Do banks care about analysts' forecasts when designing loan contracts?

• Author: Derrald Stice,Joshua Coyne
• Date: 01 May 2018
• DOI: http://doi.org/10.1111/jbfa.12304
• Published date: 01 May 2018

DOI: 10.1111/jbfa.12304

Do banks care about analysts'forecasts when

designing loan contracts?

Joshua Coyne1Derrald Stice2

1Schoolof Accountancy, University of Memphis,

Memphis,TN 38152, USA

2Schoolof Business & Management, Hong Kong

Universityof Science & Technology, Kowloon,

HongKong

Correspondence

JoshuaCoyne, School of Accountancy, Univer-

sityof Memphis, 221 Fogelman College Admin

Building,Memphis, TN 38152, USA.

Email:jgcoyne@memphis.edu

JELClassification: M41, G21

Abstract

We investigate whether banks rely on the information content in

equity analysts’ annual earnings forecasts when assessing the risk of

potential borrowers. While a long literature finds that analysts pro-

vide useful information to market participants, it is not clear that

banks, which have access to privileged information, would bene-

fit from publicly available analysts’ forecasts. If, however, banks do

rely on this information, then more precise private information in

earnings forecasts may inform banks. We focus our analysis on the

requirement of collateral because it is a direct measure of default

risk, whereas other loan terms such as interest spread and debt

covenants can also protect against other risks, such as asset misap-

propriation. The direct link between collateraland default risk allows

usto examine whether information from analysts is relevant to banks

when designing loan contracts. Consistent with our predictions, we

find that higher precision of the private information in analysts’ earn-

ingsforecasts is associated with a lower likelihood of requiring collat-

eral, and this effect is larger when a borrower does not have a prior

relationship with the lender or their accounting or credit quality is

low. We also find that this association disappears after the imple-

mentation of Regulation FD,consistent with this regulation reducing

analysts’ access to private information.

KEYWORDS

bank loans, earnings forecasts, equity analysts, information

precision

1INTRODUCTION

Market participants respond to the information contained in analyst reports, and especially to earnings forecasts (see,

e.g., Asquith, Beatty, & Weber,2005; Brown, 1993; Huang, Zang, & Zheng, 2014; Schipper, 1991). Analysts provide

information unavailable elsewhere and can also reduce information asymmetry by confirming expectations, but it is

unclear whether analysts have information that would be relevant to banks for at least two reasons. First, banks have

an information advantage over analysts because of their access to private information that is only available to firm

insiders (Fama, 1985). Analysts are considered firm outsiders and are not allowed to receive private information from

J Bus Fin Acc. 2018;45:625–650. wileyonlinelibrary.com/journal/jbfa c

2017 John Wiley & Sons Ltd 625

626 COYNEAND STICE

management, especially following Regulation Fair Disclosure (hereafter, Reg FD).1Second,banks monitor loans after

issuance and can place considerable pressure on borrowers to provide additional information on an ongoing basis,

whereas analysts do not have similar bargaining power (Bhattacharya & Chiesa, 1995; Carrizosa & Ryan, 2016). Taken

together,as firm insiders, banks have an information advantage over analysts, and the recognized relevance of analyst

information for other parties may not apply to banks.

On the other hand, analysts’ earnings forecasts generate private information from public sources by means of their

superior information-processing ability,and therefore, analysts can contribute unique information that lenders cannot

acquire directly from borrowing firms by providing information discovery and interpretation roles (Chen, Cheng, &

Lo, 2010; Francis, Schipper,& Vincent, 2002). Also, prior research indicates that managers learn from equity prices

(Chen,Goldstein, & Jiang, 2007), which provides evidence that outsiders, such as analysts, have the potential to provide

informationto insiders. Furthermore, it is not necessary for analysts to have superior information in order to inform the

lending process. Even if analysts only provide information already available to lenders, analysts’ forecasts mayreduce

information asymmetry by providing confirmation.

We investigate whether equity analysts'earnings forecasts are in the information set of banks when setting collat-

eral requirements in response to borrowers'default risk. If the information content in analysts'forecasts is associated

with loan terms after controlling for other sources of information, then we can infer that analysts'forecasts are infor-

mative for private lenders. The primary benefit of selecting the collateral requirement as our dependent variable is

that this requirement is a direct measure of default risk. Whereas other loan terms also protect against other risks,

such as asset misappropriation, the collateral requirement protects specifically against default risk by providing the

lender with a senior claim to a borrower's assets. This link allows us to form a clearer picture about the association

between forecast precision and perceived default risk.2

We investigatethe relevance of information from analysts for lenders using the precision of the private information

in equity analysts’ annual earnings forecasts as our measure of analyst information. We use forecast precision as our

measureof private information for two reasons. First, more precise accounting information about the borrower's credit

quality reduces uncertainty about default risk (Armstrong, Guay, & Weber, 2010; Bharath, Sunder, & Sunder, 2008;

Watts, 2003). Bharath et al. (2008) highlight the ability of precise information to reduce credit risk by showing that

higher quality accruals are associated with lower cost of debt. Equity analysts'earnings forecasts also provide a source

of accounting information, and the more precise the forecast, the higher its information content and resulting ability

to reduce uncertainty regarding default risk (Fried & Givoly,1982).

Second, using forecast precision allows us to focus exclusively on the private information in analysts'forecasts.

Although several studies use forecast dispersion as a measure of uncertainty and information quality, Abarbanell,

Lanen, and Verrecchia (1995) demonstrate using theoretical models that dispersion alone is insufficient to measure

uncertainty. Barron, Kim, Lim, and Stevens(1998) build on these initial findings and develop a model of the precision

of private information in analysts'forecasts. Their model is a function of three forecast characteristics: forecast dis-

persion, forecast error, and analyst following. By using private precision as our independent variable of interest, we

can measure the reduction in the cost of debt associated with an increase in analysts'ability to reduce uncertainty

surrounding borrower default risk by incorporating higher quality (i.e., more precise)private information in their earn-

ings forecasts.3

1Researchexamining the effects of Reg FD has generally concluded that it has provided a “more level playing field” for investors (Koch, Lefanowicz, & Robin-

son,2013), so it is especially unclear whether analysts will inform lenders after Reg FD. Recent research, however, has provided evidence that Reg FD has not

completely eliminated the transfer of private information from managers to analysts. Notably,Soltes (2014) states that analysts continue to have “off-line”

access to management, and Green, Jame, Markov,and Subasi (2014) provide evidence that broker-hosted investor conferences (attended by analysts and

mangers)lead to higher quality analyst forecasts and larger stock-price reactions to stock recommendation changes, evenafter Reg FD.

2Because of the simultaneous choice of severalloan terms in a single contract, we control for three other loan terms (i.e., loan spread, maturity, and number

ofcovenants) in all of our regression models.

3Thelevel of the earnings forecast (i.e., first moment) is alternative proxy for default risk because firms with higher earnings are less likely to default. We use

forecast precision (i.e., second moment) for the reasons we have listed aboveand because of two issues with using earnings forecast level. First, we cannot

measure information content or information quality with earnings level because a higher forecast is not necessarily more or less precise. In the extreme,a

very high forecast with zero precision cannot reduce anyuncertainty regarding default risk, whereas a perfectly precise forecast that projects low earnings