TAX COMPLIANCE: SLIPPERY SLOPE FRAMEWORK AND THEORY OF REASONED ACTION APPROACH

Purpose — This study aims to examine and analyze factors that influence tax compliance using socio-psychological elements. Design/methodology/approach — Tax compliance in this study was examined using the slippery slope framework and theory of reasoned action approach. This study used the quantitative method. The data were harvested through a questionnaire survey and analyzed using structural equation modelling in SmartPLS version 2.0. Findings — This study finds that tax compliance is influenced by legitimacy and coercive power of the authority, trust in the authority, and the taxpayer's intention. Practical implications — Tax authorities should retain the power to maintain trust in the tax revenue that will improve tax compliance. Originality/value — This study develops the slippery slope framework and refines it using intention as the variable, as described by the Theory of Reasoned Action.


INTRODUCTION
Tax is the source of state income used to pay for government expenditures. However, most people avoid it (Gobena and Dijke 2015), indicating non-compliance with applicable tax regulations. (Saeroji 2019) stated that tax compliance in Indonesia, especially tax reporting, is still dominated by individual employee taxpayers. Therefore, continuous efforts are needed to improve compliance with tax, which helps finance the country's development. Indonesia, as a developing country, will always strive for development in various sectors. Therefore, the government has taken ways to optimize various types of state income. framework is weak in explaining such behaviour. It only sees that tax compliance behaviour is only influenced by the authorities' power and the taxpayer's trust in them (the interaction between the taxpayer and the authority). Therefore, in explaining tax compliance, it is insufficient to focus only on the interaction. Fishbein & Ajzen (1975) explained that a person's behaviour is formed from his underlying intention. It is the intention-based behaviour that is explained in the Theory of Reasoned Action. Theory of Reasoned Action explains that a person's behaviour is determined by his faith or belief in the consequences of his behaviour (Fishbein & Ajzen, 1975). This theory places a person's intention to do (or not to do) something as a determining factor for action. Changes in a person's intention will change the resulting behaviour (Ajzen & Fishbein, 1980). This study uses the Theory of Reasoned Action to see that tax authority's legitimacy power, coercive power, and taxpayer's trust in the tax authority foster taxpayer's intention.
Furthermore, the intention produces behaviours of complying with any applicable tax regulations. (See 2011) Furthermore, Alleyne and Harris (2017) found that intention affects tax compliance. Bidin et al. (2014)reported that intention has a positive effect on tax compliance. The opposite result was found by Oktaviani & Nurhayati (2015) that intention does not influence taxpayer's compliance.
This study uses taxpayers in the Small Taxpayers Office of Batu city. The Batu City Small Taxpayer Office was inaugurated in 2017 and sought to optimize tax revenue targets by conducting socialization related to annual SPT reporting. The realization only shows 56%, while the target is 90%. The phenomenon of low tax compliance can also be seen from the low 2018 Batu City tax revenue target. The tax target for 2018 is IDR 163 billion, and the tax revenue achieved in that year is only IDR 134 billion. In 2019, the tax target was set to increase to IDR 208 billion (Richa & Helmy, 2019). Setting a higher tax target without achieving the previous year's tax target is an obstacle for the Batu City Small Taxpayer Office to work harder to optimize tax revenue. In the tax collection system in Indonesia (self-assessment system), taxpayers must pay and report their taxes to fulfil applicable tax obligations. Therefore, non-employee individual taxpayers were chosen as research subjects because the tax compliance behaviour reflected in the fulfilment of their tax obligations came from within the taxpayer.
The novelty of this study is an in-depth study of tax compliance using the slippery slope framework, with refinement to overcome its weaknesses. The weakness is that the framework only sees tax compliance from the interaction between taxpayers and tax authorities.

Vol. 29, No. 2 August 2021 © Centre for Indonesian Accounting and Management Research Postgraduate Program, Brawijaya University
This research tries to deconstruct the slippery slope framework, as described in the Theory of Reasoned Action. The purpose of this study is to provide empirical evidence regarding (1) the influence of legitimacy power on trust in authority, (2) the influence of coercive power on trust in authority, (3) the effect of legitimacy power on taxpayer's intentions, (4) the influence of coercive power on taxpayer's intention, (5) the effect of trust in the authority on taxpayer's intention, (6) and the effect of the taxpayer's intention on tax compliance. This research also contributes to confirmation of theories, i.e. slippery slope framework and Theory of Reasoned Action, based on the phenomenon of low tax compliance level. The results of this study can also be used as a reference for optimizing the performance of the tax authority (Directorate General of Tax) in increasing tax compliance and to foster taxpayers' intention to fulfil their tax obligations, given that the role of tax in improving people's welfare is crucial.

HYPOTHESIS DEVELOPMENT The Relationship between Legitimacy Power and Trust in Authorities
Legitimacy power is built upon applicable tax law, the relevance of information presented by the authorities, and the expertise of tax authorities (Tjondro, Setiabudi, and Joyo 2019). Kirchler, Hoelzl, and Wahl (2008) and Kastlunger et al. (2013) proposed that taxpayer's trust increased along with the legitimacy power of tax authorities. In Indonesia, Djajanti (2018) concluded that authorities' power is not necessarily coercive; they must make taxpayers believe that they are professional law enforcers and that they are not people you can play around with. Hence, the following hypothesis was formulated. H1: Legitimacy power positively influences trust in authorities

The Relationship between Coercive Power and Trust in authorities
Coercive power is based on supervision, penalty and refined by the tax authority. Kastlunger et al. (2013) found the negative influence of coercive power on taxpayer's trust. The same thing was stated by Gangl, Hofmann, and Kirchler (2015), Hofmann et al. (2017), and Ratmono (2014) that taxpayers are compliant because they are afraid of audits, fines, and the power of the tax authority. Hence, the following hypothesis was formulated. H2: Coercive power negatively influences trust in authorities The Relationship between Legitimacy Power and Taxpayer's Intention The intention is a determining factor of one's behaviour. Fishbein & Ajzen (1975) stated that his belief produces a person's behaviour about his consequences. Researches regarding the relationship between legitimacy power and taxpayer's intention are still very limited in number. Batrancea et al. (2019) examined tax compliance based on taxpayer's trust and authority's power. They concluded that these two factors increase tax-compliant behaviour, which can reduce intentional tax avoidance. This shows that the higher the legitimacy power of the tax authority, the higher taxpayer's intention to comply with tax rules. Therefore, the following research hypothesis was formulated. H3: Legitimacy power positively influences taxpayer's intention

The Relationship between Coercive power and Taxpayer's Intention
When authorities have a strong command in conducting an examination and giving penalties and fines to taxpayers committing illegal taxation avoidance, in this condition, taxpayers will have a common intention to commit tax avoidance because they feel that they are continuously monitored and afraid of penalties or fines due to tax avoidance. Their intention is based on compulsion and fear. There is still no research that examines the effect of coercive power on taxpayer's intentions. Kirchler, Hoelzl, and Wahl (2008) stated that authority's power in tax audit and punishment creates forced compliance to taxpayers. This shows that, in its essence, taxpayers do not have the intention to comply with tax regulations. However, the coercive power of tax authority is a form of power to increase the taxpayer's intention in fulfilling his tax obligations. Hence, the following research hypothesis was formulated.

H4: Coercive power positively influences taxpayer's intention The Relationship between Trust in Authorities and Taxpayer's Intention
The highest voluntary compliance is achieved in conditions when taxpayers have high trust in their tax authority (Kogler et al., 2013). Research about the influence of trust on taxpayer's intentions has not been conducted before. Taxpayer's intention to comply must be increased by keeping their trust in the authority's performance. The higher taxpayers 'trust in the tax authority, the greater their intention to fulfil their tax obligations. Therefore, the following research hypothesis was formulated. Ajzen & Fishbein (1980) stated that one's actions are affected directly by their intention to do or not to do something. Theory of Reasoned Action explains that one's behaviour results from his intention, which is a function of a particular belief. Alleyne and Harris (2017) have

Vol. 29, No. 2 August 2021 © Centre for Indonesian Accounting and Management Research Postgraduate Program, Brawijaya University
provided empirical evidence that intention influences illegal tax avoidance behaviour. See (2011) and Bidin, Shamsudin, and Othman (2014) showed that intention positively affects one's compliant behaviour. Those studies show that the greater a person's intention to comply with tax rules, the higher his compliance. Based on the description above, the following research hypothesis was formulated. H6: Taxpayer's intention positively influences tax compliance This quantitative research uses a survey on non-employee personal taxpayers. They were selected as the unit of analysis because they are directly related to their attitude determination and taxation decisions, both for themselves and their business. They were selected through non-probability sampling, i.e. convenience sampling. The method was used due to the researcher's limitations in accessing data or taxpayers' personal information. The sample of this study is 183 taxpayers.

Data Collection
This study uses primary data from questionnaires distributed to non-employee personal taxpayers. The questionnaires were distributed online through the Google Form link and directly at the Small Taxpayers Office of Batu city and the taxpayer's business location.

Measurement a Legitimacy Power (LP)
Legitimacy power belongs to the tax authorities based on applicable tax regulations and their skills. The indicators of legitimacy power used as the basis for making the questionnaire's statement items refer to indicators used by Kastlunger et al. (2013), Gobena and Dijke (2015), and Tjondro, Setiabudi, and Joyo (2019); they are tax authorities' competence, professionalism, and effectiveness, the implementation of good standards and services, and the provision of helpful information for taxpayers.
b Coercive Power (CP) The coercive power of authorities is strict and compelling taxpayers to comply. The indicators of coercive power in this research refer to the studies of Kastlunger et al. (2013), Gobena and Dijke (2015), Djajanti (2018), and Tjondro, Setiabudi, and Joyo (2019); they are high amount of fine for tax offenders, severe punishment for tax offenders, severe punishment if caught by the tax authority, fostering hostile feelings, enforcement of authority's objective through audits, and the assertiveness of tax authorities.
c Trust in Authorities (TIA) Trust in authorities is the taxpayer's confidence that the tax authority has a good attitude and always works for the benefit of the citizens. The indicators of trust in authorities in this study refer to Gobena and Dijke (2015) and Djajanti (2018); they are tax authority's competence and consistency, tax services, transparency in tax usage, benefits of tax authority's policy, satisfaction from tax authority's problem solving, trust in tax official's decisions, respect in tax authority, knowledge of tax authority about taxpayer groups. d Taxpayer's Intention (TI) The intention is the tendency of taxpayers to carry out their tax obligations. Taxpayer's intention determines the behaviour towards tax rules, either to comply or to not comply. The dimensions of taxpayer's intention can be measured using three indicators, which refer to Theory of Reasoned Action (Fishbein & Ajzen, 1975) and the research of Damayanti (2013), namely tendency, decisions, and desires for consistency to comply with tax rules.
e Tax Compliance (TC) Taxpayers must carry out their tax obligations, starting from calculating their taxes correctly, pay the tax appropriately, and report their tax notification letter carefully. They are forms of tax compliance. The indicators of tax compliance are based on the Regulation of the Minister of Finance Number 192/PMK.03/2007 and refer to the research of Damayanti (2013), namely 1) compliance with tax estimation, 2) compliance with tax error, 3) compliance with tax treatment, 4)

Vol. 29, No. 2 August 2021 © Centre for Indonesian Accounting and Management Research Postgraduate Program, Brawijaya University
compliance in delivering periodic tax return, 5) compliance in delivering annual tax return, 6) compliance in tax payment, 7) tax underpayment.

Data Analysis Method
This study uses descriptive statistics and inferential analysisthe former aims to describe the research respondents and variables. The respondents were described through several characteristics, namely gender, age, last education, employment, and monthly gross income. The description of the research variable is shown in minimum, maximum, mean values. The latter analysis tests the research's structural model. Based on the explanation above, this study's analysis and hypotheses testing uses Structural Equation Modelling (SEM-PLS) performed in SmartPLS version 2.0.

FINDINGS AND DISCUSSION Descriptive Statistics
From 183 processable questionnaires, it was found that most respondents are males (108 people or 59%), aged 20-29 years (98 people or 53.6%), undergraduates (93 people of 50.8%), and businesspeople (150 people or 82%). Furthermore, most of them get a gross income of below IDR 5 million per month and fill and report their tax return by themselves (166 people or 90.7%). The constructs of this study were built using a 7-point Likert scale. The mean values are from 4.985 to 5.723, with the standard deviation from 1.149 to 1.595. The descriptive statistics of trust in authorities have the lowest average value of all constructs, lower than 5. It shows that the taxpayer's trust in the authority is in the interval position. Furthermore, the descriptive statistics show that most respondents' responses to questions about legitimacy power, coercive power, taxpayer's intention, and tax compliance are higher than 5. This shows that most respondents tend to be optimistic about the power of the tax authority in both legitimacy and coercion. The respondents also have good intentions in carrying out their tax obligations and have high compliance.

Validity and Reliability Testing
The two main parameters built are construct validity (convergent validity and discriminant validity) and construct reliability. The results of the validity and reliability testing of the construct are shown in Table 1. The amount of the loading factor shows the convergent validity of each indicator in measuring the construct. The loading factor used in this study is > 0.6. Based on the results of the AVE testing in Table 1, all constructs in this study has an AVE of >0.5. The discriminant validity is calculated using cross-loading. The criterion is that if the loading factor of a corresponding indicator is more significant than an item's correlation value with other indicators, the item should be declared valid in measuring the corresponding indicator. Thus, indicators that measure each of these constructs are declared valid. The reliability test results show that the composite reliability value is> 0.7 and that the composite reliability value is >0.6. Hence, all indicators that measure the construct are declared reliable.

The goodness of Fit Model
The Goodness of Fit Model is used to determine the contribution of exogenous variables to endogenous variables. The Goodness of Fit Model in this study is presented in the following table:   Table 2 shows that, first, the R-Squared of trust in authorities (Y1) is 0.435 (43.5%), meaning that the variation of trust in authorities can be explained by legitimacy power and coercive power for 43.5%; other constructs explain the rest outside of this study. Second, the R-Squared of taxpayer's intention (Y2) is 0.438 (43.8%), indicating that the variation of taxpayer's intention can be explained by trust in authorities for 43.8%; other constructs explain the rest outside of this study. Third, the R-Squared of tax compliance is 0.248 (24.8%), indicating that the taxpayer's intention for 24.8% can explain the variation of tax compliance; the rest is the contribution of other variables outside of this study.
The predictive relevance (Q 2 ) is 0.761 or 76.1%. It shows that the variation of tax compliance can be explained as a whole by the model, 76.1%. In other words, the effect of legitimacy power, coercive power,

Vol. 29, No. 2 August 2021 © Centre for Indonesian Accounting and Management Research Postgraduate Program, Brawijaya University
trust in authorities, and taxpayer's intention at the tax compliance is 76.1%; the remaining 23.9% is the influence of other factors not discussed in this study.

Hypothesis Testing
Hypothesis testing is used to determine the effect of exogenous variables on the endogenous variables. If T statistics ≥ T-table (1.64), it can be concluded that the exogenous variables affect the endogenous variables. The hypothesis can be tested using the following table.  Table 3 shows the T Statistics of > 1.64 for all constructs. In addition, the coefficients are positive. The results of the hypothesis testing show that hypothesis 1, that is, legitimacy power has a positive effect on the trust in authorities, is accepted. With the T Statistics of >1.64 and positive coefficient, hypothesis 2, i.e. coercive power has a positive effect on trust in authorities, is rejected. Next, hypotheses 3, 4, and 5 are accepted, which means that legitimacy power, coercive power, and trust in authorities positively affect taxpayers' intentions. Finally, the results of the hypothesis testing show that hypothesis 6, taxpayer's intention has a positive effect on tax compliance, is accepted.

DISCUSSION
The empirical evidence shows that H1 is accepted, which means that the higher the legitimacy power of the authorities, which comes from their expertise and professionality, the higher the trust of taxpayers in them. This result is consistent with the slippery slope framework that authority's power (Legitimate Power) increases taxpayer's trust in the authority (Kirchler, Hoelzl and Wahl, 2008). This result is also relevant with Kastlunger et al. (2013), Djajanti (2018), and Tjondro, Setiabudi, and Joyo (2019). Thus, it can be concluded that the legitimacy power of the authority is the determining factor of taxpayer's trust in the tax authority.
The empirical evidence shows that H2 is rejected due to the positive coefficient. Therefore, the higher the coercive power of authorities, which is based on examinations and penalties or fines, the higher the trust of taxpayers in them. This is not relevant with the slippery slope framework, which explains that the coercive power of authorities reduces taxpayer's trust in them (Kirchler et al. 2008). The coercive power of authorities will create a hostile climate, a condition when the tax authorities always suspect that taxpayers will commit tax evasion. The suspicion will reduce the taxpayer's trust. Coercive power is not proven to negatively influence the trust in authorities, allegedly because Indonesia is a developing country, so that coercive power of tax authority is still needed to foster taxpayer's trust. The coercive power of authority makes taxpayers assume that tax authorities are law enforcers that cannot be played around. They can feel the assertiveness of the tax authorities in carrying out their duties. This result is consistent with Ratmono (2014). Thus, it can be concluded that the taxpayer's trust is inseparable from the coercive power of the authority.
The empirical evidence shows that H3 is accepted, which means that the greater the legitimacy power of the authorities, the higher taxpayer's intention to carry out the tax obligations. Legitimacy power has a close relationship with tax regulations (Kirchler et al., 2008). This result is relevant to Batrancea et al. (2019) finding that all components of the slippery slope framework, i.e. legitimacy power, coercive power, and trust in authorities, can increase tax compliant behaviour. If the authority has good legitimacy power, taxpayers will always have a significant intention to carry out their tax obligations. Based on the result, it can be concluded that legitimacy power affects a taxpayer's intention to carry out his tax obligations.
The empirical evidence shows that H4 is accepted, which means that the greater the coercive power of the authority, the higher the taxpayer's intention to fulfil his tax obligations. Taxpayer's intention created from the coercive power of the authorities is an intention to comply due to pressure. In this condition, taxpayers will always be supervised by the tax authority, so they are afraid of committing tax fraud. In the end, taxpayers will intend to comply due to the anxiety of being examined or punished, and tax fraud in the form of evasion decreases. Hence, it can be concluded that coercive power determines the taxpayer's intention in carrying out his tax obligations.
The empirical evidence shows that H5 is accepted, which means that the higher taxpayers' trust in their tax authority, the higher their intention to comply. In addition to the power of tax authority, taxpayers'

Vol. 29, No. 2 August 2021 © Centre for Indonesian Accounting and Management Research Postgraduate Program, Brawijaya University
trust in tax authority is also essential to improve their intention to comply. Based on the Theory of Reasoned Action, a person's belief affects his intention to behave. This belief consists of belief in consequences and normative beliefs (Fishbein & Ajzen, 1975). Kogler et al. (2013) stated that the highest intention to comply is in the condition where the community has a high level of trust in their authority. Taxpayer's intention is strongly influenced by the excellent relationship between taxpayers and tax authorities. It can make people have high trust in the authority. Thus, it can be concluded that trust in authorities affects taxpayer's intention.
The empirical evidence shows that H6 is accepted, which means that the higher the taxpayers' intention, the higher the tax compliance of a society. This follows the Theory of Reasoned Action, which explains that intention is the determinant of a person's attitude in conducting an action (Ajzen & Fishbein, 1980). High taxpayers' intention encourages them to comply with tax rules. This result is similar to the finding of See (2011);Bidin, Shamsudin, and Othman (2014); Woro, Subekti, and Baridwan (2015); and Alleyne and Harris (2017). Thus, it can be concluded that taxpayer's intention affects tax compliance.

CONCLUSION
The results of this study show empirical evidence that tax compliance increases along with legitimacy power, coercive power, trust in authorities, and taxpayer's intention. The theoretical implication of this research is that the research concept becomes a form of deconstruction and provides a new perspective on new research explanation of the theory used. This study provides additional empirical evidence on tax compliance research in terms of economic and social psychology. The analysis results also confirm the slippery slope framework and the theory of reasoned action in that most of the findings are following them. However, a uniqueness was found in the relationship between coercive power and trust; that is, trust in authorities increases along with coercive power of authority. This finding is different from the slippery slope framework. The results of previous studies have concluded that the coercive power of authorities decreases taxpayer's trust in them, but this study found the opposite. The results of this study also provide practical implications in the form of recommendations for the tax authorities to improve tax compliance by maximizing the power of tax authorities (legitimacy power and coercive power). For taxpayers, they must maintain their tax compliance intentions in order to increases tax compliance in Indonesia.
The weakness of this study is the data collection which is relatively short, the range of business locations is scattered, and some business owners are reluctant to fill out questionnaires for reasons of uncertain income, which causes them not to pay taxes during the Coronavirus pandemic. Suggestions for future research are to extend the time for distributing questionnaires to get a more significant number of samples from this study. This conceptual research can be carried out again in normal conditions after the end of social restrictions to obtain a maximum return on the questionnaire by distributing it directly to each research respondent. This research can also be done using a different research sample, using a sample of corporate taxpayers, because the tax revenue from corporate taxpayers is greater than that of individual taxpayers. In addition, further researchers can use other research variables.