An analysis of key audit matter disclosures in South African audit reports from 2017 to 2020

from KAM disclosures. Practical implications: This research will be relevant for standard-setters, regulators and users of audit reports interested in how ISA 701 is being implemented and the state of KAM disclosures in an emerging economy. Originality/value: Barring only some exceptions, relatively little is known about the auditors’ expanded reporting requirements in a South African context.


Introduction
The auditor's report has been a focal point for regulators and standard-setters in both developed and developing economies (Burke et al., 2021).One of the most recent developments effective for financial years ending on or after 15 December 2016 is the publication of ISA 701 1 by the International Auditing and Assurance Standards Board (IAASB), which requires auditors to report 'key audit matters' (KAMs).These are the 'matters that, in the auditor's professional judgement, were of most significance in the audit of the financial statements of the current period' (IAASB, 2016:para 8).
http://www.sajbm.orgOpen Access during an audit and how they were managed allows users of financial statements to understand how audits are conducted and assess the risks inherent in their investments more effectively (Christensen et al., 2014;Elliott et al., 2020;Kachelmeier et al., 2020;Ozlanski, 2019;Prasad & Chand, 2017).
In this context, several articles examine what factors determine the number of reported KAMs (e.g.Abdelfattah et al., 2020;Pinto et al., 2020;Rousseau & Zehms, 2020;Sierra-García et al., 2019), and the implication of expanded auditor reporting for legal liability (see, e.g., Asbahr & Ruhnke, 2019;Gimbar et al., 2016;Kachelmeier et al., 2020).With a few exceptions (e.g.Bédard et al., 2019; Duboisee de Ricquebourg & Maroun, 2022;Kend & Nguyen, 2020;Segal, 2019), this research has mainly focused on highly developed economies, such as the United States and the United Kingdom.Differences in the regulatory, economic and legal environments between developed and developing economies motivates the need for further studies to examine auditors' expanded reporting requirements within developing economies, which have been largely overlooked by the academic community (see, e.g., Duboisee de Ricquebourg & Maroun, 2022;Segal, 2019).This study contributes to the assurance literature by studying the disclosure of KAMs within such a setting.
South Africa adopted ISA 701 without amendments shortly after it was issued by the IAASB.The country boasts a wellestablished and regulated assurance market where audit firms have extensive experience applying international auditing and accounting standards and engaging with different external regulators (Atkins et al., 2020;Maroun et al., 2014).As a result, South Africa provides an excellent setting for examining expanded reporting requirements while providing novel insights into audit practice within a developing economy.
The article makes an important practical contribution by examining the nature and extent of KAMs included in audit reports within a developing economy.A detailed account is provided of the number and types of KAMs disclosed by South African auditors using a hand-collected sample of audit reports issued for 356 Johannesburg Stock Exchange (JSE)-listed entities for the four consecutive financial years ending between 2017 and 2020.Variations in these disclosures over time, among industries and among audit firms, are also considered.
The results of this study should be relevant to standardsetters, regulators and users of audit reports interested in how ISA 701 is implemented and the current state of KAM disclosures.The primary aim is to establish a qualitative understanding of KAM disclosures that resonates with both academics and practitioners rather than providing a detailed econometric analysis of KAM determinants.While this may be considered an inherent limitation of the study, a more interpretive review of KAMs focusing on both the nature and extent of information being reported responds to calls for practically grounded research on assurance practices, which complements, rather than competes with, studies emphasising methodological or theoretical considerations (consider Hay, 2015;Power, 2003).
The remainder of this article is structured as follows.The 'literature review' section provides a summary of the KAM literature.The 'methodology' section discusses the methodology used to collect and classify KAMs, which is then followed by the presentation and discussion of the findings in the 'results and discussion' section.Finally, the 'conclusion' section concludes, discusses implications and outlines areas for future KAM research.

Literature review
Recent corporate and auditing failures (see Berger et al., 2016;Jones, 2011) mean that the role of the auditor as a public watchdog is, once again, under intense public scrutiny (consider ACCA, 2019;Berger et al., 2016;IFAC, 2020;Maroun & Solomon, 2014;Porter, HÓgartaigh & Baskerville, 2012;Sikka et al., 1998), with additional measures being introduced to address an actual or perceived decline in the utility of audit services (Harber & Maroun, 2020;Knechel, 2016).The introduction of KAMs is one such intervention.The publication of ISA 701 is not a response to a specific audit failure but is an example of how the IAASB is working to enhance the transparency of the audit process and provide additional insights into the key financial and operating issues at auditees (Gold et al., 2020;IAASB, 2015a;Reid et al., 2019).
The standard was released to enhance the value of the audit report and users' confidence in the audit process (IAASB, 2015a).As observed by the IAASB chairman at the time: [A] standard to enhance auditor's reports is a significant step forward in providing useful and relevant information to investors and other users … Auditor's reports are no longer 'boilerplate' reports -auditors are now providing additional information about the audit, which is highly valued by users.(IAASB, 2017) Key audit matters should highlight key areas in an audit for users' attention, improve communication with management and stakeholders and renew the auditor's focus on the risk assessment and response process at the heart of contemporary audit practice (IAASB, 2015a(IAASB, , 2015b)).
ISA 701 uses a principle-based approach for determining KAMs.'Key audit matters are selected from matters communicated with those charged with governance' (IAASB, 2016:para 8).When evaluating if an issue reported to a governing body is also a KAM, the auditor considers (1) the risk of material misstatement, (2) whether or not significant estimates and judgements are involved and (3) any significant events and transactions which have taken place during the reporting period (IAASB, 2016:para 9).The KAMs, which are identified for this process, are then described in a separate section of the auditor's report (IAASB, 2016:para 11).This includes an explanation of why each matter was of most significance for executing the audit and how it was addressed (IAASB, 2016:para 13).If, however, there are no KAMs to report, then this must also be clearly stated in the audit report (IAASB, 2016:para 16).
Initial studies examining the effects of these audit reporting changes find that KAMs may have the potential to reduce information asymmetry and influence investors' decisionmaking under certain conditions (Christensen et al., 2014;Elliott et al., 2020;Ozlanski, 2019;Prasad & Chand, 2017).Moreover, KAMs may highlight additional risk factors (Kachelmeier et al., 2020), curtail the use of aggressive or overly optimistic accounting estimates (Gold et al., 2020;Reid et al., 2019) and increase confidence in some audit reports (Moroney et al., 2021).However, academics and practitioners have highlighted several challenges when it comes to applying ISA 701 and using the additional information being reported to stakeholders.

The benefits and challenges associated with key audit matters
In theory, KAM disclosures enhance the transparency of the audit process (Gold et al., 2020;IAASB, 2015a;Reid et al., 2019).This allows those charged with governance and other stakeholders to understand how auditors approach key areas in an audit (ACCA, 2018).For smaller audit firms, in particular, the use of KAMs signals improved auditor value and credibility (Moroney et al., 2021).In a business context, management will be aware of significant risk areas or those aspects of financial reporting which require the use of judgement.Increased auditor scrutiny should lead to more conservative decision-making, better application of financial controls and more reliable reporting to investors (ACCA, 2018;Asbahr & Ruhnke, 2019;Gold et al., 2020).Moreover, the process of identifying KAMs facilitates discussions between the auditor and those charged with governance, which could enhance audit quality because of a more rigorous risk identification and response process (ACCA, 2018;Segal, 2019).The increased stakeholder and public scrutiny of audit reports mean that auditors should be more focused on ensuring that audit objectives are appropriately met and documented (Kachelmeier et al., 2020).Ultimately, this may lead to a reduction in the audit expectation gap as users of audit reports obtain a better understanding of the audit function (Segal, 2019).
Yet ISA 701 also gives rise to several challenges.For example, KAMs may be misunderstood as areas of concern by the public, who do not distinguish between problems, frauds or misstatements on the one hand and significant risk areas that require additional audit work on the other (Kend & Nguyen, 2020).This is counterproductive to the goal of reducing the audit expectation gap and is the result of complex terminology and the fact that KAMs can be difficult for nonexperts to read and understand (Abdelfattah et al., 2020;ACCA, 2018;Segal, 2019;Velte, 2018).Audit firms may also provide boilerplate KAM disclosures that do not sufficiently explain the underlying risk assessment and testing processes (Abdelfattah et al., 2020;Segal, 2019).Key audit matter disclosures may also increase the threat of auditor litigation and lead to audit fee recoverability issues (see Brasel et al., 2016;Gimbar et al., 2016;Segal, 2019).Key audit matters may incorrectly be used by stakeholders as a substitute for analysing financial statements and assessing risk for investment decision-making purposes (Bédard et al., 2019).Finally, there is no guarantee that the KAMs disclosed in an audit report are accurate and complete (ACCA, 2018).The perception of what is a 'key matter' is subjective, resulting in inconsistent disclosures (Segal, 2019;Sierra-García et al., 2019).It is therefore important to understand the KAM determination process, as this can impact how the benefits from KAM disclosures are derived and how any challenges can be navigated.
This study contributes to the assurance literature by studying the disclosure of KAMs in South Africa.A descriptive analysis of the state of KAMs provides novel insights that academics, standard-setters, regulators and users of audit reports can use to better understand the context of KAMs in developing economies.

Methodology
South Africa adopted ISA 701 with effect for listed companies with financial years ending on or after 15 December 2016.The sample collection process commenced by collecting the audit reports of all JSE-listed companies for financial years ending on or after 01 January 2017 4 and ended 4 years later with the audit reports for the financial years ending up to and including 31 December 2020.The goal was to examine the KAMs reported by these companies and how they changed over the sample period.As a result, the analysis excluded all companies which were delisted or newly listed during the period, leaving a final sample of 356 JSE-listed companies with 1424 audit reports containing 2903 KAMs issued between the beginning of 2017 and the end of 2020. 5The period of review ensured that KAMs were considered from the earliest full financial year when all companies reported on KAMs (2017) to the most recently available published financial statements at the date of preparing this article (2020).The period covered, and the fact that all listed companies were considered to the extent practical, provided a large sample for examining KAM reporting in South Africa.

Collecting and categorising key audit matters
Qualitative content analysis was used to collect and analyse the data because of its suitability for dealing with material that is not consistently formatted, while highlighting trends 4.As a result, the 2016 financial year of South African-listed companies has been excluded from the analysis because not all companies early adopted the KAM requirements.Preliminary analysis of the data revealed that 54 companies reported 141 KAMs in their 2016 audit reports.The remainder did not.This is not considered to be material in the context of the total KAMs analysed.
5.The same 356 companies were analysed across the period.This was an inherent limitation, as certain KAM observations were lost by excluding entities; however, this ensured consistency and comparability in the information under review.Furthermore, each audit report over the 4 years was analysed.This may have led to certain reports having missing KAM data at the time of collection; however, this did not materially impact the results.Untabulated results indicate that inferences were not impacted by removing these companies from the analysis.
and investigating both qualitative and quantitative disclosures (Krippendorff, 2013).Each entity's audit report was read several times, focusing on the part of the report detailing the KAMs.For this purpose, each KAM was treated as the unit of analysis.The KAMs were reviewed to determine the issue being flagged by the auditor, how the auditor framed the underlying risks and the accounting for the balances and transactions under review.
The result was 2903 KAMs being coded into 50 subcategories.The subcategories were consistent with those highlighted by earlier studies (see Kend & Nguyen, 2020;Segal, 2019;Sierra-García et al., 2019) and the recognition, measurement and presentation or disclosure requirements articulated in the applicable International Financial Reporting Standard (IFRS).Details were recorded per firm and year.The 50 subcategories were then further grouped into one of 13 themes listed in Table 1, which were then differentiated between entity-and account-level KAMs.The former impact the organisation on a broad, entity-wide level, while account-level KAMs focus on specific issues with individual financial statement line items (Sierra-García et al., 2019).The examples are not intended to be exhaustive.
The classification of the KAMs into the subcategories was performed interpretively.Each KAM was assigned to one of the initial subcategories by a team of research assistants working in pairs.Rather than testing for intercoder reliability, the KAMs were classified independently by the research assistants, with differences observed and resolved.The preliminary classifications were then reviewed by the researchers for accuracy and completeness.Changes were discussed before being processed.As a final safeguard, the KAM classifications were examined, on a sample basis, by an experienced audit manager and audit academic. 6

Data analysis
Descriptive methods were primarily used to analyse the number and type of KAMs reported in total, by auditor, between industries (based on the JSE industry groupings) and over the 4 years of the sample period.The results are presented in the 'results and discussion' section.Nonparametric Mann-Whitney U tests, Kruskal-Wallis tests and Jonckheere-Terpstra post hoc assessments were used for confirmatory purposes.These nonparametric tests are less sensitive to sample-size effects and departures from normality.
The focus was on the number of KAMs reported in total and per type, with scaling for the number of clients per audit firm where applicable.These scores were treated as being at least ordinal.To provide additional insights, the readability of the KAMs was computed.The readability scores captured reading ease determined according to the Flesch Reading Ease Level.The results were corroborated using the 6.The KAM classification was piloted with a sample of 20 companies.This was used to refine the data collection instrument and determine coding buckets provisionally.As additional reports were analysed, the number of subcategories (code buckets) being used was increased.Previously coded KAMs were re-evaluated.This continued until a sense of thematic saturation was achieved, which was after the 51st company had been coded (for a single year).
Flesch-Kincaid Grade scores (untabulated).Although these use the same core measures (word and sentence length), they have different weighting factors.In the Flesch Reading Ease test, a higher score or result indicates that KAM disclosures are easy to read, while a lower score represents text which is more Finally, the reader's attention is drawn to the fact that a formal model for predicting the number and type of KAMs being reported has not been developed.This is in keeping with the article's exploratory objectives and practical focus.Generalisation of findings should be performed with caution, but the principles discussed in the 'results and discussion' section are applicable for a broad audience.

Results and discussion
The If KAMs can enhance the transparency of the audit process and reduce information asymmetry (Segal, 2019), auditors may be inclined to increase the absolute number of KAMs being reported.In keeping with a risk-based approach to assurance, auditors would also focus on the most complex and judgemental areas of a client's financial reporting (ACCA, 2018;IAASB, 2015b;Sierra-García et al., 2019), with implications for the type of issues which are being flagged as KAMs.Given that auditors become more experienced with KAM determination over time (ACCA, 2018) and that the South African assurance market is not as litigious as the American market (see Gimbar et al., 2016, dealing with auditor liability), this may also result in auditors including additional KAMs in their audit reports over time.
Yet Figure 1 shows that the total number of KAMs reported for the sample of 356 JSE-listed companies steadily declines across the 4-year period.The 2017 financial year reported the most KAMs, possibly because firms were dealing with new auditor reporting requirements and taking a more conservative approach when it came to classifying issues identified during engagements as KAMs.Since this year, the number of KAMs has steadily decreased so that, 4 years after the adoption of ISA 701, the sample of JSE-listed companies reported 8% fewer KAMs compared with 2017.As a result, the average number of KAMs reported per company has decreased from 2.11 in 2017 to 1.94 in 2020.Untabulated details show that over the sample period, 2% of audit reports reported zero KAMs, 53% reported one or two KAMs, 31% reported three to five KAMs, and 2% reported six or more KAMs.A total of 12% of audit reports had missing KAMs, either because of the audit reports being unavailable at the time of data collection or the report referring to entities in the broader group structure for KAM details.
As audit firms become more familiar with the application of ISA 701, they may be refining the number of material issues identified during an engagement, which are also regarded as KAMs, accounting for the decline in KAMs reported.Equally possible is that audit firms are standardising their disclosures and reducing the level of detail reported as a means of reducing the risk of regulatory inspections and legal exposure, particularly given the lack of material incentives to increase the number of disclosed KAMs (Brasel et al., 2016;Gimbar et al., 2016).
However, the overall decline in the number of KAMs masks a shift in the type of KAMs being reported by South African auditors over the sample period.Categorising KAMs according to the entity or account level shows that the overall decline in the number of KAMs is mainly because of fewer account-level KAMs being reported each year.While the number of entity-level KAMs increased by 14% over the sample period, the number of account-level KAMs dropped by 20% (H = 11.975,p < 5%) from 488 to 390 over the same period.The difference between the number of entity-and account-level KAMs suggests that as auditors' understanding of their clients and the requirements of ISA 701 improves, the core issues being flagged as 'key' are moving away from

Types of issues being flagged as key audit matters
Next, to provide additional insights, the KAMs are analysed according to their themes (Table 1), as shown in Figure 2.
Business combinations and goodwill (including goodwill impairment) have the highest frequency of KAM disclosures with 801 reported KAMs (28%).This is followed by 680 KAMs (23%) dealing with the accounting for nonfinancial assets, the main focus of which is the impairment of these assets.Next are KAMs concerned with the application (including impairment considerations) of IFRS 9 to financial instruments (410 KAMs, 14%).Employee benefits (4 KAMs, 0.1%), changes in accounting policies and errors (28 KAMs, 1%) and the accounting for inventory (73 KAMs, 2.5%) resulted in the fewest number of KAMs.
There are only 111 reported KAMs (4%) dealing specifically with going concern issues over the 4-year period, including the resulting impacts of coronavirus disease 2019 .This proportion appears low but, on closer examination, it was found that auditors discussed the effects of the global pandemic and related liquidity issues as part of other KAM themes, including the impairment of goodwill, nonfinancial assets and financial assets.
When comparing the results with other jurisdictions, Australian listed companies' audit reports included KAMs most frequently related to acquisitions, impairment considerations, valuation of goodwill and the valuation of noncurrent assets including property, plant and equipment and inventory (see Kend & Nguyen, 2020).This aligns with the results in this study, with more than half (51%) of all KAMs in South Africa dealing with business combinations and goodwill (including goodwill impairment) and the accounting for nonfinancial assets.The Australian KAMs also address revenue recognition as a common KAM theme (Kend & Nguyen, 2020).In a South African context, revenue is the fourth most disclosed KAM theme (182 KAMs, 6%), indicating that it is being raised with less prominence by South African auditors in contrast to other jurisdictions, where it is a more common theme (ACCA, 2018).Data from the United Kingdom, Brazil, United Arab Emirates, Kenya, Nigeria, Cyprus, Oman, Romania and Zimbabwe flag impairments as the most reported subject matter of KAMs (see ACCA, 2018), which aligns with the South African themes.
Figure 3 shows the number of KAMs per the 13 themes, as well as the associated readability scores.
Figure 3 shows how the types of KAMs being reported have changed over time.Figure 3 confirms the increased number of KAMs dealing with 'going concern and impact of COVID-19' is because of a spike in these KAMs in 2020, when the economic consequences of the COVID-19 restrictions on movement were unfolding around the world.The only other KAM theme which has an increase in the frequency of disclosure in Figure 3 is 'other issues'.These KAMs deal mainly with the adoption of new accounting standards, which became effective over the period under review, such as the IFRS dealing with leases and revenue. 8Consistent with the results in Figure 1, the frequency of other reported KAMs is either stable or falling over the sample period.
Figure 3    and the fact that the audit report is intended to be read by users with a reasonable understanding of accounting standards.

Differences in key audit matters among industries
Figure 4 examines how KAMs differ among industries after controlling for differences in the number of companies per industry.The size of an audit client and the complexity of its business model will have implications for the auditor's assessed risk of material misstatement at the overall and assertion level (IAASB, 2013).If KAMs are selected from among the most important issues identified during an engagement, it follows that there should be a direct link between the levels of audit risk and the number and type of KAMs being reported.Several complexity and audit risk proxies have been considered by the prior literature.One of the most common is industry membership (Burke et al., 2021;Sierra-García et al., 2019).
Figure 4 shows that companies in the construction (average 2.34) and public administration (average 2.32) industries reported the highest number of KAMs.On the other extreme, companies in the agriculture, forestry and fishing (average 1.53) and manufacturing (average 1.54) industries reported the fewest KAMs.Readers may have expected firms in the finance, insurance, and real estate industry to have the most complex accounting and, in turn, the greatest number of KAMs per organisation.The ISA 701, however, requires auditors to select from material issues those which were the most significant for conducting an audit.That banks, insurance providers or property funds have complex business models and accounting requirements is not unusual, and audit firms have well-developed methodologies for dealing with these types of entities.As a result, the underlying complexity of an auditee's business environment, information systems and financial statements does not automatically give rise to additional KAMs.Whether or not a material issue is also a KAM is a matter of professional judgement and the context of each client.It cannot be assumed that specific types of companies or business sectors will therefore have more or fewer KAMs in their audit reports than other entities.
The Flesch average reading ease score (untabulated) ranges between 33 and 47, and an untabulated Kruskal-Wallis test confirms that the scores among industries are similar (H = 0.110, p > 0.05).This reaffirms the fact that KAMs are client specific and not industry specific.

Differences in key audit matters among audit firms
The assurance literature has long distinguished between engagements performed by the Big N and other audit firms (Boone et al., 2010;DeAngelo, 1981).Briefly, the largest audit firms have the time and resources (both financial and intellectual) to conduct more rigorous audits and resist client pressures which may otherwise compromise auditor independence.Empirical evidence repeatedly confirms that audit quality, evaluated using different quality proxies, is  http://www.sajbm.orgOpen Access higher for engagements completed by Big N than for those completed by smaller competitors (see, e.g., Boone et al., 2010;Francis, 2004).Results on the relevance of auditor type as a KAM determinant are mixed (Abdelfattah et al., 2020;Pinto et al., 2020;Sierra-García et al., 2019), but if the largest firms conduct engagements of the highest quality, they may also be best placed to identify important issues and include these as KAMs in their audit reports.The number of average KAMs per year for the Big 4 and non-Big 4 audit firms are presented in Table 3.The averages are reported after controlling for the number of clients audited by each firm.
Table 3 shows that the Big 4 audit firms account for 71% of all KAM disclosures.This is reasonable because they audit 64% of the companies in the sample, which are also the largest and most complex clients (see Sierra-García et al., 2019).
PricewaterhouseCoopers has the highest share of reported KAMs (26%), followed by Deloitte (20%), KPMG (14%) and then EY (11%) -in line with the audit market shares of the respective firms.Compared with their non-Big 4 counterparts, the Big 4 audit firms also report slightly more KAMs per client (2.25 compared to 2.11). 9On average, KPMG reports 2.39 KAMs per client, followed by Deloitte (2.35), EY (2.29) and PwC reporting the lowest number of KAMs per client (2.11), despite having the most KAMs in total.The average number of KAMs reported by the Big 4 audit firms in South Africa is also consistent with how their international offices 10 are applying ISA 701 and may reflect the effort expended by the Big 4 to standardise audit methodologies among their offices.
A comparison of the readability scores among the Big 4 shows very little difference between these firms.But when compared with the non-Big 4 firms, the KAMs written by Big 4 firms are significantly more complex (score = 36) compared 9.That the average number of KAMs reported per client per audit firm is not materially different is confirmed by an untabulated Mann-Whitney U test (U = 1.082, p > 0.05).
with their non-Big 4 counterparts (score = 45). 11This is possibly because of the portfolio of Big 4 audit clients being larger and more complex than their non-Big 4 counterparts.
Figure 5 illustrates the average KAMs per audit report by type and audit firm from 2017 to 2020.This shows how often a KAM theme is mentioned when a particular issue is identified.As with Figure 4, results are after controlling for differences in the number of each audit firm's clients.
Figure 5 shows that KAM themes are indicative of risks impacting various aspects of that theme in an organisation at both the entity and account level.Audit reports can reinforce significant risk areas by addressing the same KAM theme multiple times in the report, as shown by themes that have more than one KAM reported for the same theme.For example, PwC reports 3.23 business combinations and goodwill KAMs per client where that issue has been identified.This indicates that the core issues are being reinforced and have impacted the client from more than one perspective.
Overall, the proportion of KAMs per theme is not consistent among the audit firms.Despite efforts to standardise how ISA 701 is interpreted and applied in other jurisdictions (consider Rousseau & Zehms, 2020), South African audit firms and individual engagement teams probably have considerable discretion when it comes to identifying KAMs at individual clients (Duboisee de Ricquebourg & Maroun, 2022).A level of consistency is observed in the accounting changes and errors theme and other issues theme, mainly because of the accounting standard changes for revenue and leases being flagged as KAMs consistently by all audit firms during this period.Employee benefits are also consistent among the three firms who flagged this as a KAM; however, as employee benefits overall represent approximately 0.1% of all KAMs, are not identified by all 11.Untabulated Mann-Whitney tests further support the observation that, on average, the largest audit firms have KAMs which are more difficult to read than those reported by the smaller firms (U = −6.218,p < 0.05).firms and are a client-specific issue, this trend is not considered to be material.
Other than employee benefits, all of the audit firms include at least one instance of the account-and entity-level KAM themes.Account-level KAMs are, on average, reported on more frequently by audit firms (2.03) compared with entitylevel KAMs (1.67).This is likely because entity-level KAMs are, by their nature, broader and probably less often applicable than account-level KAMs.Account-level KAMs may also include significant risk areas across different aspects of the accounting standard applicable to that theme, which may result in multiple KAMs being recorded for that theme for one client.An untabulated Kruskal-Wallis Test confirms that the number of entity-(H = −2.605,p < 0.05) and accountlevel KAMs (H = −5.294,p < 0.05) varies among audit firms, and these differences are statistically significant.
There is no indication of audit firms having a preference for certain types of KAMs.As observed in the 'differences in key audit matters among industries' section, KAMs are a function of the nature and circumstances at each client rather than a product of generic firm guidance, which predetermines certain KAMs.What is interesting to observe is that the presence of one KAM theme during the audit may lead to another KAM theme being consistently disclosed in the audit report because of the relationships between significant risks.The inter-relationships between KAM themes are deferred for future research.However, the next section will further explore the inter-relationship between time, audit firm and industry type.

The inter-relationship between time, audit firm and industry type
The preceding discussion deals with KAMs over time and among the audit firms and industry separately.To consider interconnections among these variables, the industries were recoded into three broader categories from 'brick and mortar'type industries to more service-based ones.An analysis of covariance (ANCOVA) was used to evaluate the interaction among the number of KAMs and their readability, on the one hand, and the type of auditor, industry and financial year on the other.The ANCOVA has been run after including controls often used by the KAM determinants literature for firm performance, levels of financial risk and complexity (see Abdelfattah et al., 2020;Bentley et al., 2021;IAASB, 2013;Rousseau & Zehms, 2020;Sierra-García et al., 2019).The first was gauged using return on assets and a dummy variable coded 1 if a firm is loss-making and 0 if this is not the case.Financial risk was measured using leverage (the ratio of debt to equity).Complexity was captured by the number of subsidiaries for each entity.In the interest of brevity, results for the control variables were excluded (Refer to Table 4).
Financial year (F = 1.491, p > 10%), auditor (F = 3.760, p > 10%) and industry (F = 0.158, p > 10%) were not associated with  differences in the number of KAMs in total.As discussed in the 'differences in key audit matters among audit firms' section and the 'differences in key audit matters among industries' section, the type of auditor (F = 5.282, p < 1%; 22.304, p < 5%) and industry (F = 6.964, p < 5%; 3.828, p < 1%) have some relevance for the number of entity-and accountlevel KAMs, but statistically insignificant interaction terms suggest that these differences are not consistent among industries and over time.This is confirmed by the estimated marginal means presented in Figure 6. 12   The estimated marginal mean of the total number of KAMs for large (range = 1.16-2.39)and second tier firms (range = 2.01-2.14)was consistent among industries and over time (untabulated).There were differences in the number of entity-and account-level KAMs among the industries.Concerning the interaction between auditor and industry, large audit firms were consistently reporting more accountlevel KAMs per industry than the second-tier firms.
Conversely, second-tier auditors reported more entity-level KAMs than the Big 4.
As discussed in the 'types of issues being flagged as key audit matters' section, readability scores have decreased from 2017 to 2020, indicating a growing complexity in KAM disclosures (F = 15.310,p < 5%).On average, readability scores were consistent among industries (F = 0.141, p > 10%) but Figure 6 shows that Big 4's KAMs were more difficult to read than those prepared by smaller audit firms (F = 4302, p < 5%).This effect was more pronounced for the retail, wholesale and utilities and the financial and other services industries, as indicated by the interaction term between auditor and industry (F = 455, p < 5%) and displayed in Figure 6.

Conclusion
This article provided a detailed overview of KAM disclosures in South Africa from 2017 to 2020.The findings suggested the most common KAM disclosures are related to business combinations and impairments of goodwill, followed by measurement and impairment considerations for nonfinancial assets.These areas deal with technical and highly complex features of financial accounting where 12.As the financial year is statistically insignificant in Table 4, it has been excluded from Figure 6.
auditors can be expected to devote more time and attention.
In particular, the use of estimates, the impact of uncertainty and the need to apply professional judgement mean that these are more likely to be higher-risk areas in an audit where a greater number of issues will be discussed with governing bodies and flagged as KAMs.Less often resulting in KAMs were the auditees' systems and internal controls, inventories and employee benefits.Key audit matters dealing with prior period errors, changes in accounting policies and going concern issues were also infrequent, except in 2020 when COVID-19 posed additional risks to business continuity.
The finance, insurance and real-estate sector and the manufacturing sector reported the greatest number of KAMs, but this was a result of the concentration of companies in these sectors on the JSE.After adjusting for differences in the number of companies in each industry, the construction industry and public administration industries had the highest number of KAMs per audit report.
Key audit matter disclosures among the Big 4 were relatively consistent at 2.25 KAMs on average per audit report.Smaller South African audit firms reported a similar number of KAMs (on average, 2.11 KAMs per audit report).Prior research argues that the Big 4 provide better quality audits than smaller firms but this may not be evidenced by the number of KAMs included in an audit report.However, this study found that smaller firms were reporting a similar number of KAMs, which may be because of smaller firms replicating more established industry members when it comes to how many KAMs they disclose.
While the results point to a degree of standardisation in how ISA 701 is being applied, important differences in KAM disclosures are emerging.The large and small audit firms neither consistently report the same type of KAMs nor do they consistently report on KAMs within the same industries.Account-level KAMs are a key focus area across all audit firms.Second-tier firms do, however, report marginally more entity-level KAMs than the Big 4 who report more accountlevel KAMs than their smaller competitors.There were differences in the readability of the KAMs.Key audit matters written by the Big 4 tended to be more complex and difficult to read than those prepared by the second-tier http://www.sajbm.orgOpen Access auditors, possibly reflecting the inherent complexity of clients of the two groups of auditors.Variations in how KAMs are worded may also reflect differences in how KAMs are identified and internalised by audit firms, particularly when contrasting the most prominent industry members with other professional accounting firms.Overall, KAM disclosures are, however, difficult to read, irrespective of the type of auditor or industry in which an auditee is located.This reflects the inherently technical nature of the KAMs and the fact that the target audience is users who are proficient in IFRS and auditrelated matters.
A related area for future research can focus on the information content of KAMs.This study was limited by a twodimensional perspective based on the relative number of KAMs and a simple measure of readability.Developing a schematic that captures the informativeness of a KAM would be useful for researchers wanting to test the value-relevance of KAM disclosure and for regulators and standard-setters keen on evaluating the costs and benefits of applying ISA 701.The mix of entity-and account-level KAMs and their readability are a reasonable starting point.These indicators or factors could be complemented by other such as the extent of generic disclosure, the number of KAMs added to or removed from an audit report and the extent to which KAMs complement the information already contained in financial statements.It would also be useful to examine the procedures used in those areas of an engagement where KAMs arose and whether or not this level of detail is useful for financial capital providers.
For scholars interested in the readability of KAMs, how this is measured should be more carefully revised.The scores generated in this study were the Flesch Reading Ease and the Flesch-Kincaid Grade Level.These are widely applied but can be tailored to deal with the technicalities of financial statements, including the terminology used by accounting and auditing standards.More refined measures of KAM readability will be useful for evaluating the information content of KAMs.They can also be used in conjunction with detailed interviews with auditors, preparers and users of financial statements to evaluate how effective KAMs are at providing insights into the audit process and to identify areas for improving KAM disclosure.
Finally, the researchers have focused only on South Africa.the jurisdiction-specific focus means that this study does not provide an international review of KAM disclosures.The findings and principles discussed in this article should be relevant to a broad audience, but additional research will be required to understand in more detail how KAMs vary among countries and the factors contributing to the homogenisation of, or divergence in, the application of ISA 701 by audit firms.

FIGURE 3 :
FIGURE 3: Key audit matters reported per year by core theme.

FIGURE 4 :
FIGURE 4: Average key audit matters per industry.

FIGURE 5 :
FIGURE 5: Average key audit matters (KAMs) by type and audit firm.

TABLE 1 :
Core key audit matter themes and examples of key audit matter detail classification buckets.
To compare KAMs among industries, the JSE industry groupings have been used.This study differentiated between firms in mining; manufacturing and agriculture; retail, wholesale and utilities; and financial and other services sectors.These were then distinguished between KAMs reported on average by the Big 4 (PricewaterhouseCoopers [PwC], Deloitte, Klynveld Peat Marwick Goerdeler[KPMG]and Ernst & Young[EY]) and other audit firms operating in South Africa 7 by the average KAMs by type per audit firm and the average KAMs by industry per audit firm.

TABLE 2 :
Flesch Reading Ease score interpretation.

TABLE 3 :
Average number of key audit matters identified per South African auditors (2017-2020).

TABLE 4 :
Analysis of variance: financial year, audit firm and industry versus key audit matter measures., key audit matters.F statistics *** significant at the 1% level; ** significant at the 5% level. KAM The choice of jurisdiction was informed by South Africa's long-standing application of International Auditing Standards and its position as a prominent developing economy.As a result, studying South African KAM reporting complements the growing body of work in this area, which is predominantly concerned with the determinants of KAM reporting in the United States and the United Kingdom.Nevertheless,