Journal of Management & Governance

, Volume 19, Issue 3, pp 655–686

Allowance for failure: reducing dysfunctional behavior by innovating accountability practices

Authors

Article

DOI: 10.1007/s10997-013-9276-3

Cite this article as:
Mahlendorf, M.D. J Manag Gov (2015) 19: 655. doi:10.1007/s10997-013-9276-3

Abstract

A common theme in current corporate governance is to increase managers’ accountability. In contrast, this paper emphasizes the need for less accountability in certain situations and consequently introduces the concept of “allowance for failure”. This concept refers to the style in which the decision maker’s environment—such as capital market, corporate governance, and management control system—deals with potential failures (e.g., project or investment failures). The argument that allowance for failure is important is illustrated by the example of failing projects, drawing on the escalation of commitment literature. It is hypothesized that allowance for failure indirectly reduces project escalation, i.e., the continuation of a failing project. The relationship is mediated by managers’ perceived threat in case of project failure. In addition, the paper suggests that capital market orientation increases managers’ perceived threat in the case of project failure and thus indirectly increases project escalation. Cross-sectional survey data were collected to test these hypotheses. The results from 320 failed projects under the responsibility of top-level managers support the hypothesized effects. Cross-validation with 109 projects terminated by lower-level managers and 133 projects terminated by company owners shows consistent results. The study’s findings highlight the importance of carefully analyzing potential consequences of promoting capital market orientation. More important, the study indicates a need for innovative accountability practices that expand allowance for failure to avoid dysfunctional consequences for decision-making, especially if strong capital market orientation is prevalent.

Keywords

Allowance for failureCapital marketDecision-makingProject escalation

JEL Classification

G30M10

1 Introduction

A common theme in current corporate governance is to increase manager’s accountability. It seems en vogue to hold managers personally responsible for failures, to increase pressure, and to tighten controls. This trend occurs in various streams, such as new shareholder activism (Brav et al. 2008; Del Guercio et al. 2008; Rose 2010; Sudarsanam and Broadhurst 2010), responses to corporate scandals such as Enron, WorldCom, Tyco, and others (Ingley et al. 2010), as well as reactions to the global financial and economic crisis 2007–2009 (Hilb 2010).

It is suggested, for example, that “shareholders may launch a media campaign e.g., showing pictures of the board in newspapers under the header ‘Badly-performing assets’ arguing why management should be replaced” Rose (2010, p. 2). In a similar vein, Del Guercio et al. (2008, p. 85) explain that “a campaign that directly targets board members could be a more effective threat than an issue-oriented shareholder proposal”.

While calls for more accountability might be appropriate reactions to many problems faced by the economy during recent years, it is important not to forget that the resulting pressure may have potentially dysfunctional consequences on decision-making (DeZoort and Lord 1997). Therefore, in contrast to the apparent trend, this paper emphasizes the need for less accountability in certain situations and therefore introduces the concept of “allowance for failure”. This refers to the style in which the decision maker’s environment—such as capital market, corporate governance, and management control system (MCS)—deals with potential failures (e.g., project or investment failures). Allowance for failure is high if the environment considers that failure is a possible and acceptable outcome. The concept of allowance for failure can be implemented by innovating a variety of management and accounting practices, as will be explicated in more detail in this paper.

The argument that allowance for failure is important is illustrated by the example of failing projects, drawing on the escalation of commitment literature. Escalation refers to the human trait to pursue the original course of action or even increase the commitment towards an ongoing project, although new information suggests to cancel the project (Barton et al. 1989; Schmidt and Calantone 2002, p. 104; Staw 1976). This issue is important because empirical evidence demonstrates that project escalation leads to the inefficient allocation of funds, higher costs, and severe organizational disruptions (Meredith 1988, p. 31). This paper suggests that project escalation is increased by pressures, e.g., due to capital market orientation, while allowance for failure has the opposite effect.

Consequently, the paper makes several contributions. First, it draws attention to a concept that may deserve more attention in today’s business environment, namely allowance of failure. This is particularly interesting, because the current trend in politics and the media seems to favor tighter controls and stronger personal responsibility, whereas this paper emphasizes that it is important to use innovative management and accounting practices in such a way that perceived threat of failure is reduced for the decision maker. Second, the topic is analyzed and empirically tested in the setting of project escalation, which is an important setting that has found widespread attention, especially in psychology and project management research. Specifically, this paper is the first which presents empirical evidence that capital market orientation can have dysfunctional effects on decisions concerning the termination of failing projects. Moreover, it is one of the first studies to show that allowance for failure significantly influences decision-making in real-world projects. In doing so, the paper corresponds to the call brought forth by Schulz and Cheng (2002, p. 82) regarding the importance to provide solutions that mitigate the escalating effects of personal responsibility. Moreover, the study adds to prior literature on innovative management and accounting practices aimed at and succeeding in reducing the escalation of commitment (Cheng et al. 2003; Ghosh 1997; Kadous and Sedor 2004).

Finally, this paper provides external validity for prior experimental studies conducted in the laboratory (Simonson and Staw 1992) and opens new terrain for examining how the decision maker’s environment can impact perceived threat and its subsequent impact on decision-making.

The remainder of this paper is organized as follows. Section 2 provides an overview of the relevant literature and derives the hypotheses. Section 3 describes the research methods while Sect. 4 presents the empirical results of the main model, the cross-validations, and exploratory extensions. Section 5 discusses the findings. Finally, in Sect. 6, conclusions are drawn, limitations are discussed, and implications for future research are presented.

2 Literature review and hypotheses development

Previous literature suggests that errors should be “dealt with in a supportive, learning manner rather than in a threatening, punitive, blame-oriented manner” (Senatra 1980, p. 597). In the same line of thought, other authors mention the need for tolerance for failure (Danneels 2008; Sullivan and Snodgrass 1991, p. 26). They emphasize that it has to be “understood that failure is a necessary part of success” (Danneels 2008, p. 542). The current study attempts to make these important thoughts more concrete and thus describes specific management and accounting practices that can be subsumed under the heading “allowance for failure”. Those practices are drawn from the literature on escalation of commitment. Simonson and Staw (1992) propose three practices that have the potential to reduce perceived threat in case of failure: being prepared to change policy if minimum target levels are not achieved, evaluating decision makers based on their decision process rather than outcome, and making negative outcomes less threatening (for a more detailed description see Chapter 2.2). The current paper proposes to subsume these practices under the concept of “allowance for failure” (for more suggestions on respective management and accounting practices see Chapter 5.2).

This paper connects the allowance for failure literature with corporate governance and capital market influences. More specifically, this study investigates failing projects as a setting, where capital market orientation may have dysfunctional effects on decision-making and suggests allowance for failure as a concept to counterbalance those effects.

Capital market orientation refers to the importance the company assigns to its relation to the capital market. The capital market represents all places where companies can raise long-term funds (i.e., capital for more than one year) (O’Sullivan and Sheffrin 2003, p. 283). In the broad view, this includes term loans and financial leases, corporate equities, and bonds. Thus, capital market actors are shareholders, analysts, and creditors. It is argued, that accounting and management control often serves the demands of these actors and provides the means to increase capital market orientation. “Surely, accounting has had a strong role in bringing about this dominance” of the finance view (Vollmer 2003, p. 366). This view is “not shaped by the necessities of production or the desire to sell more products. Instead, … [it is] focused on the corporation as a collection of assets that could and should be manipulated to increase short-run profits” (Fligstein 1990, p. 226).

It has been stated that the capital market orientation replaces the “velvet glove of the corporatist state” with the “whip of the market” (Ezzamel et al. 2008, p. 110). However, this “whip” does not necessarily seem to be positively related to companies’ economic performance (Carr and Tomkins 1998, p. 214). Graham et al. (2005, p. 35), for example, provide evidence that due to capital market pressures to meet earnings targets managers may delay the start of a new project, even if this decreases long term value.

Following this line of thought, the current study investigates a setting in which a capital market orientation may have dysfunctional effects on managers’ decision-making, namely project escalation. In accordance with prior literature (Ruchala 1999, p. 163), project escalation is defined as the decision to continue investment in a failing project. In other words, managers may increase their commitment towards the original course of action, although recent information points in another direction (Barton et al. 1989; Schmidt and Calantone 2002; Staw 1976). Research suggests that escalation of commitment is a phenomenon which leads to systematic delay of exit decisions. Hence, far more projects are terminated too late, rather than too early (Meredith 1988). Consequently, project escalation leads to higher costs, inefficient resource-allocation, and stronger organizational disruption (Meredith 1988). This is of crucial importance especially in areas such as R&D, where the majority of high risk projects are expected to fail (Mansfield et al. 1971). Moreover, escalation of commitment has also been reported for strategic decisions (Kisfalvi 2000). Examples within the accounting-related literature include Cheng et al. (2003), Chow et al. (1997), Ghosh (1997), Harrell and Harrison (1994), Harrison and Harrell (1995), Jeffrey (1992), Kadous and Sedor (2004), Kanodia et al. (1989), Schulz and Cheng (2002), Tan and Yates (2002).

2.1 Capital market orientation and perceived threat in case of project failure

Managers in capital market-oriented companies face a variety of pressures. Shareholders and banks monitor management decisions (Goergen et al. 2008, p. 175; Ramseyer 1994, p. 231) while financial analysts “lambast” those managers who fail to deliver (Ezzamel et al. 2008, p. 134). Brav et al. (2008), for example, report that CEO turnover increases by almost 10 percentage points in the year after hedge fund activism is announced, suggesting that hedge funds push the replacement of unsuccessful managers. Such pressures can lead to different consequences. On the one hand, they may exert a positive influence on performance. For example, shareholder value orientation is negatively related to cost stickiness (Calleja et al. 2006). Accordingly, it seems plausible that pressures lead to quicker responses in the case of unfavorable economic developments. On the other hand, capital market orientation can evoke dysfunctional effects. For example, capital market pressure can increase myopic behavior (Bhojraj and Libby 2005, p. 17). In addition, it has been demonstrated that managers select accruals at the time of the initial public offering (IPO) to report high earnings opportunistically (Teoh et al. 1998, p. 202).

These studies indicate that capital market orientation exerts diverse pressures on managers. These pressures can be expected to increase managers’ perceived threat in case of project failure. Although investors may appreciate the termination of failing projects in some cases where failure was already obvious, market reactions tend to be negative if project abandonment occurs unexpectedly. Moreover, financial investors view patterns in reported data as representative of future patterns and thus overreact (Lev et al. 2005). Accordingly, managers might often perceive the cancellation of a failing project as threatening. In addition, even without taking overreactions into consideration, the announcement of the failure of a large project can be extremely displeasing for the manager: “By acting as ‘traders’ rather than ‘owners’, institutional investors allegedly place excessive focus on short-term developments, leading managers to fear that an earnings disappointment will trigger large-scale institutional investor selling and result in a temporary undervaluation of the firm’s stock price” (Bushee 1998, p. 306).

Perceived threat will increase even more if the company is at risk of failing to meet analyst expectations. In such cases, it has been observed that “CFOs take potentially value destroying actions to meet such expectations” (Koh et al. 2008, p. 1967). Another factor is venture capitalists, who usually have high expectations of future growth and thus exert pressures (Granlund and Taipaleenmäki 2005) that might make project failures more threatening to responsible decision makers. Likewise, abandonment of large projects can make it more expensive to raise loans from banks, because the failure might be interpreted as signal for a lower probability of repayment.

Based on these empirical findings, the following hypothesis is suggested.

Hypothesis 1

Capital market orientation has a significant positive relation with perceived threat in the case of project failure.

2.2 Allowance for failure and perceived threat in case of project failure

Given the substantial amount of resources invested in failing projects, the decision makers’ environment should apply adequate management and accounting practices to reduce project escalation. For example, Ghosh (1997) demonstrated that unambiguous feedback, progress reports, and information about future benefits of additional expenditures mitigate escalation behavior. Kadous and Sedor (2004) have shown that consultants acquire and retain more information about threats to project viability if they are explicitly asked to provide recommendations on the (dis-)continuation of a project.

Several studies have highlighted the relations between management control systems and job-related tension or pressure (Brownell and Hirst 1986; Hirst 1983; Hopwood 1972; Merchant 1990; Shields et al. 2000). For example, financial controls are associated with increased perceived pressure to meet financial targets (Merchant 1990). If some aspects of management control systems can increase pressure, it seems likely that other aspects can reduce perceived threat (i.e., those practices that this paper subsumes under the concept of allowance for failure). In a similar vein, Laverty (2004) emphasized the importance of routines that enable managers to pursue long-term goals without having to worry too much about short-term effects. Indeed, Schulz and Cheng (2002, p. 82) found that “current de-escalation strategies based on reducing information asymmetry are less likely to be effective where escalation of commitment is the result of managers’ high personal responsibility for projects. Instead, solutions need to be developed that mitigate the effect of personal responsibility or provide incentives that counteract the motivation of these managers to escalate their commitment.”

The current study follows this call by investigating allowance for failure. To operationalize this variable, the paper focuses on a set of activities that has been recommended in the literature to reduce perceived threat in case of failure. According to prior experiments on escalation, some of the most effective de-escalation strategies have to do with the specific design of accountability. Simonson and Staw (1992) indicated that three management and accounting practices are most effective for reducing escalation: (1) being prepared to change policy if minimum target levels are not achieved, (2) evaluating decision makers based on their decision process rather than outcome, and (3) making negative outcomes less threatening. This paper refers to these management and accounting practices as “allowance for failure”. The term allowance embraces the notion that failure is an acceptable outcome of every project in which business risk is involved (Geraldi et al. 2010). If the possibility of failure considered from the beginning and if project cancellation is an acceptable result, it can be expected that managers will perceive less threat in the case of project failure. The three practices will be discussed in more detail in the following paragraphs.
  1. 1.

    According to Simonson and Staw (1992), the possibility of failure and premature project termination should be considered from the beginning. The present paper follows this line of thought, arguing that the decision maker’s perceived threat in the case of project failure will be reduced if the environment regularly raises the issue of potential failures for all projects. This is in line with Brockner and Rubin (1985, p. 193) who recommended communicating the general risk of project failure to improve decision makers’ judgments of potential project success. Furthermore, experimental evidence suggests that subjects are less likely to encounter entrapment situations if they have been provided with information about entrapment beforehand (Nathanson et al. 1982). In the same line, decision makers may consider sunk costs more appropriately if supervisors and peers express the importance to do so (Tan and Yates 1995, p. 316).

     
  2. 2.

    In risky environments, even the best decision maker cannot guarantee success. Therefore, instead of outcome, process quality should be controlled (Keil and Robey 1999; Kirby and Davis 1998; Simonson and Staw 1992). Thus, a second important element of allowance for failure refers to the evaluation of the decision makers. Although outcomes can be informative, “specific information regarding the decision maker’s decision method and its quality … provide[s] much more direct evidence on decision quality” (Tan and Lipe 1997, p. 316). In the case of failing projects, evaluation of the decision process can be expected to reduce more pressure than outcome-based measures. Thus, escalation might be reduced if evaluation focuses on the quality of the decision process instead of the outcome (Keil and Robey 1999, p. 70; Simonson and Staw 1992). To increase allowance for failure, this practice should be communicated early to the decision maker.

     
  3. 3.

    The third aspect emphasized by Simonson and Staw (1992) is to make negative outcomes less threatening. In contrast to the first two aspects, this experimental manipulation does not translate directly into an organizational setting. Therefore, the present study used the protection of the decision makers’ reputation as a proxy. Although this is not the only possible operationalization, reputation protection seems feasible as it is clearly linked with the concept of minimizing threat, which is characterized as reducing “psychological and social costs” (Simonson and Staw 1992, p. 420). Moreover, reputation considerations have been suggested to provide the incentive to escalate if project abandonment were to reveal hidden information and decrease the manager’s value on the labor market (Kanodia et al. 1989). This effect has been shown to be even stronger for junior project managers, who still have to build their reputation (Harrell and Harrison 1994). Likewise, according to the experimental study of Heng et al. (2003), providing assurance to the decision maker that the colleague’s perception of the decision maker’s efficacy remains unchanged in the case of premature project termination is a useful strategy for facilitating the de-escalation of commitment. Thus, it can be assumed that protecting the reputation reduces perceived threat in the case of project failure.

     

These propositions lead to the following hypothesis.

Hypothesis 2

Allowance for failure has a significant negative relation with perceived threat in the case of project failure.

2.3 Perceived threat and project escalation

A number of studies have reported on the effects of perceived pressures on attitudinal as well as behavioral aspects (for an overview see DeZoort and Lord 1997). Examples of affected attitudes include job satisfaction and organizational commitment, whereas behavioral aspects comprise judgment and decisions. For example, perceived demand capability imbalance reduces cognitive performance and decision-making quality (Weick 1983). In this line, financial controls have been shown to be associated with increased pressure to meet financial targets, which is in turn related to management myopia (Merchant 1990). Analogous, perceived threat might reduce decision quality related to the termination of failing projects.

In addition to cognitive limitations, opportunism can explain the effect of perceived pressures on decision-making. Empirical evidence suggests that obedience pressure increases auditors’ willingness to sign off on materially misstated financial statements (Lord and DeZoort 2001) and accountants’ willingness to create slack (Davis et al. 2006). Specifically, in regard to decisions about failing projects, it has been shown that perceived threat increases project escalation. For example, Steinkühler et al. (2013) found that decision makers who feel threatened by a failing project have biased cognitions which further increases investments in this project. Moreover, based on an agency perspective, Harrell and Harrison (1994) provided experimental evidence that individuals worried about negative consequences for their reputation or career and who have private information exhibit a greater tendency to continue a project that is expected to become unprofitable.

Based on such reasoning, a positive relationship between perceived threat in case of project failure and project escalation can be expected.

Hypothesis 3

Perceived threat in the case of project failure has a significant positive relation with project escalation.

3 Research method

3.1 Sample and survey procedure

A wide range of project types has been found to be prone to escalation of commitment, such as research and development projects (Schmidt and Calantone 2002), software projects (Heng et al. 2003; Keil et al. 2003; Tiwana et al. 2006; Zhang et al. 2003), strategic decisions (Kisfalvi 2000; Lant and Hurley 1999), market-entry decisions (Camerer 1999; Rapoport 1995; Rapoport et al. 1998), and large-scale construction and building projects (Ross and Staw 1993). Therefore, a cross-sectional approach was chosen, which is favorable for the generalizability of the results. A complete list of project types is presented in Table 1 while the number of projects in each industry is shown in Table 2.
Table 1

Summary statistics of the sample projects

Project type

Position of decision maker

Total

CEO

Lower-level manager

Owner

n

%

n

%

n

%

n

%

Implementation (e.g., software)

68

21.3

27

24.8

16

12.0

111

19.8

Optimization/process improvements

71

22.2

29

26.6

29

21.8

129

23.0

Product development

59

18.4

26

23.9

28

21.1

113

21.1

Launch of a new product

39

12.2

11

10.1

21

15.8

71

12.6

Market entry

20

6.3

0

0.0

12

9.0

32

5.7

Plant construction

22

6.9

7

6.4

8

6.0

37

6.6

Real estate construction

11

3.4

1

0.9

6

4.5

18

3.2

Outsourcing

10

3.1

2

1.8

3

2.3

15

2.7

Basic research

4

1.3

4

3.7

2

1.5

10

1.8

Other

16

5.0

2

1.8

8

6.0

26

4.6

Total

320

100.0

109

100.0

133

100.0

562

100.0

Table 2

Summary statistics of the sample industries

Industry

Position of decision maker

Total

CEO

Lower-level manager

Owner

n

%

n

%

n

%

n

%

Mechanical engineering and plant construction

53

16.6

23

21.1

26

19.5

102

18.1

Other services

41

12.8

13

11.9

23

17.3

77

13.7

Electronics, precision engineering, and optical devices

37

11.6

13

11.9

9

6.8

59

10.5

Chemicals and plastics

31

9.7

12

11.0

7

5.3

50

8.9

Automobile

18

5.6

6

5.5

2

1.5

26

4.6

Manufacture of food products, beverages, and tobacco

19

5.9

1

0.9

9

6.8

29

5.2

Trade

18

5.6

8

7.3

17

12.8

43

7.7

Consumer goods

18

5.6

3

2.8

11

8.3

23

5.7

Information and telecommunication

17

5.3

7

6.4

6

4.5

30

5.3

Building industry/construction

15

4.7

5

4.6

11

8.3

31

5.5

Energy and raw materials

14

4.4

4

3.7

3

2.3

21

3.7

Finance and insurance

13

4.1

6

5.5

1

0.8

20

3.6

Transportation and logistics

13

4.1

1

0.9

6

4.5

20

3.6

Health and biotechnology

11

3.4

5

4.6

1

0.8

17

3.0

Public administration, education, social welfare

1

0.3

1

0.9

0

0.0

2

0.4

Missing

1

0.3

1

0.9

1

0.8

3

0.5

Total

320

100.0

109

100.0

133

100.0

562

100.0

The data for this study was collected in 2007 as part of a larger research project (Mahlendorf and Wallenburg 2013). To collect the data, personal contact information was drawn randomly from the Hoppenstedt database, one of the most comprehensive databases of firms located in Germany.1 Personalized invitation letters were sent via email.

Escalation is defined as commitment to a failing course of action. Hence, this paper focuses on failed projects. This procedure is in accordance with previous studies (Dilts and Pence 2006) that attempted to investigate escalation in the field instead of the laboratory. In the beginning of the questionnaire, the participants were asked to indicate whether they had personal experience with a project that had been terminated before its goals were accomplished. The participants were requested to reply to all subsequent questions with respect to one specific failed project that they remembered well. Participants without appropriate acquaintance with a failed project did only receive a short version of the questionnaire. Therefore, while 2,006 people were willing to participate (which is equal to a response rate of 24.6 %), a large proportion of those respondents did not have the required experience with a failed project, which reduced the number of usable responses to 562. For the analyses, this remaining sample was split into three groups: projects with a CEO as decision maker (N = 320), projects with a lower-level manager as decision maker (N = 109), and projects with a company owner as decision maker (N = 133).

At the time of the project termination, the average respondent in our sample had more than 15 years of work experience, of which approximately 9 years had been spent in the company where the failed project occurred. This suggests that the respondents are experienced professionals. The three types of projects most often mentioned were implementation projects, optimization/process improvement projects, and product development projects (see Table 1).

The survey study procedures followed the literature recommendations (Dillman 2000; Van der Stede et al. 2005). The study was designed to minimize impression management and related problems (Dillman 2000; Hague 1993; Hardesty and Bearden 2004; Podsakoff et al. 2003). To reduce order response bias, the online survey instrument was programmed to present questions in random order within each block of questions.

To test for a potential non-response bias, the answers of early respondents were compared with the answers of late respondents (J. S. Armstrong and Overton 1977; Oppenheim 1966). A MANOVA, including all model variables, was performed. Results did not indicate a significant difference between early and late respondents. In addition, key figures from the database were used to compare respondents with non-respondents. T tests for annual turnover, amount of total assets on the balance sheet, and annual earnings were conducted. No significant differences were found.

Furthermore, retention times per page were analyzed. The questionnaires of the fastest respondents were scrutinized to investigate possible biases from careless and cursory replies. No noticeable patterns were identified in the answers.

In addition, to test for method variance, the correlations between measures were investigated (Lindell and Whitney 2001; Malhotra et al. 2006). Statistical analysis suggests that the lowest correlations are almost zero; thus, common method bias does not seem to be materially influencing the relations under study.

3.2 Pretests

For all variables, a thorough review of managerial and accounting constructs (Kwok and Sharp 1998; Luft and Shields 2007; Nazari et al. 2006; Schäffer 2008) and psychological test databases (Ovid Technologies 2007) was undertaken. Since the questionnaire was administered in German, items had to be translated. To improve validity, several persons translated the same items independently from each other. Differences were discussed. After agreeing on one translation, the “back translation” procedure was employed (i.e., a native speaker of English translated the German version back to English). The deviations from the original items were marginal, which supports the adequacy of the translations.

The development of new constructs followed established procedures (Bispe et al. 2007; Churchill 1979; DeVellis 2002; Diamantopoulos 2005; Nunnally 1978; Rossiter 2002). Several quality checks were conducted to improve the new scales as well as the survey as a whole. First, the variables and their respective measures were discussed with both, academics and practitioners. Once the questionnaire had been developed, the complete questionnaire was discussed with academics and practitioners from the target population. Cognitive interviewing was applied, i.e., the respondent was encouraged to think aloud during the process of answering the questionnaire (Bolton 1993; Schwarz 1999). As a result, several formulations were modified to mitigate occasional ambiguity and the length of the questionnaire was reduced.

As a quantitative pretest, the complete questionnaire was sent to 456 technical managers and 417 accountants; 84 respondents (9.6 percent) completed the questionnaire. The reliability analysis (Cronbach’s alpha and composite reliability) demonstrated satisfactory results, except for the five item construct impression management. The two reversed items of the impression management had to be dropped due to low item reliability, leaving a three item scale (see “Appendix”). Data from the pretest was not included in the following analysis.

3.3 Measures

The central variables of interest measured in this study include capital market orientation, allowance for failure, perceived threat in the case of project failure, and project escalation. All variables were measured as latent variables with multiple items. The complete text of each item is shown in the “Appendix”. For respondents who were not in the role of a key decision maker in the respective project, references to “I” and “me” were replaced with “the key decision maker”. A test of invariance (Byrne 2001, p. 187) between the two wordings revealed no significant differences at the 5 % level. The following discussion presents the measurement scales in detail.

The measurement of capital market orientation is based on a reflective three-item scale, developed by Grieshop (2008). The construct relates to the importance the company assigns to its relation to the capital market. A sample item is “Capital market orientation was of utmost importance to the company.” The Cronbach alpha coefficient of this scale is 0.89, suggesting high reliability (Bagozzi and Yi 1988, p. 82).

Allowance for failure was measured using three reflective items based on the practices proposed by Simonson and Staw (1992). Allowance for failure refers to the style in which a decision maker’s environment deals with potential project failures. Allowance for failure is high if the capital market, corporate governance, or company’s management control system takes into consideration that failure is a possible and acceptable outcome. A sample item is “I have not only been judged according to the results of my decisions, but also for the quality of my decision-making process (i.e., it has been evaluated whether the decision was justified at the particular point in time).” The Cronbach alpha coefficient of this scale is 0.64, which is satisfactory.

Perceived threat in case of project failure was measured with three items that refer to the decision maker’s concern related to a potential premature project termination (Harrell and Harrison 1994, p. 573).2 A sample item is “I worried about my reputation in case of premature project termination.” The Cronbach alpha coefficient of this scale is 0.87, indicating high reliability.

Project escalation was measured using a three-item scale based on the definition of escalation as the continuation of a failing course of action (Ruchala 1999, p. 163). The scale reflects the respondents’ perception of the extent to which the failed project was terminated too late. A sample item is “The project has been terminated too late.” The Cronbach alpha coefficient of this scale is 0.93. In the absence of an objective measure for escalation of commitment in real investment projects, this construct is a proxy for escalation of commitment in failed projects. This construct addresses escalation more precisely than prior proxies, such as the single item “willingness to continue a project” by Keil et al. (2000, p. 308), the dichotomous measurement of project escalation by Keil et al. (2003, p. 245), and the use of “executive management team tenure … divided by the change in revenue over the 5-year period” by Armstrong, Williams, and Barrett (2004, p. 369).

A number of control variables were included in the study. Listed companies are usually larger companies. Accordingly, company size can be expected to influence capital market pressures. In addition, company size is related to corporate governance practices such as the emergence and sophistication of management control systems (Aerts 2005, p. 505; Davila 2005, p. 243; Davila and Foster 2007). Sullivan and Snodgrass (1991, p. 26) found mixed results concerning the relationship between company size and failure tolerance. Consequently, company size was included as a control variable. It was measured as annual turnover and, following common practice, was represented by the base 10 logarithm in the statistical analysis (Cassar and Gibson 2008, p. 716).

The literature suggests that project size may be important to project termination decisions (Dilts and Pence 2006, p. 384). Meanwhile, invested resources may impact the decision maker’s perceived threat in the case of failure. Hence, the base 10 logarithm of the financial resources invested in the failed project was used as a control variable.

Given that this study deals with a sensitive topic like failed projects, impression management may systematically bias the answers. Impression management occurs when respondents attempt to portray themselves favorably in terms of social norms and standards (Kline et al. 2000; Paulhus 1992; Thompson and Phua 2005). When such statements are about others, the phenomenon is called leniency bias—namely, “the tendency to describe others in favorable but probably untrue terms” (Schriesheim et al. 1979, p. 1). To address impression management, an adopted version of the Paulhus’ impression management subscale (Bernardi 2006; Paulhus 1986) was included in the questionnaire.

Moreover, the study includes gender, number of years of work experience, and company tenure of the respondent, which have been shown to affect managerial beliefs, values and actions (Tabachnick and Fidell 2001).

In addition, cross-validations with different hierarchical positions of the decision maker were conducted (see Chapter 4.2).

4 Empirical results

4.1 Statistical analysis of the hypothesis

Structural equation modeling (SEM) using AMOS 17.0 (Arbuckle 2008) was used to analyze the data. SEM is considered as state-of-the-art approach to testing for indirect effects (Iacobucci et al. 2007, p. 144) and thus superior to regression analysis for the model in this paper.

In line with the suggestion of Anderson and Gerbing (1988), a two-step procedure was used. Before the structural model was analyzed, the measurement models were established using confirmatory factor analysis (CFA). The maximum likelihood procedure was applied. The results (factor loadings, χ2/df, NNFI/TLI, CFI, RMSEA, and SRMR) supported the internal consistency of the measurement instruments and suggest good fit (Hu and Bentler 1999). Discriminant validity was assessed using the procedure proposed by Fornell and Larcker (1981). Descriptive statistics and the estimates from the confirmatory measurement models can be found in the Tables 3, 4, 5.
Table 3

Descriptive statistics and confirmatory factor analysis (CFA) for the main variables of interest for projects with CEOs as decision maker (n = 320)

Measure

Mean

SD

Construct loadings*

Capital market orientation (CMO)

Allowance for failure (AFF)

Perceived threat in the case of project failure (PT)

Project escalation (PPE)

Indicator

CMO1

2.64

2.05

0.81

   

CMO2

3.27

2.15

0.82

   

CMO3

2.99

2.10

0.91

   

AFF1

4.03

2.02

 

0.44

  

AFF2

4.05

1.90

 

0.55

  

AFF3

3.95

2.26

 

0.78

  

PT1

3.15

1.93

  

0.91

 

PT2

3.16

1.81

  

0.87

 

PT3

3.54

1.93

  

0.72

 

PPE1

4.28

2.23

   

0.93

PPE2

4.34

2.13

   

0.87

PPE3

4.41

2.22

   

0.91

 

Construct covariances (standardized)†

Construct

CMO

  

1.00

   

AFF

  

−0.02 ns

1.00

  

PT

  

0.15

−0.37

1.00

 

PPE

  

0.05 ns

−0.11 ns

0.28

1.00

SD Standard deviation

* All loadings are statistically significant at the 0.001 level (two-tailed), with the smallest t value being 3.74. Common metric completely standardized estimates are given. Seven-point scales were used

† All construct covariances are significant at the 0.05 level (two-tailed), except where indicated by ns

Table 4

Descriptive statistics and confirmatory factor analysis (CFA) for the main variables of interest for projects with lower-level managers as decision maker (n = 109)

Measure

Mean

SD

Construct loadings*

Capital market orientation (CMO)

Allowance for failure (AFF)

Perceived threat in the case of project failure (PT)

Project escalation (PPE)

Indicator

CMO1

2.71

1.96

0.74

   

CMO2

3.32

2.05

0.78

   

CMO3

3.22

2.04

0.98

   

AFF1

3.66

1.90

 

0.54

  

AFF2

3.77

1.84

 

0.72

  

AFF3

3.78

2.17

 

0.84

  

PT1

3.19

1.63

  

0.97

 

PT2

3.45

1.72

  

0.92

 

PT3

3.63

1.72

  

0.68

 

PPE1

4.02

2.14

   

0.93

PPE2

3.95

2.08

   

0.88

PPE3

4.29

2.15

   

0.96

 

Construct covariances (standardized)†

Construct

CMO

  

1.00

   

AFF

  

−0.14 ns

1.00

  

PT

  

0.14 ns

−0.34

1.00

 

PPE

  

0.10 ns

−0.45

0.38

1.00

SD Standard deviation

* All loadings are statistically significant at the 0.001 level (two-tailed), with the smallest t value being 4.84. Common metric completely standardized estimates are given. Seven-point scales were used

† All construct covariances are significant at the 0.05 level (two-tailed), except where indicated by ns

Table 5

Descriptive statistics and confirmatory factor analysis (CFA) for the main variables of interest for projects with company owners as decision maker (n = 133)

Measure

Mean

SD

Construct loadings*

Capital market orientation (CMO)

Allowance for failure (AFF)

Perceived threat in the case of project failure (PT)

Project escalation (PPE)

Indicator

CMO1

2.47

1.94

0.83

   

CMO2

2.67

1.95

0.69

   

CMO3

2.45

1.75

0.77

   

AFF1

4.07

1.91

 

0.60

  

AFF2

3.97

1.94

 

0.62

  

AFF3

3.99

2.00

 

0.76

  

PT1

3.10

1.91

  

0.92

 

PT2

2.78

1.76

  

0.75

 

PT3

3.51

1.90

  

0.69

 

PPE1

3.83

2.18

   

0.96

PPE2

3.78

2.17

   

0.90

PPE3

4.10

2.12

   

0.89

 

Construct covariances (standardized)†

Construct

CMO

  

1.00

   

AFF

  

0.14 ns

1.00

  

PT

  

0.10 ns

−0.14 ns

1.00

 

PPE

  

0.04 ns

−0.25

0.22

1.00

* All loadings are statistically significant at the 0.001 level (two-tailed), with the smallest t value being 4.80. Common metric completely standardized estimates are given. Seven-point scales were used

† All construct covariances are significant at the 0.05 level (two-tailed), except where indicated by ns

SD Standard deviation

Having established the measurement models, the structural model was then estimated, again using the maximum likelihood procedure. All three groups (decision maker is CEO vs. lower-level manager vs. company owner) were estimated simultaneously. The model has a χ2 of 647 based upon 456 degrees of freedom, resulting in a normed χ2/df of 1.42. The NNFI/TLI is 0.94, the CFI is 0.96, the RMSEA is 0.027, and the SRMR is 0.05. These fit measures all indicate good model fit (Hu and Bentler 1999).

The estimated parameters for the main group (i.e., with CEOs as decision makers) are presented in Fig. 1 (the results for the other groups that serve as cross-validations are presented in Chapter 4.2). The estimated parameters show a significant positive relation between capital market orientation and perceived threat in the case of project failure. This result supports Hypothesis 1. Furthermore, in accordance with Hypothesis 2, allowance for failure has a significant negative relation with perceived threat in the case of project failure. This suggests that a higher allowance for failure significantly reduces the threat in the case of project failure. In addition, the data indicate a significant positive relationship between perceived threat in the case of project failure and project escalation. Thus, Hypothesis 3 is also supported.
https://static-content.springer.com/image/art%3A10.1007%2Fs10997-013-9276-3/MediaObjects/10997_2013_9276_Fig1_HTML.gif
Fig. 1

Fully standardized estimated parameters of the structural model for projects with CEOs as decision maker (N = 320). For purposes of clarity, only significant standardized path coefficients are shown. Indicators, error terms, and residuals are not depicted. The significant levels (two-tailed) are *p < 0.10, †p < 0.05, and ‡p < 0.01

To test the indirect effects of allowance for failure and capital market orientation on project escalation, bootstrapping was applied. This approach is considered superior to the Sobel z-test (Arbuckle 2006, p. 435; Zhao et al. 2010; Preacher and Hayes 2004, 2008). Results based on 2,000 bootstrap samples indicate that both indirect effects are significant. The standardized estimate for the indirect effect of capital market orientation is 0.03 (p = 0.042) and for allowance for failure −0.08 (p = 0.012).

Some of the control variables demonstrate significant effects on the variables under study. However, including these control variables does not eliminate the hypothesized relationships.

4.2 Cross-validation and exploratory extensions

In addition to the 320 projects with CEOs as key decision maker, which were used in the previous chapter for testing the hypothesis, data collection also yielded information from 109 projects that were terminated by second or lower-level managers and from 133 projects terminated by the company owners. As a cross-validation and exploratory extension, the same model as in Fig. 1 was tested for these decision makers.

The results for the projects that were terminated by second or lower-level managers are as follows. The standardized path coefficient for the effect of allowance for failure on perceived threat in case of project failure is −0.29 (p < 0.01, two-tailed). The coefficient between capital market orientation and perceived threat is 0.07 (not significant on a conventional level). The path coefficient for the effect of perceived threat on project escalation is 0.40 (p < 0.01, two-tailed). Thus, all hypothesized effects are similar to those reported for the main sample except the relationship between capital market orientation and perceived threat in the case of project failure, which is not significant here. Accordingly, capital market orientation does not influence the perceived threat of managers at lower hierarchical levels, although allowance for failure plays an important role in reducing perceived threat and, thus, escalation.

The results for projects for which company owners were directly responsible are largely similar to the main analysis. The standardized path coefficient of allowance for failure on perceived threat is −0.20 (p < 0.10, two-tailed). The effect of capital market orientation is 0.17 (p < 0.10, two-tailed). The effect of perceived threat on project escalation is 0.23 (p < 0.05, two-tailed).

In summary, the cross-validation and exploratory extensions with the decision maker coming from different hierarchical levels show results that are consistent with the hypotheses and the findings for the main analysis, with the plausible difference that the effect of capital market orientation is not significant for lower level managers.

As an additional exploratory analysis, the relationship between the two variables capital market orientation and allowance for failure was investigated in a separate model (not depicted). The reason for this additional analysis is that one could argue that capital market orientation might have an influence on allowance for failure. However, no significant relation was found.

In addition, the potential for the two variables’ interaction effect on perceived threat in the case of project failure was tested. The procedure proposed by Ping (1995) to test an interaction among continuous variables was applied. No significant effect was found.

The structural equation model in this paper investigates linear relationships. However, it may be possible that some of the relationships are non-linear. It could be argued, for example, that the relationship between perceived threat and project escalation follows an inverted U-shaped function. In this case, an optimal level of pressure would exist. To test for such a non-linear relationship, the items for perceived threat and project escalation respectively were summed up and regression analysis was applied using SPSS Statistics 17.0 (2008). First, a regression with a linear term only was calculated. Second, a regression with a linear and a quadratic term was calculated. Adding the quadratic term does not increase the explained variance significantly. Thus, a linear relationship seems to be an appropriate assumption in the investigated range of values.

5 Discussion

5.1 Capital market orientation as a source of perceived threat

In line with hypothesis 1, the results indicate that capital market orientation has a significant positive relation with perceived threat in the case of project failure. In other words, decision makers in companies with strong capital market orientation are more afraid that their reputation and personal success will be negatively affected by project failure. This in turn is positively associated with project escalation. Thus, decision makers that feel more threat are more hesitant to cancel a failing project and continue investing for too long. Moreover, the test of the indirect effect of capital market orientation via perceived threat on project escalation was significant, supporting the conclusion that on average stronger capital market orientation leads to higher investments into failing projects—a side effect of capital market orientation that is certainly intended neither by capital market players nor by the company.

The results of the study may encourage reevaluating some existing innovative corporate governance practices. New shareholder activism, for example, emphasizes personal accountability of top managers (Rose 2010; Del Guercio et al. 2008). Along these lines, Brav et al. (2008, p. 1740) provide a drastic example for attacking individuals: “We attribute these missteps to CEO Peter Socha’s lack of operating experience within the coal industry and to the Company’s lack of a CFO … We are now convinced that the Company’s senior management team is simply not up to the task of achieving such goals.” Such statements will most likely increase perceived threat in case of project failure, not only for the individuals that are blamed but also for other managers who witness the affront. In the light of the results in this paper, such practices might have some serious downsides. Thus, instead of blaming individuals more intensely for company failures, the optimal level of personal pressure (e.g., exerted by important shareholders) should be considered, because too much pressure might actually be bad for shareholder value, as the current study suggests.

While the projects with CEOs as key decision makers supported hypothesis 1, the exploratory analysis with a sample of projects that were decided by lower level managers did not find a significant relationship between capital market orientation and perceived threat in case of project failure. This result can be explained in terms of lower level managers being less affected by capital market pressures than top-level managers. Marginson and McAulay (2007), for example, suggest that “middle-level managers may be sufficiently removed from the influence of stock markets to be able to focus on actions that improve long-term performance.”3

Further exploratory analysis did not find a significant relation between capital market orientations and allowance for failure. This does not necessarily imply that there is no connection at all between the two variables. Rather, it might be the case that positive and negative effects of capital market orientation cancel out each other. On the one hand, capital market orientation might increase allowance for failure. For example, stronger capital market orientation may be associated with a more salient message to terminate failing projects. This view is in line with the manager statement: “If they see they have a dog, and they are never going to reach their targets with this business, they want to get rid of it,” as quoted by Radcliffe et al. (2001, p. 149). Moreover, companies with stronger capital market orientation can be expected to have professional staff that evaluates projects and investments and thus helps to focus on economic reasoning (e.g., considering sunk costs as irrelevant) which may foster allowance for failure. In a similar vein, Fiss and Zajac (2004, p. 502) state that firms with a shareholder value model have “superior capabilities … to abandon inefficient investments more rapidly.”

On the other hand, capital market orientation may also put more emphasis on results than on decision process quality. This can be expected to decrease allowance for failure. In accordance with this perspective, the empirical results of Sullivan and Snodgrass (1991) suggest that firms’ emphasis of a financially oriented performance measure (e.g., ROI) is negatively associated with failure tolerance. In addition, Mian (2001) demonstrated that CFO turnover is often punitive and proceeded by poor stock price and operating income performance. This also supports the perspective that strong capital market orientation is associated with lower allowance for failure. Future studies should try to disentangle these potentially competing effects.

5.2 Allowance for failure as an innovative concept for reducing perceived threat

In line with hypothesis 2 and 3, the results consistently suggest that allowance for failure reduces perceived threat in case of project failure and in consequence helps to reduce project escalation. Thus, innovative management and accounting practices that contribute to the concept of allowance for failure are beneficial in certain circumstances. This idea challenges the current trend to make managers personally more responsible for failure.

In the exploratory extension, the effect was also tested for projects terminated by the company owner. As one would expect, the relationship between allowance for failure and perceived threat is less strong in this case. Nevertheless, it is still marginally significant. It is interesting to note that for those projects also the perceived threat is significantly related to escalation, which suggests that even company owners feel perceived threat and are not immune to escalation behavior, despite their powerful position.

This study used three specific practices to measure the concept of allowance for failure. The three practices are: (1) Making clear to the decision maker that he has to abandon the project if the goals can no longer be reached. (2) Judging the decision maker not only according to the results of his decisions, but also for the quality of his decision-making process. (3) Communicating to the decision maker that his reputation will not be harmed from premature project termination. These practices have an escalation reducing effect, as the results of this study have shown.

However, there are more innovative management and accounting practices that have the potential to foster allowance for failure. For example, providing objective ex ante criteria for failure has the potential to shift the focus from ex post blaming of people to ex ante awareness that failure is an inherent element of all risky business decisions (Keil and Robey 1999). Likewise, several empirical studies have shown that predetermined, publicly known stopping rules significantly reduce escalation of commitment (Bazerman 1986; Boulding et al. 1997; Brockner et al. 1979; Keil and Robey 1999). Furthermore, allowance for failure might be increased by sharing responsibility so that a second person (e.g., CFO) has to sign major decisions and thereby signals to the primary decision maker that someone else supports his decisions (Heng et al. 2003). Sharing responsibility will make it easier to terminate a project, if the earlier decisions turn out not to have the desired results. Finally, it can be helpful, if decision makers are made aware of the concept of sunk costs and are primed that sunk costs should be treated as irrelevant for decisions about future investments (Dzuranin 2010). In sum, these examples indicate that the concept of allowance for failure can be generalized beyond the specific practices investigated in this study. This idea is important, because it may help to connect findings that are currently investigated in different streams in the literature. In consequence, the unifying umbrella term of allowance for failure may help to increase the awareness of the overall concept and its potential consequences.

5.3 The relevance of the setting for allowance for failure and capital market orientation

The data for this study was collected in Germany. Germany’s capital market has been labeled as bank-centered (Black and Gilson 1998; Edwards and Fischer 1996), which means that firms traditionally rely on banks for a large proportion of their financing (Gedajlovic and Shapiro 1998; Fligstein 1990, p. 308) and financial intermediaries “tend to have much closer, longer-term relationships with many of their industrial clients than do their counterparts in the Anglo-American family of countries” (Kester 1994, p. 181).

Compared with Germany, Anglo-American countries have a stronger orientation towards shareholder value and more intense capital market pressure (Fiss and Zajac 2004). Moreover, individualism is stronger in the Anglo-American corporate culture (Crossland and Hambrick 2007), which may in case of failures go along with more blame on the individual responsible decision maker. Both arguments suggest that perceived threat in case of project failure might be even more relevant in Anglo-American countries. In consequence, the need for innovating management and accounting practices to increase allowance for failure might also be stronger.

In contrast, one could also argue that the more personal and long-term relationships in Germany in some situations increase perceived threat in case of failure. This effect may occur, because people will most likely pay more attention to the opinion of those people they know well. Thus, issues such as face-saving and self-justification may become more relevant with more personal and long-term relationships. In addition, uncertainty avoidance is considered to be stronger in Germany than in the USA (Hofstede 1980). This may result in lower tolerance for unexpected executive actions (Crossland and Hambrick 2007) such as terminating a failing project. These arguments would suggest that the need for allowance for failure is higher in Germany.

Finally, some studies state an international convergence of capital market orientation which shifts Germany’s corporate governance regimes towards the Anglo-American model (Sudarsanam and Broadhurst 2010). Other studies find that Germany and the USA are pretty similar with respect to failure tolerance and granting a second chance but are different from countries such as Sweden or Greece (Hughes and Burchell 2007, p. 17). In conclusion, different contingency effects can be postulated for different countries. Thus, additional studies that investigate allowance for failure in different countries seem desirable.

It should also be stated, that the usefulness of allowance might depend on other characteristics of the setting. For example, it is to be expected that trust (Barretta et al. 2008; Busco et al. 2006; Six 2007) interacts with allowance for failure. Likewise, leadership style may affect perceived threat in the context of failing projects (Otley and Pierce 1995). Moreover, job-related tension (Dunk 1993) represents a variable that may be of interest for future studies on allowance for failure. Drawing analogies from face-saving research, it can also be expected that the effect of allowance for failure increases with the individual’s social anxiety (Brockner et al. 1981) and with the cultural importance of social standing (Chow et al. 1997). Furthermore, the reward and incentive system may interact with the effects observed in this study. For example, the proportion of performance dependent bonuses might help to explain why capital market orientation was found to be significantly related to perceived threat in case of failure for CEOs, but not for lower level managers. The respective influence of each of the different potential sources of allowance for failure (e.g., capital market, corporate governance, and management control system) could also be investigated in more detail in future studies.

The data were collected in 2007, and thus before the recent financial crisis. It can be expected that the relations presented in this study are even more relevant in times of capital shortages and increased pressures. Moreover, as explained in the introduction, the accountability of managers seems to have increased since the time of the data collection, making the emphasis on allowance for failure even more important today. While the specific accountability practices may change over time, the underlying psychological mechanisms for the effects observed in this study will not change in the foreseeable future.

This study empirically investigated allowance for failure in the setting of escalating projects. However, the conceptual relevance is not limited to this setting. Allowance for failure will have consequences in a variety of settings. Allowance for failure can be regarded as an antidote to a variety of pressures that have been shown to be damaging to both the company and the individual professional’s career (DeZoort and Lord 1997). Pressures have been shown to be associated with job dissatisfaction, propensity to leave a job and reduced work performance (Rebele and Michaels 1990). In a similar vein, allowance for failure can be expected to reduce stress—a factor that has substantial negative consequences for management as well as accounting (Weick 1983).

6 Conclusions

This paper has sought to make two main contributions. First, it presents empirical evidence that capital market orientation can have dysfunctional effects on decisions concerning the termination of failing projects. The explanation for this unintended effect is that capital market orientation causes the decision maker to perceive higher threat in case of project failure. This in turn leads to higher investments into failing projects. Taken together, the results suggest that capital market orientation increases managers’ perceived threat in the case of project failure and thus indirectly increases project escalation. These findings provide an additional potential explanation for the fact that capital market orientation is not always related to superior economic performance.

Second, the paper draws attention to the concept of allowance for failure. This concept refers to the style in which the decision maker’s environment—such as capital market, corporate governance, and management control system (MCS)—deals with potential failures (e.g., project or investment failures). While current trends in corporate governance seem to advocate increasing managers’ accountability, this paper emphasizes the need for less accountability in certain situations. The empirical results from failed projects support the hypothesis that allowance for failure reduces perceived threat in case of project failure. Since perceived threat in turn is associated with higher project escalation, allowance for failure provides an opportunity to improve decisions on failing projects. In other words, allowance for failure leads to earlier termination of failing projects because it reduces the decision makers’ fear that their reputation and personal success will be negatively affected by the decision to cancel the project.

The results of the study have relevant implications for practitioners. The findings highlight the importance of carefully analyzing potential consequences of promoting capital market orientation. While the “whip of the market” might often be beneficial for company performance, it can also have dysfunctional effects on decision making, as this study has shown. Consequently, capital market players and designers of corporate governance and management control systems should think about ways that promote capital market orientation without increasing psychological pressure to be successful too much. To foster early termination of failing projects, practitioners may want to consider innovating their formal and informal control system in a way that strengthens allowance for failure. Elements like making the possibility of premature project terminations clear, judging decision makers according to their decision process (rather than merely focusing on results), and protecting decision makers’ reputation in case of project termination seem helpful to decrease perceived threat in case of project failure. Thus, those practices provide means that can help to counterbalance psychological pressures induced by capital market orientation.

In conclusion, the results indicate that allowance for failure ties together different management and accounting practices that might deserve more attention in today’s challenging and sometimes threatening business environment. At the same time, the desirable consequences of organizational allowance for failure shown in this study should not be interpreted as a carte blanche for practitioners to generally reduce accountability. Accountability has many functions and benefits, which are not covered by this study. Thus, desirable and dysfunctional consequences should be considered holistically to find the optimal balance.

The study also has implications for future research. Given the empirical evidence of the current study, corporate governance and accounting research should further investigate how certain aspects of control increase or decrease perceived threat and, in turn, influence decision quality. In particular, it might be worth strengthening the links among capital market pressures, management control, and the psychological literature on the escalation of commitment.

Concerning the model’s power to explain project escalation, it should be kept in mind that the model only comprises a limited subset of influencing factors and focuses exclusively on a single psychological aspect (i.e., perceived threat in the case of project failure). In real-world projects, a plethora of other factors are likely to influence project termination and escalation, such as other psychological aspects, organizational aspects, and external factors. Moreover, the company’s lifecycle stage, for example, may play an important role, as the consequences of a failing project will be more severe for the funding of a start-up compared to a more mature company. In addition, it may be of interest to investigate the relationships among capital market orientation, continuous monitoring, and project escalation, as continuous monitoring significantly reduces managers’ willingness to stick to risky projects (Hunton et al. 2008). More generally, capital market orientation, allowance for failure, and perceived threat might provide interesting perspectives on the issue of how to motivate managers to take more risks and push boundaries.

Finally, capital market orientation and allowance for failure can be expected to impact many other variables in addition to perceived threat in the case of project termination. The escalation of commitment literature has shown the importance of variables such as selective perception, sunk costs, optimism, and decision scope. Thus, investigating the effects of innovative management and accounting practices on these variables might be worthwhile endeavors for future studies.

Footnotes
1

Both listed and unlisted companies were addressed for this study (cp. Aerts 2005). The reason for this choice was to get sufficient variance in the variables of interest, especially concerning the variable “capital market orientation”. If all companies in the sample had an extremely high value on “capital market orientation”, it would be very difficult to find correlations with other variables. In an extreme case, if all companies had the highest value on capital market orientation, this would be a constant and not a variable. In this extreme scenario, it would not be possible to detect any significant correlations. In other words, variance in the variables of interest is a required prerequisite for the analysis and a broad sample increases the variance. Furthermore, although it can be assumed that listed companies have a stronger capital market orientation, unlisted companies can rely on the capital market as well. For instance, they may be subsidiaries of listed companies, they may intend to pursue an IPO in the future, they may issue bonds, they may use or strive for venture capital, and they may raise long-term loans from banks. Even “small businesses rely on financial institutions for over a quarter of their total financing. In contrast to conventional wisdom, this is true even for the smallest small firms” (Berger and Udell 2002, p. F36). If none of these arguments applied, the respondent could choose the answer option “not correct at all” when asked for the intensity of capital market orientation of their company (see “Appendix”).

 
2

In previous research, similar items have also been used to capture involvement (Biyalogorsky et al. 2006) and need for self-justification (Steinkühler et al. 2013), which are conceptually very close to perceived threat in case of project failure.

 
3

However, one might also have expected to find a significant effect of capital market orientation, because increasing pressures from the capital markets lead to “a reconfiguration and rearrangement of forms of management accounting, management control, management information systems, operational control, and so on” (Hopwood 2008, p. 10). Management accounting and control systems thereby provide the means to pass capital market pressures down to managers on different hierarchical levels (Fligstein 1990; Vollmer 2003). These pressures reshape organizations (Zorn et al. 2006) and are reflected by performance evaluation systems, bonuses and payments (Healy and Wahlen 1999), and dismissals (Mian 2001). As such, financial discipline is not only imposed on top-level managers; rather, “the number of involved individuals is much greater than many would suspect. This deep programming of corporate/individual wealth harmony is an essential element in the construction of a culture of performance. In these plans, a heavy reliance on financial accounting measures is typical” (Radcliffe et al. 2001, p. 150).

 

Acknowledgments

The author would like to thank Joan Luft, Wim A. Van der Stede, Jürgen Weber, Utz Schäffer, Erik Strauß, Florian Herschung, Sigrid Gschmack, Maximilian Margolin, Mary A. Malina, participants of the Management Accounting Section Research and Case Conference 2010 and workshop participants at London School of Economics 2008, WHU-Otto Beisheim School of Management 2008, and University of Innsbruck 2009.

Copyright information

© Springer Science+Business Media New York 2013