“Corporate Governance Code Compliance and financial performance: the case of Austrian stock listed companies”

This article analyzes the correlation between compliance to the Austrian Code of Corporate Governance and financial success of Austrian stock listed companies. It uses a sample of 52 Austrian companies that are listed on the Vienna Stock Exchange and corporate data collected from company publications such as annual reports, financial reports, corporate governance reports and company websites. Three accounting measures – return on assets, return on equity and net profit margin – were chosen in order to proxy the financial performance of a company. The period under review ranges from 2008 to 2016, whereas particular attention is given to the years 2010 to 2016. A corporate governance compliance score has been established on the comply or explain basis and recommendation rules of the Austrian Code of Corporate Governance in order to measure a company’s ability of implementing ‘good’ corporate governance practices. In line with research for other countries, this study finds no statistical evidence that a correlation exists between high compliance to the Austrian Code of Corporate Governance and financial success of companies listed on the Austrian Stock Exchange. The paper highlights the uniqueness of the Austrian Corporate Governance system when compared to other systems and gives arguments why companies comply with corporate governance recommendations.


INTRODUCTION
Corporate governance revolves around the question on how investors are getting a return on their investment (La Porta et al., 2000). The Organization for Economic Cooperation and Development (OECD) principles of corporate governance (OECD, 2004) state that corporate governance deals with the relationship between the management of a company, the directors (in the board), shareholders and other stakeholders. According to the principles (OECD, 2004), corporate governance also tackles the question of monitoring performance.
Cadbury defined corporate governance as being the system by which companies are directed and controlled (Cadbury, 1992). A shareholder's view definition on corporate governance suggests that it deals with the ways, in which company investors are rewarded a return on their investment (Shleifer et al., 1997).
SAHA Rating assessed Austria as being among the 22 countries, which scored best on their World Corporate Governance Index (WCGI) in The results attest that the framework of self-regulatory standards and corporate governance provisions provided by the Austrian Code of Corporate Governance (ACCG) is solid internationally. The ACCG supplements the Austrian law on stock markets and capital markets by a set of voluntarily comply or explain rules (companies either comply to a corporate governance rule, or -in case of non-complianceexplains the deviations) in order to support Vienna Stock Exchange (VSE) companies in the establishment of good corporate governance and control systems. This paper analyzes if there is a correlation between high compliance to the ACCG and financial performance of companies listed on the Vienna Stock Exchange. There is no similar research for Austria. The question is whether investors of Austrian Traded Index (ATX) companies can expect a greater return on their investment by being shareholder of a high-complying-with-the-ACCG Austrian listed company than being one of a 'low-complier'.
In a global context, one of the most cited works about the topic has been an analysis by Gompers et al. (2003), who developed a Corporate Governance Index (CGI) discussing shareholder rights and the effects on firm value. The methodology of Gompers et al. has been applied and adjusted in many studies with different results. This paper targets the correlation between compliance to a corporate governance code and the financial situation of Austrian companies listed on the Vienna Stock Exchange.
For that, a Corporate Governance Performance Score (CGPS) for each company is produced, which proxies the respective company's ability to apply practices of 'good' corporate governance within the company. Three accounting measures have been chosen in order to proxy the financial performance of a company: ROA (return on assets), ROE (return on equity) and NPM (net profit margin). All three measures were observed for the period of nine years (2008-2016), an average value has been drawn in order to include both years of financial turbulences at the beginning of the observation period and years of economic recovery. The same procedure was applied to a further period under review (2010-2016). The period under review ranges from 2008 to 2016, whereas particular attention is given to the years 2010 to 2016 in order to exclude the effects of the recessionary years 2008 and 2009. A sample of 52 Austrian companies listed on Vienna Stock Exchange has been evaluated in the analysis and the results have been discussed and compared to those of other studies.

LITERATURE REVIEW
A correlation between compliance with a code and financial performance has been widely discussed in literature. One approach to deal with the topic has been made by La Porta et al. (2000) who examined the effects of the different legal systems on shareholder protection and company value. They concluded that shareholder protection in common law countries is higher than in civil law. In a further study, La Porta et al. (2002) underpin their observation in a cross country comparison. A better investor protection led to a higher valuation of firms incorporated in common law versus companies from civil law countries. They based their analysis on Tobin's Q (La Porta et al., 2002). The correlation is found between independence of boards and ROA (Bhagat & Black, 2002), board committee composition on ROA (Klein, 1998), as well as the correlation between the use of external mechanisms and financials measured by Tobin's Q and ROA (Gompers et al., 2003;Bebchuk et al., 2009), between the size of boards and profitability measured by ROE (Sanda et al., 2006), and last but not least, between the frequency of board meetings and performance (Vafeas, 1999;Carcello et al., 2002), and between proportions of independent board directors and Tobin's Q (Yermack 1996). Ashraf et al. (2017) examined the impact of different corporate governance attributes on firm performance on the basis of a sample that includes 19 companies of the Pakistan cement industry. The results have been different ranging from a positive relation between large board size and ROA to a negative influence of firm size on ROE.
An analysis that also examined the influence of specific corporate governance attributes on firm performance was made by Obradovich and Gill (2013). A sample of 333 New York Stock Exchange (NYSE) manufacturing firms had been chosen, which provided a negative impact of board size on the value of American manufacturing firms (Obradovich & Gill, 2013). Along the same lines, Rouf (2011) stated that the benefits of monitoring are offset by the poor results of bigger groups.
Consequently, compliance to corporate governance indices has been discussed in various studies (Ntim, 2009) Another aspect was provided by Beiner et al. (2006), who used a comprehensive set of corporate governance provisions in a sample of listed companies on the Swiss Stock Exchange (SXW), excluding investment companies, to analyze the relationship. Beiner et al. (2006) found that the output of the established corporate governance index (CGI) and the chosen Tobin's Q as an indicator of firm performance supports those findings in the literature that identified a positive relationship between the above factors. Henry (2008), in a similar study for Australia, also supports the hypothesis on a positive relationship. Therefore, the non-evidence statement about developed countries by Arcot and Bruno (2006) cannot be applied to Switzerland and Australia based on the chosen parameters by Beiner et al. (2006) and Henry (2008).
All these approaches exemplify the importance of this topic and show the various levels of overlaps between the financial performance of companies and having a focus on corporate governance initiatives.  Some companies have introduced a fiscal year that deviates from the calendar year. The values of such fiscal years have been accepted for the calculations of an average value for both periods under review, a seven-year period and a nine-year period. Two companies have been assessed in regards to their CGPS, yet could not be included into the financial success analysis, as their publications do not encompass all the years of the period under review. Subsequently they have also not been taken into account for the correlation analysis. Two other prime market companies have not been assessed at all. One of the respective companies is French and consequently not relevant for this analysis due to the focus of this paper. The other company's first day of trading on the VSE was in 2014. Therefore, the company did not disclose any financial reports for the years before 2014. A continuous change of members in the mid-market led to only three companies, which have disclosed the necessary financial reports for all years under review. One of them had to be excluded from the analysis as it neither published a Corporate Governance report nor discloses issues, which are asked for by the CG report (rules of incorporations, shareholder structure, etc.).

METHOD
There have been companies who left the standard market within the period under review and others who had their first listing on it without publishing financial reports of the respective previous years. Again, one company could not be part of the analysis due to the fact of not being an Austrian company. Companies that were listed on the Vienna Stock Exchange's global or the other securities market have been excluded from the analysis as they do not commit themselves to the ACCG. In total, the analysis of the financial figures consisted of 35 prime market, two mid market and 15 standard market companies.
The Pearson correlation coefficient (PCC) has been used in order to discuss the research question and the hypotheses if a normal distribution of the values of the variables in question is given. The existence of normal distribution within the respective variables has been derived from the tests of normality by Kolmogorov-Smirnov and Shapiro-Wilk. In case of absence of normal distribution, Spearman's rank correlation coefficient test has been utilized. The variables, which have been part of the correlation coefficient tests, have been the CGPS (categorized in C-rule CGPS, r-rule CGPS and a combined value of both rules, the total CGPS) and the operating performance indicators (NPM, ROE, ROA).
In the first step, the tests were carried out on an ATX prime market level, respectively on a level for all evaluated companies, regardless on the market categorization. In the second step, the test was done on an industry-level for both periods under review. The industry categories that are given by the VSE (VSE 2018a) have been applied and adjusted. Consumer Products and Consumer Services have been combined to one industry (Consumer Products and Services). This adjusted industry as well as the categories Financials (with ROE and ROA only), Basic Industries, and Industrial Goods and Services make up the four industries, on which the industry-level analysis has been carried out. 'Utilities' and 'Technology & Telecom' have been excluded from the industry-level analysis as each category includes only two companies. Cohen's categorization of effect size has been used to interpret the coefficients of Pearson's correlation coefficient (PCC) and Spearman's rank correlation coefficient (SRCC) test: • there is no correlation if |PCC/SRCC| is below 0.10; • a low effect size is given if |PCC/SRCC| is 0.10 or higher but below 0.30; • a medium effect size exists if |PCC/SRCC| is 0.30 or higher but below 0.50; • a large effect size exists if |PCC/SRCC| is 0.50 or above.
The p-value of a 2-tailed significance test provides information if the correlation coefficient is significantly different from zero. There is no significant difference from zero if the p-value is > 0.05.

Financial performance parameters
The return on assets (ROA) is applied in order to measure the amount of profit that is generated by a company and expressed as a percentage of their average total assets: .

Net income interest on debt
Average total assets + The Net income corresponds to the profit of the profit-and-loss-statement that is distributed to the shareholders of the parent company after taxes and minority interests. In case of non-existing profitand-loss-statement-profits, the profit after taxes of the statement of comprehensive income is drawn upon. The interest on debt is drawn upon the companies' consolidated financial statement. The average total assets correspond to the balance sheet total. It is an average annual value of the opening stock (OP) from the opening balance sheet (OBS) and a final stock (FS) from the final balance sheet: from from . 2

OP OBS FS FBS +
The Return on Equity (ROE) is applied in order to measure the companies' operational performance and to determine the interest ratio of the capital invested by the shareholders within one year: .

Net income Shareholder s equity ′
The net income corresponds to the NPM's net profit. The shareholder's equity is compiled by annual average value of the opening stock (OP) from the opening balance sheet (OBS) and a final stock (FS) from the final balance sheet (FBS). from from . 2

OP OBS FS FBS +
In case of non-existing shareholder's equity, the group equity of the balance sheet is drawn upon.
The net profit margin (NPM) is applied in order to analyze the companies' ability of producing net profits from its net revenues. It is expressed as a percentage resulting from the following ratio: .

Net profit Total revenue
Net profit corresponds to the ROA's net income. The net profit margin could not be applied on companies of the financial industry, as banks, real estate companies, and insurance groups do not generate revenues that are comparable to those of the formula.

Corporate governance
performance score (CGPS) The corporate governance compliance checklist consists of 46 C rules and four R rules. Some rules encompass more than one question, which leads to 104 C rule questions, and eight R rule questions that are evaluated within the corporate governance compliance checklist.
Publications that have been taken into consideration for the check against the corporate governance compliance checklist are the corporate gov-ernance report of the year 2016, the annual report of 2016, the supervisory board report/letter, the corporation's website, the published external reports on the findings of the compliance with the C rules of the ACCG (if available on the website), the internal rules of procedure (if available on the website), the articles of incorporation/association, the signed declarations of supervisory board members and the quarterly reports 2016.
The section "Change of Significant Voting Rights Thresholds" on the website of the VSE (VSE 2018b) was visited in order to verify whether corporations that do not disclose changes to voting rights on their website did not have such changes in the reporting period respectively in the past years. L rules of the ACCG refer to mandatory legal requirements. Consequently, it is assumed that they have to be fulfilled by each corporation of the evaluation anyway. Therefore, L rules are not part of the corporate governance compliance checklist.
The term for the compliance degree of a company, which is used in this paper, is the CGPS. The CGPS arises out of the companies' accordance to C and R rules. One point was assigned for each fulfilled question. No point was assigned if a rule has not been discussed at all in any published document.
"Comply or explain" characterizes how C rules are treated in the ACCG. If a company does not comply with the C rule, the reason for being incompliant has to be explained.
Statements vary from being reasonable and comprehensive to statements that just question the necessity of the rule without providing any reason for being incompliant. Therefore, the model narrows down C rule deviations: a point for a non-complying-with-C-rule question can only be received if the rule is explicitly discussed in the corporate governance report. If the respective corporate governance report just mentions the existence of a deviation to the rule, half a point is assigned to the company. The whole point is awarded for well comprehensible explanations. A critical questioning of the respective rule as an explanation such as "For business policy and competition reasons, the object and terms of such contracts are not published in the Annual Report as stipulated in rule 49" (Agrana Beteiligungs-AG 2017a) or "The company does not have compliance with the C Rules evaluated by an external institution. Based on the company's circumstances, the Management Board and Supervisory Board do not consider it to be sensible to commission a company to perform such an evaluation" (S Immo AG 2017) does not lead to the "whole" point.
The mentioning of a deviation to the rule combined with a well comprehensible explanation cannot lead to more than one point per question. All questions of the corporate governance compliance checklist are respectively all achieved points, which are equally weighted in the CGPS.
The CGPS formula is given by: .

Achieved points in total Number of applied questions
Some questions have not been applied to all evaluated VSE corporations, as they depend on certain events. For instance, questions one and two regarding C rule 38 are not applicable if no management board appointment took place in 2016. Another example is question one regarding C rule 36, which asks for additional supervisory board meetings it necessary. Those rules are excluded from the assessment in such cases.

RESULTS
The  Hypothesis H0 cannot be rejected on the basis of Spearman's rank correlation coefficient test for AVROA and the three types of CGPS for the evaluated companies of all markets in both periods under review.
There is no statistically significant correlation between high compliance to the Austrian Code of Corporate Governance (ACCG) and financial success within all evaluated ATX companies based on the utilized parameters CGPS and AVROA.

Return on equity (ROE)
Spearman's rank correlation coefficient test has been applied between the C-rule CGPS and AVROE, as the first variable does not fulfil the criteria of normal distribution as discussed in 3.4.2.1. The output shows a statistically significant medium effect size for SRCC if the test is done for all evaluated companies for the period under review from 2008 to 2016. No correlation has been pro-vided by the same test for the period under review from 2010 to 2016. No further evidence about the relationship of the variables is given by the simple linear regression, as the results have not been statistically significant according to the corresponding ANOVA-tables of both periods under review.
Spearman's rank correlation coefficient test between R-rule CGPS and AVROE has been applied again as the values of all CGPS variables have not been normally distributed according to the tests of normality. There is no correlation between R-rule CGPS and AVROE in both periods under review according to the output of the applied tests. The results of the simple linear regression for these variables are statistically not relevant according to the corresponding ANOVA-tables. The all markets output for Spearman's rank correlation coefficient test between the total CGPS and the AVROE provides a low effect size of the correlation coefficient, which is statistically not significant in the period under review from 2008 to 2016. No correlation has been provided by the same test conducted for the period under review from 2010 to 2016. The results of the simple linear regression for the same variables do not have statistical significance according to the corresponding ANOVA-tables.
Hypothesis H0 cannot be rejected for both periods under review if Spearman's rank correlation coefficient test is done for AVROE and the three types of CGPS for the evaluated companies of all markets.
There is no statistically significant correlation between high compliance to the Austrian Code of Corporate Governance (ACCG) and financial success within all evaluated ATX companies based on the utilized parameters CGPS and AVROE.

Net profit margin (NPM)
18 'non-Financials' companies are excluded from the evaluation of this category.
Normal distribution does not exist for the values of the variables C-rule CGPS and NPM according to the tests of normality of Kolmogorov-Smirnov and Shapiro-Wilk. Spearman's rank correlation coefficient test for the variables C-rule CGPS and AVNPM for 34 'non-Financials' companies provides a statis-tically not significant low effect size of the correlation coefficient for both periods under review.
The statistically significant output of the simple linear regression for the same variables of both periods under review has to be seen critically as normal distribution does not exist for both variables. Spearman's rank correlation coefficient test between the variables of R-rule CGPS and AVNPM shows a low effect size of the correlation coefficient, which is statistically not significant for the period from 2008 to 2016. No correlation is provided by the test for the period from 2010 to 2016. No statistical significance is given for the results of the simple linear regression for the same variables according to the corresponding ANOVAtables. Spearman's rank correlation coefficient test between total CGPS and AVNPM shows a low effect size of the correlation coefficient, which is statistically not significant in both periods under review. The statistically significant results of the simple linear regression for these variables have to be seen critically due to the absence of normal distribution of the values within the variables. Hypothesis H0 cannot be rejected for both periods under review on the basis of Spearman's rank correlation coefficient tests for AVNPM and the three types of CGPS for the evaluated companies of all markets.
There is no statistically significant correlation between high compliance to the Austrian Code of Corporate Governance (ACCG) and financial success within all evaluated ATX companies based on the utilized parameters CGPS and AVNPM.

DISCUSSION
The findings of the empirical study highlight the uniqueness of the Austrian Corporate Governance system. A comparison with literature focuses on corporate governance that especially examines corporate governance performance via established corporate governance indices, provides mixed results for each financial performance measure.
Cengiz (2016) evaluated the correlation of Turkish companies' corporate governance ratings and five financial ratios (ROA, ROE, NPM, EPS, and MBV). Cengiz (2016) also examines the difference in finan-cial performance between listed companies in the Corporate Governance Index 'XKURY' and those listed on the BIST 100 without being part of XKURY.
According to Cengiz (2016), licensed rating agencies determine the corporate governance rating of a company by its compliance with Capital Markets Board of Turkey (CMB) governance principles. An overall corporate governance score of 7/10 and a minimum score of 6.5/10 in each of the four main sections (shareholders, stakeholders, public disclosure, directors) are the conditions for being indexed on the XKURY (Cengiz, 2016). Cengiz (2016) states that the financial performance of XKURY Index companies have been statistically significant higher in terms of ROA, ROE, and NPM than in non-index companies, whereas no significant differences existed in terms of earnings per share and market book value between the groups. Consequently, Cengiz (2016) concluded that an efficient adoption of corporate governance principles provides higher financial performance.
Another view is provided by Todorovic (2013), who examined if application of corporate governance correlates with higher share earnings and higher net profit margin. The study included an assessment of 19 companies of the Banja Luka Stock Exchange (BLSE) by applying a Scorecard analysis (Todorovic, 2013).
The achieved scores of the companies regarding corporate governance implementation, as well as the corresponding ratios of EPS and NPM were compared to the results of a similar analysis for companies listed on the Vienna Stock Exchange (Todorovic, 2013). Todorovic (2013) refers to Kaufmann (2004) regarding the overall Austrian corporate governance implementation score of 78.44 percent and provides an overall corporate governance implementation score of 53.66 percent for the companies of the Banja Luka Stock Exchange. The given NPMs of the study have been 3.04 percent for the companies of the VSE and minus 7.11 percent for the companies of the BLSE, whereas the given EPS has been 1.58 EUR for the VSE companies and 0.82 EUR for the BLSE. Todorovic (2013) concluded that the lower scores of BLSE companies in all three measures compared to the VSE companies indicates that companies with higher level of implementation of corporate governance principles and better prac-  Achim et al. (2016) stated that the strongest relationship (with a statistically significant correlation of medium intensity) has been identified between corporate governance score and ROA, whereas no statistically significance existed between corporate governance score and ROE within the sample, which underlines the results of this paper. Ntim (2009) examined the correlation of governance structures and financial results of companies listed on the South African Corporate Governance Index (SACGI). It consisted of 50 provisions that were categorized The respective company's annual reports provided the information on which Ntim's (2009) analysis has been based. Each fulfilled provision assigns one point to the evaluated company, no point is assigned to the respective company if the criteria of it has not been met leading to an SACGI score range from zero (0%) to fifty (100%). The SACGI was applied for a five-year (2002 to 2006) observation period and proxied the quality of internal organizations. ROA and Tobin's Q have been used as a measure for financial performance in the analysis (Ntim, 2009). Ntim (2009) concluded that firms with a better governance reported higher financial returns. Contrary to the findings of this paper, the results of the analysis provide a statistically significant positive relationship between the SACGI and firm financial performance according to Ntim (2009).
Another analysis is provided by Price et al. (2011), who established a corporate governance index (CGI) with the dimensions of board composition, committee structure, and transparency. Mexican firms, which were required to report on their compliance with the Mexican Code of Best Corporate Practices. As in this paper, the authors did not find evidence for a correlation between corporate governance and financial performance.
Other authors such as Arcot and Bruno (2006) suggest that companies that depart from the onesize-fits-all approach of corporate governance can outperform all others in genuine circumstances. More methodological work is, therefore, needed on the effects of particular parts of corporate governance on the financial performance of VSE companies as well as on the question whether companies that increased their compliance to the ACCG through the years could also improve their financial performance.

CONCLUSION
As discussed in this paper, the study did not find a correlation between high compliance to the Austrian Code of Corporate Governance and financial success of companies listed on the Austrian Stock Exchange.
This statement has to be limited to the utilized parameters of the model, which are the corporate governance performance score and the three accounting measures -ROA, ROE and NPM. Further limitations to the analysis have been the relatively small sample size, the inequality in the number of companies between the four industries on the industry-level analysis, the lack of a representative number of companies per industry, the manual assessment of the publications that on occasion includes decision-making on individual cases (which results in a certain amount of subjectivity), the chosen 'transparency and disclosure' approach (meaning that a rule can be fulfilled by a company without referring to it in a publication), as well as the one-size-fits-all assumption of the corporate governance compliance checklist in respect of the applied valuation-scheme.