REGULATION AND DISHONEST BEHAVIOR: HOW CONTROLLING REGULATORY MECHANISMS PROMPT DISHONESTY IN TAX CONSULTING.

• Author: Gavious, Ilanit

1. INTRODUCTION 700 II. BACKGROUND 705 A. Transfer Pricing Regulations and Incentives 705 for Manipulation B. Literature on Dishonesty and Hypotheses 708 Development III. EXPERIMENTAL DESIGN 711 A. Procedure 711 B. Basic Payment Structure 713 C. Treatments 714 D. Definitions of the Variables 715 IV. RESULTS 717 A. Descriptive Statistics 717 B. Gender Effect 718 C. Audit Effect 719 D. Multivariate Analyses 721 V. SUMMARY AND CONCLUSION 723 Every man, as long as he does not violate the laws of justice, is left perfectly free to pursue his own interest in his own way, and to bring both his industry and capital into competition with those of any other man or order of men. Adam Smith An Inquiry into the Nature and Causes of the Wealth of Nations 1776 I. INTRODUCTION

   From Adam Smith's "Invisible Hand" to the fundamental theorems of welfare economics, the search for a stable equilibrium is usually conducted using a set of underlying assumptions, some of which are stated, while others are considered obvious. In his book, Adam Smith explicitly noted that one necessary requirement for the existence of trade is obeying the law. (1) Economic theory has developed based on the existence of this requirement. Obviously, not all market participants comply with the law. However, for a stable equilibrium to be maintained, the number of those who do not comply must be too small to impair the effectiveness of the economy. Various examples in economic history indicate that even a small group of people acting illegally or merely unethically can have a significant impact on people, institutions, and the market as a whole. Examples from just the recent past include the accounting fraud of Enron (among other firms), Bernie Madoff's Ponzi scheme, the dot-com bubble, and the sub-prime mortgage crisis.

   Dishonest behavior in business is usually prompted by a conflict of interests. Taxes are a classic example of such a conflict, where the desire to minimize one's own tax burden harms the interests of others. Allingham and Sandmo suggest that taxpayers avoid paying taxes the more profitable and less risky it is for them. (2) Their theoretical work is supported empirically (3) and experimentally. (4) Minimizing tax payments while complying with regulations is considered legitimate. However, when done illegitimately, tax avoidance by one party harms society, which benefits from the taxes collected.

   This study examines the behavior of tax consultants who have to determine whether their client is in compliance with regulations (and thus exempt from additional tax payments) or whether the client is not in compliance (and hence must pay additional taxes). We use the setting of transfer pricing (TP) consulting for multinational enterprises to study experimentally how the conflict of interest faced by tax consultants--between telling the truth and pleasing their clients--affects their behavior. Our analyses account for factors potentially affecting the consultant's decision-making process including gender, consultation fee, and the level of regulatory mechanisms controlling the consultants' work.

   Our experimental design simulates a situation where a consultant is required to provide TP documentation showing whether the client is in compliance with the regulations. The requirement for TP documentation was established by regulators in many countries worldwide as a means of preventing multinational enterprises from shifting profits from higher tax rate countries to lower tax rate countries to reduce their overall tax burden. When TP regulations became too complicated and time consuming, large multinational enterprises opted to outsource the preparation of the TP documentation to external independent consultants. While consultants are expected to work ethically and maintain their integrity, they also have an incentive to please their clients in order to ensure a long-term relationship with them. In some cases, there may exist a conflict of interests that might drive the consultant to behave dishonestly. This conflict is between providing the most objective, honest assessment of the tax owed, and trying to minimize the tax illicitly to please the client, notwithstanding the social costs of such tax evasion. Indeed, simulations with tax consultants show that their judgement is biased towards their clients' favor. They overweight supportive information, (5) are sensitive to framing effects, (6) and are attentive to their clients' preferences. (7)

   Moral and professional standards may be an important factor in preventing dishonesty. However, tax authorities also try to prevent dishonesty by conducting audits. An audit serves several purposes, one of which is to make dishonesty risky for consultants. Given the possibility of an audit, the potential benefit from cheating may be offset by the likelihood and the results of getting caught. As such, the possibility of being audited by the tax authorities serves as a controlling regulatory mechanism. As indicated, in the study we examine the effect of such monitoring on the consultants' behavior.

   Our results reveal dishonest behavior by most participants (80%), as captured by the inconsistency between their true opinion and the opinion they submitted to the tax authorities. Despite their initial choice that would preserve their professional integrity, the vast majority of participants opted to change that choice in order to ensure that the client was satisfied. A set of forensic-type tests conducted to identify the pattern of dishonest behavior also reveals that the possibility of being audited by the tax authorities significantly affects the decisionmaking process of participants when acting dishonestly. Specifically, given the possibility of being audited, the participants' output in the initial step of their decision-making process is of higher quality. However, once the participants realize that their initial choice results in an additional tax burden for the client, the possibility of an audit increases their efforts to change their initial choice toward a result that would please the client. As the risk of an audit increases (as captured by the probability of being audited), the participants make more frequent and opportunistic changes until they reach their desired end result. These findings hold regardless of the consultant's gender or fee. (8)

   Overall, our findings demonstrate the significant impact of the risk of being audited (and potentially caught) on the initial, but not on the end, result of the consultation, with the discrepancy between the starting and ending points increasing with the increase in the probability of being audited. This risk affects the decision-making process of the participants, consistently resulting in an outcome designed to please the client. Remarkably, as the audit risk increases, the participants become even more willing to go the "extra mile" for their client.

   Our findings suggest that while control-type regulatory approaches can have an immediate effect on behavior, evident in the risk of an audit improving the consulting quality at the beginning of the process, its effects may not last long, and the behaviors it seeks to enforce may not persist over time. Moreover, more regulation can actually increase, rather than decrease, the level of dishonesty among regulatees. These outcomes are supported by the psychological underpinnings of individuals' defiance of stringent external controls. Rather than inducing compliance, controlling systems often prompt individuals to do the opposite of what the regulation demands. (9)

   The main contribution of our study is in describing the decision-making process when engaging in dishonest behavior. Our work, being forensic, focuses on how manipulation is exercised, and how it is affected by the monitoring mechanism of an audit. Existing literature on dishonest behavior deals mostly with its motives and circumstances. (10) Moreover, the experimental design in studies of dishonesty is such that subjects generally make one decision that can either be honest or dishonest. We employ a different design where the subjects are allowed to revise and change their decisions as many times as they want throughout the experiment. In this unique design, subjects can start with dishonest behavior and correct it later on or, alternatively, start with an honest decision and "modify" it gradually. This approach allows us to study the dynamics of the decision-making process. It is reasonable to assume that regulation would affect this process, as we document.

   Lastly, we acknowledge that our participants--masters and undergraduate business students--are not professional consultants and, in our experiment, are serving an imaginary rather than a real client. Nevertheless, as an experimental study, it illustrates the behavior of people in general in the situation under examination. We believe that our experimental design incorporates the main incentives facing TP consultants. Therefore, our conclusion can shed light on the incentives that result in certain behaviors, including those of TP consultants. Our empirical evidence and behavioral insights are of relevance to policy setters, regulators, tax authorities, managers, and academics.

   The rest of the Article proceeds as follows. Part II provides background on transfer pricing, reviews the literature, and develops the hypotheses. Part III describes the experimental design and lists the variables used in the study. Part IV discusses the results, while Part V concludes.

2. BACKGROUND

   1. Transfer Pricing Regulations and Incentives for Manipulation

      Transfer pricing is an accounting practice that refers to the setting of prices for the goods or services exchanged among related entities such as subsidiaries, affiliates, or commonly controlled legal entities that are part of the same large enterprise. When related transactions are conducted internationally, transfer pricing may have significant tax...