assets_Valued Admin Marketer
Variables Coef. (t-stat) Coef. (t-stat) Coef. (t-stat) Coef. (t-stat)

Investor Monitoring
Ln_Owners 0.0099 (0.745) 0.3974 (0.409) 0.0177 (0.382) -0.0173 (-0.858)
%Insider 0.0006 (0.427) -0.0609 (-0.680) 0.0021 (0.598) -0.0008 (-0.527)
%Own_Fund_of_Funds -0.0001 (-0.154) 0.0006 (0.015) -0.0014 (-0.530) -0.0010 (-0.962)
%Own_Fgn 0.0000 (0.011) 0.0624 (0.699) 0.0023 (0.832) 0.0034 (1.185)
Ln_Min_inv 0.0003 (0.196) -0.0445 (-0.353) 0.0076 (1.207) 0.0076** (2.379)

Auditor Involvement
Big4 -0.0250 (-0.953) -3.2319 (-1.364) 0.0747 (0.967) 0.0376 (0.794)

Third-Party Involvement
Admin 0.0800 (1.300) 9.1064 (1.501) -0.0578 (-1.185)
Marketer -0.0134 (-0.885) -0.7149 (-0.425) -0.1493 (-1.212)
Valuation 0.3938* (1.956) -0.0255 (-0.914)

Fund Attributes
Ln_Ttl_Committed_Capital -0.0008 (-0.100) -0.0782 (-0.110) -0.0409 (-0.892) -0.0191 (-1.098)
Experience -0.0024 (-0.754) -0.2601 (-0.868) 0.0125 (0.747) 0.0459*** (3.107)
Fundraising 0.0297 (0.607) 3.1961 (0.669) 0.0353 (0.382) -0.0080 (-0.142)
Quartile 0.0111 (1.014) 1.3678 (1.288) -0.0235 (-0.709) 0.0020 (0.095)
Exempt 0.0957 (0.910) 10.4442 (1.038) 0.0583 (0.451) 0.0205 (0.368)

Market Attributes
SP500_Ret_12mo 0.0167 (0.632) 0.0481 (0.024) -0.0815 (-0.894) -0.0518 (-0.789)
SP500_Ret_12mo_Lag 0.0126 (0.516) -0.9118 (-0.480) -0.0698 (-1.517) 0.0007 (0.018)

Constant -0.0578 (-0.817) -3.6313 (-0.662) 0.1158 (0.614) 0.4352** (2.118)

Observations 856 856 856 856

Adjusted R-squared 0.034 0.083 0.050 0.168
Fund Age FE Yes Yes Yes Yes
Clustered SE By Fund By Fund By Fund By Fund

52
 Table 6 (Cont'd)
Multiple Regression Analyses of Use of Third-Party Service Providers
Panel B: Venture Capital (VC) Funds
DVit = α0 + α1*Ln_Ownersit + α2*%Insiderit + α3* %Own_Fund_of_Fundit + α4*%Own_Fgnit + α5*Ln_Min_Invit + α6*Big4it +
[α7*Adminit] + [α8*Marketerit] + [α9*Valuationit] + α10*Ln_Ttl_Committed_Capitali + α11*Experienceit + α12*Fundraisingiq +
α13*Quartileit + α14*Exemptit + α15*SP500_Ret_12mot + α16*SP500_Ret_12mot-1 + αt*ΣFund_Ageit + eit
(1) (2) (3) (4)
Dependent Variable Valuation Perc_Assets_Valued Admin Marketer
Variables Coef. (t-stat) Coef. (t-stat) Coef. (t-stat) Coef. (t-stat)

Investor Monitoring
Ln_Owners -0.0587* (-1.705) -4.7494 (-1.553) -0.0515 (-0.885) 0.0291 (1.125)
%Insider -0.0034 (-1.662) -0.2542 (-1.589) -0.0021 (-0.711) 0.0035 (1.012)
%Own_Fund_of_Funds 0.0000 (0.123) 0.0010 (0.031) -0.0005 (-0.771) 0.0000 (0.073)
%Own_Fgn -0.0001 (-0.081) -0.0702 (-0.481) -0.0026* (-1.917) 0.0012 (1.135)
Ln_Min_inv 0.0015 (0.937) 0.0932 (0.699) 0.0059 (1.329) 0.0044 (1.143)

Auditor Involvement
Big4 -0.0310 (-0.639) -3.9247 (-0.843) -0.1015 (-1.392) 0.0356 (1.152)

Third-Party Involvement
Admin -0.0681 (-1.080) -6.9506 (-1.119) 0.0482 (0.958)
Marketer 0.0611 (0.665) 7.6202 (0.877) 0.1430 (1.240)
Valuation -0.1577 (-1.434) 0.0477 (0.703)

Fund Attributes
Ln_Ttl_Committed_Capital -0.0187 (-1.009) -1.4059 (-1.023) 0.0348 (0.690) -0.0208 (-1.074)
Experience -0.0140 (-0.786) -1.5856 (-0.934) -0.0488* (-1.794) 0.0586 (1.227)
Fundraising -0.0506 (-1.318) -4.0305 (-1.319) 0.0090 (0.117) -0.0609 (-0.904)
Quartile -0.0039 (-0.162) -1.1496 (-0.543) -0.0072 (-0.193) 0.0310 (1.219)
Exempt -0.0351 (-1.007) -3.4420 (-1.039) -0.0128 (-0.340) 0.0669 (1.137)

Market Attributes
SP500_Ret_12mo -0.0573 (-1.075) -5.4964 (-1.067) -0.0370 (-0.514) 0.0218 (0.588)
SP500_Ret_12mo_Lag -0.0236 (-0.690) -1.3005 (-0.581) 0.0505 (0.937) 0.0120 (0.664)

Constant 0.4246 (1.266) 40.1690 (1.216) 0.3067 (1.487) -0.4050 (-1.217)

Observations 384 384 384 384

Adjusted R-squared 0.075 0.085 0.115 0.178
Fund Age FE Yes Yes Yes Yes
Clustered SE By Fund By Fund By Fund By Fund
This table presents OLS regression estimation results for determinant models predicting the usage of third-party service providers
based on external monitoring, third-party provider, and fund and market attributes. For ease of presentation, we group attributes
into five categories, Investor Monitoring, Auditor Involvement, Third-Party Involvement, Fund Attributes, and Market Attributes.
Fundraising is calibrated to be set to 1 for fundraising in the next quarter at any point in the year (rather than only in the next
quarter, as it is elsewhere), and 0 otherwise. Panel A (B) presents results for BO (VC) funds. The dependent variable in Column 1
is Valuation. Column 2 uses Perc_Assets_Valued as the dependent variable. Columns 3 and 4 use Admin and Marketer as
dependent variables, respectively. The Third-Party Involvement group excludes the indicator for the respective dependent
variable in each regression. Appendix A provides variable definitions. Robust t-statistics are in parentheses with ***, **, and *
indicating significance at the one percent, five percent, or ten percent level, respectively.

53
 Table 7
Multiple Regression Analyses of NAV Adjustment for Buyout (BO) Funds
Panel A: Buyout (BO) Funds Only
>10 Pct Adj or Pos Adjiq = α0 + α1*Ln_Ownersit + α2*%Insiderit + α3* %Own_Fund_of_Fundit + α4*%Own_Fgnit +
α5*Ln_Min_Invit + α6*Big4, Q1-3it + α7*Big4, Q4iq + α8*Non-Big4, Q4it + α9*Perc_Assets_Valuedit + α10*Adminit +
α11*Marketerit + α12*Ln_Ttl_Committed_Capitali + α13*Experienceit + α14*Fundraisingiq + α15*Quartileit +
α16*Exemptit + α17*SP500_Retq + α18*SP500_Retq-1 + αt*ΣFund_Ageit + eiq
(1) (2)
Dependent Variable > 10 Pct Adj Pos Adj
Variables Coef. (t-stat) Coef. (t-stat)

Investor Monitoring
Ln_Owners -0.0256 (-1.502) -0.0372** (-2.134)
%Insider 0.0001 (0.085) 0.0026 (1.296)
%Own_Fund_of_Funds -0.0012 (-1.344) -0.0004 (-0.413)
%Own_Fgn 0.0002 (0.126) 0.0005 (0.543)
Ln_Min_inv -0.0047** (-1.980) -0.0006 (-0.389)

Auditor Involvement
Big4, Q1-3 0.0284 (0.728) -0.0153 (-0.378)
Big 4, Q4 (audited) 0.1125*** (2.706) 0.0519 (1.177)
Non-Big4, Q4 (audited) 0.1156*** (3.317) -0.0023 (-0.038)

Third-Party Involvement
Perc_Assets_Valued 0.0004 (0.371) 0.0010** (2.149)
Admin -0.0174 (-0.544) -0.0542 (-1.422)
Marketer 0.0508 (0.766) -0.1415*** (-2.783)

Fund Attributes
Ln_Ttl_Committed_Capital 0.0081 (0.470) 0.0294* (1.926)
Experience 0.0107 (1.120) 0.0299*** (5.053)
Fundraising 0.0554 (0.812) 0.0297 (0.565)
Quartile 0.0134 (0.991) 0.0382*** (2.775)
Exempt -0.0298 (-0.304) -0.1206 (-1.254)

Market Attributes
SP500_Ret 0.0829 (0.784) 0.4207*** (3.400)
SP500_Ret_Lag -0.1564** (-2.027) 0.5225*** (5.632)

Constant 0.5372*** (3.698) 0.2120 (1.406)

Big 4: Audited - Unaudited 0.0841*** 0.0669***

Observations 3,135 3,135
Adjusted R-squared 0.028 0.050
Fund Age FE Yes Yes
Clustered SE By Fund By Fund

54
 Table 7 (Cont'd)
Multiple Regression Analyses of NAV Adjustment for Venture Capital (VC) Funds
Panel B: Venture Capital (VC) Funds Only

>10 Pct Adj or Pos Adjiq = α0 + α1*Ln_Ownersit + α2*%Insiderit + α3* %Own_Fund_of_Fundit + α4*%Own_Fgnit +
α5*Ln_Min_Invit + α6*Big4, Q1-3it + α7*Big4, Q4iq + α8*Non-Big4, Q4it + α9*Perc_Assets_Valuedit + α10*Adminit +
α11*Marketerit + α12*Ln_Ttl_Committed_Capitali + α13*Experienceit + α14*Fundraisingiq + α15*Quartileit + α16*Exemptit +
α17*SP500_Retq + α18*SP500_Retq-1 + αt*ΣFund_Ageit + eiq
(1) (2)
Dependent Variable > 10 Pct Adj Pos Adj
Variables Coef. (t-stat) Coef. (t-stat)

Investor Monitoring
Ln_Owners 0.0016 (0.046) 0.0534 (1.435)
%Insider 0.0002 (0.082) 0.0028 (1.232)
%Own_Fund_of_Funds -0.0005 (-0.531) 0.0001 (0.133)
%Own_Fgn 0.0005 (0.301) 0.0011 (0.720)
Ln_Min_inv 0.0111** (2.156) 0.0012 (0.204)

Auditor Involvement
Big4, Q1-3 -0.0520 (-1.110) 0.0295 (0.573)
Big 4, Q4 (audited) 0.1144** (2.018) 0.0287 (0.501)
Non-Big4, Q4 (audited) 0.1877*** (2.722) 0.0594 (1.001)

Third-Party Involvement
Perc_Assets_Valued 0.0005 (0.377) 0.0002 (0.264)
Admin -0.0602 (-0.989) -0.0490 (-0.750)
Marketer -0.1135 (-1.269) -0.1449** (-2.241)

Fund Attributes
Ln_Ttl_Committed_Capital -0.0169 (-0.478) 0.0449 (1.531)
Experience 0.0109 (0.417) -0.0077 (-0.364)
Fundraising -0.0971 (-0.859) 0.2260*** (2.649)
Quartile -0.0059 (-0.274) 0.0454** (2.144)
Exempt -0.0968 (-1.220) -0.0782 (-1.102)

Market Attributes
SP500_Ret 0.0792 (0.511) 0.5071*** (3.214)
SP500_Ret_Lag -0.0968 (-0.860) 0.2537** (2.126)

Constant 0.3911 (1.095) -0.3441 (-1.382)

Big 4: Audited - Unaudited 0.1664*** -0.0008

Observations 1,367 1,367
Adjusted R-squared 0.032 0.035
Fund Age FE Yes Yes
Clustered SE By Fund By Fund
This table presents OLS regression estimation results of NAV adjustments on external monitoring, third-party provider, and fund and
market attributes. > 10 Pct Adj (Column 1) is an indicator set equal to 1 in fund-quarters making NAV adjustments with an absolute
value greater than 10% of NAV, and 0 otherwise. Pos Adj (Column 2) is an indicator set equal 1 in fund-quarters with positive NAV
adjustments, and 0 otherwise. For ease of presentation, we group attributes into five categories, Investor Monitoring, Auditor
Involvement, Third-Party Involvement, Fund Attributes, and Market Attributes. Panel A (B) presents results for BO (VC) funds. We
cluster the standard errors by fund. Appendix A provides variable definitions. Robust t-statistics are in parentheses with ***, **, and *
indicating significance at the one percent, five percent, or ten percent level, respectively.

55