DOES MANAGERIAL OPPORTUNISM EXPLAIN THE DIFFERENTIAL PRICING OF LEVEL 3 FAIR VALUE ESTIMATES?

• Author: Dahlia Robinson,Adrian Valencia,Thomas Smith
• Date: 01 June 2018
• Published date: 01 June 2018
• DOI: http://doi.org/10.1111/jfir.12146

DOES MANAGERIAL OPPORTUNISM EXPLAIN THE DIFFERENTIAL

PRICING OF LEVEL 3 FAIR VALUE ESTIMATES?

Dahlia Robinson and Thomas Smith

University of South Florida

Adrian Valencia

Florida Gulf Coast University

Abstract

Using hand-collected Level 3 data, we find that banks near key capital ratios report

higher unrealized gains in Level 3 assets, consistent with managers using Level 3

valuations to boost capital ratios. Additionally, we document an incremental pricing

discount for Level 3 assets among these firms, suggesting that the pricing discount

observed by prior research may not be entirely due to measurement error inherent in

Level 3. Furthermore, we observe that this opportunistic behavior diminishes under the

new Financial Accounting Standards Board disclosure regime, suggesting that the

expanded disclosures may increase the value relevance and reliability of fair value

measurements.

JEL Classification: G14, G15, G30, K22, M41, M42

I. Introduction

Accounting standard setters, academicians, and professionals have been interested in

improving the understanding of the financial reporting characteristics of fair value

measurements. Statement of Financial Accounting Standard (SFAS) 157 “Fair Value

Measurements”provides guidance for the measurement of assets and liabilities at fair

value through a hierarchical framework and proposes that “the fair value hierarchy gives

the highest priority to quoted prices (unadjusted) in active market for identical assets or

liabilities (Level 1) and the lowest priority to unobservable inputs (Level 3)”(Financial

Accounting Standards Board [FASB] 2006). The FASB’s rationale for this ordering is

that quoted prices in an active market provide the most reliable evidence of fair value and

as such should be used whenever available as valuation inputs to measure fair value.

The 2008 financial crisis coincided with the effective date for SFAS 157, which

was required for financial statements issued for fiscal years beginning after November 15,

2007. Therefore, Level 3 valuations of assets and liabilities were particularly important

during this period when the markets for certain financial instruments were inactive (Laux

The authors would like to thank Stephen Penman, Douglas Hanna, Martin Loeb, Larry Gordon, and Geert

Bekaert, as well as workshop participants at Florida State University and the 2012 American Accounting

Association annual meeting. They would also like to thank Tom Linsmieir and John Archambault for sharing their

institutional banking knowledge.

The Journal of Financial Research Vol. XLI, No. 2 Pages 253–289 Summer 2018

253

© 2018 The Southern Finance Association and the Southwestern Finance Association

and Leuz 2009). Furthermore, financial institutions appeared to rely on management-

derived valuation models (allowed under SFAS 157) rather than on dealer quotes that

were potentially distorted by illiquidity. Consistent with this scenario, Laux and Leuz

(2010) document that several banks reported significant decreases in the Level 1 asset

classification and compensating increases in Level 2 and Level 3 assets, and substantial

net transfers into the Level 3 category early in the crisis (FASB 2006).

1

Although there is no clear consensus on whether the discretion allowed under

SFAS 157 resulted in more reliable financial statements, proponents argue that during

periods when orderly markets do not exist, managers have a distinct advantage over

inactive markets as their valuation models incorporate private information, which

potentially increases the accuracy and reliability of estimates of assets and liabilities.

Some research suggests that managers do indeed provide superior estimates of fair values

because of this private information (Barth, Landsman, and Rendleman 1998), and that

investors consider fair value estimates reliable enough to be priced (Barth 1994; Song,

Thomas, and Yi 2010; Kolev 2008; Goh et al. 2015). In contrast, researchers, investors,

and auditors have expressed concern that managerial discretion in determining Level 3

valuations may lead to distorted and unreliable values (Laux and Leuz 2009; Penman

2007; Earley Hoffman, and Joe 2014; Kohlbeck, Smith, and Valencia 2017).

2

However,

existing research does not directly address this issue.

Song, Thomas, and Yi (2010) and Goh et al. (2015) find that investors price

Level 1 and Level 2 values higher than Level 3 values, consistent with a perceived lower

reliability in the Level 3 estimates. However, these studies provide mixed evidence on

whether the lower valuations for Level 3 assets may also be due to managerial incentives

to manipulate these estimates. For example, Song, Thomas, and Yi do not directly

examine whether management uses the discretion in Level 3 opportunistically, but

instead show that firms with stronger corporate governance experience a lower valuation

discount for their Level 3 assets. Goh et al. provide a more direct test of management

opportunism by examining whether Level 3 assets are priced differently for banks with

Level 3 gains, but the authors are unable to find any differential pricing among these

firms. Given the mixed evidence in the literature, our study examines whether banks

appear to use Level 3 reporting opportunistically, and whether the capital market

valuation discount is partially explained by this opportunistic bias.

3

In this study, we first investigate whether bank managers use their Level 3

discretion to achieve capital adequacy targets. We focus on financial institutions for

1

For example, by the first quarter of 2008, Bear Stearns, Lehman, and Merrill Lynch reported cumulative

Level 3 transfers of more than 70% of the precrisis balance, and Citigroup transfered $53 billion into Level 3 from

the third quarter of 2007 to the first quarter of 2008 alone (Laux and Leux 2010).

2

Under SFAS 157, managers have discretion over (1) the choice of the valuation model, (2) the inputs to that

model, and (3) whether financial instruments are classified as Level 3 (vs. Level 1 or 2). The CFA Institute, in a

position paper supporting the reporting of financial instruments at fair value, notes existing concern about the

potential for managerial opportunistic manipulation of Level 3 valuations through the marking-to-model process.

3

Fiechter and Meyer (2009) also examine Level 3 opportunism. Their evidence is consistent with “big-bath”

behavior with respect to unrealized gains/losses of Level 3 instruments. Our study differs from theirs in that we

focus on beating benchmarks instead of big-bath behavior. Additionally, we consider capital adequacy targets.

Finally, we examine a larger sample covering a longer time span, which allowsus to also examine the impact of the

increased required disclosures.

254 The Journal of Financial Research

several reasons. First, banks are required to report a large portion of their assets and

liabilities at fair value. For example, Khurana and Kim (2003) report that the assets and

liabilities that are subject to fair value disclosures at bank holding companies are on

average 87% and 88% of total book value of assets, respectively.

4

Second, Level 3

instruments represent an important subset of these fair value disclosures, potentially

making Level 3 instruments more important to banks relative to other companies. Third,

the financial industry was among the first to adopt the new disclosure requirements under

SFAS 157, implying a larger number of available firm-quarter observations. Finally,

focusing on banks allows us to examine strong capital adequacy incentives during a

particularly stressful time for the industry, marked by a significant increase in bank

failures largely driven by severe contractions in capital.

5

We use Berger et al. (2008) to calculate a bank-specific Tier 1 risk-based capital

ratio target, and examine whether firms near that ratio report higher (lower) levels of

Level 3 unrealized gains (losses).

6

We use the Berger et al. target capital ratio rather than

a regulatory minimum ratio because in their review of bank literature, Beatty and Liao

(2014) state that banks tend to hold more capital than required. For example, our sample

banks have an average Tier 1 risk-based capital ratio of 10.8%, which is well above the

minimum level of capital adequacy. Furthermore, consistent with Beatty and Liao (2011)

and Van den Heuvel (2009), we suggest that banks had even greater incentives to

maintain higher internal capital ratios during the financial crisis period because the

ability to raise new capital at reasonable prices would have been severely curtailed.

Perhaps more important, banks with high capital ratios had greater opportunities to

participate in Federal Deposit Insurance Corporation (FDIC)–assisted acquisitions for

failed banks, which often resulted in substantial gains for acquirers (Dunn, Kohlbeck,

and Smith 2016). We focus on unrealized gains or losses on Level 3 positions because the

potential variability of Level 3 fair value estimates is inherently large, and as the market

for certain financial instruments became inactive during 2008, both the importance and

measurement uncertainty of Level 3 fair values increased (Ryan 2008; FASB 2009).

Using a sample of 266 U.S. bank holding companies (which we refer to

throughout as banks), we find evidence consistent with managers increasing their

estimates of Level 3 valuations to report higher (lower) unrealized gains (losses) to boost

capital adequacy ratios when they report ratios near the Berger et al. (2008) target. After

observing higher levels of unrealized gains in Level 3 among firms near their key capital

benchmark, we then examine whether investors appear to differentially price these

4

Beatty and Liao (2014) note that “only 20% of bank assets are recognized at fair value in the bank balance

sheet . . . the use of fair values in the income statement is even lower since changes in fair values are often reported

in other comprehensive income rather than being reported in the traditional income statement”(p. 351). However,

during the financial crisis, a significant percentage of assets were subject to other than temporary impairments often

classified under Level 3 (Gonzalo 2008; Badertscher, Burks, and Easton 2012). In these cases, the income statement

(and regulatory capital) is significantly affected.

5

Since the beginning of the recession in January 2008 until the end of our sample period (December 2010), the

FDIC reported 322 bank closures compared to only 23 bank closures from 2000 to 2007.

6

We rely on Berger et al. (2008) to estimate target Tier 1 risk-based capital ratios because firm-specific target

capital ratios are not available to the public. Because of the inherent error in our target capital measure, we include

firms that fell just below or just above our predicted target ratio (as opposed to those meeting or just exceeding it) to

identify those with a high incentive to opportunistically boost Level 3 valuations.

Managerial Opportunism 255