Results Proceeding E Book 4A Turky

159 FIGURE 3 Path Coefficients FIGURE 2 Path Coefficients Panel 2 Scenario 2 Notes: Figures in parentheses are path coefficient for tax complexity on fairness perceptions. R 2 = 0.538 a p0.001 b p0.005 c p0.010 d p0.025 e p0.050 f p0.100 Attitude towards compliance Subjective Norms Intention to comply Tax knowledge Tax complexity Overall Fairness 0.577 a Vertical fairness Exchange fairness Horizontal fairness Admin. fairness Perceived Behavioural Control -0.153 d -0.041 0.141 d 0.135 d -0.082 f 0.036 0.275 a 0.097 0.086 -0.183 b 0.201 b -0.027 0.113 f -0.068 0.184 e -0.128 f 0.076 0.037 0.351 a 160 The model also describes the path coefficients for tax knowledge and tax complexity on fairness perceptions. The results show that tax knowledge had negative direct effects on exchange fairness and administrative fairness at the 0.10 and 0.005 levels, respectively. Other variables were not significant. For tax complexity, the results reveal that horizontal fairness and vertical fairness were not significant. The other paths had a significant positive influence on fairness perceptions at the 0.001 level, except for overall fairness with significant level of 0.05. Tax complexity but not tax knowledge had a significant positive influence on perceived behavioral control at the 0.10 level.

7. Discussion

The purpose of this study was to examine the fairness perceptions of New Zealand taxpayers on the income tax system and how their perceptions influence the compliance behavior. In so doing, I used a well-established model of the TPB. The TPB model provides a theoretical framework of behavioral determinants consisting of attitude, subjective norm and perceived behavioral control. For the purpose of this study, fairness perceptions were included to extend the existing TPB model, particularly in the tax compliance environment. The TPB model was tested using the data obtained from actual taxpayers. Overall, the results suggest that the TPB model fits the data well. This study reveals that taxpayers view fairness of the income tax system from various perspectives, namely overall fairness, exchange fairness, horizontal fairness, vertical fairness and administrative fairness. This is consistent with previous studies which contend that fairness perceptions are multidimensional Gilligan Richardson, 161 2005; and Gerbing, 1988. Thus, the findings provide support for Hypothesis 1 that fairness perceptions are multidimensional. Hypothesis 2 predicts that fairness perceptions will positively influence compliance behavior. Specifically, the hypothesis suggests that the fairer taxpayers perceive the tax system, the more likely they will comply with their tax obligations. However, this is only correct for overall fairness and vertical fairness. The findings support the concept in Equity Theory and are consistent with previous studies for example, Gilligan Richardson, 2005; Turman, 1995; Roberts, 1994; and Porcano Price, 1992. Other fairness dimensions, however, are not significant. Surprisingly, exchange fairness is found to have a negative effect on compliance behavior. A possible explanation for this is that the mean for the exchange fairness variable is below 4.0, suggesting that most respondents perceived the tax system as unfair. Similarly, the inverse relationship between administrative fairness and compliance behavior may be attributable to the fact that the respondents are neutral about administrative fairness as indicated by its mean of 4.0. Tax knowledge proves to have an influence on exchange fairness and administrative fairness. The findings, however, do not provide support to Hypothesis 3 which predicted that tax knowledge will positively influence fairness perceptions. Also, the findings are not consistent with previous studies Fallan, 1999; White et al., 1990; and Harris, 1989, which claimed that tax knowledge will increase fairness perceptions. The negative relationship, on the other hand, suggests that taxpayers with a higher of level of tax knowledge found these fairness dimensions to be unfair. This result may have arisen because the majority of respondents have a good