Abnormal Audit Fee And Audit Qualiy: A Moderating Effect Of Firm Characteristics

The aim of this research work is to investigate the moderating effects of client characteristics on the association between abnormal audit fee and audit quality. Fifty-two quoted companies were randomly selected for study and the periodical scope of the study is fifteen years (2001-2015). Ordinary least regression technique was employed to analyse the data collected from field. The results revealed that abnormal audit fee has no significant relationship with audit quality. The result also showed that client complexity and client profitability have negative moderating influence on the relationship between abnormal audit fee and audit quality. The study further revealed that client risk has no significant moderating influence on the audit quality and abnormal audit fee relationship. The study recommended that statutory bodies should further securitize the audited financial reports of firms that declare huge profit.


INTRODUCTION
Auditing is as old as civilization and can be linked specifically with Greek and Babylonian civilization (DeAngelo, 1981).Auditors were hired to protect the interest on the owners and check the activities of the business on behalf on the owners.However in modern times the auditor are saddled with more responsibilities.Auditors are needed as check mechanisms to curtail the excesses of management (DeAngelo, 1981).Audit as scrutinizing device implies that it is a mechanism that enhances the dependability of the financial reports prepared by managers.Despite the engagements of auditor for quality assurance there have been audit expectation gaps-users expectations have not been met.Anecdotal evidence shows that some corporate entities went under shortly after they were certified financially healthy by the auditor (for example, Enron, Worldcom) DeAngelo (1981) argues that the inducements for auditors to give up their autonomous judgments are interconnected to the worth of the client.The worth of a client is normally calculated as the auditor remuneration paid by a given client divided by the summation of auditor remuneration collected from entire clientele in a given accounting year.An auditor has to decide whether to mortgage his independence by expressing an unqualified opinion in the presence of poor earnings quality, in return for quasi-rents in order to retain a key clientele that is possibly managing or manipulating earnings or to give unbiased judgment in order to protect his brand name.
The consistent demand for audit services and the amount charged in audit market are the reflections of energy exerted by the auditor on the audit work and the moral hazards connected to audit engagement (Simunic, 1980;Choi et al. 2006Choi et al. , 2010)).The various prices for audit services that prevailed across the audit market essentially are pointers to the differences in audit efforts exerted by auditors and the client-specific risk.The employment of audit pricing as a gauge for measuring the degree of dependence on client can initiate nontrivial measurement errors on the correlation between auditor's remuneration and the quality of audit except the cross-sectional disparities in costs of effort and the risk of being sued are properly under surveillance.It is unlikely that the extraneous relationship that exists between assorted calculated audit fee and qualitative audit reported in extant literature are driven by the failure to put into cognizance some vital externalities rather than by the lack of an underlying relationship.Prior studies employed anticipated abnormal auditor remuneration a long side with actual auditor remuneration metrics to investigate the audit quality and abnormal audit fee relationship.
The audit reforms that took place post-enron era gave insight to the fact that are some external factors that determine the extent to which abnormal audit influences audit quality.Some studies (Carcello, Hermanson, Neal & Riley, 2002;Hay et al., 2006;Tsui, Jaggi & Gul, 2001) show that corporate governance is a veritable moderating factor that determines the effect of excess audit pricing on audit outcome.. However nothing was said about factor within the client's quarter.Factors in the client quarter are known as client specific characteristics and they include: client profitability, client risk, client complexity and client size.
This study majorly focuses on the moderating influence of specific client characteristics on audit quality and abnormal audit fee relationship.The introduction of firm characteristics as a moderating variables makes this study novel among prior studies.To the best of our knowledge this is the first Nigerian study to use client characteristics as moderating variables to moderate abnormal audit fee and audit quality relationship.

Audit quality
Audit quality is certainly a multifaceted and vague concept.This concept has generated lots of controversies but not much has been understood about the subject matter.Over the past two decades, many scholars have tried to give an exact definition of audit quality.Nevertheless, consensus has not been reached by scholars on the exact definition of the subject matter.This is due to fact that audit quality depends on opinions and also these opinions are functions of whose stand points we are looking from.Bazerman, Loewenstein and Moore (2001) argue that the result of an audit is often unobservable and for this reason they attempted to explain that audit quality from that which it is "not" quality audit.They employed indirect but quantifiable proxies to represent audit outcomes.They further reported that abysmal audit quality is caused by recasting previously prepared accounts, existence of anomalous accumulations, or when there are irregularities in the audit exercise.De Angelo (1981) defines audit quality as the probability that accounting restatement or misrepresentation will be reported when discovered.She stresses that audit quality depends on the ability of the auditor to spot out errors in the reports placed before him by manager and his capacity to act independently as perceived by stakeholders.Teoh and Wong (1993) define auditor quality as perceived credibility of earnings reports.Palmrose (1988) defines audit quality with respect to the degree of reassurance because the reason for an audit is primarily to give an assurance that an accounting statement is free from significant errors.In fact, this definition employs the outcome of the audit, which is, reliability of financial statements and reflection of true and fair view.Titman and Truman (1986) define audit quality also in respect to the precision of reports given by the auditor to shareholders.AICPA (1992) opine that quality audit to a large extent is a function of personality built on integrity and honesty being the cornerstone.Arens, Elder and Beasley (2012) opine that audit quality is the auditor's capacity to spot out and disclose significant errors in the financial reports prepared by management.The spotting out feature is an indication of auditor's proficiency, whereas disclosure is an evidence of high moral standard or truthfulness of the auditor and his autonomy in expressing audit opinion.

Abnormal audit fee
Choi,Kim and Zang (2010) define abnormal fees as the variance amid actual audit fees paid to auditors and the anticipated normal level of audit fees.From a broader perspective, abnormal audit fees can be seen as "client-specific quasirents".Chung and Kallapur (2003) opine that charging the client certain amount as audit fee will entice the auditor and becloud his objective reasoning about a particular client.The authors further explained that audit fee is made up of two components, namely: normal and abnormal.The first component reflects the actual fee paid by a client.It is calculated by the differential of residual values for all variables that clients have in common, like size, complexity and risk.The second component shows the excess auditor remunerations that are peculiar to an auditor-client relationship.Krishnan, Zhang and Sami (2005) report that abnormal audit fee is a measurement for under or over payment for audit services.Ahmad et.al (2006) assert that the mixed results gotten from the impact of positive and negative abnormal audit fees on audit quality from prior put the subject matter in the accounting spotlight in recent times.Auditors are more likely to give in to pressure mounted on them by a client that pays excessively high audit fee.Auditors are also likely to compromise audit quality when a client pays abnormally low audit fee.Hope, Kang, Thomas and Yoo (2009), Choi et al.(2010) and Frankel et al. (2002) measure abnormal audit as the residuals values from a regression of total audit fees on variables that affects audit fee.Positive residuals connote overpayment while negative residuals imply underpayment.However estimation may drastically affect the value of abnormal audit fee if later method is used

Monitoring hypothesis
The monitoring hypothesis explains agency relationship.Jensen and Meckling (1976) postulate that cost of agency is probably a response to principal-agent discrepancy and this can be resolved by engaging the services of an independent auditor.Wilson (1983) opines that the supervisory responsibility of an auditor is to curtail moral hazard and undesirable problem of adverse selection that may occur due to information asymmetries.He further explains that managers are in a vintage position that give them access to more privileged information on the operations of firm.This in turn enables them to manipulate information for their own personal gains.In other words, the independent auditors are expected to guarantee stakeholders that the information given by managers is reliable.It is important to know that the auditor can be demanded for by the owners (monitoring cost) or by management (bonding cost).Humphrey (1997) argues that agents can demand for an independent audit and when they do so the agent/audit relationship will thrive on bonding cost.The implication of this is that agents will pay high auditors remuneration to induce the auditor to conceal vital information.This is scenario is common when a firm declares artificial huge profit to attract potential investors.Based on bonding cost model this study predicts that profitability will negatively influence excess auditor remuneration and audit quality relationship.

Reputation hypothesis
Reputation hypothesis gives an explanation for the purchase of the diverse levels of audit quality.Moizer (1992) a proponent of this hypothesis asserts that the inability of sellers to build a reputation creates two major agency major problems namely: problem adverse of selection and problem arising from moral hazard.Moizer (1992) further argues that in a situation where buyers are unable distinguish one level of audit quality from the other, they may presume that all audit services in market are the same hence offer the same amount for all available audit services in the market.Auditors cannot in any way convince the buyers to acquire his services in preference to others.This situation will definite lead to moral hazard because audit firms will freely sell substandard product for meagre amount to be enable them maximize profit.

Audit quality, abnormal audit fee and client specific characteristics
The argument in recent times in post Enron's era is that abnormal audit fee on its own does determine audit quality (Sawan & Alsaqqa, 2013).There are some situations where firms went under few months after they were declared to have clean audit reports.Francis and Yu (2007) suggests that there are some inherent factors in the client's quarters that influence the direction of the audit quality and abnormal audit fee relationship.It is perceived that auditee profitability determines the amount charged by the auditor for audit services rendered.This implies that higher profitability will attract high audit fee hence high audit quality vice versa.Anecdotal evidence shows that firms that made losses sometimes pay higher audit fee in order to buy clean audit reports.
Chen and Elder (2001) using ROI as proxy for profitability to ascertain the effect of profitability on audit quality and abnormal audit fee relationship.The results reveal that profitability has significant moderating influence on audit quality and abnormal audit fee relationship.Stanley (2011) investigates the link between observed auditor remuneration and profitability employing data collected from US public company engagements for seven year.The result shows an inverse relationship between clients' operating performance and abnormal audit fee while audit quality is compromised in the long run.Joshi and Al-Bastaki (2000) perform a study in Bahrain to ascertain the influence of profitability on audit quality and excess audit pricing relationship.The result shows that firms that declare huge profits are usually subject to high degree public scrutiny and auditors of such firms have to carry out audit testing of such clients' revenues and expenses with all dexterity in order to maintain their integrity and reputation.This will demand much audit efforts, which also will translate into higher audit fees.Much effort will lead to improved audit quality Another factor that emanates from client's quarters is client risk.It is believed that riskier firms pay higher audit fee but whether this translate into higher quality has become a contentious issue over the years.Events in the Nigerian banking industries in the past twenty years revealed that banks with high portfolio at risk (PAR) sometimes pay very high audit fee which at times does not translated into high audit quality (Abubakar et al., 2005 ).Sometimes in the bid to keep the client the audit firm collects abnormally low audit fee from banks with high PAR.This has also become a controversial topic in auditing discourse (Abubakar et al, 2005).Fields et al. (2004) investigate impact of client risk on auditor remuneration using standard audit fee model to measures of risk and complexity.Their result shows that banks with large volume of various banking activities, high level of risk, high level of inefficiency and high level of Portfolio at Risk (PAR) pay higher auditor remuneration.They also find that financial institutions that engage in hire purchase and firms that are statutorily obligated to keep high level of risk are charged higher audit fee.Walker and Casterella (2000) using data from companies in the United States, found that auditors are doing all they can to manage their exposure to audit risk arising from auditee's risk by adjusting the amount they charge as audit fees.Consequently they found positive correlation between these two.
Asare, Haynes and Jenkins (2007) examine the relationship between client risk and excess auditor remuneration and it influence the decision of the auditor.Their result reveals that more effort is exerted by the auditor to check statement of financial reports of a client with high risk.This implies that excess audit fee and audit quality relationship is positively influence by client risk.Previous studies failed to put into consideration the fact that audit is done in an ecosystem and such the factor within the system invariably will affect the audit outcome.

METHODS
The population of this comprises all listed firms between 2001 and 2015 excluding banking institution because of their peculiar nature (the proxy for audit quality for banks is abnormal loan loss provision).The population of this study is 168 companies quoted on the floor of the NSE as December 31, 2015.Yamane (1967) scientific technique was employed to determine the sample size.This technique was considered due to its lucidness and it is seen as the furthermost and widely used scientific method for calculating sample size.The formula is given as: Where n = the sample size n= Population e = level of precision (error limit on the basis of 10% confidence level).n = 168/ 1 +168 (0.01) 2 =52 companies Based on the above, a sample of 52 firms chosen randomly from the filtered population of 168 companies.

Data Source
The study gathered data from secondary source.Data kept in the archive will be employed for this study.Data were extracted from audited financial statement firm selected for 2001-2015:

Model and measurement
AQ= f( Abnormal auditfee* firm specific characteristics)

FINDINGS
The result using OLS technique shows that there is a negative relationship between RISK (-0.056) and DAC.This implies that risk has no significant effect on audit quality at 5 % level of significant (p=0.2).The impact of COMP on DAC is also negative (-0.051) but this impact is also not significant at 5 % (p=0.12).This infers that audit quality is not significantly impacted by client complexity.PAT (6.84E) on the contrary, has a positive effect on DAC.This effect is not significant at 5% (p=0.19).This implies that PAT does not have significant effects on audit quality.ABAFEE (4.58) has a negative but not significant influence on audit quality at 5% (p=0.94).Client risk (RISK*ABAFEE) as a moderator has a positive (-5.69) influence on the relationship between abnormal audit fee and audit quality at 5% (0.07).Firm complexity (COMP*ABAFEE) as a moderator, negatively influence (3.63E) the relationship between abnormal audit fee and audit quality but this influence is not significant at 5% (p=0.062).Profitability (PAT*ABAFEE) as a moderator, negatively and significantly influences (8.35) the audit quality and abnormal audit fee relationship at 5% (0.000).
The model parameters are as follows; coefficient of determination (R 2 ) = 60%, ADJ R 2 = 56%.These values suggest that the dependent variables explained about 60% of variations in audit quality.The F-stat=18, P(f-stat) = 0.00 and D.W=1.3.The F-values reveals that the can hypotheses use to establish linear relationship among the variables and cannot be rejected at 5% level while the D.W statistic infers that the presence serial correlation presence in the residuals is unlikely.
The robust estimation was also conducted using the fixed and random effects estimations.The results for the fixed effects estimation reveals that RISK has a positive (-0.68) but insignificant relationship with audit quality at 5 % level of significance (p=0.59).This infers that client risk as a single variable has no significant influence audit quality.The influence of COMP is also positive (-0.035) but not significant at 5% (p=0.24) this is in line with a priori expectation, which predict that COMP will positively influence audit quality.The effect of PAT on audit quality is negative (5.72E) though not significant at 5% (p=0.3).The effect of ABAFEE as a single variable is negative (3.38E) though not significant at 5% (p=0.55).
Client risk (RISK*ABAFEE) as a moderating variable has no significant influences (-3.90E) on the relationship because audit quality and abnormal audit fee at 5% (p=0.17).Client complexity (COMP*ABAFEE) as a moderator negatively influence (4.93E) the audit quality and abnormal audit fee relationship at 5% (0.002).This implies that more subsidiaries lead to poor audit quality when abnormal audit fee is charged.Profitability (PAT*ABFEE) as moderator, negatively influences (5.03E) the relationship between abnormal audit fee and audit quality at 5% level of significance (0.0001).This implies clients that make huge profit pay abnormal audit fee, and audit quality is lowered this is contrary to reputation theory that presupposes that auditors in bid to protect their brand name will exert more audit efforts on firms that declare huge profit because these firms are subjected to extra scrutiny by stakeholders and regulatory bodies.
The model parameters are as follows; coefficient of determination (R 2 ) = 66%, Adj R 2 = 63%.These values suggest that about 64% of variations in audit quality is explained by the independent variables and moderating variables The F-stat=23.9,P (f-stat) = 0.00 and D.W=1.5.The F-values reveals the fitness of the model hence can used to establish the relationship between the dependent at 5%, while the D.W statistic infers that the presence of serial correlation in the residuals is unlikely.

CONCLUSION
The robust regression result using the fixed effects estimation reveals client risk has no significant influence on audit quality.This implies that the present of client risk has no significant effect on audit quality.Furthermore, the result shows that client risk a moderating variable has no significant influence on the audit quality and abnormal audit fee relationship.This result is at variance extant positive gotten by Asare, Haynes and Jenkin (2007).Consequently the null hypothesis that client risk has no significant moderating influence on the relationship between abnormal audit fee and audit quality is not rejected.
Furthermore the robust estimation results for the fixed effects estimation reveals that firm complexity has as single variable has no significant influence on audit quality.This implies audit quality does not depend on the number of foreign branches and subsidiary a client has.This is at variance to with anecdotal evidence which presupposes the more complex a firm is the higher the chances of auditor committing material error.The result further shows that complexity as a moderating variable has negative influence on the audit quality and abnormal audit fee relationship.This implies that with firm many foreign branches has large volume audit work, which attract high audit fee and in-turn lead to low audit quality.Consequently, null hypothesis that client complexity has no significant moderating influence on the audit quality and abnormal audit fee relationship is rejected.
Finally the robust estimation results for the fixed effects estimation reveals client profitability as a single variable has no significant influence on audit quality.This result is contrary Naser et al (2007) which shows that profitability and audit quality are positively related.The result further shows that profitability as a moderating variable negatively moderates audit quality and abnormal audit fee relationship.This result is at variance with Joshi and Al-Bastaki (2000) which reveals that profitability has a positive influence the relationship between abnormal audit fee and audit quality.It is also at variance with Chen and Elder (2001), which shows that profitability has no significant influence on audit quality and abnormal audit fee relationship..This result is in line with Stanley (2011) which shows that profitability has a negative influence on the relationship between abnormal audit fee and audit quality.Consequently the null hypothesis that client profitability has no significant moderating influence on audit quality and abnormal audit fee relationship is rejected.
The study recommended that statutory bodies should set up a committee that will further scrutinize the books of quoted companies that declare huge profit.
The study suggests that researchers who intend to go into this area of study in future should replace the moderating variable client specific characteristics with auditor attributes.Furthermore, the study suggests that researchers who intend to veer into the subject matter in future should focus on the banking sector only.
A major limitation of this is with the proxy adopted to capture audit quality.The study assumes that discretionary accruals is an appropriate measure of audit quality and is also inversely related with audit quality.Despite the widely accepted use in prior accounting research, discretionary accruals is often criticized as a noisy proxy for the quality of the audit conducted.Though abnormal audit fee was computed using an audit fee estimation model that appears to be well specified and in line with the results of prior audit fee studies, we cannot rule out the possibility of an unknown degree of model misstatement; due to endogeneity and correlated omitted variables.Finally, our sample composition is based on quoted non-financial firms.Therefore, the results cannot be used for generalization for banking sector.

Table 3 . Regression Assumptions Test
From table 3, it is observed that the variance inflation factor (VIF) that measures the degree of collinearity among the variables also reveals the extent of the deviation of each variable.Fundamentally, VIFs above 10 call for concern.RISK reported a VIF of 1.04;COMP (1.02); PAT (1.05); ABAFEE (2.6).Since, the VIFs of the variables are less than 10, it implies multicollinearity is unlikely amongst the variables and hence the variables can be analysed to establish the relationship ,among dependent variable and independent variablesThe ARCH test for heteroskedasticity was carried out as a precautionary measure.The results reveal likelihoods of excess of 0.05.hencewe reject the presence of heteroskedasticity in the outstanding.The Lagrange Multiplier (LM) test for higher order autocorrelation shows that the hypotheses of zero autocorrelation in the residuals were not rejected.This is because p-values (Prob.F, Prob.Chi-Square) are greater than 0.05.The LM test did not, show serial correlation problems for the model.. Ramsey RESET test reveals large pvalues that larger than 0.05, which implies miss-specification problem is not emblematic.