The History Of The Agency Theory Accounting Essay

Published: Last Edited:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.


This chapter describes the general theories aligned with the research. The research attempts to focus on the Agency theory, Stewardship theory and the Stakeholder theory which create a sound theoretical framework.

The Agency theory discusses the relationship between the agent and the principal and issues arise while working towards the same objective.

The Stewardship theory on the other hand is a substitute view of Agency theory, in which managers are expected to act in their own self-interests at the outlay of shareholders.

Stakeholder theory suggests that the drive of a business is to generate as much value for stakeholders. According to Drucker (1990), today corporations have to balance their responsibilities towards society in which they function, while meeting the ever increasing demands of their stakeholders to be competitive in a global economy.

Agency Theory

The concept of Agency theory is an important economic theory of accountability is defined as 'one or more people appoint another person to perform a service on their behalf which involves assigning some decision making authority to the agent' (Jensen & Meckling, 1976). Although this is mainly destined to be the relationship among the shareholders and directors, it has also said by Jenson & Meckling that there are other deals which could be measured within the framework such as directors and other interest groups and such relates to realise the role of external auditors.

The allocation of authority includes the belief among the agent and the principal whereby the Agency theory takes a view that people cannot be trusted to act in the public good and in the interest of the shareholders. This has resulted in rise of monitoring and controlling of the act of agents and mechanisms such as the external audit to strengthen this trust.

External auditors are subject to provide a reliable and a liberate opinion on the financial statements to the company shareholders but at the same time they are expected to work autonomously.

The separation of management and ownership leads to the problem of agent and principal. This is further explained by Watts and Zimmerman in 1978 by the positive accounting theory to find motives as to why managers cook their own books. In practical world, this is considered where the management have a full switch on the capitals of shareholders where they tend self-serve by maximising their own pockets.

When understanding the operating model of listed companies in Sri Lanka the most important basis of Agency theory is that the managers are usually motivated by their own personal gains and work to exploit their own personal interests rather than considering shareholders' interests and maximising shareholder value. For example, managers may be attracted to buying lavish offices, company cars and other extravagant items, since the cost is borne by the owners. However the concept of auditing and control carried out by a reputed audit firm will ensure that the operating model comes under proper scrutiny.

Thus, the key difficulty indicated by Agency theory is ensuring that managers pursue the interests of shareholders and not only their own interests. Eisenhardt (1989, p. 58) explains that agency problems commence when "The goals of the principal and agent conflict and it is difficult and costly for the principal to verify what the agent is actually doing".

Controversy occurs because principals are unable to monitor the performance of agents (Jensen & Meckling 1976). The pursuit of self-interest increases costs to the firm, which could include the costs of the formation of contracts, loss due to decisions being taken by the agents and the costs of observing and controlling the actions of the agents. Leuz et al. (2003) assert that the effects of such behaviour ultimately reflect in the company earnings. It is now evident that the investment in an audit needs to be apportioned to ensure that the performance of the agents or managers is monitored and proper feedback is provided.

It is also important that the management has an reason to manage the company's reported earnings in order to meet or beat earnings targets and, thus, to receive any bonuses that may be tied to the company's earnings (performance-related pay). This creates an information asymmetry in that managers can exercise the discretion they have on accruals, which in turn reduces the relevance and reliability of reported earnings, and the whole financial statements. Davidson et al. (2004) argues that when management provides inaccurate financial reporting information, it introduces earnings management as a type of agency cost. To counter this situation an internal audit may not be sufficient and the importance of an external audit must be conducted to understand the variances and changes that happen in the operating model of a company. The strict monitoring of managers by the principals or their representatives, such as the firm's board, is seen as fundamental to protecting shareholders' interest from being compromised when managers maximise their self-interest at the expense of the organisation's profitability.


In order to limit agency costs triggered by the separation of ownership and control, Fama and Jensen (1983b, p.309) propose that firms need a system that can separate decision management from decision control. This would limit agency costs by controlling the power of management and ensuring the proper consideration of shareholders' interests. It is clearly mentioned that the role of the external auditor is to reduce agency costs by cutting information asymmetry in financial reporting (Poit, 2001). This constitutes the importance of external audit and how it can positively impact the transparency in financial reporting.

Agency theory recognises external auditing as the most important monitoring mechanism because it controls conflicts of interest and diminishes agency costs. Watts and Zimmerman (1983) confirm that high quality external auditing will undermine the opportunistic behaviour cost (agency cost) introduced by management. 

These monitors act on behalf of the shareholders. As a result, high audit quality involving specialised independent auditors like the Big 4 audit firms can decrease opportunities for managers to pursue self-interest at the expense of owners and, thus, principals obtain more favourable returns. Hence the agency theory can be looked pon as a basis to develop theories for determining the relevance and composition of audit fees for companies public quoted companies in Sri Lanka.

Stewardship Theory

Unlike the Agency theory, the Stewardship theory, based on a psychological and sociological approach, maintains that the interests of corporate executives (as stewards) are aligned with those of the organisation and its owners (Albrecht et al., 2004). The stewardship theorists focus on structures that empower and facilitate rather than monitor and control. They reject the highly individualistic model of Agency theory that promotes a suspicious "policeman's" approach, thinks that agents are egocentric and whereby both the principal and agent share unalike interests. This is a theory that contradicts and challenges to a certain extent the concept of external auditing for all intents and purposes of monitoring management and employees of firms. The theory also goes on to reject the view that principals need to invigilate the opportunistic agents by monitoring them and apply sanctions or incentives as means of control.

Stewardship theory takes an opposite perspective in suggesting that the agents are trustworthy and good stewards of the resources are entrusted to them, which makes monitoring unnecessary (Donaldson, 1990; Donaldson & Davis, 1994; Davis et al., 1997). Since managers are not opportunistic and act in the finest welfare of the owners, they should also be given independence grounded on belief, which reduces the cost of governing their conduct. This means that according to the Stewardship theory managers are considered loyal and their conduct does not need to be scrutinised. However Donaldson and Davis (1994, p. 51) witness, "organisational role-holders are regarded being motivated by having to gain responsibility and then to gain appreciation from colleagues.

In most listed companies in Sri Lanka we see a significant proportion of dividend paying companies (i.e. Hayleys PLC, John Keels Holdings and etc.). From a Sri Lankan perspective a dividend paying company is considered stable and attractive to most investors. As per the Stewardship theory, the conduct of the steward is co-operative, because the steward looks out to accomplish the organisation's goals (e.g. profitability). This, in turn, benefits the principals through the positive effects of profits on dividends and share prices (Davis et al. 1997). In most local companies managers believe that their interests are aligned with those of the firm's owners. Thus, Stewardship theory maintains that the optimum governance structures are those that enable effective coordination in the enterprise. The stewardship perspective sees directors, as well as managers, as stewards of the. Davis et al. (1997) suggest that stewards gain more satisfaction by achieving corporate objectives than chase behind to fulfil their own pockets.

Davis et al. (1997) argue that achieving organisational success also satisfies the personal needs of the stewards. Thus, the stewardship theory considers that managers' decisions are also influenced by non-financial motives, such as need for achievement and recognition, the intrinsic satisfaction of successful performance, and respect for authority and the work ethic. However it is interesting to analyse, if this would be more prominent as opposed to the financial incentives a manager would visualise. This may well contradict the Stewardship theory in the Sri Lankan context.

However, according to this theory, it makes sense for corporate governance to be based on the view that the directors, on behalf of stakeholders, want to be good stewards of the corporate assets, and there is no conflict of interest or opportunistic behaviour at the expense of stakeholders. They work diligently to gain high levels of corporate profit and shareholder return. These concepts have been documented in organisational studies, such as in Muth and Donaldson (1998). To this effect the importance of an audit procedure may seem far-fetched and the value proposition may not be as clear, if this theory were to be considered.


When understanding the impact of the Stewardship theory from a level of the board, it considers the board of directors as an instrument of assistance to a steward chief executive officer rather than a controlling mechanism (Albrecht et al., 2004). It also considers that management is less likely to practice earnings management. However, the problem lies in the extent to which the management aspires to attain a good corporate performance. This could also take a turn where chief executive officers' may aspire for personal incentives & benefits over broader shareholder expectations.

The concept of identifying key people and drivers is common in most local companies. This enables a selected bunch of mangers to take on more responsibility and authority to drive a firm towards success. Davis et al. (1997) recommend that key managers identified within the firm leads to a personal relationship with achievement or failure. Daily et al. (2003) reason that management personals also want to guard their statuses as proficient decision makers. As a result, managers run businesses in a manner that increases performance in terms of profitability, while gaining a good return on shareholder funds.


The Stewardship theory goes on to argue that the firms which are managed by expert managers allow them to make the most of profits. Also, the dominance of executive directors on board is much desirable as they are capable in making better decisions due to their expert knowledge, dedication and accessibility to information (e.g. Boyd, 1995). In the listed companies a committee is appointed to make decisions of the audit practices both internal and external. This would mean that the committee would be a cross section of managers, directors and the board who would collectively take decisions on behalf of the audit practice within the company and ensure best practice.

In summary, the Stewardship theory argues that the responsibility and authority of executive managers provides a better focus on company objectives, leadership and implementation of operational decisions, leading to more effective corporate governance and corporation. Donaldson & Davis (1994) contend that the Stewardship theory remains the theoretical foundation for better regulation and legislation in corporate governance. Muth and Donaldson (1998) relate the likelihoods of Agency theory with those of Stewardship theory and treasure backing for the latter as a good model of reality. The question of how important this this theory is for determining external audit is somewhat unclear. Yet it adds a basis from which external audit is to be determined in the context of the audit practice and good governance.

Stakeholder Theory

The stakeholder theory revolves around the concepts that managers should on the one hand manage the corporation for the benefit of its stakeholders in order to ensure their rights and the participation in decision making and on the other hand the management must act as the stockholder's agent to ensure the survival of the firm to safeguard the long term stakes of each group. However the definition of a stakeholder, the purpose and the character of the organisation and the role of managers are very unclear and contested in literature and has changed over the years.

Bisset (1708) primarily defined the term stakeholders as "a person who holds the stake or stakes in a bet". Freeman (1984) quotes as "any group or individuals who can affect or is affected by the accomplishments of the organisation goals" and supports that the firm's actions should be anticipated in a wider outlook and also it should be useful to all relevant stakeholder rather limiting it only to the share owners. Moreover it includes the safeguarding of the benefits of employees, shareholders, customers, government, society, suppliers and prospective shareholders. The feature of "enlightened Stakeholder theory" has been suggested by Jenson (2001) as a result of incapability to protect the stakeholder interests in a business. This proposes that shareholder return growth is interlinked and therefore initially the company should maximise the returns of shareholders and that would routinely lead to the profit maximisation of stakeholders as a whole.

In answering the stakeholder prospects directors have a larger deal of accountability on meeting their prospects whereas auditors provide an assurance on the information provided by the directors through statutory audits. The concept of a Stakeholder theory around 60 years ago was proposed by Mary Parker Follett (Schilling, 2000) and it re-emerged in the 1980's. Freeman (1984, p.52 quoted in Schilling 2000, p.225) defines a stakeholder as "any group or individual who can influence or is influenced by the achievement of the organisation's objectives". The term 'stakeholder' may, therefore, include a large group of participants, in fact anyone who has a direct or indirect 'stake' in the business (Carroll 1993, p.22 quoted in Schilling 2000, p.225).

In a Sri Lankan context it is important to identify direct stakeholders and what their impact is on the performance and operating models of a firm. This is very important in the context of the listed companies since shareholders are particularly concerned about the direct stakeholders of a listed entity. Direct stakeholders are shareholders, employees, investors, customers and suppliers whose interests are associated with the organisation. An example of an indirect stakeholder is the government, which is indirectly affected by the company's role (Kiel and Nicholson 2003).


In a nutshell the Stakeholder theory supports that "companies and society are co-dependent and that the business aids a wider social drive than its duties to shareholders" (Kiel & Nicholson, 2003a, p. 31). Donaldson and Preston (1995, p. 85) recognise stakeholders as "persons or groups with genuine interests in practical aspects of corporate activity". Wheeler and Sillanpaa (1997) identify the stakeholders that needs to be taken into attention in the governance structure as investors, managers, employees, customers, business partners (e.g. suppliers and subsidiaries), local communities, civil society and the environment. Mitchell et al. (1997) claim that stakeholders can be identified by a few characteristics of:

Power to impact on the firm

the legitimacy of relationship with the firm

the urgency of their claim on the firm

This allows managers to pay attention and react to several forms of stakeholders.

Upon gaining strength during the 1970s and 1980s the Stakeholder theory would reflect the fear at a societal level that large national corporations were becoming too powerful and beyond accountability to stakeholders, including governments. With the time, the topic raised increased social awareness and the tendency toward triple bottom line reporting. In the context of Sri Lankan companies it is vital that this theory be looked upon critically. The reason for this is that the Colombo Stock Exchange is dominated by a few large conglomerates and ownership rests with the government and a few high net-worth investors.

Given the monopolistic nature of the market the Stakeholder theory is more of an addition of the agency theory, which believes boards of companies to take care of the benefit of shareholders. However, this fine focus on shareholders has been prolonged to take into account the interests of many diverse stakeholder groups (Freeman, 1984; Donaldson & Preston, 1995; Freeman et al., 2004). However, contrary to this Margaret Blair (1995) argues that, although stakeholder theory has more substantial historical roots, practical applications, and intellectual appeal than agency theory, it has had much less impact on thinking and policy-making about corporate governance.


Additionally it is important that the managers' incentives may not necessarily be associated with the interests of shareholders. However, managers who claim that this is due to consideration of other stakeholders' objectives "may be using stakeholder claims as a smokescreen to obscure what is really their inability to deliver value to the company's shareholders" (Healy 2003, p.24). Freeman et al. (2004) suggests that managers should attempt to produce as much value as possible for stakeholders by solving clashes among them so that stakeholders do not withdraw the deal.

The theory summarises external audit as effective monitoring systems that could protect all stakeholders' interests. In terms of audit quality, Baker et al (2002) suggest that the role of external auditor as monitoring mechanisms is not only directed for shareholders' benefit, but also for the interests of all Stakeholders. This change in the role of the boards has directed to the improvement of Stakeholder theory. Stakeholder theory can be seen as not necessarily supporting the view that amplifying shareholder value is the top priority for a business. Managers may legitimately follow objectives that do not increase shareholder wealth. Finally Sundaram and Inkpen (2004a) argue that shareholder value strengthening matters because it is the only objective that leads to decisions that boost results for all stakeholders. They claim that recognising a large number of stakeholders and their main morals is an impractical duty for managers. Supporters of the stakeholder viewpoint also argue that shareholder value maximisation will lead to an expropriation of value from non-shareholders to shareholders. This discussed the importance of the Stakeholder theory as an extended version of the Agency theory and lays a distinct foundation of theoretical research.


Among the findings it can be concluded that the Agency theory recognises external auditing as the most important monitoring mechanism because it controls conflicts of interest and diminishes agency costs. The allocation of authority includes the trust between the agent and the principal, however the agency theory takes a view that people cannot be trusted to act in the best interest of the shareholders. This has given rise to the need of monitoring and controlling of the act of agents and therefore need to put in place mechanisms such as the external audit to strengthen this trust. The Stewardship theory argues that the effective control held by professional managers empowers them to maximise the firm's performance and corporate profits. Subsequently, boards which are conquered by executive directors are desirable because of their proficiency and awareness of subject matters, access to critical information and commitment to the firm. Stakeholders' theory views external audit as effective monitoring systems that could protect all stakeholders' interests. Stakeholder theory goes on to propose that the stress of managerial activity should be on the evolution and conservation of all stakeholder relationships, not just that with shareholders. Based on the theoretical framework it is obvious that the role played by the external auditor in the listed companies is one of the great sensitivities and needs further research for developing an audit fee model.



This chapter mainly illustrates the philosophical framework which describes the research strategy, the process, data collection and research approach. Once the methodology is defined it then develops the empirical model in line with the hypotheses by identifying the independent variables (i.e. size, complexity and risk). These independent variables are analysed against the dependent variable which is the audit fees.

Research Philosophy

An essential step in conducting social science research is to determine and justify the chosen research philosophy adopted by the researcher.

Research philosophy has two paradigms, these are positivistic and the interpretative paradigms (Hussey and Hussey, 1997; Patton, 1990). These paradigms represent the end of a continuum in social science research which illustrates the links between these ontological, epistemological and methodological assumptions. Research philosophy can also be largely alienated into three views (i.e. positivism, interpretivism and realism) subject on the researcher's philosophical thinking (Saunders et al. 2003).

The philosophy of positivism desire 'working with an visible social reality and that the end product of such research can be law-like generalisations similar to those produced by the physical and natural scientists' (Remenyi et al. 1998 cited in Saunders et al. 2003, p.83). This model adopts that an apprehensible realism exists that is driven by absolute natural laws and mechanisms (Guba and Lincoln, 1994, p. 109). This model is defined by quantitative or scientific styles to clarify, forecast and analyse testable hypotheses relating to associations between measurable variables since it assumes that reality is `objective' or independent of observers.

The interpretative paradigm, on the other hand, assumes that reality is `subjective' or dependent on observers as they are actually part of what is being observed (Patton, 1990). This paradigm is therefore described by qualitative approaches. Interpretivism has a variety of different names including hermeneutic, qualitative, phenomenological and inductive research (Ticehurst and Veal 1999, p.20).

Realism approach recognizes the 'social reality' like positivism but it emphases on variation and conflict which may not always seem. It is more concerned with opinions, norms and ethical values, thus researchers need to understand history, implement a set of values, and know where to look for fundamental structure to interpret the facts (Neuman 1997). Therefore, the realistic paradigm recognises the subjective reality and attempts to understand people's socially constructed interpretations at the same time (Saunders et al., 2000).

This research will adopt the positivism approach due to its relevance to this type of research. Positivism is taken to indicate the following ideologies:

Knowledge confirmed by the senses can sincerely be acceptable as knowledge.

The purpose of theory is to create hypotheses that can be tested.

This research is hypotheses based; and these hypotheses will be tested using quantitative methods.

Knowledge is arrived through assembling of evidences that provide the basis for laws.

The research uses a random sample of companies and collects facts from their annual reports to test the hypotheses.

Science must be led in a way that is value free.

Sample selection will be by systematic random sampling method; and data will be tested to see if there are any relationships between the variables using correlation techniques.

Research design

The deductive approach tends to be preferred more by positivist researchers than interpretivist (Ticehurst and Veal 1999). The deductive research procedure includes the development of a theory or hypothesis to test the hypothesis. The inductive approach is used when data is collected first, and a theory is developed as a result of the data analysis.

The deductive research approach is also built on the universal idea to reach at the precise state and it is linked with the positivism model, whereas, inductive approach works over a specific idea to generalise the situation as per the research topic, is linked with the interpretivism paradigm (Crowther and Lancaster 2009).

Positivism is linked with systematic, tentative, quantitative and deductive frameworks where researchers seek specific quantifiable observations thus often using statistics and experiments to test their hypotheses (Neuman 1997). Thus, this research uses a deductive approach where the general information is worked down towards more specific situation which is to determine factors affecting audit fees.

Research process

Upon identification of the research design, the next step being the research process is to identify ways of producing the knowledge. This takes three different forms; empirical research, constructive research and exploratory research.

Empirical research

Empirical research is constructed on pragmatic and measured phenomena. It reports researches based on real observations or trials using quantitative research methods and it may produce mathematical data between two or more variables. Some of the purposes of the empirical research are to go beyond simply reporting observations and promote an environment for a better understanding while combining wide-ranging research with detailed case studies too.

Constructive research

Constructive Research covers the area of theory, and doesn't need that the research be based on firmness. It is more based upon case studies, theories and hypotheses. It is used to test theories and suggest solutions to a problem or question.

Exploratory research

Exploratory research is important to research methods as it helps to outline a new problem or query. Exploratory research is conducted for a problem that has not been clearly defined.

The data was collected or produced in a way consistent with the recognised practice in the field of study. When choosing a research approach the empirical, constructive and exploratory approaches were considered. Exploratory research is conducted for a problem that has not been clearly defined. This is not consistent since the approach involves a continuous process of inquiry and investigation that is likely to be open-ended and may produce indefinite results. Constructive research on the other hand is used to test the theories and propose a solution which is not in line with the outcome of the research. Further the research is based on quantitative data and statistical analysis.

It is logical that the empirical research approach be used to collect new data on the proposed topic which is fairly understood. The data collected is quantitative and suitable for analysis using statistical methods. Hence empirical research approach was selected as it tests the feasibility of a solution using empirical evidence which is appropriate to the objectives of the study. This is supported by Simunic 1980) who established an important "optimistic model of the course by which audit fees are determined". Simon et al (1986) found similar results similar to previous research in India. Again, Size and Complexity were key determinants of audit fees.

Research type

The two main data gathering methods discussed are primary research and secondary research.

Primary Research

There is no data available and data needs to be gathered from primary sources such as questionnaires and surveys.

Secondary Research

Secondary research gathers existing information through available information sources such as published records, annual reports, journals and prior research papers.

It is remarkable to note that primary research is costly to conduct as it comprises primary sources and the time taken to conduct such is typically extensive when compared to the time taken to perform a secondary research. For all intents and purposes secondary data has been selected for this study as it is valid, audited, clearly documented and easily accessible.

All data contained within in this study is achieved from published annual reports which were found on Colombo Stock Exchange website for the years 2009, 2010 and 2011. The companies were selected on a random basis spanning all twenty sectors (i.e. Banks, financial institutions and insurance, Food beverage and tobacco, diversified holdings, hotels, plantations, IT, energy, oil palms, manufacturing, etc.)

The process involved an examination of the balance sheet, income statement, auditor's independent report and the director's report to identifying the required variables.

Research approach

Research methodology is well-defined by Leedy & Ormrod (2001) as "the general method the researcher takes in carrying out the study".

Two main methods of research approach are quantitative and qualitative.


Quantitative research is conclusive in its purpose as it tries to quantify the problem and understand how prevalent it is by looking for projectable results to a larger population. Quantitative research involves the collection of data so that information can be quantified and subjected to statistical treatment in order to support or refute "alternate knowledge claims" (Creswell, 2003, p. 153). Quantitative research approaches use a standard layout, with a few minor inter-disciplinary alterations, of creating a hypothesis to be verified or not. This hypothesis must be demonstrable by mathematical and statistical means, and is the basis around which the whole study is designed.


Qualitative research has no clear intention to count or quantify the conclusions, which are instead describes in the language employed during the research process (Leach, 1990). The main objective of qualitative research is to gain an understanding of underlying reasons and motivations and provide insights into the setting of a problem for later quantitative research.

However in quantitative approach the findings are definite and usually expressive in nature. Quantitative research design is an outstanding way of finalising results and validating a stated hypothesis. The approach will calculate data and generalise results from a sample to the population of interest and is the preferred approach for this study.

5.3 Development of Hypotheses

In identification of the hypotheses, it is influenced by a body of scholarly work that has reviewed audit fee determinants. Overall past scholarly work & prior research (e.g. Simunic, 1980 and Taffler & Ramalingam, 1982) has identified that there are two key factors "size" (measured in balance sheet and/or income/expense terms of the entity under audit) and "complexity" (assessed in terms of the scope and related risk of the various operations of the entity under audit) that influence the audit fee. Hence this research identified relationships mainly upon client size, client complexity and the risk faced by the client.

While there have been exceedingly many contributors to the literature on audit fee modeling, it is generally agreed that the seminal contribution has come from Simunic (1980). For this reason, the research by Simunic (1980) is explained in some detail in some of the following paragraphs. Later, that seminal contributions were supported by contributions from, Firth (1985), Palmrose (1986), Haskins et al (1988), Low et al (1990), Chan et al (1993) and Brinn et al (1994).

The three hypotheses are developed accordingly to answer the research question which is addressed earlier in this research. These three hypotheses will be tested using the audit fee model whereas the null hypothesis will be tested for the purpose of statistical significance by using graphical and statistical methods and these will be discussed in the next section and results will be examined in the latter part.

The three hypotheses are stated as follows:

H1 There subsists a positive relationship between the size of the firm and audit fees.

H2 There subsists a positive relationship between the risk and the audit fees.

H3 There subsists a positive relationship between the complexity of the client and the audit

Ho There subsits homogeneity of variance (homoscedasticity) in the data.

Measurement of the dependent variable

The Securities and Exchange Commission of Sri Lanka requires disclosure of audit fees and it is this disclosed figure, which is used as the dependent variable. This figure can be relied upon according to the Accounting body and the Colombo Stock Exchange. Thus with consideration to the above, an initial list of variables was extracted from the relevant financial statements.

Audit fees consist of the entire fees salaried by the company to its auditor be it domestic and overseas including all subsidiaries. It take account of all the expenses concerning to the audit work, but eliminates any other special work which is not related to the audit such as management consultancy services. The value of audit fees should be revealed in the company's annual report as auditors are required to reveal it for constitutional purposes, as stated in the fourth schedule of the Companies Act 2007.

Measurement of the independent variables

This section provides detailed information about the measurement of each independent variable. The variables are grouped into three main classes, where each incorporating discrete variable/s representing detailed attributes. The individual variables for each of these categories are discussed below.

Firm Size

The most dependable result in all the previous research has been the size of the company is by far the most substantial explanatory variable in determining audit fees. Several of these studies have used 'total assets' as a measure of size. This is mainly a balance sheet based measure.

Firms' Risk

The majority of audit firms in Sri Lanka base their auditing on some kind of audit risk model. There is an association between audit risk and audit fee. Higher the risk higher the audit fee either as a result of more audit testing or as an 'insurance premium' (Chan et al, 1993). Firm risk was measured by the two ratios; long term debt to total assets and current assets to total assets.

Firm Complexity

This variable is not particularly relevant in the case of Sri Lanka, as financial statements are not required to be consolidated. But this was still used due to the verification of schedules of subsidiaries. A much more relevant variable might have been the number of associated companies - and this could be considered within future studies.

Empirical Model

The empirical model with the main variables of this study is formed as follows:

AF it = α + β1 TA it + β2 QUICK it + β3 LTD it + β4 CATA it + β5 SUB it +

ε it

AF : Natural log of audit fees

TA : Total Assets

QUICK : Quick Ratio

LTD : Long term debt to total assets

CATA : Current assets to total assets

SUBS : Number of subsidiaries held by the individual company

Analytical procedure

This section will discuss the statistical methods employed for data analysis. The data sample used in this research was identified by systematic random sampling method. A sample of 95 companies out of a population of 285 across three years; 2009, 2010 and 2011 were selected in order to give a high degree of confidence in the statistical testing and will be adopting the regression equation in a generalised manner to the selected sample of 95 companies. The data for this study is analysed using the software package SPSS. Once the data was collected it was screened and cleaned, using SPSS functions to ensure the data was accurate and contained no input faults. In order to define the relationship between the audit fees and the identified independent variables; the proposed hypotheses need to be tested with the help of suitable and appropriate statistical tests. Therefore the choice for right statistical testing is vital.

The statistical methods for analysing data are classified into two broad categories: one is parametric and the other is nonparametric. In general, the nature and characteristics of the study data will determine which method should be used.

Gujarati (2003) suggests four critical assumptions that must be met before utilising parametric tests:

1) Assumption of normality: Under this assumption, samples must be drawn from normally distributed populations. A normal distribution is an idealised sample which is based on a population of an infinite number of cases and which takes the form of a bell.

2) Assumption of Linearity: This assumption suggests that the model should have linear parameters.

3) Assumption of homoscedasticity: This assumption requires the variance or standard deviation of the dependent variable within the group to be equal.

4) Assumption of independence of error terms: Under this assumption the error terms are independent from one another and therefore no serial correlation exists.

In general, parametric tests are more powerful when all assumptions are met and when the variables under analysis are measured on at least an interval scale (Judge et al. 1985). However, if any of the previously mentioned assumptions are violated by the nature of data; non-parametric tests become more appropriate (Balian, 1982).

Therefore, the assumptions of the parametric tests will be tested in the following chapter using as suggested by Mark (2008) through the graphical presentation of a histogram, Kolmogorov-Smirnov test and Shapiro-Wilk test and Normal Q-Q plot to check for the normality assumption. To test the homosedasticity assumption, the most common test is used, namely, visual inspection of the residuals. According to Mark (2008) the residuals are plotted in a graph against the independent variable that is suspected of causing the problem of heteroscedasticity. To test for multicollinearity, this study applies correlation coefficient and variance inflation factors (VIF) tests. The tolerance factor and variance inflation factor of each variable is then calculated. A tolerance factor close to 0, and a value of the variance inflation factor greater than 10, shows the presence of multicollinearity in the models and further to this Hair et al. (1998) and Kennedy (2008) suggests that a VIF of more than 10 indicates multicollinearity.

Given the above discussion, the various tests are conducted to test the data against the OLS assumptions. Parametric tests are adopted in this study to analyse as the data of this study does meet the conditions required for the parametric tests.

For the purpose of this study, it is felt that a linear regression model should be used. Out of the two basic linear regressions - simple linear regression and the multiple linear regression the researcher concentrated on the multiple linear regression. Simple linear regression is used to evaluate the relationship between a dependent variable and 'one' independent variable where as a multiple linear regression is carried out to assess a relationship between a dependent variable and 'two or more' independent variables. Therefore according to the nature of the dependent and independent variables, researcher has followed the multiple regression technique on SPSS which allows investigating how a set of independent variables do associate with a dependent variable of interest.


Selecting the appropriate research methodology and data collection techniques is a very critical stage in conducting any research because they ensure that the research objectives are achieved. This chapter has provided a detailed description of the steps taken to prepare for the analysis phase of this study which is discussed in the next chapter. These steps include illustrations of the research philosophies, dependent variable measurement, the measurement of the independent variables and the selection of the appropriate analytical methods.

To summarise the methodology described in this chapter, this research uses corporate disclosure data collected from annual reports of the companies which are listed in the Colombo Stock Exchange website to empirically test for the determinants of audit fees. Firms were randomly selected for this study and the years 2009, 2010 and 2011. This research will adopt the positivism approach and a time series approach is applied to test the hypotheses, and model is tested using pooled panel OLS regression.

The following Research findings chapter presents the results of the tests selected to analyse the data gathered according to the steps described in this chapter.