Preference Of Australian Citizens For Tax Agents Or Self-preparers For Lodging Tax Returns

Methods and Data Collection

Lodging tax returns is the prime responsibility of each and every citizen of any nation. Taxes should be payed to the government (Gallemore and Labro 2015). Despite of this, all the citizens are not supposed to pay the income taxes. There is an income limit. Any individual earning more than the income limit in a year is liable to pay the taxes (Bitler, Hoynes and Kuka 2016). Thus, on the occasion of lodging the tax returns, an individual can opt for two different ways. The first way is to appoint a registered tax agent and pay him or her to do the work for the peoson. The other way is to prepare the lodgement file by himself or herself. This study is mainly based on Australia and the preference of the Australian citizens for tax agents or self-preparers will be assessed in this study.

Two different types of data has been collected for conducting this study. The first data has been obtained from the Australian Taxation Office (ATO) website. The data on tax lodgement for the year 2013-2014 has been obtained and a sample of 1000 data has been extracted from the original for the purpose of this study. The data thus collected is a secondary data as it is obtained from a website. The variables involved such as gender is a categorical variable, age range is a categorical variable and lodgement method is a categorical variable. Other two variables involved such as total income and total deduction are both quantitative variables. In table 1 below are the first five cases of the first dataset.

Table 1.1: First five cases of the ATO sample dataset

Gender	age_range Save Time On Research and Writing Hire a Pro to Write You a 100% Plagiarism-Free Paper. Get My Paper	Lodgment_method	Tot_inc_amt	Tot_ded_amt
1	0	S	2389	0
1	6	S	4936	0
1	0	S	2462	0
1	9	A	29448	425
1	0	A	49665	684

Unlike the first dataset, the second dataset contains information about the preference of tax lodgement methods by 150 international students studying in Australian Universities. This data is thus primary data and the variable involved in this dataset is a categorical variable. 

Table 2.1 shows a summary of the number of people of Australia who prefer tax agents and who prefer themselves for the lodgement of tax returns. It can be seen that 73.2 percent of the people prefer tax agents. The proportion of preference for tax agents is shown diagrammatically in figure 2.1. It has also been obtained from the analysis that 70 – 76 percent of the people of Australia prefer tax agents. This can be claimed with 95 percent confidence.

Results and Analysis

Table 2.1: Summary of Lodgement Methods of ATO Dataset

Row Labels	Count of Lodgment_method	Count of Lodgment_method2
A	732	73.20%
S	268	26.80%
Grand Total	1000	100.00%

 Figure 2.1: Pie chart showing proportion of Lodgement Methods for ATO Dataset 

Table 2.2: Proportion of Tax Agents of ATO Dataset
Sample Size	1000
Count of Successes	732
Confidence Level	95%
Sample Proportion	0.732
z Value	1.9600
Standard Error of the Proportion	0.014006284
Margin of Error	0.0275
Calculations for Computation of Confidence Interval
Interval Lower Limit	70.45%
Interval Upper Limit	75.95%

Table 3.1 shows a summary of the number of international students who prefer tax agents and who prefer themselves for the lodgement of tax returns. It can be seen that 74 percent of the students selected prefer tax agents. The proportion of preference for tax agents is shown diagrammatically in figure 3.1. It has also been obtained from the analysis that 71 – 85 percent of the international students prefer tax agents. This can be claimed with 95 percent confidence. 

Table 3.1: Summary of Lodgement Methods of Students Dataset

	Values
Lodgement Method	Frequency	Proportion
A	134	0.74
S	46	0.26
Grand Total	180	1

Figure 3.1: Pie chart showing proportion of Lodgement Methods for Students Dataset

Table 3.2: Proportion of Tax Agents of Students Dataset
Sample Size	150
Count of Successes	117
Confidence Level	95%
Sample Proportion	0.78
z Value	1.9600
Standard Error of the Proportion	0.033823069
Margin of Error	0.0663
Calculations for Computation of Confidence Interval
Interval Lower Limit	71.37%
Interval Upper Limit	84.63%

To test whether the proportion of Australian people (p₁) and International students (p₂) preferring tax agents are equal, z-test has to be computed (Park 2015). The null and the alternate hypothesis for this test are defined as follows:

Null Hypothesis (H₀): p₁ – p₂ = 0

Alternate Hypothesis (H₁): p₁ – p₂ ≠ 0

The difference p₁ – p₂ is denoted by p. From the results of the test given in table 3.3, it can be seen that the p-value is 0.212 which is greater than the level of significance (0.05). Thus, the null hypothesis is accepted. Thus, there is no difference in the proportion of Australian people and International students preferring tax agents for lodgement of tax returns.

Table 3.3: Test for equality of proportion of people appointing Tax Agents from the two datasets
Null Hypothesis H0:	p	0	0%
Alternative Hypothesis HA:	p	<>	0%
Test Type	Two
Level of Significance	0.05
Number of Samples for Group 1	1000
Number of Successes for Group 1	732
Number of Samples for Group 2	150
Number of Successes for Group 2	117
Hypothesized Difference	0
Proportion for Group 1	0.732
Proportion for Group 2	0.78
Average Proportion	0.738261
Difference in Two Proportions	-0.048
Critical Z value	-1.24709
p-value	0.212364
Result of the Analysis
Do not reject Ho

The comparison of different types of lodgement methods with respect to different ages is summarized in table 4.1. The comparison is also shown clearly in figure 4.1. From the analysis, it can be seen that people for all the age ranges prefer lodging their tax returns with the help of tax agents.

Table 4.1: Comparison of Age Range and Lodgement Method for ATO Dataset

Count of Lodgment_method	Column Labels
Row Labels	A	S	Grand Total
0	40	21	61
1	33	6	39
2	52	13	65
3	65	14	79
4	69	17	86
5	92	15	107
6	82	24	106
7	74	37	111
8	66	35	101
9	72	35	107
10	55	34	89
11	32	17	49
Grand Total	732	268	1000

Figure 4.1: Bar Graph comparing Lodgement method according to Age Range

The association between the two variables age range and lodgement method has to be tested. This test can be done using the chi square test of association (Gilbert and Prion 2016). The expected frequencies that are necessary for performing the test is given in table 4.2. The null and the alternate hypothesis for this test are defined as follows:

Null Hypothesis (H₀): There is no existence of significant association between the two variables.

Alternate Hypothesis (H₁): There is existence of significant association between the two variables.

From the significance given in table 4.3 which is obtained from the analysis, it can be seen that the value is less than 0.05. Thus null hypothesis is rejected. There is existence of relationship between age group and lodgement methods.

Conclusion

Table 4.2: Expected Frequency table for Age Range and Lodgement Method

Expected Frequency
Row Labels	A	S	Grand Total
0	44.652	16	61
1	28.548	10	39
2	47.58	17	65
3	57.828	21	79
4	62.952	23	86
5	78.324	29	107
6	77.592	28	106
7	81.252	30	111
8	73.932	27	101
9	78.324	29	107
10	65.148	24	89
11	35.868	13	49
Grand Total	732	268	1000

Table 4.3: P-Value for the test of Association

Chi Square Significance Value

0.000156

It can be seen from table 5.1 that the average of the total income is higher for the people appointing tax agents. Figure 5.2 shows the variation in the total income of the individuals. It can be seen that there is huge variation in the incomes. The incomes of the individuals are not at all close to the average income. A large number of people are earning more than the average income. There are 55 people preferring tax agents whose income is much higher than the usual income of the people belonging to that group and 16 people preferring self-preparers whose income is much higher than the usual income of the people belonging to that group

Table 5.1: Average Income for each type of Lodgement Methods

Row Labels	Average of Tot_inc_amt
A	70547.08607
S	46670.23881
Grand Total	64148.091

Figure 5.1: Bar graph comparing the average income of different lodgement methods

Table 5.2: Summary of total income of different lodgement methods

Measures	Tax Agents	Self-Preparer
Mean	70547.0861	46670.23881
Standard Error	6665.38537	2624.92117
Median	46564.5	38863.5
Mode	16778	0
Standard Deviation	180335.324	42971.81156
Sample Variance	3.25E+10	1.85E+09
Kurtosis	364.396143	7.003206861
Skewness	17.0865485	2.10474629
Range	4153282	311090
Minimum	-6234	0
Maximum	4147048	311090
Sum	51640467	12507624
Count	732	268
First Quartile	23185.75	16859.5
Third Quartile	75541.5	62481.75
Interquartile Range	52355.75	45622.25
Calculation of Outlier Range
Lower Outiler Range	-29170	-28762.75
Upper Outlier Range	127897.25	108104
Number of Outliers
Number of Outliers	55	16

Figure 5.2: Boxplot showing shape of the distribution

Positive relationship exists between total income and total deduction of the people preferring tax agents but it can be seen from the r square value given in table 6.1 that the relationship is very weak. Only 2 percent of the variations in deductions can be explained by income. The relationship can be expressed with the help of the following equation:

Deduction = (0.0057 * Income) + 2431.8

Figure 6.1: Relationship between Income and Deduction for Tax Agents

Table 6.1: Regression Statistics (Tax Agents)
Multiple R	0.13
R Square	0.02
Adjusted R Square	0.02
Standard Error	7661.26
Observations	732

Table 6.2: ANOVA (Tax Agents)
	df	SS	MS	F	Significance F
Regression	1	7.76E+08	7.76E+08	13.229	0.000
Residual	730	4.28E+10	58694945
Total	731	4.36E+10

Table 6.3: Regression Coefficients (Tax Agents)

	Coefficients	Standard Error	t Stat	P-value	Lower 95%	Upper 95%
Intercept	2431.813	304.093	7.997	0.000	1834.813	3028.814
Tot_inc_amt	0.006	0.002	3.637	0.000	0.003	0.009

Positive relationship exists between total income and total deduction of the people preferring themselves for lodgement of tax returns but it can be seen from the r square value given in table 6.4 that the relationship is moderate. Only 26 percent of the variations in deductions can be explained by income. The relationship can be expressed with the help of the following equation:

Deduction = (0.049 * Income) + 779.63 

Figure 6.2: Relationship between Income and Deduction for Self-Preparers

Table 6.4: Regression Statistics (Self-Preparers)
Multiple R	0.51
R Square	0.26
Adjusted R Square	0.26
Standard Error	3543.75
Observations	268

Table 6.5: ANOVA (Self-Preparers)
	df	SS	MS	F	Significance F
Regression	1	1.18E+09	1.18E+09	94.211	0.000
Residual	266	3.34E+09	12558162
Total	267	4.52E+09

Table 6.6: Regression Coefficients (Self-Preparers)

	Coefficients	Standard Error	t Stat	P-value	Lower 95%	Upper 95%
Intercept	-779.628	319.903	-2.437	0.015	-1409.491	-149.765
Tot_inc_amt	0.049	0.005	9.706	0.000	0.039	0.059

Conclusion

From all the analysis conducted in the above sections, it can be concluded that both the people of Australia and the International students has a preference for appointing tax agents for the lodgement of tax returns. There is no difference in the proportion of the Australian people and the International students preferring tax agents. Relationship has been found between age range and preference of lodgement methods. The people who are tax payable earn more than the average income. There is positive relationship between income and deduction amounts for both the types of lodgement methods but the relationship is weak.

The effect of gender on the lodgement methods have not been analyzed so far in this research. This can be conducted as further research. 

References

Bitler, M., Hoynes, H. and Kuka, E., 2016. Do In-Work Tax Credits Serve as a Safety Net?. Journal of Human Resources.

Gallemore, J. and Labro, E., 2015. The importance of the internal information environment for tax avoidance. Journal of Accounting and Economics, 60(1), pp.149-167.

Gilbert, G.E. and Prion, S., 2016. Making Sense of Methods and Measurement: The Chi-Square Test. Clinical Simulation in Nursing, 12(5), pp.145-146.

Park, H.M., 2015. Hypothesis testing and statistical power of a test.