Opening the Black Box of Upper Echelons: Drivers of Poor Information Processing During the Financial Crisis

• DOI: http://doi.org/10.1111/j.1467-8683.2010.00796.x
• Published date: 01 May 2010
• Author: Margit Osterloh,Katja Rost
• Date: 01 May 2010

Opening the Black Box of Upper Echelons:

Drivers of Poor Information Processing During

the Financial Crisiscorg_796212..233

Katja Rost* and Margit Osterloh

ABSTRACT

Manuscript Type: Empirical

Research Question/Issue: Why did the majority of directors prior to the f‌inancial crisis not have the foresight to predict the

problems of taking on too much risk? We analyzewhether executives’ characteristics affect strategic choices due to bounded

rationality, as proposed by the theory of upper echelons. The literature has thus far not empirically opened this black box.

Relying on psychological economics, we develop hypotheses under which conditions expertise and gender can lead to

biased information-processing.

Research Findings/Insights: To test the hypotheses, we propose a two-study methodology and take the f‌inancial crisis as

a natural experimentalsetting. In Study 1, we analyze individualphenomena and show that under conditions of uncertainty,

the processing of information by f‌inancial experts and men is worse than by non-f‌inancial experts and women. In Study 2,

we test these f‌indings for organizational phenomena. We show that banks with a higher percentage of f‌inancial experts

within TMTs perform better in stable environments, but are more negatively affected by the f‌inancial crisis.

Theoretical/Academic Implications: An important moderator within the theory of upper echelons is f‌inancial market

discipline during turbulent periods, explaining why the performance of homogenous TMTs is volatile and why the

performance of diverse TMTs is sustainable. The theory can be strengthened by including the insights of psychological

economics as a micro-foundation.

Practitioner/Policy Implications: For a sustainable performance, greater TMT diversity in public companies should be

instituted by the board of directors.

Keywords: Corporate Governance, Theory of Upper Echelons, Bounded Rationality, Psychological Economics, Infor-

mation Processing, Uncertainty

INTRODUCTION

The f‌inancial market crisis that spread around the globe

raises the question as to why a many directors within

the banking and insurance industry did not have the fore-

sight to predict the problems of taking on too much risk. In

the banking industry, even when problems became appar-

ent, few bank managersor directors of spoke up (Steverman

& Bogoslaw, 2008-10-11).

This paper argues that an important reason for these fail-

ures may be a lack of heterogeneity and a lack of differenti-

ated viewpoints in top management teams (TMTs). The task

of the executive TMT is to suggest an overall strategy with

respect to risky investments, while the non-executive TMT

has either to agree or disagree with this suggested strategy.

The TMT of a company thus has a major inf‌luence on risky

investments. The theory of upper echelons (Hambrick &

Mason, 1984) proposes that, within turbulent environments

and under great job pressure in particular, differences in

information, knowledge and perspective may benef‌it

decision-making and thus enhance f‌irm performance (Ham-

brick & Finkelstein, 1987; Hambrick, Finkelstein, & Mooney,

2005). In contrast, homogeneous TMTs might be more

advantageous within stable environments. Homogeneous

groups make faster decisions and are more capableof debat-

ing complex issues (Knight et al., 1999; Pelled, Eisenhardt, &

*Address for correspondence: Institute of Organization and Administrative Science,

University of Zurich, Universitätsstrasse 84, 8006 Zürich. Tel:+41 44 634 29 17; Fax: +41

44 634 49 42; E-mail: katja.rost@iou.uzh.ch

212

Corporate Governance: An International Review, 2010, 18(3): 212–233

© 2010 Blackwell Publishing Ltd

doi:10.1111/j.1467-8683.2010.00796.x

Xin, 1999; van der Walt & Ingley, 2003). The theory of upper

echelons explains the inf‌luence of executives’ characteristics

by referring to bounded rationality. Experience, values and

personalities are ref‌lected in the characteristics of executives

and affect their f‌ield of vision, selective perception and inter-

pretation. “In very few studies, however, have researchers

attempted to conf‌irm whether executive characteristics

affect information processing in this way. As a result, the

psychological and social processes, by which executive

prof‌iles are converted into strategic choices, still remain

largely a mystery – the proverbial black box” (Hambrick,

2007: 337).

Our article addresses this gap by developing a research

design that opens the black box of information-processing

(for an excellent overview on the gaps see also Carpenter,

Geletkanycz, & Sanders, 2004). We rely on psychological

economics, which is sometimes referred to as “behavioral

economics.” As pointed out by Simon (1985), the term

“behavioral” is however misleading, since it may be con-

founded with the “behaviorist” approach in psychology. We

therefore prefer the expression “psychological economics.”

We combine psychological economics with the theory of

upper echelons to develop hypotheses for information-

processing, depending on gender and expertise. Both char-

acteristics may be important in explaining risk estimates of

TMTs in stable situations and in turbulent environments like

the aforementioned f‌inancial crisis. In most countries, TMTs

are strikingly homogeneous (Brammer, Millington, &

Pavelin, 2007; Campbell & Minguez-Vera, 2008; Francoeur,

Labelle, & Sinclair-desgagne, 2008; Kang, Cheng, & Gray,

2007; Rose, 2007). Within the f‌inancial sector in particular,

most TMTs mainly consist of males with a f‌inancial back-

ground, i.e., of males with a degree in management, econom-

ics, or f‌inance (Daily,Certo, & Dalton, 1999; Rost & Osterloh,

2008).

Our research contributes to the theory of upper

echelons in two ways. First, we connect research on micro-

processes with research on macro-organizational phenom-

ena. To open the black box of information-processing in the

upper echelon theory, we use a student sample and

compare the results with observable measurements of

TMTs. Within both samples, the same external stimuli, i.e.,

the f‌inancial crisis, affected personalized constructions of

the situation. Second, we treat the f‌inancial crisis as a

natural experiment. Natural experiments help in sorting

out reverse causality, i.e., when it is not clear whether an

event is a cause or an effect. In natural experiments, the

chronology between cause and effect is clear because the

stimulus operates strictly exogenously and cannot be

manipulated.

Section 1 introduces the theory of upper echelons and

psychological economics, and the resulting research design.

Section 2 develops hypotheses on which conditions of

expertise and gender systematically inf‌luence information-

processing of individuals. Section 3 discusses the conse-

quences of biased information-processing for decision

making groups. Section 4 presents empirical f‌indings on

information-processing of individuals. Section 5 shows

empirical consequences for groups. Section 6 discusses the

f‌indings and points out the limitations, implications and

further research.

THEORETICAL BACKGROUND

The next section introduces the theory of upper echelons

and psychological economics. Wepoint out the limitations of

the theory of upper echelons, and suggest that some can be

overcome by including insights from psychologicaleconom-

ics. Building on both theories, we develop a research design

to study information-processing of individuals and TMTs.

Theory of Upper Echelons

Building on the premise of bounded rationality (March &

Simon, 1958), the theory of upper echelons emphasizes the

importance of understanding the mentalmodels of top man-

agers in explaining the choices they make. Leaders are typi-

cally confronted with a vast amount of information that

demands attention (Mintzberg, 1973). Theydecide on appro-

priate responses to important stimuli and discard informa-

tion that is less important (Weick, 1979), according to their

interpretation of the situation, applying their beliefs, knowl-

edge, assumptions and values (Finkelstein & Hambrick,

1990; March & Simon, 1958). Different mental models can

also operate on the group level, e.g., as shared cognition

(Klimoski & Mohammed, 1994) or dominant logic (Prahalad

& Bettis, 1986).

Three facets of the theory of upper echelons have signif‌i-

cantly stimulated major streams of research (Hambrick,

2007). First, most studies focus on TMTs (for an overviewsee

Certo, Lester, Dalton, & Dalton,2006), since management is a

shared effort in which a dominant coalition collectively

shapes organizational outcomes (Cyert & March, 1963).

Second, demographic characteristics of executives are used

as valid (but incomplete) indicators of executives’ cognitive

frames. The linkage between TMT characteristics and f‌irm

performance is used as a valid (but incomplete) indicator of

real psychological and social processes that mediate

between executives’ demography and their behavior. This

approach is useful for executive research because “black

box” problems, i.e., unknown, underlyingpsychological and

social processes of individuals, may be bypassed. Third, the

theory’s predictive strength is dependent on certain mod-

erators. For example, managerial characteristics have an

impact on strategy and performance under the following

conditions: (a) there exists a great deal of means-ends ambi-

guity (Eisenhardt, 1989), i.e., managerial discretion is high

(Hambrick & Finkelstein, 1987); (b) executives operate

under great pressure, i.e., executive job demands are high

(Hambrick et al., 2005); (c) executives have considerable

inf‌luence on f‌irm outputs (Miller & Droge, 1986); (d) behav-

ioral integration characterized by intense interaction that

produces open information exchange and collaboratively

based solutions and decisions is high, i.e., a TMT weighs

multiple approaches against each other (Li & Hambrick,

2005; Simons, Pelled, & Smith, 1999).

Psychological Economics

Psychological economics broadens the standard economic

model in three ways: in terms of bounded rationality,

bounded self-interest and bounded utility concept (Frey &

Benz, 2004). It has been shown, for example, that markets

OPENING THE BLACK BOX OF UPPER ECHELONS 213

Volume 18 Number 3 May 2010© 2010 Blackwell Publishing Ltd