Methodology and hypothesis.
This chapter describes the methodology of the study, by firstly provide information about the sample selection and data collection. It is then followed by providing specifications of the different measurements, lastly describing the hypotheses development and explanation of the logistic regression model.
Data for this study has been gathered by using a quantitative approach through observations of annual reports and statistics (Bryman 2012). The study selects of 49 company-year observations during the years 2005-2018. 19 of the observed company-years are for the period when the company committed the breach, 15 are from one year post the breach and 15 are from the year prior to the breach. The observed companies are from 8 different industries and are listed on either Nasdaq Stockholm or Nasdaq First North. Information for which company got sanctioned and the reason for it was manually collected from the Nasdaq Stockholm webpage where documents about which companies got sanctioned and the reason for the decision regarding the sanctions are available.
The necessary data in order to measure audit independence were collected manually from each of the sanctioned company’s annual report for the given year. Data regarding the audit firm revenue was collected manually from allabolag.se. However, all necessary numbers were not available on allabolag.se, and in those cases the data was collected directly from the audit firms’ annual reports. In some cases, older annual reports were not available online and were thereby collected by contacting the audit firm by email for the necessary information. To investigate the hypotheses the data were later used in a logistic regression model.
This study focused on companies that were listed on Nasdaq Stockholm and Nasdaq First North, which received either a warning or sanction by the Nasdaq Stockholm Exchange during the period 2008-2018.
The sample selection for this study consisted of 5 different steps. The first step was to identify all sanctions or warnings that were issued during the selected time period of 11 years (2008-2018). The amount of sanctions/warnings that were issued for each year is explained in table 1. Table 1 further presented that during these time period, there were in total 95 sanctions or warnings issued by the Exchange. That is an average of around 8,6 decisions per year, and as seen in table 1 most decisions were issued during 2016 (18). However, a big majority of the breaches were of a nature that was not relevant for this study, such as the sanctioned company published information on the wrong channel or failed to provide information in the right time to the Exchange.
Therefore, step 2 in the sample selection was to identify those sanctions or warnings that were of financial nature, such as accounting errors or incorrect financial statement. 0,832 of the total sample were excluded, due to the fact that they were not based on breaches of financial kind, such as accounting errors or incorrect financial statements. The total sample of 95 sanctions got reduced to 16 breaches which presents the final sample this study investigated. The distribution of final sample is presented in table 1, explained which years the sanctions or warnings were issued and how many of them were of financial nature and also that 2014 was the year which had the most breaches that were of financial kind (4).
Measure for Audit independence
As mentioned in section 2.1, the definition of audit independence is the likelihood of that the auditor discover a breach and report it (DeAngelo, 1981) and there are four main threats towards audit independence;
Client importance, 2. Non-audit services, 3. Auditor tenure and 4. Client affiliation with audit firms (Tepalagul & Lin, 2015). However, as explained in section 2.1, the threat of client affiliation is not as great as the other three elements. Therefore, it is a variable that is not commonly used when measuring audit independence, thus the proxy variables that was used in this study to determine audit independence will be Client importance (CI), Non-audit services (NAS) and Audit tenure (AT).
1.4 Structure of the thesis
2 Literature Review
2.1 Audit independence
2.5 Agency Theory
3 Methodology and hypothesis.
3.2 Sample selection
3.3 Data collection
3.4 Measure for Audit independence
3.5 Control variables
3.7 Hypotheses development
3.8 Logistic regression model
4 Empirical Findings
4.1 Descriptive statistics for variables used in logistic regression
5.1 Suggestions for Future Research
6 References .
GET THE COMPLETE PROJECT
Does audit independence have an impact on issued sanctions?