Podpora rízení rizik peclivým výberem protikorupcních opatrení - kritické zhodnocení cinností pri rízení kompliance (dodrzování predpisu)/Support of Risk Management Through Careful Selection of Anti-corruption Measures - a Critical Evaluation of Compliance Management Activities

Abstract

Organizations tend to apply Compliance Management System (CMS) as means to reduce risk. The fraud or corruption risk can be reduced by applying anti-corruption measures. The usefulness of these anti-corruption measures is not always questioned. Therefore measures might be applied which have little or no value in reality. Compliance Management Systems might be enhanced if hints can be found which measures were helpful for the organizations. In a laboratory experiment possible cause-effect relations for anti-corruption measures were tested in an international setting. Populations from different countries were selected to test also differences in the national settings. The effectiveness of CMS depends on a careful selection of the right activities. Nevertheless elements could be identified which are likely to have a stronger impact on the decision making than others. These should be in the focus when designing, evaluating and/or implementing a CMS. The approach should be embedded in an overall risk analysis to make sure that the relevant compliance risks are covered.

Headnote

Abstract:

Purpose of the article: Organizations tend to apply Compliance Management System (CMS) as means to reduce risk. The fraud or corruption risk can be reduced by applying anti-corruption measures. The usefulness of these anti-corruption measures is not always questioned. Therefore measures might be applied which have little or no value in reality. Compliance Management Systems might be enhanced if hints can be found which measures are helpful for the organizations.

Methodology/methods: In a laboratory experiment possible cause-effect relations for anti-corruption measures were tested in an international setting. Populations from different countries were selected to test also differences in the national settings.

Scientific aim: The research aims at finding most effective anti corruption mechanisms and makes suggestions which kind of measures have a positive impact and which are negligible. This research presents the results of an experimental approach to further strengthen activity in the area of CMS. It highlights the use of CMS in order to conduct business through persons which are ethically sensitized.

Findings: The research highlights that not all applied measures do have a positive impact on decisions winch have to be taken in ethical dilemma situations. Furthermore for single areas it became obvious that only specific measures do have an impact, general measures do not have an impact.

Conclusions: The effectiveness of CMS depends on a careful selection of the right activities. Nevertheless elements could be identified winch are likely to have a stronger impact on the decision making than others. These should be in the focus when designing, evaluating and/or implementing a CMS. The approach should be embedded in an overall risk analysis to make sure that the relevant compliance risks are covered.

Keywords: anti-corruption measures, Compliance Management Systems, laboratory experiment

JEL Classification: C91, D73

1. Introduction

The transferred Milton Friedman logic of "the business of business is business" is no longer paramount. What Friedman defined as social responsibility was narrowly paraphrased as he argued the social responsibility of business is to increase its profits (Friedman, 1970). In the 21st century this perspective has changed. It is assumed that business must be done in responsible ways. The social responsibility of business is therefore ethical business conduct with the purpose of profit maximization - which is not a contradiction. Compliance Management enhances the business conduct and supports accordance with laws and regulations. Additionally a constant increase in the request for transparency can be noted: "Around the world ethical expectations, if not behaviors, are leveling up." (Hamel, 2012, p. 5) The research elaborates on how business can be managed sustainably by applying CMS. It is worthwhile to discuss tlie question if CMS matter or not. In an survey about ethics at work winch was conducted in continental Europe with approx. 3.000 participants it was revealed that honesty is practiced by a majority of 77% in daily routines (Basran, 2012, p. 4). However 22% percent of the respondents felt pressured at work to compromise on ethical standards. Tins pressure might lead to unethical decisions winch could have a negative impact for the company in the long run regarding for example fines and the image. A functioning CMS could possibly reduce this percentage. The reduction of the percentage would decrease the risk of image damages consequently. Corruption has been researched using principal-agent theory (Jacquemet, 2005). In his research studies, Jacquemet focuses on the delegation aspect winch means that the agent has a conflicting interest when he is paid fairly by the principal. For this research this effect can be neglected, because monetary incentives are not given to the participants.

Fadzil, Haron and Jantan also focus on principal-agent theory in thier theory (Fadzil, Haron, Jantan, 2005). The research question winch leads the research work postulates that specific CMS activities enhance the conduct of busmess and have therefore a positive impact for the business. Palazzo et al. use another scheme to explain corruptive behavior. According to them corruption takes place in a morale vacuum winch is called the ethical blind spot. This means that individuals who would not engage in illegal transactions normally can commit crimes if the circumstances lead to a situation hi winch morale standards are not applied anymore. This ethical blindness is created by proximal context in interaction with the sense making which is done by individuals and organizations (Palazzo, 2012, p. 324).

The object of this research is the study of certain CMS activities and their application to decrease corruptive behavior. The aim lies in an evaluation of the theoretically available means of a CMS regarding their pertinence, test possible combinations of activities and judge which activities do have an impact and should therefore be followed. Challenges arise from the fact that the model was tested in a laboratory experiment. The generalization of results must be done with deliberate restrictions and general deductions for parts of the work should be done cautiously. Constraints such as the practical handling of the test population entailed a laboratory experiment with a relative small sample group.

2. Methodology

The question of the value add of CMS was tested in a laboratory experiment winch was developed incorporating CMS activities winch are applied in practice and could be applied theoretically. In order to test the theory of value add of specific anti-corruption measures a controlled enviromnent is necessary to separate variables. The laboratory experiment was chosen as research method as the cause-effect relationship can be studied with a high degree of control. Furthermore there is a chance to advance research in this particular field. Other researchers see experiments as an important instrument in this regard as "(...) laboratory experiments could become one of the most effective tools to study corruption." (Armantier, O., Boly, A., 2008, p. 21)

The experiment was not used to gain new data. It was used to test whether assumed cause-effect relationships are valid. In order to reduce general distortion to a minimum, tire experiment was set in the context of a business game. The participants were told that they are about to participate in a business game experiment, but they did not know the intention of the game. The business game represents a context which allows for empirical variables that can be tested.

The hypothesis underlying the experiment states that specific CMS activities enhance the conduct of business and have therefore a positive impact for the business.

In order to test the hypothesis the model underlying is based on the assumption that the return on Compliance depends on the effect applied Compliance measures have on the decision making of individuals.

The model was used to research the behavior in contexts where Compliance measures had been applied and in contexts where no Compliance measures were applied. As shown in Figure 1, the results of the decision making process were gained through questions (possible answers include "Certainly", "Almost certainly" etc) with a treatment group and were compared with the results of the decision making progress by a control group who also participated in the experiment. If there is a difference between the two groups in the decision making process and the decisions reached it can be concluded that certain Compliance measures do have an impact. The theoretical conception implies that human beings act better if they have been sensitized ethically For the 5 variables items were used to test the underlying hypotheses. These items were deducted from a list of 30 elements for Compliance activities which had been clustered into three groups: Institutional elements, strategic elements and operational elements. The items were tested in questionnaires which were given to the probands. As it is not possible to attract professional business managers for a laboratory experiment all at the same time, it was decided to conduct the experiment with students of management science. The set-up is therefore a simulative environment of entrepreneurship. The experiment with a simulation of a business environment is also a learning environment for future business manager. Therefore it can have an impact on their behavior as the intention of the whole procedure was also explained to the participants after the experiment was finished. It must be noted that students of management in the future will be working as business managers and they might be put into business dilemma situations which they faced in the experiment. The experiments can be judged as preparation and sensitization for the real fife scenarios which await them after university. The results can be generalized as the decisions taken in the role as business development managers did not include a very specialized role. The experiment was set in a class of students attaining Bachelor and Master degrees at two universities. These are the University of Latvia in Riga and the University of Applied Sciences in Fulda. The profile description of the persons who participated in the experiment are classes of BSc and MSc students, both male and female, internationally mixed. The age is in the beginning 20s mainly. The logic of sampling required the choice of a sample of students to participate in the experiment. The sample size was determined by scientific convention and group size of 29 students respectively 30 students for the first experiment and 29 and 40 students for the second experiment. In total 58 persons in the treatment group were tested with 70 persons in the control group. Two classes of students were separated into groups. In the first part of the experiment 59 persons were tested. The treatment group consisted of 29 persons, the control group of 30 persons. In the second part of the experiment 69 persons were tested. The treatment group consisted of 29 persons and the control group of 40 persons. Overall the population for the laboratory test consisted of 128 probands, whereas 58 had a treatment and 70 did not receive a treatment. The separation criteria for the laboratory experiment were a class with information on compliance management activities and a class without information on compliance management activities. A group of individuals was separated into two subgroups at random. The behavior of the experimental group with treatment is compared with the control group without experimental treatment. A significant difference in the mean between these two groups leads to the conclusions of a causal influence of the independent variables in the experimental conditions.

The variables which were defined were the variables

* Control,

* Information,

* Instruction,

* Reporting and

* Sanction.

The variables were used in questions winch were presented to the probands. The answers were analyzed using SPSS software. The phenomenon compliance is complex and characterized by many influences. Compliance in this context depicts measures against corruptive practices. A commonly accepted methodology to test theories has not been developed yet. Nevertheless it is the postulation from business that the applied measures have to be researched in terms of effectiveness: "More research on ways to measure effectiveness of instruments and practices to prevent, detect and investigate corruption is thus recommended. (...) it is difficult to draw clear-cut conclusions on the effectiveness of anti-corruption measures." (PWC, 2013, p. 10)

The data gathered in experiment sessions shows the answering behavior of test persons and was used to test the hypotheses of different variables. These variables are used in the sub-hypotheses which stated that if specific compliance activities in regard to the variables were performed than the tendency towards corruptive behavior would decrease. The variables Information, Control, Instruction, Sanction and Reporting were tested in 4 items each. The variable Information was tested in 5 items. Additionally the variables Information and Sanction were also introduced to the treatment group with general notes at the beginning of 3 different rounds. Totally, 30 items were tested in specific scenarios requesting decisions from the participants. Furthermore dummy items were introduced to distract from the real aim of the announced business game.

Table 1 shows the independent variables and the description for the experiment.

In total three questionnaires were handed out to the participants. The first questionnaire covers 24% percent of the variables (excluding the variable Instruction), the second and third questionnaire each cover 38% of the variables. The reason for a smaller coverage in the first questionnaire results from the necessity to use the first questionnaire also to convey general information and to make the participants familiar with the set up. Therefore a smaller population of scenarios was chosen.

3. Results

A non-paramateric test procedure was chosen as the ordinal scale and the not proven normal distribution did not allow for parametric test procedures. For the Mann-Whitney test the values of the treatment group and the control group winch are compared are placed in one group. Following this each value is given a rank according to the position in the group. In the next step the sum of ranks is calculated for both groups separately. Based on the calculated sinus the hypothesis that both groups are the same is tested. The null hypothesis postulates that there is no difference between the two groups. If the null hypothesis is true the average values of the middle ranks of the two groups should be similar. This means that the value of the treatment group and the control group in the rank order should be distributed approximately equally.

3.1 Group analytics

For the rejection of the null hypothesis the significance level of 0,05 is chosen as scientific convention. A rejection of the null hypothesis on a significance level of 0,05 can be made for 9 items. These 9 items relate to 4 variables of the 5 variables tested. 2 out of the 4 variables showed a strong influence on the item. The results of the item analysis are presented as follows:

* Variable Information: Invitation to go skiing with competitor (0,009).

* Variable Control: Making a donation (0,004), Payment to "black listed" external partner (0,000), Payment to supplier (0,004).

* Variable Instruction: Facilitation Payment (0,024), Mergers and Acquisitions (0,000), Video Beamer (0,010).

* Variable Reporting: Reading News (0,034), Documentation of invitations (0,017).

* Variable Sanction: For no item the null hypothesis could be rejected.

For all other items relating to the hypotheses the group means can not be judged as significantly different. Therefore the conclusions drawn can not be stated without a risk of mistake. The specific scenarios are explained in the following.

Variable Information

It was tested if the variable Information had an influence on the decision of the test persons regarding ethical decisionmaking. People who received specific information regarding the dealing with competitors decided ethically better than those who did not receive such information. The treatment group was informed via a separate document which stated in the core principles to be careful when interacting with the competition. In the treatment group a middle rank with a value of 73,71 was reached, indicating that they would not engage in illegal activities with competitors contrary to 56,87 in the control group. The proposed activities included a weekend-trip skiing paid for by the competition to get to know each other and get a common understanding of the market and the customers. The agreeing to accept an invitation could be explained by missing knowledge or deliberate acting against the core principles in the control group. Concerning actively engaging in illegal activities such as adapting the prices it can be concluded that a middle rank with an average value of 66,85 in the treatment group was reached indicating that they would choose rather not to engage in an activity compared to 62,55 of the control group. It can be concluded that there is a common understanding that agreements on prices are illegal. The difference of 4,34 between both groups could be explained by a CEO statement which was given to the treatment group. In the statement it was demanded to gain market share in a fair way from the treatment group. The control group had only the information go gain market share. No special CEO statement was given.

Considering the engagement of a dubious export agent the treatment group decided slightly better than the control group. A middle rank with an average value of 65,35 of the treatment group shows rather non engagement in business transactions compared to a middle rank of 63,79 in the control group. In the scenario of the invitation of top customers, the test persons were asked whether they would spend parts of the budget which is left over to invite externals for roughly 5.000 euro per person. As also public customers were invited this business conduct can be clearly marked as illegal behavior. The scenario was conceptually designed to put pressure on the participant as in the situation the superior who was a strong influence on the individual supported the idea of an invitation. Overall the tendency to invite these customers was rather high. With an average middle rank of 65,41 the willingness to invite these external guests is slightly higher ranked in the treatment group. In the control group an average value of 63,74 and a rank sum of 4462,0 indicate almost certain invitations. The last scenario requested a decision of whether to accept obvious lying or not. The treatment group was provided with the values of the company which include as one value honesty. The average value of 69,00 of the treatment group indicates a strong rejection of lying, but also the control group strongly objects lying with a middle rank with an average value of 59,80. Therefore the conclusion is that a common sensitization has taken place already.

Variable Control

For the variable Control one test person in the treatment group and another person in the control group did not answer all questions. Therefore the answers are concluded from 57 to 58 test persons in the treatment group and from 69 to 70 persons in the control group. It was tested if the variable Control had an influence on the decision of the test persons regarding ethical decision making. In the scenario "Making a donation" the treatment group was informed that spending activities are checked by the internal audit function to evaluate sponsoring activities and donations. Therefore unethical if not illegal spending activities are scrutinized by an audit. With this information an middle rank of 73,89 was reached concerning the question if the donation which a friend requested would be accepted. On the other hand an middle rank of only 55,69 was reached of the control group. Hence it follows that donations would rather be made than rejected on average in the control group which is not in scope of an audit which is executed by the audit department. In the treatment group the knowledge about an audit function which might check spending behavior leads to a higher rejection rate. In another scenario the two groups were confronted with an apparently legitimate payment request by an external partner. The treatment group knew that a specific control disapproved the payment to the external partner. With this information a strong rejection of the payment request was reached with a middle rank of 81,65. Compared to that the control group with no indication of disapproval was on average less sure about a rejection. A middle rank of 48,51 indicates a position with an inclination to rejection in evaluating whether a payment should be made or not. The announcement of a future audit to control the activities and scheduling of a concrete date for the treatment group had little influence on the decision behavior. The groups were asked about simplifying documentation in the sense of stopping the documentation of several decisions therefore decreasing transpar ency in their area of work. In the context of the increased controlling activity by the audit function, the treatment group shows a slightly higher tendency to keep the existing documentation with a middle rank of 67,34 while the control group remains slightly below in that scenario with a value of 62,14. The scenario which asked for a payment to a supplier differentiated between an existing control for the treatment group by the execution of the so called 4 eye principle and the non-existence of a control. The so called 4 eye principle is based on a segregation of duties concerning the persons responsible for the handling of transactions. It means that the person who transfers the money or makes the payment can not be the same person who oversees and judges if the payment is allowed. This separation of roles guarantees that only approved payments are executed. With a middle rank of 73,85 the treatment group shows a higher value than the control group with a middle rank 55,72. Both groups rather disapprove the payment, also a tendency towards indifference can be concluded.

Variable Instruction

For the variable Instruction one test person in the treatment group and another person in the control group did not answer all questions for all items. Therefore the answers are concluded from 57 to 58 test persons in the treatment group and from 69 to 70 persons in the control group. For the scenario to employ a business consultant, the mere existence of an instruction has an impact. In the treatment group a middle rank of 66,18 indicates that the business consultant will rather not be employed. The control group which was unaware of the instructions gained a middle rank of about 61,28. This average is still rather rejecting, but slightly below the treatment group. The decision to borrow a company asset for private reasons was rejected by both the treatment and the control group. The existence of an instruction led to a middle rank of 72,03 in the treatment group which would reject, the non-existence of an instruction in the control group led to a middle rank 56,46. In a scenario concerning new cooperations and partnerships in South America a Mergers and Acquisitions (M&A) situation was depicted. A few proposals need an evaluation. In the treatment group an instruction helped the test persons to estimate the risk level. While in the control group a middle rank of 43,04 indicates a high probability to engage in the deal in the treatment group the existence of an instruction led to the discernment of unforeseen risk which resulted in a rather negative response. With a middle rank of 90,36 concerning a rejection of the M&A deal, a participation is unlikely. In a situation of interaction with authorities at the airport, the threat of contractual fines (up to 2 Mio. US) made both the treatment and control group accept the payment of a facilitation fee which can be illegal. While in the control group an average of rejecting the request of such a payment with a middle rank of 57,03 is low, it is interestingly that the treatment group rejects the payment with a middle rank of 73,52. Even the written hint "The company instructions forbid such kinds of payments." does not have a bigger influence. It is therefore concluded that business needs combined with contractual threats do override company given instructions.

Variable Sanction

For the variable Sanction one test person in the treatment group did not answer all questions for all items. Therefore the answers are concluded from 57 to 58 test persons in the treatment group and 70 persons in the control group. The first item testing of the variable Sanction asks the participants about their estimations in regards to employee behavior according to the internal rules. In the treatment group it is pointed out that behavior which is not in accordance with the company rales will at least be sanctioned with an informal warning. In the treatment group the majority of participants guesses that employees will not breach the rales if sanctions are known to them. For the treatment group a middle rank of 67,06 is reached. The control group shows a middle rank of 62,38 for the statement that employees will not breach the rules. It can therefore be concluded that the deterring effect of sanctions is limited regarding to the obedience to formal rules. Classical misconduct as theft of company assets is neither tolerated by the control group nor the treatment group. A specific treatment for that item was not chosen as it can be characterized as common knowledge. The only influence from the core principles given out as additional information in the first round is the statement "... we honor the law". In the treatment group a middle rank of 62,07 was reached indicating that the dismissal of the employee as a sanction would be the light action. For the control group a middle rank of 66,51 was reached indicating the same conclusion even slightly above the value of the treatment group. The situation of the notification that an employee is giving business to a friend breaching company rules yielded surprising results. The treatment group indicates with a middle rank of 65,39 that the employee should rather not be kept in the workforce. The control group comes to the same conclusion, with a middle rank of 62,86. It appears that although the treatment group has been given a clear statement on giving advantages to individuals ("No decision must be made on personal interest but only for the best of the company.") there has been no effect on the treatment group. The control group which did not receive any information on the topic at all scored almost identical. The situation of passive bribery of a purchase manager who accepted a gift was seen by the treatment group negatively with a middle rank of 69,42, whereas the control group was headed in the same direction with a middle rank of 60,42. It must be noted that a direct treatment was only given for the first item. The other items were modified by general information which informed about sanctions, but they were not linked to specific scenarios. It is concluded that sanctions which are directly linked to actions do have a deterring effect, but the activities per se are rather weak instruments to stir etlucal right behavior.

Variable Reporting

For the variable Reporting one test person in the treatment group and another person in the control group did not answer all questions for all items. Therefore the answers are concluded from 57 to 58 test persons in the treatment group and from 69 to 70 persons in the control group.

The report on company rules violation is mentioned for the treatment group. Although the treatment group is informed about the existence of such a reporting it does not have an impact on their acting which would lead to ethically better acting. The treatment group has a middle rank of 55,69 rejecting circumventing internal rules. That means that a big population does consider breaking the rules knowing that it will be reported about them. On the contrary, the test persons who did not read the report (the control group) have an ethically better acting with a middle rank of 70,14 rejecting circumventing internal rules. Apparently a reporting about etlucal wrong behavior in general does not lead to a change in decision making. Furthermore, it is concluded that the fact that the reporting states a constant number of persons breaking the rules, the test persons can use the reporting for the justification of their acting. The fact that personal reporting requirements are created does lead to a little difference in acting. Together with the announcement that the reporting will be shown to the head of the department, a middle rank of 72,16 in the treatment group is reached concerning rejecting the question if customers are invited for the 3rd time in a month. It can be concluded from the answer pattern that the majority would still invite customers even if a reporting requirement is given. In the control group were such requirement is not given the middle rank of 57,14 considering rejection indicates that there is no concern of misconduct for this group. In the scenario theft of company assets, reporting requirements do have a positive impact on problem consciousness. In the treatment group which does have a monthly reporting a middle rank of 65,85 is reached rejecting the question that theft can not be controlled. In the control group tlus opinion is also gained, but to a lesser degree with a middle rank of 61,56. The situation of a business lunch which costs the test persons are asked to cover led to wide acceptance. In the treatment group a middle rank of 66.99 was reached indicating that with certainty or almost certainty the costs for the lunch would be covered by the test persons. The fact that the treatment group was informed that the costs had to be reported to their superior did apparently only have a minor impact. The control group scored similarly with a middle rank of 62,44 indicating that in the control group the costs would even be borne more readily. It is therefore concluded that the variable did not have a high influence.

3.2. Inter group analysis

Both in research and reality a tendency exists to either make the allegation that compliance is helpful for modern business conduct or to reject compliance in total. In reality, the opinion that in several countries or industries corruption is necessary and therefore compliance activities are in vain is also articulated. The country argument also encompasses the question whether specific cultures are more prone to corruption than others. While analyzing data from the laboratory experiment it was compared how the Latvian population stood in comparison to the German population. An ethnological research could enhance the research in tins regard as a cross border managing approach is perceived to be more complex than the managing of conduct with a homogeneous culture (Martin, Altizer, Barrera, 2014, p. 14). For theoretical purposes it is task of research to find out if certain activities do have an impact or not. Possible criticism could be that the conception about compliance activities might be perceived as trivial and logic. On the other hand critics allege that all compliance measures are in vain. Therefore the question winch is the most promising activity is worth to be discussed. A CMS does not guarantee from failure in challenging environments, but it is worth to look at areas for improvement as in such a way the risk can be reduced. The empirical part shows that the conception is not as trivial as it might be perceived by critics and that only in certain cases significant differences were made.

From the 5 independent variables not all could be attributed a high influence to stir ethical better behavior. In a further analysis the similarities between the German and Latvian population were researched. In the context of the variable Information for the Latvian population the provision of information material in general does not have an effect for changed decision making. For the German population general information material does have in one item an effect, namely the acceptance of an invitation by a competitor. The provision of general information material overall is limited for the German population, too. For the variable Control the analysis revealed for two out of four items the Latvian population shows a significant difference between the treatment and the control group. These two items are the "Payment to blacklisted external partner" with an asymptotic significance of 0,002 and the item "Payment to a supplier" with an asymptotic significance of 0,001. The German population showed also a significant difference between the control group and the treatment group for two out of four items. One of the items is the "Payment to blacklisted external partner" for winch an asymptotic significance of 0,000 is reached. For the other item "Making a donation" an asymptotic significance of 0,014 is reached. Apparently both groups disapprove the payment of blacklisted external partners which is positive and shows that the groups are sensitized. For further activities winch had the announcement of checking of sponsoring and donation activities has a stronger influence on the German than the Latvian population. On the other hand the Latvian population would not make a payment to a supplier which hasn't been double checked with another person. In the total population 3 items showed a significant difference. Concerning the variable Instruction for two out of four items the Latvian population shows a significant difference between the treatment and the control group. These two items are the "Videobeamer" with an asymptotic significance of 0,033 and the item "Mergers & Acquisitions" with an asymptotic significance of 0,000. The German population showed also a significant difference between the control group and the treatment group for two out of four items. One of the items is the "Mergers & Acquisitions" for winch an asymptotic significance of 0,000 is reached. For the other item "Facilitation payment" an asymptotic significance of 0,024 is reached. In the total population 3 items showed a significant difference. An 100% identical answer is reached concerning the engagement of an M&A transaction. It can be concluded that there is total agreement between both groups. Clear instructions about borrowing of company assets are consciously refused by one part of the population. A reason might be that the probands judge themselves superior to written instructions. On the other hand facilitation payments are more likely to be not to be made by the German population when treatment is given. For the variables Sanction and Reporting no discrepancies were shown. As there is no difference regarding the variable Sanction between the two groups it can be concluded that general sanctions or mentioning of activities does not have an influence on the groups. Only specific activities might have an influence. For the variable Reporting neither the Latvian nor the German population does show strong differences between treatment and control group. In fact the overall significance levels are the weakest with a value of 0,034 for "Reading News" and 0,017 for "Documentations of invitations".

Overall, the hypothesis HO could be rejected for 4 of 5 variables, as a significant difference between the treatment group and the control group was calculated m specific items. For one variable the significance level does not allow for rejection of the hypothesis HO. From the item analysis it can be shown that not all items led to this conclusion. A few of the items apparently do not have a significant influence. It appears that given information must be very precise and concrete otherwise it does not have the desired effect. General information is not helpful. Sanction mechanisms obviously play a minor role in improving ethical behavior. The reporting schemes or requirements do also not have a major impact on the sensitization of ethical acting. Practical conclusions are that CMS do have a positive influence generally. The highest influence can be attributed to activities winch are based on Information, Controls, Instructions and Reporting. However these activities must be designed specifically to individual needs.

4. Discussion/Conclusions

From the 5 individual variables not all could be attributed a high influence to stir ethical better behavior.

For that reason as general advice it can be stated these activities could be preferred in a holistic approach. The approach should be embedded in an overall risk analysis to make sure that the relevant compliance risks are covered. Furthermore the aspect of control should be evaluated. If control is meant to be decreased in the future and also work places become more virtual, organizations have to rely on employees with a sound etlucal understanding who propel the business. This change must be lived by the managing function: "If practitioners prefer to remain as managers who "do things right", aim for stability and control, and thereby enact compliance roles, they are unlikely to change their organizational culture. However, when practitioners flunk of themselves as leaders who "do the light thing", then it is possible for them to engender change in their respective organizations." (Sison, 2010, p. 333). Control activities must therefore be tailored to the specific risk in order to decrease corruptive tendencies.

Confounding variables could be created by a person during the experiment who interferes with the decision process. In an office environment using a field experiment approach this is likely to happen, but in a laboratory test tliis situation was avoided and did not happen. Another factor might be the influence which is given by copying answers from another test person. Tliis influence is deemed little as the test persons are told that they are not competing for the right answer. Furthermore the test persons are under observation by the experimenter. The attitude of students might be a confounding variable as they might judge the experiment as artificial. Tlus is a limitation which is inherent to laboratory experiments. Furthermore another limitation might be the age as confounding variable as the students do not have the same level of maturity, experience and personal history as managers. These possible influences are accepted as the general reduction from the complex problem corruption to a model is decided on purpose to advance research knowing that only a small step is possible within the context of tliis work.

Additionally it must be noted that the task to establish an organizational framework which allows for compliance has to be set by the management (Pies, 2008, p. 45) as they present the organizational leaders (Thoms, 2008, p. 421). They are foremost challenged to implement a culture which combines high etlucal standards with high business performance.

In the long run, the pressure on employees will rather augment than diminish. Partly the reason for tliis is a striving for profit maximization, sometimes enforced by the shareholder-value-concept. A pure focus on profit and satisfaction of monetary requirements imposed by shareholders will turn out to be risky in the long term.

If management wants to improve their etlucal standing than it must take the time to flunk about its position: "Without adequate direction and well thought-out plans based on existing research findings, it is difficult for those responsible to manage and enhance the etlucal behaviour of employees." (McDonald, G. M., Zepp, R. A. (2007), p. 55). The deployment of a CMS might lead to new alternatives which were not even thought about before. It is not the focus of a CMS to forbid all activities, but to govern the compliance program for the organization in order to ensure behavior in accordance with the law and to promote etlucal business. For tliis reason activities must carefully be selected for a tailored approach. CMS are scalable. The choice of the scope and the scale of a CMS depends on a variety of factors. A holistic CMS also has the task to find out what motivates employees to act illegally or non-comphant. It must analyze the company, its reward schemes and check whether it is appropriate to incentive the employees to act in a legal framework and in the moral standards which are set. Furthermore it is also necessary to look at company values to align the CMS with them: "Values are the key to understanding how to shape employee rule-oriented behavior and are particularly important in motivating employees to voluntarily adopt company values as their own." (Tyler, Dienhart, Thomas, 2008, p. 34) The challenge of a holistic compliance management lies in its reduction to address necessary operational and strategic requirements from the business. A scope which is too narrow can be limiting, a scope which is too wide leads to a waste of resources. It can not be expected that a CMS covers all possible risks an organization encounters. In order to address future risks CMS need innovative developments (Wieland, 2010). Working on compliance management in this way, it can become a competitive advantage. Compames understanding compliance management holistically chooses CMS elements carefully. The sudden end of corruptive behavior is an illusion. But due to media coverage and information access via the internet it is likely that more cases in the future than nowadays will become transparent. The reason is an increase in awareness and increased transparency. As reflection of society businesses with all business functions should acknowledge tins development and incorporate a holistic compliance management in then activities to proactively prevent cases of corporate misdeed.

References

References

Armantier, O., Boly, A. (2008). Can corruption be studied in the lab? Comparing a field and a lab experiment Scientific Series, pp. 1-29. Retrieved from: http://papers. ssrn.com/sol3/papers.cfm?abstract_id=1324120.

Basran, S. (2012). Employee Views of Ethics at Work: 2012 Continental Europe Survey. Institute of Business Ethics, pp. 1-42. Retrieved from: http://www.ibe.org.uk/ userfiles/euethicsatwork2012.pdf.

Fadzil, F. H., Harón. H., Jantan, M. (2005). Internal auditing practices and internal control system. Managerial Auditing Journal, 20(8), pp. 844-866.

Friedman, M. (1970). The Social Responsibility of Business is to Increase its Profits. The New York Tunes Magazine, September 13. Retrieved from: http://www. colorado.edu/studentgroups/libertarians/issues/friedmansoc-resp-bu sin ess .html.

Hamel, G. (2012). What matters now: How to win in a world of relentless change, ferocious competition and unstoppable innovation. San-Francisco: Jossey-Bass.

Jacquemet, N. (2005). Corruption as Betrayal: Experimental Evidence on Corruption Under Delegation. GATE Groupe d Analyse et de Théorie Economique Working Papers, 05-06, pp. 1-29. Retriev'd from: http:// ideas.repec .org/p/gat/wpaper/05 06 .html.

Martin, A., Altizer, L., Barrera, V. G. (2014). Understanding Compliance Risk Across Emerging Markets. Journal of Business Compliance, 1(14), pp. 12-22.

McDonald, G. M., Zepp, R. A. (2007). Business Ethics: Practical Proposals. Journal of Management Development, 8(1), pp. 55-66.

Palazzo, G, Krings, F., Hoffrage, U. (2012). Ethical blindness, Journal of Business Ethics, 109, pp. 323-328.

Pies, I. (2008). Markt und Organisation. Programmatische Überlegungen zur Wirtschafts- und Unternehmensethik. In: W. Kersting (Ed.), Moral und Kapital Grundfragen der Wirtschafts- und Unternehmensethik, Paderborn, Mentis, pp.27-58.

Price Waterhouse Coopers (2013). Public procurement, costs we pay for corruption. Utrecht, DC: Author.

Sison, M. (2010). Recasting public relations roles: agents of compliance, control or conscience. Journal of Communication Management, 14(4), pp. 319-336.

Thoms, J. C. (2008). Ethical Integrity in Leadership and Organizational Moral Culture. Leadership, 4(4), pp. 419-442.

Tyler, T, Dienhart, J., Thomas, T. (2008). The Ethical Commitment to Compliance: Building Value-Based Cultures. California Management Review, 50(2), pp. 31-51.

Wieland, J. (2010). Compliance Management als Corporate Governance - konzeptionelle Grundlagen und Eifolgsfaktoren. In: J. Wieland, R. Steinmeyer, S. Grüninger (Eds.), Handbuch Compliance Management, Berlin: Erich-Schmidt Verlag, pp. 15-38.