Public Information Precision and Coordination Failure: An Experiment

• Published date: 01 September 2016
• Author: SANJAY BANERJEE,MICHAEL MAIER
• DOI: http://doi.org/10.1111/1475-679X.12124
• Date: 01 September 2016

DOI: 10.1111/1475-679X.12124

Journal of Accounting Research

Vol. 54 No. 4 September 2016

Printed in U.S.A.

Public Information Precision and

Coordination Failure: An

Experiment

SANJAY BANERJEE

∗AND MICHAEL MAIER

∗

Received 30 April 2014; accepted 6 April 2016

ABSTRACT

More precise public disclosure reduces uncertainty about economic funda-

mentals, but it can increase uncertainty about other agents’ actions, leading

to coordination failure. We conducted a laboratory experiment to study the

effects of public information precision and strategic complementarity on co-

ordination failure. Information precision is operationalized in terms of “gran-

ularity” (level of detail). We found that (1) granular public disclosure, which

is disaggregated and precise, increases the likelihood of coordination failure

and decreases coordination efficiency when public information is pessimistic

about future economic prospects; (2) the deleterious effect of granular dis-

closure is stronger when strategic complementarity is high; and (3) higher

levels of strategic complementarity decrease coordination efficiency. Overall,

the observed likelihood of coordination failure is higher and coordination

∗University of Alberta.

Accepted by Haresh Sapra. We appreciate constructive comments from two anonymous

referees. Part of this research has been funded by an SSHRC (Social Sciences and Humani-

ties Research Council) grant. We thank Efstathios Avdis, Douglas DeJong, Jeffrey Hales, Steve

Huddart, Karim Jamal, Chandra Kanodia, Lisa Koonce, Mark Penno, Hong Qu, Florin Sabac,

Douglas Skinner (discussant), Jack Stecher (discussant), and conference participants at the

University of Alberta Accounting Research Conference, Banff, 2014; the Indian School of

Business (ISB), Hyderabad, 2014; the Indian Institute of Management (IIM), Calcutta, 2013;

the University of Calgary, 2015; and attendees at the European Accounting Association An-

nual Congress 2015, Glasgow; the Ninth Accounting Research Workshop, Zurich, 2015; and

American Accounting Association (AAA) Annual Meeting, Chicago, 2015 for their helpful

comments and insights. We also thank WeiTu for his excellent research assistance.

941

Copyright C, University of Chicago on behalf of the Accounting Research Center,2016

942 S.BANERJEE AND M.MAIER

efficiency is lower than predicted by theory. Our findings have implications

for the Federal Reserve’s decision to publicly disclose detailed stress test re-

sults for distressed banks, and the debate on whether the Public Company

Accounting Oversight Board should publicly release reports on firm-specific

quality-control deficiencies of audit firms.

JEL codes: C92; D82; D84; E58

Keywords: coordination failure; transparency; information granularity;

global game; heterogenous information; strategic uncertainty; experiment

1. Introduction

Public disclosure of financial information is a fundamental issue in account-

ing. To the popular press, regulators, and politicians, greater transparency

through increased public disclosure of information is almost always desir-

able. From the Sarbanes–Oxley Act of 2002 to the recent Dodd–Frank Wall

Street Reform and Consumer Protection Act of 2010, and current legisla-

tion (U.S. House of Representatives [2011], H.R. 3503) requiring the Pub-

lic Company Accounting Oversight Board (PCAOB) to publicly disclose

firm-specific quality-control deficiencies of audit firms, regulations have

mandated greater levels of transparency.

The argument in favor of greater transparency is that more disclosure im-

proves price efficiency and market discipline. While more information can

result in better decisions when they are made by a single agent (Blackwell

[1951]), greater transparency may not necessarily be desirable when deci-

sions are coordinated among multiple parties (Morris and Shin [2004]). In

economic settings with strategic complementarities (i.e., when each agent’s

incentive to act increases as other agents take the same action), more pre-

cise public information can lead to coordination failure1and economic in-

efficiencies. A more precise public signal can induce agents to “overreact”

to public information and “underreact” to private information. Such an

overreaction to public news can trigger a coordination failure and reduce

economic efficiency in the sense that agents place too much weight (com-

pared to Bayesian updating) on public information and too little weight on

private information.

In this paper, we study how transparency, in a setting of strategic com-

plementarities, affects coordination failure and economic efficiency. Un-

like prior experimental studies (e.g., Anctil et al. [2004, 2010]) in which

the transparency of private information was considered, we consider the

transparency of public information, defined as its level of precision. We

operationalize information precision by its “granularity” (level of detail):

disaggregated, firm-level information is more precise (granular) than

1Coordination failure occurs when agents fail to coordinate to attain the more efficient

Pareto-superior equilibrium among multiple equilibria (e.g., Cooper et al. [1990, 1992], Van

Huyck, Battalio, and Beil [1990, 1991]).

PUBLIC INFORMATION PRECISION AND COORDINATION FAILURE 943

aggregated information. Specifically, we ask: How do disclosure granular-

ity and strategic complementarity affect coordination failure?

We ran an experiment designed to test the theory of the role of public

information precision in a coordination game with strategic complementar-

ities. We consider an economic setup similar to the bank-run model2used

by Morris and Shin [2002]. The main innovation is that we include more

than one bank in our model. There are multiple creditors, and Mnum-

ber of banks. Each creditor is endowed with Munits of a loan, one loan to

each of Mbanks. Each bank then invests one loan unit in a risky project.

All creditors individually but simultaneously decide whether to roll over or

foreclose on their loans. Each creditor’s decision-making process is identi-

cal. If a creditor forecloses on a loan, she gets a fixed positive amount, f;

however, if she rolls over the loan, she receivesr>fif the risky project suc-

ceeds, and nothing if the project fails. Creditors do not know the economic

fundamentals3of projects with certainty. Each creditor receives a private

signal (i.e., observed only by that particular creditor), and a public signal

(i.e., observed by all creditors) about the fundamental value.

For any project, if the economic fundamentals are very strong, the

project will succeed even if all creditors foreclose on their loans, and rolling

over a loan is the unique equilibrium. If the economic fundamentals are

very weak, a project will fail even if no creditors foreclose, and foreclos-

ing is the only equilibrium. In the intermediate fundamentals, beliefs are

self-fulfilling. There are two pure-strategy Nash equilibria4: (1) every credi-

tor forecloses on her loan thinking that other creditors will also foreclose,

leading to a self-fulfilling run on a project, or (2) every creditor rolls over

her loan thinking that other creditors will do the same, leading to another

self-fulfilling, but more efficient equilibrium of project success.

We test the key theoretical result that granular, firm-level disclosure (vs.

aggregate, industry-level disclosure) increases the likelihood of a bank run

when the public signal is low (pessimistic), but decreases the likelihood

when the public signal is high (optimistic). A pessimistic public signal has

two effects: it reveals negative information about a bank’s economic funda-

mentals, and informs each of the bank’s creditors that every other creditor

is also aware of the pessimistic news. While more precise public informa-

tion reduces uncertainty about economic fundamentals (fundamental un-

certainty), it can increase uncertainty about other agents’ actions (strategic

2There are several other examples of coordination games with strategic complementarities,

including those that model speculative attacks and political regime changes. We chose a bank-

run model for its ease of exposition, and the relevance of bank runs given the recent financial

crisis. However, we acknowledge that we do not specifically model a bank, which has short-

term debts and long-term loans that are subject to regulatory constraints. Our setup is more

general—we study economic environments that are prone to strategic complementarities, and

banks are examples of such environments.

3One way to interpret the economic fundamental is that it represents the total net present

value of all cash flows generated by the project.

4There is also an unstable mixed-strategy Nash equilibrium.