The effect of media on tax compliance: Hypothetical scenarios study
Introduction
It is well known that people do not always behave as predicted by the rational choice theory, this applies especially in the case of tax compliance. Tax compliance in general is higher than the classic economic model predicts. Behavioural tax compliance models are useful in explaining tax behaviour and improving predictions of classic economics models. This study aims to analyse a media effect, especially the news in the newspaper about taxes on tax compliance decisions of people. In order to analyse the effect of different news on peoples behaviour, five different scenarios are randomly presented to a total of 820 students. In the first section, the basic model of tax evasion is explained briefly and behavioural models and some nudges are discussed. The second section presents the methodology and the model of the study. Results are discussed in the final section.
Tax Compliance Models
Tax evasion is a risky and illegal activity. In economic terms, taking risks in a legitimate activity is not very different from taking risks in an illegal activity. Therefore, it seems reasonable to employ methods used in other fields of public economics to tax evasion. So, the standard economic model of tax evasion uses expected utility maximisation.
Expected utility models assume a particular form of deterrence theory, in which the probability of audit and the severity' of fine act as evasion control policy. Different alternatives are assessed by considering the likely consequences or outcomes. The utility of each outcome is evaluated and probabilities are attached to uncertain outcomes, leading to the expected utility of the alternatives. The individual chooses the alternative with the highest expected utility.
Allingham and Sandmo’s (1972) model is quite simple. They assume that behaviour is affected by factors such as the tax rate, the penalties for tax evasion, and the probability of detection. In each time period, an individual has to decide how much income to declare to the tax authority. They derived results about the influence of tax rates, penalties, and detection rates on tax evasion (Allingham & Sandmo, 1972).
The basic models consider the individual’s decision to evade as a simple gamble that is affected mainly by the probability of expected fine and the tax rate. However, most empirical studies point out that there are some other factors such as perception, attitudes and moral judgements, and demographic variables. Aim, McClelland, and Schulze (1992) state that in most countries the audit rate and the penalty for tax evasion are very small. Therefore, according to the basic economic model, everybody should evade taxes and the amount of tax evasion should be much more than it is.Then, the question that researchers should try to answer is not “Why do people evade taxes?” but “Why do people pay taxes?” Dean, Keenan, and Kenney (1980) were highly critical of the way economic theory tries to explain tax evasion.They argued that the assumptions of analytical models “are naive and far-removed from reality” (Dean, Keenan, & Kenney, 1980, p. 29). The authors claimed that although there may be a relationship between the expected fine and evasion, criminological and social studies found more important factors influencing criminal acts such as moral beliefs and lack of peer involvement of an individual, etc. (Dean, Keenan, & Kenney, 1980). Frey (1992, 1997) indicates that a higher enforcement by tax authorities may actually result in lower tax compliance by individuals, when the increased enforcement crowds out the intrinsic motivation that causes people to pay their taxes. Dean, Keenan, and Kenney (1980, p. 29) indicated that these types of studies “have long abandoned the primitive sanctions/ deterrence model”. They claimed that “how taxpayers might act if they were condemned to being entirely rational, utility maximising automatons can only serve to postpone the emergence of realistic tax theories and useful policy insights” (Dean, Keenan, & Kenney, 1980, p. 44).
Some limitations of the basic economic models are that they ignore the roles of tax practitioners, cost of compliance, ongoing interaction between taxpayers and tax authority and different sources of income, the effect of media on tax behaviour, etc. Now there is considerable amount of empirical work on the subject of income tax evasion, which finds these and many other variables to be relevant to income tax evasion.The Internal Revenue Service (IRS) in the USA has detailed 64 potential compliance factors (Jackson & Milliron, 1986) ranging from “income” and “age” to “mental health”.
The criticisms of the expected utility model lead to development of an alternative model: prospect theory model.The theory was developed by Kahneman andTversky (1979) and it is reviewed and discussed by several authors such as Lattimore and Witte (1985), Jackson and Milliron (1986), Roth, Scholz, and Witte (1989), Tversky and Wakker (1995) amongst others. Prospect theory explores how decisions are made and it is developed in order to overcome some problems associated with expected standard utility theory.
Kahneman and Tversky (1979, 1984) suggested that there is a two-phase assessment process when an individual chooses between risky alternatives. First, the editing phase, and second, the evaluation phase (in which the edited prospects are evaluated, and the highest value prospect is chosen). In the editing phase, the outcomes are first coded, that is, valued as gains or losses relative to current asset level. When gains are involved, in general people avoid risks, while they take risks in order to avoid losses (Kahneman & Tversky, 1979,1984).
The relevance of prospect theory for tax compliance is indicated by authors such as Jackson and Milliron (1986), Loftus (1985), Spicer (1986), Smith and Kinsey (1987), Chang, Nichols, and Schultz (1987), Roth, Scholz, and Witte (1989), Carroll (1987, 1989, 1992),Webley et al. (1991), Elffers and Hessing (1997), Robben et al. (1990), and Yaniv (1999). These studies indicate that framing is very important in tax decisions. For example, an analysis of focus- group discussion indicates that taxes which have to be paid with tax returns at the end of the year, as well as tax amounts owed in general, especially those due on income from secondary sources, have greater utility than taxes that are withheld, especially from the primary salary (Ekstrand, 1980). Accordingly, the overwhelming majority of taxpayers in the USA (75%) prefer to have more withheld than is necessary (Smith & Kinsey, 1987) and there was a storm of protest in the USA against the 1986 tax reform because the act intended to lower advance tax payments for many, thus lowering their tax refunds (Elffers & Hessing, 1997). Loftus (1985) noted that withholdings lead to a shift of the reference point and higher withholdings thus decrease the motivation to reduce taxes at the time of filings both by legal and illegal means. However, as Carroll (1989) pointed out, increases in withholdings reduce tax evasion only if: (i) the reference point is zero taxes owed at filing time; and (ii) the reference point on withheld income sources also shifts to new reduced levels. A taxpayer who has paid /(5,000 in taxes and owes another /(200 at the filing time can take his/her reference point as a /(200 loss, a /(5,200 loss, or, say, /(500 gain if s/he paid a total of /(5,700 in taxes last year. The taxpayer could also take the reference point as what his/her colleagues, neighbours, or other people similar to him/her in some aspects (e.g., income sources, amount of income, spending level, etc.) have paid. Thus, the reference point can be almost anything, and it is arbitrary.
Carroll (1989) indicates that a discontinuity in the decision weight function at the reference point in prospect theory causes a difficulty to base policy on its existence, since the exact location of the discontinuity is unknown. Accordingly, if the probability' of audit has been doubled and taxpayers have been told about it, there may have been four different outcomes: (i) it might have no effect, if taxpayers consider the probability was too low anyway; (ii) it might have a small effect, if the probabilities were in the shallow part of the curve; (iii) it might have an important effect, if the probabilities were in the steep part of the curve; (iv) it might have a very significant effect, if the increase in probabilities resulted in a shift in decision weight from certainty of no risk to uncertainty of risk. The author noted that an annual 1% probability of getting caught at tax evading can be thought of as a 50% chance when continued over a working lifetime of 37 years. If a taxpayer coded the probability of audit as in the second case, it would produce a much greater deterrence effect. Moreover, a taxpayer who perceives 2% probability of an audit and 50% probability of being fined (if audited), tends to weight the second stage probability more strongly than the first stage (Carroll, 1989).This indicates that decisions depending separately on the probability of audit and the probability of fine can lead to more compliance than would be expected from the joint probability of being fined. Furthermore, it implies that more deterrence effect could be obtained from the same policy by advertising a highly probable conditional probability (fine rate) rather than a low probability event (audit rate).The basic effect was named as the pseudo certainty effect byTversky and Kahneman (1981).Johnson and Payne (1985) gave another example to demonstrate this: (i) for every event of serious tax evasion, 1 in 500 evaders is sent to jail to spend some time; (ii) over the average life time of paying taxes (50 years), evaders who continue to evade taxes seriously faces a 1 in 10 chance of being detected and convicted, and when convicted s/he is certain to spend time in jail. Of course, the second way of explaining the outcomes would have a much higher deterrence effect.
If a taxpayer frames taxes as government waste, a loss to the individual and to the society as a whole, then the taxpayer may have strategies for identifying loopholes and the various ways to evade. On the other hand, if the taxpayer frames taxes as contribution and sharing, s/he may have strategies for identifying legal obligations and fulfilling them. Although prospect theory indicates the importance of editing and framing, there is not a theory about how the editing processes work exactly (Carroll, 1989).
Elffers and Hessing (1997) suggested that the incentive for tax evasion could be eliminated by the tax authority’s deliberately setting taxpayers’ advance tax payments slightly higher than their true tax liabilities. However, using prospect theory and assuming the reference point is income after the payment of tax advance (as it was assumed by Elffers and Hessing, 1997),Yaniv (1999) showed that although increases in advance tax payments would encourage more compliance, they would not completely eliminate the incentive for tax evasion. This would only be possible if taxpayers significantly overestimate the low probability of audit, which is prevalent in many countries. Empirical findings by Chang, Nichols, and Schultz (1987), Carroll (1992), and Varma and Doob (1998) also indicated that taxpayers are mainly concerned about out-of-pocket gains and losses at the time of filing.
Aim and Beck (1990) pointed out that an individual is unlikely to participate in an amnesty if his/her reference point is the initial level of previously unreported income.Tax evasion for this individual is a norm and paying taxes in the amnesty is a loss. The individual in this case is a risk-seeker rather than a risk-averter. However, if the individual takes the reference point as “the level of amnesty income less full amnesty taxes”, then s/he will be much more likely to declare any unpaid taxes from previous years. In this case, the individual regards paying taxes as a norm and considers participating to the amnesty as a gain. So, it is very important to change the individual’s reference point by indicating that paying taxes is the normal and accepted form of behaviour (e.g., indicating tax evasion is a very serious crime, honest taxpayers are victimised by such behaviour and stress that tax authorities intend to catch and punish tax evaders).Thus, an amnesty increases compliance only if the amnesty makes the individual recognise that paying taxes is the norm.
Tversky and Kahneman (1973) suggested that events are judged to be more probable if they come more readily to mind. In the literature this is named as “availability'” heuristic. Availability' identifies the subjective sense of the probability of an event’s occurrence. Thus, easily imaginable or recallable events are in general thought to have high likelihood of occurrence. People depend heavily on salient information, which is easily retrievable from memory. Spicer and Hero (1985), Benjamini and Maital (1985), and Webley (1987) found that tax compliance was higher among individuals who had been previously audited. Moreover, information about friends and family members of being audited should affect the perceived risk of evasion positively. Accordingly, recent news on the media about a tax evasion case might temporarily increase the perceived probability' of punishment, especially when there are similarities between the taxpayers and the convicted evaders (such as occupation, income, living in the same city', etc.). Carroll (1989) noted that “the IRS seems to save some juicy fraud convictions for late March to take advantage of the availability' heuristics just before taxes are due” (Carroll, 1989, p. 244). However, it is also possible that the news of a convicted evader may have the opposite effect; the taxpayer might think that the IRS is busy with more important tax frauds, which involve large amounts, and therefore an average evader might feel safer. The taxpayer may also consider the audit resources are fixed, so every known case of evasion decreases his/her risk of audit, or else the taxpayer might be concerned about the probabilities, in which situation every known case of tax evasion increases the perceived risk. People usually interpret information in terms of their prior understanding to support their views. It might also be possible that an audited person might reduce the perceived audit rate, thinking the low rate of occurrence of an audit has just happened for him/her, and it will take at least a while for another to happen.
Scholz and Pinney (1995) noted that in real life taxpayers do not have precise information about the probability of audit and the fine for evasion. They indicate that duty heuristics provide both a direct and indirect effect to tax compliance. According to this, taxpayers subconsciously observe the performance of the government relative to the amount of taxes they pay'. The taxpayers summarise their observations in a judgement, which authors call “tax duty”, and take decisions about compliance depending on this core of previous experience. So, apart from the direct effect of tax duty' in providing a motivation to comply, it also indirectly biases self-interest beliefs in the same direction.Thus, the duty heuristics inadvertently reduce the conflict between self-interest and collective rationality by diminishing the conflict between duty and fear for compliance decisions. Accordingly, taxpayers’ sense of duty' to pay taxes significantly affects the perceived probability' of the expected fine for tax evasion. Taxpayers with duty heuristics will therefore perceive the expected fine for evasion as high, whereas taxpayers with little duty will perceive a low risk of expected fine.
These hypotheses are confirmed in the empirical studies carried out by Scholz and Pinney (1995) and Scholz and Lubell (1998a, 1998b). Moreover, they found that duty heuristics have more effect on perceived expected fine than the objective probabilities of risk. This was the case even for taxpayers who are facing the greatest temptation to evade because of the high potential gains from tax evasion.
Scholz and Lubell (1998a, 1998b) note the importance of trust as a heuristic in tax compliance. According to this, when trustworthy democratic institutions succeed in issuing socially beneficial laws that are willingly followed by trusting citizens, the trusting relationship between citizens and state benefits society in two ways: (i) trust between citizen and government (vertical trust) can expand the range of collective problems that legal authorities are able to tackle. Trusted government institutions will be able to extend the benefits of social cooperation to collective problems, which are too costly to resolve in the absence of trust. In order to achieve this, legal authorities use law to define “focal points” for cooperative solutions; (ii) trust among citizens (horizontal trust) diminishes the enforcement cost of maintaining collective solutions. The higher the trust, the less the need for monitoring and punishment mechanisms, which are costly. So, trust increases social capital by decreasing the enforcement cost involved in maintaining collective solutions.
In the absence of trust, the potential benefits of collective action rely either on altruism or on expected punishments. The effect of trust in large-scale collectives, however, will not necessarily provide compliance, unlike small-scale collectives. A club member who avoids his/her obligations is very likely to be found out and to lose the benefits of membership. On the other hand, since the cost of monitoring and punishment are much higher in large-scale collectives, a taxpayer who evades taxes is less likely to be found out or lose benefits even if caught. Nevertheless, Scholz and Lubell (1998a, 1998b) point out that the relationship between trust and trustworthiness can be developed in large-scale collectives, even though there is no credible institutional foundation. Authors indicate that individuals learn different heuristics for assessing trust in small- scale collectives, in which the relationship between trust and trustworthiness is strengthened through potential punishments.
Positive experiences of obtaining benefits from a collective increase in trust, and so increase the probability of complying with obligations to that collective, while negative experiences decrease both trust and compliance. Therefore “attitudes of trust serve as an action related summary measure of the benefits of a given collective that enables citizens to cooperate in the most beneficial collectives while avoiding exploitation in less beneficial ones” (Scholz & Lubell, 1998a, p. 401). It is possible that taxpayers generalise this heuristic, use of trust, to large-scale societies. In this case, there will be a positive relationship between trust and tax compliance, even though the probability of potential punishments will be much lower in a large-scale collective. The empirical study by Scholz and Lubell (1998a, 1998b) indicated that trust changes according to changes in net benefit (taxes paid minus public goods obtained). So, trust increases significantly when the amount of taxes paid decreases, and it reduces when taxes increase. Moreover, both vertical and horizontal trust positively and significantly affect the probability' of compliance. However, in contradiction to the compliance heuristics explained above, a study by Steenbergen, McGraw, and Scholz (1992) did not find a significant effect of tax changes on attitudes.
Kirchler (2007) explained the slippery slope framework of tax compliance. According to this, trust in tax authorities and power of tax authorities are two interacting dimensions that explain tax compliance. Kasper, Kogler, and Kirchler (2015) carried out a questionnaire-based experiment with 487 employees which found that media coverage affected both indicated trust in tax authorities and the perceived power of governmental institutions.Therefore, media coverage has an important effect on intended tax compliance.
Nevertheless, media news can be used as a nudge for the people to behave as they are expected to behave. This chapter investigates how newspaper affects the intended decision to declare taxes.
