The influence of ethics, experience, and competence on auditors’ professional skepticism (Study at BPKP representative of Lampung Province)

Purpose: This study aims to prove empirically the influence of ethics, experience, and competence on auditors’ professional skepticism. Research Methodology: In testing the hypotheses, 63 auditors at BPKP Representative of Lampung Province took part in responding to the questionnaires distributed. The criteria for auditors sampled: 1). Auditors who are directly involved in the assignment, especially in the process of collecting and testing audit evidence, like audit team members, audit team leaders, audit supervisors and audit quality control, and 2). Auditors who are experienced in auditing for at least one year. Partial Least Square analysis method was used and processed with SmartPLS version 3.0 software and Microsoft office excel 2007 program. Results: As hypothesized, the results revealed that ethics, experience, and competence had a positive influence on auditor professional skepticism. Discussion, limitations, and recommendations for future research are discussed. Limitations: This research is limited only to the government auditors working at BPKP Representative of Lampung Province so that the research result can not be generalized to the auditor as a whole. The auditor's tight schedule caused the returning time of questionnaires to shift from the set timeline previously. The use of questionnaires also was the last limation in terms of inaccurate answers, dishonest answers, and missinterpreted questions. Contribution: This study is beneficial for BPKP Representative of Lampung in improving auditors' professional skepticism, as well as one of literature resources for conducting further studies.


Introduction
Based on the Corruption Perceptions Index (CPI) 2016 which was released by Transparency International (TI), over two-thirds of the 176 countries in 2016 fall below the midpoint of the scale of 0 (highly corrupt) to 100 (very clean). Indonesia placed the 90th position with a score of 37 (TI, 2017). With the global average score of 43 and the Asia Pacific average score of 44, Indonesia is in the emergency of corruption. Indonesian Corruption Watch (ICW) revealed that poor performance could be attributed to unaccountable governments. Mapping research of corruption case handling trend 2016 by ICW revealed corruption cases based on government institutions as follows (ICW, 2017): 1). District government as many as 219 corruption cases with the number of stated losses as much as Rp478 billion; 2). City government as many as 73 cases with the number of stated losses as much as Rp247 billion; 3). Village Government as many as 62 cases with the number of stated losses as much as Rp206 billion; 4). Ministry as many as 28 cases with the number of stated losses as much as Rp38 billion; 5) Regional owned entity as many as 20 cases with the number of stated losses as much as Rp18 billion. Since the era of reformation, the demand for transparency and accountability for public financial management have been stronger (Butar Butar and Perdana, 2017). Government internal auditors took part in this case. In Indonesia, the function of internal auditors become the focus agenda of change, therefore internal auditors' role are being center of attention (Butar Butar and Perdana, 2017).
An auditor is demanded to generate good quality and reliable audits. Auditors also must be able to detect fraud which may happen during the process of audit that can affect the auditors' decision making. To do so, auditors must have an attitude of auditor professional skepticism. Professional skepticism is the basis of high-quality audit performance (Hurtt et al, 2013). Due professional care requires an auditor to carry out professional skepticism, an attitude that involves a questioning mind and critically evaluates the audit evidence (BPK RI, 2007). In applying professional skepticism, auditors must not be satisfied with the unconvincing evidence, even though they think that the management of an entity is trustworthy (BPK RI, 2007).
Research conducted by the SEC (Securities and Exchange Commission) for 11 periods (January 1987to December 2007 found that a third of the causes of audit failure is the inadequate level of auditor professional skepticism. Of the 40 audit cases examined, 24 cases (60%) included audit failure because the auditors did not apply the adequate level of professional skepticism (Beasley et.al. (2001) in Noviyanti, 2008). This inadequate level of auditor professional skepticism can influence public trust over auditors' ability. Enofe, Ukpebor, and Ogbomo (2015) explained that when auditors were unable to appropriately apply the professional skepticism, they might not obtain sufficient evidence to support their opinions.
The attitude of auditors' professional skepticism can be governed by the auditors' compliance of code of ethics. Pflugrath et.al. (2007) found that compliance with the code of conduct had a significant influence on the professional accountant's audit decision making. When auditors are able to uphold the ethical principles during the process of audit, the audit failure can be avoided. Minaryanti and Ridwan (2015) showed that the auditors who were responsible for detecting fraud would be able to prevent the failure audit since they were able to apply good audit methods, uphold the objective professional judgment and have the attitude of professional skepticism.
On the other hand, auditors' experience gives a contribution to the improvement of auditors' professional skepticism. Anugerah et.al. (2012) explained that the government internal audit body needs to develop various programs to enhance the knowledge and experience of their internal auditors. Libby and Frederick (1990) found the more experience, the more the auditors can generate various kinds of assumptions in explaining the audit findings. Auditors' professional skepticism can be also governed by auditors' competence. Baybutt (2014) found that an auditors' competence was essential to ensure audit quality. Auditors who have good competence within themselves are able to detect fraud during the process of audit. This research is inspired by previous studies. The previous studies tested how ethics, experience, and competence influenced the auditors' professional skepticism. Enofe et.al. (2015), Oktarini and Ramantha (2016), Silalahi (2013), Suraida (2005), Winantyadi and Waluyo (2014) agreed that ethics, experience, and competence had a positive influence on the auditor professional skepticism.
Contrary to previous studies, this research is conducted to the auditors in the governmental circumstance because there are still few studies on the auditor professional skepticism in the government circumstance. This research is also more focused only on some internal factors that influence the auditor professional skepticism, which are ethics, experience, and competence. This research is conducted at the BPKP Representative of Lampung Province.

Research questions
The research will address these following questions: This research will address the issue of auditor professional skepticism, focused on some internal factors from government auditors. In general, the findings of the research are expected to: 1. Theoretical significance a. The researcher expects that this research can contribute benefits in terms of knowledge, understanding, and experience development of researcher as well as giving contributions to the development of audit knowledge. Besides, the researcher also expects that this research can provide sufficient evidence of the influence of ethics, experience and competence of the auditors' professional skepticism. b. The researcher expects this research can add and develop literature related to the auditors' professional skepticism for government auditors, remembering that there are still few numbers of researches related to the auditors' professional skepticism for government auditors.

Practical Significance
The researcher expects this research can give motivation for the auditors, particularly for auditors at the BPKP Representative of Lampung province to continuously improve their attitude of professional skepticism, so that they can do their duties professionally.

Policy Significance
For BPKP and other public accountant offices, this research is expected to give input in making the policies, like education and training policy, to improve the auditors' professional skepticism.

Theoretical review
The use of theory as a foundation in this research, there are some previous studies that can be used as references in this study. Past researches are more conducted in the public accounting firm. However, there are also some studies that use government auditors as respondents. Suraida (2005) tested the influence of ethics, competence, experience audit, and audit risk on the auditors' professional skepticism by involving 100 public accountants who authorized to sign the auditor's report. Using the Structural Equation Model (SEM) model equations, it showed that ethics, competence, experience audit, and audit risk influenced the auditors' professional skepticism. Silalahi (2013) conducted a study to test the influence of ethics, competence, experience audit, and audit situation on the auditors' professional skepticism. There were 84 respondents involved in the study. They were working in public accountant firms located in North Sumatra and Riau, which were registered in 2009 IAI Directory Books. The research hypotheses were tested using multiple regression with SPSS. The results showed that ethics, competence, experience, and audit situations influenced the auditors' professional skepticism.
Winantyadi and Waluyo (2014) conducted a study involving 37 auditors at KAP in Yogyakarta. Hypothesis testing is executed using simple regression and multiple regression tests. The results showed that the experience, expertise, audit situation, and ethics gave a positive influence on the Auditors' professional skepticism. Anugerah et.al. (2012) conducted a study of 40 auditors from BPK RI Representative of Riau Province. The hypotheses were tested using the method of Partial Least Square (PLS). The result showed that there was a positive and significant relationship between ethics, experience, situational factors audits on auditors' professional skepticism in BPK RI Representative of Riau Province.

2.1
The framework This study uses the theory of cognitive dissonance as a theoretical base. Cognitive dissonance theory is used to explain the influence of interaction between auditors' professional skepticism in BPKP with the internal factors that influence the professional skepticism, which are ethics, experience, and competence of the audit. Based on the explanation above, this study is compiled by the framework as shown in figure 2.1 below:

Hypothesis development The influence of ethics on auditors' professional skepticism
The compliance of ethics will be reflected in the auditors' attitudes and actions. Auditors who have a high awareness of professional ethics and uphold the code of ethics will work professionally. Professionals auditors will apply the attitude of professional skepticism in each performance of his/her duties. Code of ethics for auditors at BPKP refers to Asosiasi Auditor Internal Pemerintah Indonesia (AAIPI) Number: KEP-005/AAIPI/DPN/2014 on the date of 29th of April, 2014 about Pemberlakuan Kode Etik Auditor Intern Pemerintah Indonesia, Standar Audit Intern Pemerintah Indonesia, dan Pedoman Telaah Sejawat Auditor Intern Pemerintah Indonesia. This rule is used as guidelines for auditors at BPKP in carrying out their responsibilities. The aim of this code of ethics is to prevent unethical behavior, fulfill the principles of accountability, and implement the controlling in order to create credible auditors with optimum performances (AAIPI, 2014). Codes of ethics are the norms, which rule the moral behavior of a profession through the provisions that must be met and adhered by every member of the profession (Anugerah et.al., 2012).
Research conducted by Enofe et.al.(2015) concluded that accounting ethics would play an even more important role in improving auditors' professional skepticism. This result is in line with studies conducted by Oktarini and Ramantha (2016), Silalahi (2013), Suraida (2005), and Winantyadi and Waluyo (2014) that revealed ethics had a positive influence on the auditors' professional skepticism. Thus, this lead to the following hyphothesis: H1 : Ethics positively influence auditors' professional skepticism.

The influence of experience on auditors' professional skepticism
In a study conducted by Hurt et al (2013), it was explained that the experience of an auditor could strengthen the auditor's attitude of professional skepticism. The more experienced, the higher auditor will have the attitude of auditors' professional skepticism. The more experienced, the more they can generate various kinds of assumptions in explaining the audit findings (Libby and Frederick, 1990). An experienced auditor will have objective decisions in every assignment.
Haynes (1999) in Kaplan et.al (2008) examined the interaction between auditor experience and the source credibility of the evidence. In her study, the high-experience group was comprised of governmental auditors with an average of more than seven years of audit experience whereas the lowexperience group was comprised of MBA students with no audit experience. She manipulated management credibility by changing the probability that management would truthfully report certain information. Findings suggested that the persuasiveness of information obtained from management was jointly influenced by auditor experience and source credibility. The credibility of management had a stronger influence on judgments of the high experience group compared to judgments of the lowexperience group.
According to Noviyanti (2008), the distinction between experienced and inexperienced auditors is the length of auditors working as auditors in the office. The experienced auditors will be more aware of any errors happened during the process of audit. The studies conducted by Anugerah et.al. (2012) revealed that experience positively influenced the auditors' professional skepticism. This result is in line with studies conducted by Gullkvist and Jokipii (2015), Hurtt et al (2013), Oktarini and Ramantha (2016) and Silalahi (2013) that showed that experience had a positive influence on the auditors' professional skepticism. Thus, this lead to the following hyphothesis: H2 : Experience positively influences auditors' professional skepticism.

The influence of competence on auditors' professional skepticism
Auditors must have the necessary competence in doing their tasks. One of the factors identified by auditing standards in determining the strength of the internal audit function is internal auditor competence (Seol, 2011). Gamayuni (2016) stated that internal auditor competence meant that in carrying out the task, auditors must have the knowledge, skill, personal attribute, experience, and also code of conduct that had to obey. The knowledge can be obtained either formal or informal. The skill can be seen by certificates received by auditors. After joining technical training, seminar, or consortium, auditors will get certificates to certify their skills. Competence of auditors through their general knowledge and special expertise can improve their professional skepticism. Wood (2004) added that competence will increase the confidence of the auditor. Research conducted by Silalahi (2013) stated that competence had a positive influence on auditor professional skepticism. This result is in line with Suraida (2005) and Winantyadi and Waluyo (2014) that revealed competence had a positive influence on the auditors' professional skepticism. Thus, this lead to the following hyphothesis: H3 : Competence positively influences auditors' professional skepticism.

3.1
Scope of the research This study seeks to explain the causal relationship between independent variables and the dependent variable by testing the hypothesis. The independent variables in this study are ethics, experience, and competence of the audit, while the dependent variable is the auditors' professional skepticism. The study plan is a simplification of the research conducted by Silalahi (2013) because the researchers would like to focus on internal factors that influence the auditors' professional skepticism. The research will be conducted by a survey method using a questionnaire distributed to the auditors at BPKP Representative of Lampung Province. The questionnaire will be written in Bahasa to simplify respondents in answering. The period used is the instantaneous cross-section which gives a certain fact that is only once collected in a period of the survey.

3.2
Population and sample The subjects of the population in this study are all auditors at BPKP Representative of Lampung Province. This study uses primary data by directly distributing questionnaires to the respondents. The collecting data technique used in this study is purposive sampling. Criteria of auditors sampled, namely: a. uditors who are directly involved in the assignment, especially in the process of collecting and testing audit evidence, like audit team members, audit team leaders, audit supervisors, and audit quality control. b. Auditors who are experienced in auditing for at least one year. Based on the researcher's judgment, relying on the researcher's own experience working at BPKP for more than 8 years, one year is enough for auditors at BPKP to have experience in audit.

3.3
Data sources This research is quantitative, using primary data by directly distributing questionnaires to the respondents. Questionnaires can be used to obtain data associated with thoughts, feelings, attitudes, beliefs, values, perceptions, personality, and behavior of respondents with 5 Likert scales. The data collected then will be processed using statistical software SmartPLS 3.0 and excel Microsoft office applications. Besides, to strengthen the result of this research, the researcher will have an interview with a qualified respondent. In this time, the researcher will be having an interview with the Head of BPKP Representative of Lampung Province.

3.4
Variable operational definition This study uses four variables: ethics (X1), experience (X2), and competence (X3) as the independent variable and the auditors' professional skepticism (Y) as the dependent variable. The answers to the questionnaires are collected and measured by a 5-point Likert scale ranging from "strongly disagree" (1) to "strongly agree" (5). Some questions are arranged in negative forms. On any reversed indicators, a low score on the measurement scale indicates a high collision value and vice versa. Each variables are measured in the following manner: 3.5 Data analysis methods 3.5.1 Descriptive statistics Descriptive statistics aims to explain and describe the research data profile (Abdillah and Jogiyanto, 2015). Descriptive statistics is used to analyze data by describing the data that has been collected and processed as it is without intending to make general conclusions or generalization (Sugiyono, 2011). The statistics technique that is commonly used is central tendention (mean, median, mode) and variability (Abdillah and Jogiyanto, 2015). Statistics measurement in this research used Microsoft Office Excel 2007 program.
3.5.2 Measurement model (Outer model) Data were analyzed by using the Structural Equation Modeling (SEM) model with Partial Least Square (PLS) approach, using SmartPLS application Version 3.0. SEM essentially offers the ability to perform path analysis with latent variables (Chin (1998) in Ghozali and Latan, 2015). PLS is a variance-based SEM statistical method designed to solve multiple regressions when specific data problems occur, such as small sample size measurements, missing values, and multicollinearity, to predict the effect of variable X on Y and explain the theoretical relationships Between the two variables. PLS is a reliable tool for testing prediction models (Abdillah and Jogiyanto, 2015).

Validity test
Validity shows how much a test measures what should be measured (Jogiyanto, 2016). Validity testing includes: 1. Convergent Validity scored based on the correlation between Average Variance Extracted (AVE) item score calculated by SmartPLS 3.0 application. Measurement scale of loading value > 0.5 is considered to be practical significance (Abdillah and Jogiyanto, 2015). 2. Discriminant Validity, by comparing the root square of the AVE score for each construct with the correlation between the construct and other constructs in the model. The model has sufficient discriminant validity if the AVE root for each construct is greater than the correlation between constructs and other constructs (Ghozali and Latan, 2015).

Reliability test
Reliability relates to the consistency of measuring instruments (Jogiyanto, 2016). The construct reliability test is measured by two criteria: composite reliability and Cronbach's alpha from the indicator that measures the construct. The construct is considered reliable if the value of composite reliability and Cronbach's alpha is above 0.70 (Ghozali and Latan, 2015).

3.6
Structural model (Inner model) The researcher tested the structural model by measuring the Coefficient of Determination (R 2 ), which is a goodness-fit model and Path Coefficient (ß) (Ghozali and Latan, 2015).

Determination coefficient (R 2 )
The determination coefficient (R2) aims to measure how far the ability of the model explains variations in the dependent variable. The coefficient of determination is between 0 and 1. The small R2 value indicates that the ability of independent variables explain the variation of the dependent variable is limited.

Path coefficient (ß)
This test is executed to ensure that the relationship between constructs is strong. The construct has a strong relationship if the coefficient test value of the path is greater than 0.10 and the latent relationship is said to be significant if the path coefficient test is at the 0.05 level (Ghozali and Latan, 2015).

3.7
Hypothesis test Hypothesis testing uses the comparison between the result of the path coefficient generated by T-Statistic with T-table. If the T-statistic value is higher than the T-table value, the hypothesis is supported.
The researcher uses a 95% confidence level (alpha 5%), then the T-table value for the one-tailed hypothesis is > 1.64.

4.1
Data collection result The data collected was primary data from distributed questionnaires. Questionnaires were distributed to the employees of the BPKP Representative of Lampung Province who matched with two criteria, 1) namely auditors involved in the assignment of the audit and had the role of team members, team leaders, technical controllers and quality controllers and 2) auditors who have conducted audit assignments for more than one year.
Distributed questionnaired were 76 questionnaires, following the number of auditors at BPKP Representative of Lampung Province. Of the 76 questionnaires, 65 questionnaires were returned but only 63 questionnaires could be used, with the following details:  (2017) The demographic of respondents obtained following the questionnaire that meets the criteria as follows:

1) Gender
Based on the gender category, 40 people (63.49%) of respondents are men and 23 people (36,51) are women. This data shows that most of the auditors at the BPKP Representative Office of Lampung Province are men.

2) Age
Based on the age category, 25 people (39.68%) of respondents are under 30, 13 people (20.64%) are aged 31-40, 18 people (28.57%) are in the range of 41-50 and as many as 7 people (11.11%) are aged over 51 years old. This data shows that most respondents are in productive ages.

3) Educational background
Based on the educational background category, there are 10 people (15,87%) have a diploma's degree background, 50 people (79,37%) have bachelor's degree background (S1) and the rest 3 people (4,76%) have master's degree background (S2 / S3). From this information, it can be stated that most respondents have adequate general knowledge.

4) Roles in audit team
Based on their roles in the audit team, 32 people (50.79%) roled as audit team members, 24 people (38.10%) served as audit team leaders, 4 people (6.35%) as Audit Supervisors and 3 people ( 4.76%) roled as Audit Quality Control.

5) Years of service
Based on the respondent's years of service, 18 people (28.57%) have less than 5 years and more than one year of service, 13 people (20.64%) have 6-10 years of service and 32 people (50.79%) have more than 11 years of service. It can be concluded that all respondents have enough experiences.

Descriptive statistics result
Descriptive statistics analysis is applied to 63 respondents who match with criteria. This measurement simplifies the observation, as described below:  (2017) Table 4.3 above shows there are 63 respondents took part in this research. The minimum and maximum scores explain the answer for indicators in the questionnaire. Respondents gave 1 for the minimum answer for ethics variable, whereas experience and competence variables, respondents gave 2 for the minimum answer. For the maximum score, respondents gave 5 for all three independent variables. Mean is used to find out the average opinion given by respondents on every indicator for each variable. Data above shows that mean scores for all variables are more than 4,00, which means that all respondents agree with the whole indicators. The mode shows the score that mostly appears in every variable. From the data above, most respondents gave 4 for every variable. The standard deviation score reflects a deviation measurement. Based on table 4.3, the standard of deviation scores are not more than the mean or median scores for every variable, which means that every respondent in this research have a good attitude of professional skepticism.

Measurement model (outer model) testing result
Data analysis of the outer model includes validity and reliability test. This whole test was executed by using Partial Least Squares (PLS) analysis with SmartPLS 3.0 application. A validity test could be known from the value of convergent validity, discriminant validity, and average variance extracted (AVE). Scoring in convergent validity test is based on average variance extracted (AVE) value > 0,50 and outer loading value > 0,70. However, at the scale of the development stage, loading value 0,50 to 0,60 is still acceptable (Ghozali and Latan, 2015). The researcher took the value of 0,50 as the minimum value for outer loadings. Furthermore, in determining whether the discriminant validity of the construct is sufficient, it could be known from the value of cross-loadings by comparing the correlation of indicators of a construct with other constructs. .1 shows the output of the initial measurement model among constructs. It could be seen that some constructs had outer loadings < 0,50. These indicators could be removed from their construct because they were not loaded to the represented construct. Furthermore, re-estimation was executed on the initial measurement model by removing the indicators that did not meet the criteria of outer loadings > 0,50.  (2017) After removing the indicators that did not match with the outer loadings > 0,50 criteria, the next step was re-estimating the measurement model. The result of the final measurement model is shown in the following figure: Figure 4.

Final Measurement Model (Outer Model) Output
Source: Processed primary data (2017) From the result of the final construct measurement model above, it is known that all the indicator variables have matched the criteria of convergent validity, whose outer loadings must be > 0,50. Another parameter to assess the convergent validity is checking the average value of variance extracted (AVE).  (2017) Based on table 4.4 above, it can be seen that all four variables have an AVE value > 0,50. This shows that the four variables have a good convergent validity and prove that the measuring instrument accurately measures the variables. Based on table 4.5 below, through the outer loadings, it is finally known that all the indicator values of each construct are above of the criteria (> 0,50) so it can be stated that all indicators and variables match the criteria, so it can be declared valid.  (2017) Furthermore, the discriminant validity test can be seen by comparing the square root of AVE for each construct with the correlation among the constructs in the model. The model has sufficient discriminant validity if the AVE root for each construct is greater than the correlation among constructs (Ghozali and Latan, 2015).  (2017) Discriminant validity can also be seen from the value of cross-loadings by comparing the indicator correlation of a construct with other constructs, as in table 4.7 below:  (2017) Based on the value of cross-loadings above shows that the correlation of construct indicators for each variable has a higher value compared to other indicators. Thus all indicators show good discriminant validity.
The last step of the outer model analysis is the reliability test. The reliability test is executed by using two methods, namely Cronbach's Alpha and Composite Reliability. The construct is considered reliable if the value of Cronbach's alpha and composite reliability is above 0.70 (Ghozali and Latan, 2015).  (2017) Based on table 4.7 above, it is known that the four variables have the value of Cronbach's alpha and composite reliability above 0,70 which means that all constructs in this study have a good reliability.

Structural model (inner model) testing result
The researcher tested the structural model by measuring the Coefficient of Determination (R 2 ), which is a goodness-fit model, and Path Coefficient (ß) (Ghozali and Latan, 2015), with this following structural model output:  Table 4.8, it can be seen that the value of R2 from the variables X1, X2, and X3 to variable Y is 0,597. This means that the effect of the variables X1, X2, and X3 on the variable Y is 0,597 or in other words, the research model of variation of the dependent variable changes that can be explained by the independent variable is 59.7%, while the remaining 40.3% is explained by other variables outside the research model. The value of R2 is said to be good if it has a value > 0.10.
Then to see the significance of influence between variables can be done by checking the value of the coefficient path and significant value of T-statistics in the following table 4.9. The construct is said to have a strong relationship if the path coefficient value > 0,10.  (2017) Based on Table 4.9 above, the relationship among variables can be explained as follows: -Path coefficient from Ethics (X1) to Skepticism Professional Auditor (Y) is 0.306 and it can be said that the construct has a strong relationship; -The path coefficient from Experience (X2) to Skepticism Professional Auditor (Y) is 0.177 and it can be said that the construct has a strong relationship; -The path coefficient of Competence (X3) to Skepticism Professional Auditor (Y) of 0.508 and it can be said that the construct has a strong relationship.

4.3
Hypothesis testing result Hypothesis testing is executed by using the bootstrapping procedure, using PLS analysis. Then to see the significance of hypothesis support is using the comparison of T-table and T statistics. If the value of T-statistics is higher than the T-table value, then the hypothesis is supported or accepted. The researcher uses a 5% of significance level so that the T-table value used is 1.64. Based on the value of the path coefficient in Table 4.6, it can be obtained the result of hypothesis testing result as follows: -Variable X1 to Y has T-statistics value of 2,368 > T- of an auditor's conduct for every audit task. Auditors who have ethics compliance and act based on auditors' code of ethics will be having auditors' professional skepticism in carrying out every audit task.
Due professional care, one dimension of ethics compliances, is one of the basic attitudes of auditors' professional skepticism. AAIPI (2014) states that auditors must use their due professional care carefully in every intern audit task. Pflugrath et.al. (2007) suggests that the compliance of the code of ethics has a significant influence on the professional accountant's auditing decisions. Based on the path coefficient from ethics to skepticism professional auditor, it is known that the construct has a strong relationship, which means that ethics influences auditors' professional skepticism. The more compliance of code of ethics, the higher auditors' professional skepticism. The less compliance of code of ethics, then the less auditor professional skepticism. Code of ethics rules every conduct and attitude. This finding is in line with cognitive dissonance theory about formation and change of attitude The result of this study is in line with the research result conducted by Enofe et.al. (2015), Oktarini and Ramantha (2016), Silalahi (2013), Suraida (2005), and Winantyadi and Waluyo (2014) which reveals that ethics positively influence auditors' professional skepticism.

The influence of experience on auditors's professional skepticism
Based on the hypothesis testing result, it is concluded that hypothesis 2 (H2) is supported. This means experience has a positive influence on auditors' professional skepticism. The result of this study is in line with the research result conducted by Oktarini and Ramantha (2016) and Silalahi (2013). The result of this study is also following the research result conducted by Anugerah et.al. (2012), Gullkvist and Jokipii (2015) and Hurtt et.al. (2013) which suggested that experience positively influenced the auditors' professional skepticism.
Unfortunately, some indicators must be removed from the outer model and only remain one indicator. The only indicator used in this research said that experience was an important element in the audit. Based on the researcher's limited knowledge, other indicators did not quite reflect the attitude of auditors' professional skepticism. Nevertheless, the experience still has a positive influence on auditors' professional skepticism. Experience helps an auditor in predicting and detecting problems encountered when auditing. Hurtt et.al. (2013) explains that the experience of an auditor can improve the attitude of auditors' professional skepticism. The more experienced, the more an auditor can generate assumptions in explaining the audit findings (Libby and Frederick, 1990). Pflugrath et.al. (2007) stated that greater general-auditing experience will significantly improve the auditor's judgments. Based on the path coefficient from experience to skepticism professional auditor, it is known that the construct has a strong relationship, which means that experience influences auditors' professional skepticism. The more experienced, the higher attitude of auditors' professional skepticism. And vice versa.
A total of 50.79% (32 people) of the total respondents in this study (63 people) have been auditors in BPKP for more than 11 years. Experience in auditing various types of audits and auditees can certainly improve the attitude of auditors' professional skepticism. Besides, experience also helps an auditor to be able to cope with various types of problems that exist during auditing. To strengthen this finding, the researcher interviewed The Head of BPKP Representative of Lampung, Sally Salamah. Sally explained, "when auditors have more experience in auditing, there will be two trends of auditors. The first, the more experienced, the more expert auditors would be. The more expert, the higher curiosity. It means the expert and experienced auditors will have a higher attitude of auditors' professional skepticism. The second, the more experienced, the more careless, which means those auditors will have a tendency to underestimate the task given due to experiences they have".

The influence of competence on auditors's professional skepticism
Based on the hypothesis testing result, it is concluded that hypothesis 3 (H3) is supported. This means competence has a positive influence on auditors' professional skepticism. Based on the path coefficient from competence to skepticism professional auditor, it is known that the construct has a strong relationship, which means that competence influences auditors' professional skepticism. The more competence, the higher attitude of auditors' professional skepticism. And vice versa.
Competence may include good quality of personal auditor, reasonable general knowledge, and also appropriate specialized skills in carrying out the audit task. Broad-minded curiosity is a part of the personal quality of an auditor. Understanding of Financial Accounting Standards (SAK) and Professional Public Accounting Standards (SPAP), as well as the ability to conduct analytical reviews, becomes part of the reasonable general knowledge. Competence is discussed in the section "Ethics Principles" and "Audit Standards" on KEP-005/AAIPI/DPN/2014 on the date of the 29th of April, 2014 about Pemberlakuan Kode Etik Auditor Intern Pemerintah Indonesia, Standar Audit Intern Pemerintah Indonesia, dan Pedoman Telaah Sejawat Auditor Intern Pemerintah Indonesia. It shows the importance for government internal auditors to have competence within themselves. AAIPI (2014) states that an auditor is expected to apply and uphold the ethical principles, which are integrity, objectivity, confidentiality, competence, accountability, and professional behavior. General Standard of Indonesia Government Internal Audit Standard clearly states that auditors must have the education, knowledge, skill and expertise, experience, and other competencies needed to do their responsibilities (AAIPI, 2014).
The certificates obtained by auditors at BPKP after attending the education and training related to the audit assignment add their special expertise. Although competency, thus, cannot be observedthough it can be inferred from task performance, or individual attributes, or both (Institute of Internal Auditors (1999) in Seol et.al., 2011). The result of this study is in line with the research result conducted by Silalahi (2013), Suraida (2005) and Winantyadi and Waluyo (2014) which revealed that competence has a positive influence on auditors' professional skepticism. Baybutt (2014) found that the competence of an auditor was essential to ensure audit quality.