Influence of Decision Responsibility and Consequence on Escalation of Commitment in Corporate Investment

The study examined decision responsibility (high versus low) and decision consequence (positive versus negative) on escalation of commitment. The study utilized a 2 X 2 factorial experimental design. Participants were 348 Banking and Finance graduate students randomly selected and assigned to four treatment groups. Escalation of Commitment was in two forms : Intent to Escalate Commitment (IEC), measured with a standardized scale and Escalat ion of Commitment Behaviour (ECB), measured by the amount of money a decision maker is willing to invest in a business scenario. Results of 2 X 2 (ANOVA), revealed that decision responsibility (DR) and decision consequence (DC) had significant main effects on ECB. The interaction effects of both variables showed that ECB was highest under low decision responsibility and positive decision consequence condition. Only decision responsibility had significant main effect on IEC. IEC positively influenced ECB. Importance of decision responsibility has implication for escalation of commitment. It is therefore recommended that management should be tactical in engaging individuals with decision responsibility in decision making involving in itial loss.


Introduction
In the course of decision making, it is often observed that organisations continue to invest in their original course of action even after receiving substantial negative information concerning it. An impo rtant organisational context in which organisations may take concrete actions to justify their original decision following negative feedback is that of investment decision making.
Co mmit ment to investment decision is when resources are allocated to a course of act ion and in which the level of resources can be increased at the discretion of the decision maker. W ithin investment decision context, negative feedback may actually cause the organisation to increase the commit ment of resources. An increase in commit ment of resources to a previous investment decision, in spite of negative feedback, is escalation of co mmit ment.
Escalation of co mmit ment is an irrational co mmit ment of increased resources to investment decision to bring about rationality to the in itial decision [1].
The introduction o f Structu ral Ad justment Programme (SA P) in cou nt ries such as Nigeria in 1985 led to the liberalisation of the banking license and the deregulation of the economy. However, in the early 1990's clear signs of distress started to emerge in the financial institution especially banks, probably due to increased and heightened competition fro m the new banks [2][3][4]. Distress in financial institutions may be tied to escalation o f co mmit ment [5]. Organisations including financial institution exhib it escalation of co mmit ment in some decisions and may become locked unto escalation situation. This type of situation is called "syndrome of decision error" [6]. This unwarranted persistence with a losing course of action could be avoided if a change in the course of action was instituted. Continued investment in a losing course of action may lead to great losses.
Escalation of co mmit ment may be explained in terms of behaviour such as amount of money invested in a losing course of action [1]. Ho wever, in literature it has been argued that behaviours are preceded by intentions. In the postulations of theory of reasoned action [7] and theory of planned behaviour [8] behavioural intent was posited to be the most important pred ictor of behaviour wh ich suggest, therefore, that intention to escalate can be relevant.
Every escalation situation concerns at least two decisions, the initial decision and a subsequent decision to stick to the initial course of action. Certain psychological factors have been imp licated to cause escalation of commit ment. So me of these are decision responsibility and decision consequence. Staw postulated that the interaction of decision responsibility and decision consequence is a necessary condition to infer self-justification for escalation of commit ment. Individuals of Commitment in Corporate Investment may be placed with the decision responsibility of allocating resources to a course of action in an investment decision context.
The individuals with personal responsibility for a decision may be saddled with the subsequent outcome which is the decision consequence that resulted from the allocation of resources to that course of action. Therefore they escalate in order to justify prior decision and to demonstrate the ultimate rationality of a losing course of action. Justification theory [1] and self-perception theory [9] posited that individuals restore rationality to their own behaviour. Self-perception theory suggested further that individuals avoid the self causality to negative consequences which cannot be attributed to an external source.
It is presumed that having the decision responsibility to freely make the init ial decision on a course of action that subsequently produced a negative decision consequence, decision makers felt personally responsible for the outcome. They escalate commit ment in an effo rt to justify to self for ego defense and justify to others for public justification. The escalation problem has been characterized as being born by the decision responsibility of the decision maker on the initial decision and the decision consequence especially when such outcome is unfavourable.
Researchers like [10][11][12], manipulated personal responsibi lity for initial decision and found that high personal responsibility increased escalation of commit ment. [13] in a replicat ion of Staw's experiment revealed that there is no escalation of commit ment following negative feedback. [14], [15] reiterated that escalation is not a universal reaction to a failed investment decision. In a laboratory study [16][17][18] revealed that high responsibility individuals reduce their commit ment to previously chosen courses of action rather than escalate follo wing negative feedback Researchers [19], [20] revealed that responsibility for decision consequence induces escalation of co mmit ment. Other researchers [21], revealed that escalation of commit ment did not occur in the h igh involvement group and negative feedback group for tax preparers. [17] found that even though people tend to escalate immediately after negative decision consequence, this tendency reversed over time. Furthermore, [22] reported that the more negative feedback in the early stages of a course of action, the greater the propensity of the decision maker to withdraw.
The escalation of commit ment makes it difficult for people to evaluate the outcomes of their decisions. The decision maker does not consider other options but sticks to an init ial decision wh ich resulted in negative decision consequence. This situation may adversely affect the investment fortunes of the financial institutions leading to distress of such financial institutions. This obvious lapse has prompted this study. There has been a dearth of research on the influence on decision responsibility and decision consequence on intent to escalate commit ment in general, and on escalation of co mmit ment behaviour specifically in Nigeria.
Based on the foregoing, the present study sought to emp irically co mpare intent to escalate commit ment and the escalation of commit ment behaviour of those who receive high decision responsibility and those who receive low decision responsibility. The present study will also examine the predictive ability of intent to escalate commit ment on escalation of commit ment behaviour.
The hypotheses stated are: 1. There will be significant independent and joint effects of decision responsibility and decision consequence on intent to escalate commit ment 2. There will be significant independent and joint effects of decision responsibility and decision consequence on escalation of commit ment behaviour.
3. Intent to escalate commit ment will positively pred ict escalation of commit ment behaviour.

Research Design
An experimental research was conducted, using a 2 X 2 factorial design. The independent variables in the design are decision responsibility (high versus low) and decision consequence (positive versus negative). Generally, the dependent variables are intent to escalate of commit ment (IEC) and escalation of commit ment behaviour (ECB).

Participants
A total of 348 participants who are Master of Business Admin istration (Banking and Finance) students of a University also took part in the main study. In the main study, 210 (60.3%) male and (138) (39.7%) female part icipated in the study. Their average age was 37.62 with standard deviation of 6.16. Tenure on the job had an average of 11.36 and standard deviation of 6.14. The participants were working with (185) (53.2%) in banking, (26) (7.5%) participants in other financial institutions, (115) (33.0%) participants in finance and accounts departments of non-financial organisations and 22 (6.3%) in related departments. In line with [23] pattern of study.
Each participant was randomly assigned into one of 4 treatment groups. High decision responsibility and positive decision consequence treatment group comprised of 88 participants with 55 (60.50%) male and 33 (37.50%) female. High decision responsibility and negative decision consequence treatment group was made up of 86 participants with 54 (67.79%) male and 32 (37.21%) female. Low decision responsibility and positive decision consequence treatment group co mprised of 90 participants with 52 (57.78%) male and 38 (42.22%) female. Low decision responsibility and negative decision consequence treatment group had 84 part icipants comprising of 49 (58.34%) male and 35 (41.66%) female.

Instruments
The intent to escalate commit ment (IEC) scale is constructed for the purpose of this study. The scale is composed of 27 items Likert format and rated as Strongly Agree (5) to Strongly Disagree (1). Higher values in the IEC scale reflect higher intention to escalate commit ment. It originally had 34 items that described the extent of intent to escalate commit ment, in terms of predisposition and tendency to stick to an in itial course of action despite convincing proofs received fro m negative feedback that the initial decision was effective. The 34 items were g iven to some psychologists as experts fo r content validation following [24]. The experts were requested to indicate whether or not they considered each item as measuring intent to escalate commit ment using a 5-point Likert type response ranging fro m Strongly Disagree (1) to Strongly Agree (5). Out of the 34 items, 27 items that were above the average score of 3 were accepted as valid.
The 27 items were subjected to item-total correlat ion. The items with coefficient that ranged between 0.37 and 0.94 were retained [25]. The reliability for the scale is coefficient alpha of 0.97 The 27 items were subjected to factor analysis using the varimax rotation leading to three factors accounting for 71.9% of the total variance. Based on literature [1], [23], the three factors were named external justification, ego defense and biased belief update. The coefficient alpha for the three factors are: 0.97, 0.96 and 0.84 respectively.
Escalation of co mmit ment behaviour is measured in an experimental study and utilised scenario for investment decision developed by Staw (1976). The scenario co mprised the first investment decision and the second investment decision where the independent variables decision responsibility (h igh versus low) and decision consequence (positive and negative) were manipulated. There are 4 treatment conditions of investment decisions following the 2 X 2 factorial design. Escalation of co mmit ment behaviour is measured as the difference between the amount of money initially invested to a unit and the amount invested after feedback on the outcome of in itial investment. It is obtained by subtracting the amount invested on the first investment decision (a constant) fro m the amount invested on the second investment decision. A positive score is obtained when the investment amount at first decision is lower than the investment amount on second decision. A negative score is obtained when the investment amount at first decision is higher than the investment amount on second decision. A positive score indicates escalation of commit ment. A negative score indicates de-escalation of commit ment.
For manipulat ive check of variab les, decisionresponsibili ty was measured with one item 7-point bipo lar ad jective scale. Participants indicated the extent to which they felt responsible for the first and second decision. The bipolar adjectives ranged between low responsibility to high responsibility. Decision consequence was measured with a one item 7-point bipolar adjective scale. Part icipants indicated their perception of the first investment decision outcome. The bipolar adjectives ranged between negative consequence to positive consequence.

Procedure
Prospective participants were drawn fro m the class list using systematic random samp ling. This grouping was done separately for males and females. The experiment took place in the auditoriu m. The part icipants were willing and were assured of confidentiality.
All the participants were presented with a prior ten-year financial condition of A & S Co mpany fro m 1987 to 1997. The first investment decision in which the participants made an initial decision to allocate ten (10) million naira Research and Development (R and D) funds to either of two corporate divisions (consumer or industrial).
In decision responsibility treat ment, participants in high decision responsibility condition sequentially made two investment decisions. After making the first investment decision, only high decision responsibility participants were asked to write a paragraph to defend the allocation. In low decision responsibility condition, half of the total participants under this treatment were made to receive investment decision for consumer div ision while the other half received for industrial div ision. Participants were made to understand that another financial executive of the company had made an earlier Research and Develop ment funding decision in 1997 and had decided to invest all the Research and Develop ment (R and D) funds in either the consumer or industrial corporate d ivisions. Part icipants in the low decision responsibility condition made only a second decision without having made a prior choice as to which corporate division was most deserving of the R and D funds.
In decision consequence treatment, half of the total participants under this treatment were made to receive investment decision for positive consequence condition while the other half received for negative consequence condition. In the positive decision consequence condition, participants received financial data, wh ich showed that the chosen corporate division had returned to profitable levels while the unchosen corporate division continued to decline. In a parallel manner, part icipants in the negative decision consequence condition received financial data, which showed a deepening decline in the profitability of the chosen corporate division but an imp rovement in the unchosen corporate division.
All the part icipants were presented with a five-year financial condit ion of A & S Co mpany fro m 1998 to 2002. In the second investment decision, another twenty (20) million naira Research and Development (R and D) funds was made available to be shared between the two corporate divisions. The participants were required to write their responses specifying the amount of money they would allocate to the two corporate divisions. After the second investment decision the participants were asked to write a paragraph to defend their decision. Intent to escalate commit ment was measured as a composite of the score on its three subscales which are external justificat ion, ego defense and biased belief update.

Results
A justification for the use of composite score is observed in Table 1 where the three subscales are significantly intercorrelated and each correlates significantly with intent to escalate commit ment, as can be seen in the correlation matrix.
Two-way Analysis of variance was employed to test the first two hypotheses. The result in Table 2 reveals main and interaction effects of the independent variables. Result in Table 3 reveals mean values of interaction between the independent variables. In hypothesis 1, the results reveal significant main effects of only decision responsibility on intent to escalate commit ment (F (1, 344) = 20.46, p < .01). The result reveals that participants in low decision responsibility are significantly h igher on intent to escalate commit ment ( x = 95.22) than high decision responsibility participants ( x = 84.22). Other serendipitous findings were made on the main and interaction effects of decision responsibility and decision consequence on the three subscales of intent to escalate commit ment. There is a significant main effect of decision responsibility on external justification for escalation of commit ment (F (1,344) = 23.77, p < .01).
Participants in low decision responsibility are significantly higher on external justification for escalation of commit ment ( x = 47.36) than high decision responsibility participants ( x = 39.51). The results also reveals significant main effects of decision responsibility on ego defense for escalation of co mmit ment (F (1,344) = 19.54, p < .01). Participants in low decision responsibility are significantly higher on ego defense for escalation of co mmit ment ( x = 32.85) than high decision responsibility participants ( x = 28.16). The result further reveals significant main effect of decision responsibility on biased belief update for escalation of co mmit ment. (F (1,344) = 19.94, p < .01). Part icipants in low decision responsibility are significantly higher on biased belief update for escalation of co mmit ment ( x = 20.64) than high decision responsibility participants ( x = 17.45). The hypothesis is not confirmed.
In hypothesis 2, the result of the 2 X 2 A NOVA on Tab le 2 reveals significant main effect of decision responsibility on escalation of co mmit ment behaviour (F (1, 344) =28.68, P < .01). Participants in low decision responsibility are significantly higher on escalation of commit ment behaviour ( x = 1.50) than high decision responsibility participants ( x = 1.24) (see Table 3).
The result of the 2 X 2 ANOVA shown in Table 2 reveals significant main effect of decision consequence on escalation of commit ment behaviour (F (1, 344) = 16.84, P < .01). Participants in positive decision consequence condition were significantly higher on escalation of commit ment behaviour ( x = 1.47) than negative decision consequence participants ( x = 1.26). The 2 X 2 ANOVA result further shows significant interaction effect of decision responsibility and decision consequence on escalation of commit ment behaviour, (F (1, 344) = 6.52, P < .01).
A multip le co mparison analysis test using Scheffe's analysis in table 4 reveals that escalation of commit ment behaviour is significantly higher amongst participants that received low decision responsibility and positive decision consequence ( x =1.66) than those that received high decision responsibility and positive decision consequence ( x =1.27), high decision responsibility and negative decision consequence ( x = 1.20) and, low decision responsibility and negative decision consequence ( x = 1.33). The hypothesis is not confirmed.
However, other serendipitous results show that there was no significant difference among other paired co mparisons which are high decision responsibility and positive decision consequence, high decision responsibility and negative decision consequence, low decision responsibility and negative decision consequence. In addition, the interaction between decision responsibility and decision consequence is depicted graphically in Fig. 1. It can be observed that escalation of co mmit ment behaviour is higher for low decision responsibility than high decision responsibility under both positive and negative consequence situations. However, the difference between low decision responsibility and h igh decision responsibility is h igher under positive decision consequence than under negative decision consequence as can be observed in the sloppiness of the curve.
In hypothesis three, the result in Table 5 reveals that intent to escalate commit ment positively predicts escalation of commit ment behaviour F (1, 346) = 63.22; B = 0.39; P <.01). Participants who exh ibit intention to escalate commit ment also exh ibit escalation of commit ment behaviour. The hypothesis is confirmed.

Discussion and Implication of Fi ndings
In hypothesis one the result revealed significant main effect of decision responsibility on intent to escalate commit ment and on each of the three subscales which are external justificat ion, ego defense and biased belief update. The hypothesis is not confirmed for intent to escalate commit ment and the three sub scales. Those low in decision responsibility escalate commit ment mo re than those high in decision responsibility. This result supported those of some other researchers such as [17]; [18] where h igh personal responsibility group reduced escalating commit ment. The result did not supporting other researchers like, [10][11][12] who man ipulated personal responsibility for init ial decision and found that high personal responsibility increased escalation of commit ment.
There is no significant main effect of decision consequence on intent to escalate commit ment and its three subscales which are external justification, ego defense and biased belief update. [13] in a replication of Staw's  [14]; [15] reiterated that escalation is not a universal reaction to a failed investment decision. Contrary to expectation, there is no significant interaction effect of decision responsibility and decision consequence on intent to escalate commit ment and its three sub scales which are external justificat ion, ego defense and biased belief update. This result supported prior review e.g. [21] which revealed that escalation of commit ment did not occur in the high involvement group and negative feedback group for tax preparers.
The hypothesis two was not confirmed as participants in low decision responsibility are significantly higher on escalation of co mmit ment behaviour than high decision responsibility participants. The result obtained is consistent with the results obtained in a laboratory study by [16]. They revealed that high responsibility individuals reduce their commit ment to previously chosen courses of action rather than escalate follo wing negative feedback.
Participants in positive decision consequence were significantly higher on escalation of commit ment behaviour than negative decision consequence participants. The result confirmed p revious studies, [17] found that even though people tend to escalate immediately after negative decision consequence, this tendency reversed over time. Other researchers [22] reported that the more negative feedback in the early stages of a course of action, the greater the propensity of the decision maker to withdraw.
There was a significant interaction effect of decision responsibility and decision consequence on escalation of commit ment behaviour. A mult iple co mparison analysis of mean scores revealed that escalation of co mmit ment behaviour is significantly higher amongst participants that received low decision responsibility and positive decision consequence than those that received high decision responsibility and positive decision consequence, high decision responsibility and negative decision consequence, and low decision responsibility and negative decision consequence. The hypothesis is not confirmed.
The assumption of [1] justification theory reiterated that decision makers with high decision responsibility and negative decision consequence escalate commit ment to a losing course of action. However, unlike in earlier research, findings fro m the present study suggest that neither high decision responsibility nor negative decision consequence are necessary conditions for escalation of commit ment to occur. So me researchers [19]; [20] revealed that responsibilit y for decision consequence induces escalation of commit ment. [19] found that escalation of commit ment also occurred for indiv iduals in the negative consequence group and those in the positive consequence group.
Result of the qualitative report highlighted other reasons why the participants increase or decrease commit ment of resources to investment decision. Co mmit ment of resources was based on some reasons such as the functions of the products to the end users. Some participants' commit ment was based on the previous year's financial performances of the two divisions and the consequences of loss or profit indicating that the value of the allocation determined proportion of turnover. Other participants committed resources to the division in loss position after the init ial allocation in order to resuscitate performance and improve earnings of the division because it was unwise to totally ignore a business in existence that is distressed. They therefore strategically motivated the div ision by increasing commit ment of resources. The findings of the qualitative report may be the probable reason why the result of hypothesis one and hypothesis two though significant, did not support previous study by [1].
In hypothesis three, results revealed that intent to escalate commit ment positively influences escalation of commit ment behavior. Part icipants' intention to escalate commit ment accurately predicted their actual escalation of commit ment behaviour. The result is supported by the postulation of [7] in theory of reasoned action and [8] in theory of planned behaviour that behavioural intent is the most important predictor of behaviour.

Implication and Recommendation
The study identified that decision responsibility and decision consequence influence both intent to escalate of commit ment and escalation of co mmit ment behaviour. The report that intent to escalate commit ment predicts escalation of commit ment behaviour is confirmed by literature.
With the result obtained from the research management should endorse rational decision making in staff especially those staff who have high intent to escalate commit ment so as to cub them fro m exh ibit ing the actual escalation of commit ment behaviour.
The present study used laboratory experiment with fictit ious scenarios. This is not peculiar to the present study alone but is general to all laboratory experiment. Future studies should use field experiment and or longitudinal study to collect corporate data to further assess the generalisability of the findings and a longer period of study.