KATHOLIEKE UNIVERSITEIT LEUVEN
How effective are performance audits? A multiple case study within the local audit office of Rotterdam Paper for the 5th International Conference on Accounting, Auditing & Management in Public Sector Reforms Amsterdam, 3‐5 September 2008
Katrien Weets K.U.Leuven – Public Management Institute This text is based on research conducted within the frame of the Policy Research Centre on Governmental organization in Flanders (SBOV II ‐ 2007‐2011), funded by the Flemish government. The views expressed herein are those of the author(s) and not those of the Flemish government.
How effective are performance audits? A multiple case study within the local audit office of Rotterdam
1. Introduction Performance audits can be broadly described as audits that focus on the efficiency, the effectiveness and the economy of public sector activities.1 As such they go further than the more traditional types of audit, like financial audits, whereby the truthfulness of financial accounts is investigated, and compliance audits, whereby the conformity of financial and operational controls and activities with the prevailing laws, directives, procedures and standards is verified. Furthermore, performance audits are quite public sector specific, in that an equivalent for them does not exist in private sector, commercial auditing. In addition to this, as a rule, they are conducted by auditors of Supreme Audit Institutions (SAIs) and so they take place within a very specific institutional context (Pollitt et al., 1999). Although thus far only few academic studies were conducted in which the effectiveness of performance audits has been examined, it is often contended that they impact upon and contribute to the effectiveness, the efficiency and the accountability of public sector. Frequently claims are made that performance audits support parliamentary democracy by assisting the legislative power in controlling the executive. Moreover, the information they produce would inform citizens about the ins and outs of government. Despite the limited evidence of these beneficial effects, a large number of governments heavily invest in the development and expansion of their performance audit functions. Therefore it is remarkable that the question whether or not performance audits are truly effective thus far only received limited attention from the academic world (Morin, 2004). As Bowerman et al. (2000) stated earlier: “In analysing the shape and scale of the ‘audit society’ , it is a striking paradox that the process of audit itself has been audited or evaluated in only limited ways.”
Within this general framework our explorative research had three main ambitions. Firstly, we sought to explore the extent to which performance audits are truly effective. Secondly, we aimed at uncovering the various factors that impact upon and provide an explanation for the degree of effectiveness of a performance audit. Thirdly, and no less important, we wanted to reopen the debate on performance audit effectiveness on the one hand and on the way in which it should be measured on the other. With these ends in mind, we first undertook an extensive review of the existing literature on the use, the impacts and the effects of performance audits on the one hand and policy evaluations on the other hand. Thereby we identified various factors that, according to literature, foster performance audits’ success. The results of this review are described in the following section. The third paragraph then sheds more light on some performance audit effectiveness measures and particularly on the more comprehensive measurement model we used in order to assess the degree of effectiveness of three performance audits conducted by the local audit office of Rotterdam, the so‐called ‘Rekenkamer Rotterdam’. Subsequently the results of the research are presented and discussed in the fourth section. The article concludes with an enumeration of limitations to the performance audit effectiveness measurement model used in this study and to research dealing with the effectiveness of performance audits in general. Based upon these limitations, areas for further research are identified. 1
There is however no general consensus about the proper role and scope of performance audit in the field of state audit. For a review of the different definitions of performance audit and a mapping of the frontiers of the practice of performance audit with some of the neighbouring concepts like for instance financial audit, evaluation & private audit, see Pollitt et al. (1999). Next to this Pollitt (2003) has also stated that in practice the most commonly used criterion in performance audits tends to be some notion of good management practice/good administration instead of one of the so‐called three E’s. Moreover earlier research on a transnational level has shown that various SAIs interprete the concept of performance audit in different ways. As a result performance audits conducted by different Supreme Audit Institutions (SAIs) are not at all similar (Put, 2005). For all of these reasons it can be stated that there is not really one correct and generally accepted definition of performance audit. Performance audit should conversely be seen as an umbrella term that encompasses different types of audit activities.
Katrien Weets
1
How effective are performance audits? A multiple case study within the local audit office of Rotterdam
2. Literature review As only few academic studies deal with the impacts and effects of individual performance audits and as the practice of performance audit shares a lot of characteristics with the field of policy evaluation2, we decided to broaden our literature review to include some basic concepts and findings brought to the fore by the research on the utilisation of policy evaluations. These are briefly set out in the first part of this paragraph. Thereupon the second part presents an overview of the specific literature on the impacts and effects of performance audits. Finally, the third part deals with potential side effects of the practice of performance auditing. The paragraph is rounded off with a conclusion.
2.1. Some basic concepts and findings from the research on the utilisation of policy evaluations Already in 1979 Weiss introduced the so‐called enlightenment model of research utilisation. By referring to this model, she called attention to the fact that the concept of research utilisation does not necessarily have to relate to a directly assignable and instrumental use of research results. After all, research results, theoretical perspectives and conceptions can also permeate the policy‐making process in an indirect and diffuse way, by the influence they exert on the views, the ideas, the beliefs and the attitudes of civil servants, politicians, etc. As such it is possible that their administrative or political views on certain policy issues change, without them being aware of this influence (Weiss, 1979). In ‘Social science research and decision‐making’ Weiss & Bucuvalas (1980) presented the results of their study on the usefulness of social science research for decision‐makers. Of the 255 interviewees, which were all mental health decision makers at federal, state and local levels, a large majority reported high levels of research use. They did not only claim to read a good deal of research but they also considered that their knowledge of research did have an impact upon their work. Weiss & Bucuvalas concluded that the use of social science research in policy making particularly depends on two factors: the receptivity and openness of decision makers for social science research on the one hand and the frames of references decision‐makers use with regard to research on the other hand. They described these so‐called frames of references as implicit filters that are used by decision‐makers to screen the research that they read or hear about, in order to handle the information overload they are confronted with on a daily basis. By means of a factor analysis the authors succeeded to derive an understanding of the frames of references that decision makers use in screening research.
2
There has been a lot of discussion and even controversy about the transferability of policy evaluation approaches and techniques to the field of performance audit (see for example Pollitt & Summa, 1996 & 1997; Gunvaldsen & Karlsen, 1999; Pollitt, 2003). Yet it is recognized that with the rise of performance audit over the last fifteen to twenty years there has been a shift towards a more evaluative stance within SAIs. Pollitt et al. (1999) for example stated that certainly auditors in some countries (particularly the Netherlands, Sweden, the USA & the UK) are showing an explicit interest in evaluation. As our research was conducted within a Dutch local audit office, the ‘Rekenkamer Rotterdam’, where the auditors’ work is strongly inspired by evaluation methodologies and approaches, we believed including findings and concepts placed in the forefront by the research on the utilisation of policy evaluations, would not only be useful but also enriching to our analysis.
Katrien Weets
2
How effective are performance audits? A multiple case study within the local audit office of Rotterdam
According to Weiss & Bucuvalas the criteria that are employed by decision‐makers to describe and judge studies, are the following: • relevance to issues their office deals with; research quality: the technical quality, objectivity and cogency of a study; • • conformity with user expectations: its plausibility given their prior knowledge, values and experience; • action orientation: the explicit guidance it provides for feasible implementation; • challenge to the status quo: its challenge to existing assumptions, practice and arrangements. Next to this it is worth mentioning that the political acceptability of research results only correlated to a lesser degree with the interviewees’ assessments of the usefulness of a study than was assumed in advance by the authors. However, this view was contested in later years by Korsten (1983) who stated that the political context of a research actually is a factor of great importance in explaining its use. Leviton & Hughes (1981) drew in their work on the distinction between the instrumental, the conceptual and the persuasive use of evaluation research. The concept of instrumental use alludes at a direct and assignable processing of the research results. Hence, policy makers are aware of the fact that they used a specific study to support the policy process, for instance in order to formulate different policy alternatives. That is not usually the case when research results are used in a conceptual manner, for then ideas, conceptions, and theoretical perspectives affect the policy‐making process only in an indirect way (cfr. supra). At last, contrary to the two previous categories, the persuasive use of evaluation research implies interpersonal influence. This kind of use may for instance occur when one has to defend a political decision that was already made before or when one wants to bring others to take up a certain political stand. Academic studies are then employed as political munition, what eventually may lead to a selective use of research results. Besides all this, the authors identified five clusters of variables that according to them encourage the use of evaluation research by policy makers: • the relevance of an evaluation to the needs of potential users: with respect to the content as well as with regard to the timing of the evaluation; direct and flexible communication between the potential users and the producer(s) of the • evaluation; • clear and accessible communication of the research results to policy makers (translation of evaluations into their implications for policy and programs); a positive perception of the credibility of the evaluator by policy makers; • commitment and advocacy by certain individual users in the policy making process. • Finally, in a more recent work Karen Kirkhart (2000) brought to the fore a so‐called ‘integrated theory of influence’. In doing this she proposed a radical rethink of the concept of ‘use of evaluation information’, for, in her opinion, the earlier conception of ‘use’ was too limiting and lacking in theoretical basis. Kirkhart therefore argued for a change in focus from ‘use’ to ‘influence’. She proposed this change in focus in order to broaden the framework and to move away from the unidirectional, episodic, intended and instrumental connotations of the term use. Hence, within this context the concept of ‘influence’ is considered as less oriented towards the direct results of evaluation studies and viewed as a more continuous process than the concept of ‘use’. Next to this, Kirkhart also advanced an integrated theoretical framework. As such, Kirkhart reconceptualised the findings of past research and developed a more coherent picture of the impact of evaluation studies (Cummings, 2002).
Katrien Weets
3
How effective are performance audits? A multiple case study within the local audit office of Rotterdam
According to Kirkhart (2000) then, three different dimensions have to be studied in order to determine the influence of evaluation research: • source (impacts may arise as a consequence of the process of evaluating on the one hand and as a consequence of the results of an evaluation on the other hand); • time (impacts can occur during the evaluation process (immediate impacts), at the end of an evaluation cycle (end‐of‐cycle impacts) and in the long term (long‐term impacts)); • intention (an evaluation research can have both intended and unintended consequences). It is important to mention that these categories are not mutually incompatible. Hence they do not exclude each other.
2.2. Literature on the impacts & effects of performance audits Pollitt et al. (1999) devoted a chapter of their book to the impact of the performance audit work undertaken by five European SAIs (the Dutch Algemene Rekenkamer, the French Cour des Comptes, the UK National Audit Office, the Swedish Riksrevisionsverket and the Finnish Valtiontalouden tarkastusvirasto). In this chapter they considered amongst other things what SAIs themselves stated with regard to the impact of their performance audit work, whether and how SAIs measure their impacts, which efforts SAIs make to enhance the impact of their work, if there any negative (side) effects to performance audit work, etc. In seeking for answers to their questions, the authors were able to bring a number of interesting findings to the fore. Altogether their evidence suggested that performance audits do have an impact. After all, in each of the countries they examined, performance audits did led to changes to government activities. In many cases the reactions to auditors’ recommendations also led to substantial savings in public funds. In addition, the authors drew attention to the possibility that performance audits also produce a deterrent effect arising simply from the fact that they exist, as well as a more general educational impact on the audited organisations and on public sector in general. Furthermore the authors gave a rough sketch of a number of changes that were implemented by some national SAIs in the past so as to raise the impact of their performance audits: the introduction of new methods and techniques, the recruitment of staff with new skills, the development of better planned studies, longer term audit strategies, efforts to work more closely with auditees, the development of greater publicity for reports, the move from consolidated reports to single, stand‐alone publications and efforts to obtain broader and deeper rights of access to bodies spending public money. Finally, Pollitt et al. stated that the impact of a performance audit report is largely determined by environmental factors, which cannot be easily altered by SAIs. The following environmental factors were briefly discussed: • the political context in which auditors work; • the means a SAI has at its disposal; • the way in which auditees react on a performance audit report; • the interests of the auditees; • the reaction of the public opinion on a performance audit report; the reaction of the parliament on a performance audit report; • luck (for example with regard to the timing of publication). • In one of his more recent works Pollitt (2006) made an analysis of the existing literature on the use of performance information by ministers, parliamentarians and end users. He concluded that, in spite of the big statements that are generally made about the importance of performance information for democracy, there is few evidence that these end‐users do indeed use this type of information. On the contrary, the patchy evidence that is available seems to indicate that they only rarely make use of the volumes of performance information now thrust upon them. Hence the author argued for a broader definition of performance information use, so that it would not only encompass the direct and instrumental use of
Katrien Weets
4
How effective are performance audits? A multiple case study within the local audit office of Rotterdam
performance information, but it also would cover other kinds of uses, like for example conceptual use (cfr. supra). Besides this, Pollitt referred to the possibility that the existence of performance information also has a preventive effect. Thereby he argued that the regular production of reports is more important than their regular consumption, for the need to ‘keep public servants honest’ does not require everything to be read and digested in order to achieve its effect. Johnsen et al. (2001) for their part examined whether and in what way performance audits contributed to performance improvement in Finnish and Norwegian local governments. Based on nine interviews, the authors concluded that the practice of performance audit served two functions. On the one hand, the process of performance audit appeared to be effective in identifying those organisational areas that were in need of review. On the other hand in general local civil servants and politicians did seem to act upon the auditors reports and recommendations. Furthermore the authors stated that the information included in the reports was especially used to develop management systems and to improve the quality of the municipal budget. According to a number of Scottish auditors, who where interviewed by Lapsley & Pong (2000), performance audits are indeed useful, particularly with regard to the management of public organisations. The benefits were generally said to be of an operational nature, rather than of a strategic nature. The auditors referred for example to the improvement of systems and processes, the adoption of best practices, the identification of redundant jobs and the improvement of risk management within audited organisations. In his doctoral thesis Vanlandingham (2006a) examined the strategies that were employed by US legislative oversight offices in order to stimulate the use of their research results in the legislative process. The author distinguished between two overall strategies: developing strong networks with decision makers and astute marketing of research products. He particularly examined how key stakeholders utilised and assessed the audits conducted by legislative oversight offices, and to what extent their assessments cohered with the network‐ & marketing initiatives taken by these offices. His research indicated firstly that offices located within a legislative unit were more effective in promoting use of their work than offices located within an auditing unit. Secondly it showed that offices that had adopted research standards which stress utility to stakeholders were more successful than offices conforming to the Government Auditing Standards, which strongly emphasize organisational independence (Vanlandingham, 2006b). Torgler & Schaltegger (2006) examined how audit courts and local autonomy affect political discussion. The goal of their research was to explore to which extent citizens’ level of discussion depends on institutions, focusing specifically on audit courts and local autonomy in Switzerland. Their results indicated that a higher audit court competence and a lower level of centralisation correlated with a higher level of political discussion. According to these authors, the results in Switzerland suggest that such institutions help improve citizens’ willingness to acquire information costs and discuss political matters. Furthermore, Torgler (2005) found that in the Swiss case higher audit court competences had a significantly positive effect on tax morale. He suggests that these help improve taxpayers’ tax morale and thus citizens’ intrinsic motivation to pay taxes. Eichenberger & Schelker (2007) considered in their research the Swiss local finance commissions as elected competitors to government. The authors particularly analysed to what extent the power of local finance commissions affected the tax burden and the public expenditures within 26 different Swiss cantons. On the basis of an econometrical analysis they found that the power of a local finance commission did have an economically relevant, statistically significant and robust negative effect on the tax burden and on public expenditures within a canton. The first study ever in which the economic effects of SAIs were assessed on a cross country basis was conducted by Blume & Voigt (2007). These authors particularly estimated the effect of SAIs on three
Katrien Weets
5
How effective are performance audits? A multiple case study within the local audit office of Rotterdam
groups of variables, namely fiscal policy, government effectiveness and productivity. Two surveys (one from the World Bank / OECD, the other one from INTOSAI) were used to generate a number of variables that proxy for various aspects of the institutional and organizational structure of SAIs. The variables that were used to proxy for the independence, the mandate, the transparency and the stylized model of the SAIs did not display any significant impact on the economic variables mentioned above. There was only one exception to this general rule: the perceived levels of governments’ graft were significantly higher if SAIs were structured along the court model of auditing. Problematic to this study however is that the authors only made use of indicators that cover the legal situation of SAIs. Hence, as is acknowledged by the authors, they did not any include information on the factual situation of SAIs. In the middle of the eighties the U.S. General Accounting Office (U.S. GAO) was very successful with regard to the number of recommendations adopted by audited organisations. Aiming at advising evaluation bodies in order to raise their effectiveness, Johnston (1988), relying on organisation theory literature, sorted out some possible explanations for the success of this ‘congressional watchdog’. The author successively distinguished between three explaining variables that according to him possibly could account for the success of the U.S.GAO: the types of change the U.S.GAO put forward in its recommendations, its role as an external evaluator, and the activities set up by the U.S. GAO in order to promote its work. Regarding the types of change the U.S. GAO put forward in its recommendations, Johnston stated that these were for the greater part intended to act upon the behaviour and the rules within the audited organisations. Based on organisation theory literature he concluded that such proposals for change are in general more easily accepted than recommendations directed at affecting the goal or the structure of an organisation. Concerning the role of the U.S.GAO as an external evaluator, the author indicated that this position offered it a certain flexibility in associating with other organisations. Besides this he pointed out that the U.S.GAO had good relations with the congress and the media, which put pressure on the audited organisations to adopt its recommendations. At the same time the activities set up by the U.S.GAO in order to promote its work, like for example follow‐up activities and a utilisation tracking system, contributed, according to Johnston, to the success of this audit office. Van der Meer (1999) for his part endeavoured to answer the following research question: “When, how and what do government agencies learn from evaluations?” To that end he developed a structural constructivist theoretical framework and applied it to two performance audits both conducted by the Dutch ‘Algemene Rekenkamer’. The author focused in his research on the degree to which learning processes were generated and enhanced in the audited organisations as a consequence of these performance audits. So as to formulate his hypotheses he made use of the concept of ‘repertoire’, which he defined as follows: “Repertoires are defined as stabilized ways of thinking and acting (on the individual level) or stabilized codes, operations and technology (on other levels).” It appeared that the performance audits had hardly any direct effect that could be unequivocally ascribed to them. Rather they seemed to support or counteract debates, tendencies and options that were already present, or ‘under construction’, in the interaction among the actors concerned. Van der Meer concluded that the success of performance audits is dependent on the degree to which and the manner in which the repertoires of the actors involved, in this context auditors and auditees, match. According to the author one of course has to take into account that a performance audit takes place within a given context, whereby the success of an audit can also be influenced by the repertoires of other actors.
Katrien Weets
6
How effective are performance audits? A multiple case study within the local audit office of Rotterdam
Likewise, de Vries (2000) relied on a structural constructivist approach in his doctoral study on the impacts of two performance audits of the Dutch ‘Algemene Rekenkamer’. He included the following concepts in his multivariate explanatory model: • cognitive coupling: the degree to which the recommendations of the auditor match the repertoires of the auditees (cfr. Van der Meer); • social coupling: the social‐organisational distance between the auditor and the auditees and the extent to which the auditor is perceived as external by the auditees; • embedment: the degree to which an audited organisation has “embedded” her repertoires, so that she is not open to external impulses and influences; multiplicity: on three different levels: • o regarding the internal differences within an organisation; o regarding a third actor which is considered important by the audited organisation; o regarding the policy networks in which the audited organisation participates. According to de Vries, a performance audit would be more successful if: • the recommendations of an auditor match the repertoires of the auditees; • the social‐organisational distance between auditor and auditees is smaller and the auditor is perceived less as from the outside by the auditees; • the repertoires of the auditees are less strongly embedded within the audited organisation and the auditees are open to external impulses and influences; the repertoires within the audited organisation are more diversified, and the inclusions in the • collective repertoire differ widely; the recommendations of the auditor match the repertoires of a third actor which is considered • important by the audited organisation; the recommendations of the auditor match the ‘ideas‐in‐construction’ of a number of actors of a • policy network in which the audited organisation participates. Van der Meer (1999) & de Vries (2000) both considered performance audits as social influence processes. So did Morin (2001), who perceived a performance audit as a social influence process which takes on its full meaning through the relationship arising between the source of influence, i.e. the auditor, and the target of influence, i.e. the auditee. By means of a multiple case study, in which the influence attempts in six performance audits performed by the Auditor General of Quebec and the Auditor General of Canada in 1995 and 1996, were examined closely, and using qualitative content analysis techniques, Morin could identify fourteen performance indicators, which indicate whether or not an influence attempt has been more or less successful. These were grouped into three categories: • perceptions and reactions of auditees with regard to auditors’ influence attempts; • impacts generated by the performance audit; • the contribution of a performance audit to the public debate. The performance indicators which were put in the forefront by Morin are listed in table 1.
Katrien Weets
7
How effective are performance audits? A multiple case study within the local audit office of Rotterdam
Table 1. Performance indicators for performance audits, according to Morin (2001) Category Performance Indicator Auditees’ perceptions and reactions with regard to auditors’ influence attempts
• • • • •
Auditees’ feelings towards auditors Sources of auditees’ dissatisfaction with regard to performance audit auditors’ work Auditees’ reaction to auditors’ influence attempt: internalization Cooperation offered by auditees to auditors (in auditors’ opinion) Auditees’ misgivings of auditors’ influence attempt legitimacy
Impact on the audited organisation
• • • • • • •
Auditees’ perception of added value of the performance audit Auditees’ evaluation of auditors’ findings Willingness of auditees to follow‐up on auditors’ recommendations Evaluation by auditees of auditors’ overall performance Auditees’ perception of the usefulness of the performance audit for the audited organisation Changes made by auditees to management practices Auditees perceptions of the overall effect of the performance audit
• •
Stimulation of debates in parliament Coverage by the press
Contribution to the public debate
Besides this Morin also highlighted fifteen success factors in order to find out the reasons why influence attempts either had succeeded or failed. These factors fell into two categories: • factors linked to the performance audit process; • factors linked to the existence of environmental conditions during the performance audit process. The fifteen success factors that were provided by Morin are presented in table 2. Table 2. Succes factors in influence attempts, according to Morin (2001) Category Success Factor
Factors linked to the performance audit process
• • • • • • • • •
Factors linked to the existence of environmental conditions
• • • • • •
Auditees’ perception of a participating leadership style in auditors Auditees’ perception of a preference for collaboration on part of auditors Auditees’ perception of power relations between auditors and auditees Credibility of auditors in eyes of auditees Auditees’ perception of connotation of auditors’ modes of influence and type of message Auditees’ degree of influenceability Auditees’ level of commitment Auditees’ level of tolerance to criticism Degree of fluidity in communications between auditors and auditees The will at staff level and in the central authority of the organisation being audited Political will Timing of the performance audit Major reorganisation in the body being audited Reform at the government level Place of the activity audited and of the recommendations within the priority scale of the audited organisation’s management
Morin concluded that four of the six performance audits studied, were not very successful. As a result of this she pleaded in favour of a more sceptical attitude towards performance audits’ effectiveness. However, the author did soften her findings few years later on the basis of a survey research through which she tried to measure the impact of performance audits, conducted between 1995 and 2002, on the management of organisations in Quebec. Surveys were exclusively directed at the auditees. The number of respondents amounted to a total of 99. First, the impact of performance audits was studied in ten
Katrien Weets
8
How effective are performance audits? A multiple case study within the local audit office of Rotterdam
possible variations, frequently invoked in literature. They can be listed as follows: value added by performance audits, the relevance of the recommendations made by the auditors, the preventive effect exerted by performance audits on the auditees, the influence exerted by the performance audit on auditees’ management practices, the influence exerted by the performance audit on the audited organisation’s relations with interest groups, the perceived utility of auditor general reports, the concrete actions taken by audited managers following a performance audit, organisational consequences of audits, personal consequences of audits and finally the global impact on the management of the organisation audited. Morin’s survey research indicated that the influence of auditors was perceptible for all of these ten elements. She concluded that at least in part performance audits, as practised in Quebec governmental organisations from 1995 to 2002, had led to a better performance. As the audited managers saw it, performance audits did in general have a positive influence on the management of their organisations. Subsequently Morin tried to discover the factors that determined the outcome of auditor’s attempts to make their influence felt. The research indicated that, as a rule, the relations that auditors established with the managers of audited organisations during a performance audit seemed to strengthen their influence significantly. Furthermore, it seemed that environmental conditions also exerted a favourable impact on the work of auditors, with a positive effect ranging from minimal to relatively strong. Next to this, the intervention of parliamentarians did not seem to have had any significantly negative effect. Judging from auditees’ perceptions, the work of auditors was, on average, even favourably influenced by the involvement of parliamentarians. Finally the effect of press coverage was present and was perceived as contributing to auditors’ influence, but it registered only weakly. Alon (2007) for his part sought to determine the qualitative and quantitative differences between the outcome of investigative journalism and the outcome of state audit’s efforts. His research focused on defects that have a substantial effect on a relatively large portion of the population. It indicated that 38 percent of the defects exposed by the media were completely rectified, in comparison with 46 percent of the defects pointed out by the State Comptroller. On the other hand, the media was found to be more successful with respect to defects that the establishment completely ignored (31 percent vs. 24 percent). Furthermore the author found that it took an average of 2.7 years to correct the defect in cases reported by the media, compared to 3.8 years for defects exposed by the State Comptroller. In addition to this the research showed that, of the defects exposed by the media that were ultimately rectified, sixty percent were corrected within the first year after the publication, whereas this percentage only came to twenty percent for the State Comptroller.
2.3. Literature on potential side effects of performance audits With regard to potential side effects of performance audits, Pollitt et al. (1999) noted that the presence of auditors may considerably enlarge the workload of audited departments. In addition to this, their evidence suggested that some of the auditees, when informed about the possibility of being subject to an audit in the near future, made themselves guilty to an hasty introduction of new management practices, in anticipation of the auditor’s visit. Other authors (like for example Leeuw, 2006) have contended that performance audits may impede innovation and renewal in government by the strong emphasis they put on compliance with procedures (‘ossification’). Another possible unintended side effect of performance auditing is denoted in literature by the term ‘manualisation’. It refers to an excessive use of manuals and guidebooks in order to cope with the recommendations of auditors and the procedures they prescribe, while one can ask oneself if such an approach truly leads to a more efficient and effective public sector. Moreover auditors, overstressing administrative rationality, would sometimes pay too much attention to trivial details, and thereby lose sight of truly significant inefficiencies.
Katrien Weets
9
How effective are performance audits? A multiple case study within the local audit office of Rotterdam
Furthermore report has it that performance audits can bring about tunnel vision. Tunnel vision relates to the situation whereby phenomena that are quantified and measured in the performance measurement system are overstressed at the expense of the unquantified and unmeasured aspects of performance (Smith, 1995; Dolmans & Leeuw, 1997). There are other concepts too, like for example audit society and performance paradox, that allude indirectly to the statement that performance auditing does not always result in a more efficient and effective policy, and that one also has to take into account the unintended side effects of performance measurement. After all, performance measurement on itself does not automatically result in a better policy and management. On the contrary, according to some authors, an overload of performance auditing may eventually lead to a situation of ‘analysis paralysis’ within audited organisations. Subsequently it is claimed that overemphasizing the use of performance indicators may engender ‘short‐ termism’. Auditees then become inclined to focus primarily on achieving good results with regard to the performance indicators established in their organisation, and they lose sight of the long‐term policy perspectives (Neely, 1999; Jackson, 2005). Next to this certain authors contend that auditing is a real trust‐killer (Dolmans & Leeuw, 1997). According to them, some auditors assume that they have to distrust auditees for it is plausible that these put their own interests or the interests of their organisation first. Therefore they would consider a controlling method of working more effective than a method of working which is based on mutual trust. By using ‘command and control instruments’ one of course punishes the auditees that do act for the common good. This way of acting would then decrease instead of increase the degree to which an audit is effective. Elliott (2002) for example indicates that performance auditing may harm the basis for trust on the shop floor, and in that way can be contra productive for an organisation. Apart from this he states that a performance audit does not make a clear distinction between results on the one hand and effects on the other hand, and that it is therefore not capable of fully grasping the complexity of the relations between activities and their consequences. Finally Elliott claims that performance audits do not sufficiently take into account time and context of audited activities, and that as a consequence of this they do not offer a clear picture of the actual impact of certain practices. Other authors, like for instance Shore & Wright (1999), even contend that the practice of performance auditing gradually creates ‘a climate of fear and tension’, wherein ‘coercive accountability’ strongly predominates. At last, according to some other authors, a potential macro level side effect of performance auditing is the transfer of institutional power that takes place between legislative bodies and audit institutions as a result of performance auditing. By means of performance audits audit institutions would have obtained possibilities to take part in deciding about certain policy options, that are actually outside their scope, and that go far beyond their legal audit mandate. Power (2005) for example qualified audit institutions as ‘de facto policy makers’. He argued that SAIs, in spite of their so‐called independence, continuously contribute to public sector reforms by exerting pressure on public organisations to implement new systems through which they become more ‘auditable’.
2.4. Review of the literature – conclusions In conclusion of this paragraph, table 3, that can be found in appendix A of this paper, summarizes the factors that, according to the literature discussed above, have an impact upon the degree of effectiveness of a performance audit. The potential side effects of performance audits are not included in this table, although it is acknowledged that they too eventually could exert influence on the extent to which a performance audit is effective. However, contrary to the factors that were incorporated in the table, it can be assumed that they only affect the degree of effectiveness of a performance audit in an indirect manner.
Katrien Weets
10
How effective are performance audits? A multiple case study within the local audit office of Rotterdam
The findings in the table set up the starting point for our study. After all, they were our main source of inspiration in building up the research design and selecting the appropriate research techniques. These are both discussed in the next section.
3. Research design and research method This paragraph elaborates on the research design and the research method we used. In a first part some commonly used performance audit impact and effectiveness measures are introduced and criticised. The second part discusses the more comprehensive measurement method we used in order to explore performance audit effectiveness. Finally, the third part outlines the research method we applied.
3.1. Measuring performance audit impacts Pollitt et al. (1999) recognised that measuring the precise scale of impact of a performance audit is difficult, if not impossible. According to these authors, the following means were most commonly used by the five SAIs in their study to determine the impact of their performance audit work: changes made by government (number of recommendations accepted by government and/or number of recommendations acted upon), financial savings (amount of cost savings achieved), impact on parliament (amount of attention a performance audit gained in parliament) & internal measures (including for instance comments from the audited body and the amount of press coverage a performance audit received). All of these indicators can be criticised however for they only offer a very limited view on performance audit effectiveness. For the purpose of illustration, we will apply Kirkharts’ (2000) ‘integrated theory of influence’, which was outlined in paragraph 1, to the measures mentioned above. Taking a close look at the first indicator, ‘changes made by government’, one can demonstrate on the basis of Kirkharts’ framework that this indicator does not take fully into account the three dimensions that should be studied in order to determine the impact of a performance audit. After all, he only partly grasps the dimensions that were brought to the fore by Kirkhart. As such, he provides an indication for the impacts that arise as a consequence of the results of a performance audit, that occur at the end of an audit cycle and that were intended. Impacts that arise as a consequence of the process of performance auditing, during the evaluation process or in the long term, or unintended impacts are on the contrary fully neglected by this measure for performance audit effectiveness. The figure below represents in diagram form how this first indicator relates to the integrated theory of influence proposed by Kirkhart. Figure 1. Relation between the measure ‘changes made by government’ and the integrated theory of influence (Kirkhart, 2000)
Katrien Weets
11
How effective are performance audits? A multiple case study within the local audit office of Rotterdam
As is shown in figure 1, by mapping the number of recommendations that were accepted by government and/or the number of recommendations that were acted upon, one measures performance audit effectiveness only in a very limited way. Furthermore by using this particular measure one exposes oneself to the risk that auditors display ‘gaming behaviour’ while conducting an audit. After all, the number of recommendations accepted by auditees can be quite easily manipulated by auditors (see for example Johnston, 1988). The same goes for the other indicators. Moreover they all give a very one‐sided representation of what constitutes performance audit effectiveness. The amount of cost savings achieved for example does not grasp any qualitative improvements that were brought about by a performance audit. In addition this measure does not sufficiently take into account the specific characteristics and needs of public sector organisations. The amount of attention a performance audit gained in parliament on the one hand and the amount of press coverage a performance audit received on the other hand can also be criticised for these indicators only focus on particular aspects of performance audit effectiveness, and they thus are very narrow in scope.
3.2. Towards a more comprehensive model for measuring performance audit effectiveness The literature review has shown that so far examining performance audit effectiveness is a still relatively uncultivated field. The academic studies that do exist often lack a sound empirical and methodological foundation. Moreover the available publications on potential side effects of performance audits are only too frequently based on anecdotal evidence. As was demonstrated above, a sound, valid, reliable and comprehensive measure for performance audit effectiveness does not yet exist up to now. In order to measure performance audit effectiveness on the one hand and to uncover the various factors that impact upon and provide an explanation for the degree of effectiveness of a performance audit on the other hand, we opted for the method which was supplied by Morin (2001) as the point of departure for our research. We preferred to use this method for several reasons. Firstly, it takes into account both the impacts of the performance audit results as the impacts of the performance audit process and it pays attention to the impacts in the short term as well as the impacts in the long term. As such, it grasps two of the three different dimensions advanced by Kirkhart, which is an improvement in comparison with the more limited indicators discussed above. Secondly, Morin succeeded in giving her study a sound scientific basis and in putting forth a structured framework to examine performance audit influences on public administrations. The research framework advanced by Morin was already discussed in the first paragraph of this paper. On the one hand fourteen performance indicators, which indicate whether or not an influence attempt has been more or less successful, were put in the forefront. On the other hand, Morin brought forward fifteen success factors in order to find out the reasons why influence attempts have either succeeded or failed. Inspired by Morin’s classification, we selected the following performance indicators for our research: • auditees feeling towards auditors; • auditees’ satisfaction with regard to the performance audit auditors’ work; • cooperation offered by auditees to auditors (in auditors’ opinion); • absence of doubts about the legitimacy of auditors’ influence attempt; auditees’ perception of added value of the performance audit; • auditees’ evaluation of auditors’ findings; • willingness of auditees to follow‐up on auditors’ recommendations; • auditees’ perception of the usefulness of the performance audit for the audited organisation; • changes made by auditees to management practices; • auditees perceptions of the overall effect of the performance audit; • stimulation of debates in parliament (in this context: the municipal council); •
Katrien Weets
12
How effective are performance audits? A multiple case study within the local audit office of Rotterdam
•
coverage by the press (in auditees’ perception).
Conversely, we did not include the following two indicators in our research design: • auditees’ reaction to auditors’ influence attempt: internalization; evaluation by auditees of auditors’ overall performance. • Within the limited scope of our research, it was not feasible to develop a sound and valid measurement for internalization. Next to this it can be assumed that the different performance indicators partly overlap, so that this particular indicator is already covered to a certain degree by the others. The same goes for the evaluation by the auditees of the auditors’ overall performance. Furthermore we added two performance indicators to the framework which was advanced by Morin: • attention from other institutions (like for example interest groups, universities). • attention from the audited organisation We included the first indicator primarily because of its relative importance in the Dutch context. Besides this we aimed at, as far as that is possible, giving readers a complete picture of performance audit effectiveness. The same goes for the second indicator. The success factors selected for our research were also mainly based on Morins’ work. They can be listed as follows: • auditees’ perception of a preference for collaboration on part of auditors; auditees’ perception of power relations between auditors and auditees; • credibility of auditors in eyes of auditees; • auditees’ degree of influenceability; • auditees’ level of commitment; • auditees’ level of tolerance to criticism; • degree of fluidity in communications between auditors and auditees; • the will at staff level and in the central authority of the organisation being audited; • political will; • • timing of the performance audit; • major reorganisation in the body being audited; • reform at the government level; • place of the activity audited and of the recommendations within the priority scale of the audited organisation’s management. We did not include the following success factors advanced by Morin: • auditees’ perception of a participating leadership style in auditors; • auditees’ perception of connotation of auditors’ modes of influence and type of message. Conversely, we replaced these indicators by one variable that focuses on: • the role and function adopted by the auditors during the performance audit process3
3
See Pollitt et al. (1999) for an elaborate discussion of performance auditors’ roles.
Katrien Weets
13
How effective are performance audits? A multiple case study within the local audit office of Rotterdam
Finally, independently of Morins’ framework, five other variables of which is stated in literature that they impact upon the degree of effectiveness of a performance audit, were included in our research: • auditees’ perception of the relevance of a performance audit to issues the audited organisation deals with; • degree of cognitive coupling between auditors and auditees; • auditees’ perception of the technical quality of a performance audit; • challenge to the status quo: its challenge to existing assumptions, practice and arrangements; • presence of potential side effects (ossification, tunnel vision, short termism, reduction of auditees’ motivation and disruption of the daily work scheme of the audited organisation). Based on the variables mentioned in this paragraph, we developed an elaborate research questionnaire.4
3.3. Research method Our research applied a multiple case study design. This methodology permitted us to disentangle and explore the different factors that have an impact on the effectiveness of performance audits (Yin, 1994). Besides this, earlier research on a transnational level has shown that performance audits conducted by different Supreme Audit Institutions (SAIs) are not at all similar (Put, 2005). Therefore, in our opinion, a case study design can be considered as most appropriate. In particular, three performance audits, which were all conducted by the local audit office of the city of Rotterdam (“the Rekenkamer Rotterdam”), were selected for our research and were studied in depth. The following criteria were used in order to select the cases: • timing of the performance audit: ideally the degree of effectiveness of a performance audit is examined about three years after he was conducted. After all, such a term offers a reasonable assurance that both the impacts in the short term and both the impacts in the long term have already occurred. the subject of the performance audit: ideally an audit should deal with a topic of continuous policy, • in order to examine its degree of effectiveness. Auditees then have had the opportunity to implement the recommendations of the auditors, which is not always the case if an audit concerns a once‐only project. • availability of the auditees: from a practical point of view it was important that a sufficient number of auditees still worked for the audited organisation. Therefore the audit should not have been conducted too many years ago. Based on an explorative interview with an auditor of the ‘Rekenkamer Rotterdam’, we were able to select three audits that met our selection criteria. For studying in depth the degree to which these performance audits were effective, we used different research techniques. Firstly, a documentary analysis was carried out. Secondly, 15 semi‐structured interviews were set up with both auditors (n=6) and auditees (n=9). These interviews were structured along established lines. First we requested the respondents to score a certain item on a five point Likert scale. Next we invited them to motivate and/or to specify their answer.
4
To obtain this questionnaire, please contact the author of this paper:
[email protected]
Katrien Weets
14
How effective are performance audits? A multiple case study within the local audit office of Rotterdam
Furthermore, we decided to interview the auditees within the administration (like for example public managers) instead of the auditees within the political system (like for example the mayor and the aldermen). This had several reasons: • the explorative interview with the auditor of the ‘Rekenkamer Rotterdam’ showed that in general the auditors within this local audit office had much more contacts with public servants than with politicians during the performance audit process (for example to collect data). Since Morin considered a performance audit primarily as a social influence process, the choice for interviewing the auditees within the administration arised logically from her approach; • in addition, one has to take into account street‐level bureaucracy theory. After all, it is not because the recommendations of auditors were adopted at the political level that they actually permeated within the shop floor practices of public organisations. Therefore we assumed that, interviewing the people who are closest to the policy implementation process, would provide a more complete picture of performance audit effectiveness than interviewing those who are politically responsible for a certain policy area.
4. Results Table 4, which can be consulted in appendix B of this article, presents the average scores given by both auditors and auditees to the different performance indicators which were advanced by Morin (2001), by performance audit. The table shows that, when attributing the same weight factor to each performance indicator, performance audit 2 was most effective, performance audit 1 came off second best and performance audit 3 was least successful. Furthermore the high scores that were assigned by the auditees to the performance indicator ‘absence of doubts about the legitimacy of auditors’ influence attempt’ immediately strike the eye. They lead us to state that the greater part of the auditees considered the influence attempts of the auditors as legitimate. Moreover none of the auditees indicated to have doubts about performance audits’ legitimacy. In addition to this the semi‐structured interviews showed that a wide majority of the auditees supported the idea of a local audit office. The three performance audits also achieved moderately positive to positive scores with regard to the attention they gained from the audited organisation and from the municipal council. A large majority of the interviewed auditees declared that the performance audits indeed were considered important by their organisations, especially because of the potential reprisals they could have for the audited divisions. Only one auditee stated that the organisation he worked for had paid little attention to the performance audit it was subject to. According to him this had several reasons. Firstly, he pointed to the fact that the conclusions of the audit were not at all shocking. Secondly, he indicated that the bureaucratic structures of the organisation he worked for did not allow a quick diffusion of the results of the performance audit. A number of auditees also spontaneously relegated to the additional workload the performance audit had caused. Next to this, one can contend that the performance audit reports also gained quite a lot of attention from the municipal council, for they all were discussed in one of the council committees as well as in one of the plenary sessions of the council. Another striking point of similarity between the three selected performance audits is the low score they all got for the attention they gained from other institutions. During the interviews, the auditors referred to their efforts to reach the general public by using diverse media. Yet they had to acknowledge that they did not consciously supplemented these with an active communication strategy directed towards the immediate stakeholders of the policy area subject to a performance audit. With regard to the other performance indicators, it can be stated that their scores differ widely along the three cases. Table 5 offers a global overview of the results by category of performance indicators.
Katrien Weets
15
How effective are performance audits? A multiple case study within the local audit office of Rotterdam
Table 5. Global overview of the results by category of performance indicators, advanced by Morin (2001) Category of performance indicators PA 1 PA 2 PA 3 Auditees perceptions and reactions with regard to 13.75/20.00 13.50/20.00 12.17/20.00 auditors’ influence attempts Impacts generated by the performance audit 17.13/30.00 23.50/30.00 14.08/30.00 Contribution to the public debate 12.51/20.00 13.25/20.00 13.34/20.00 Total score 43.39/70.00 50.25/70.00 39.59/70.00 Based on these results, one could easily assume that the higher degree of effectiveness of performance audit 2 directly follows from its relatively high score on the second category, ‘impacts generated by the performance audit on the audited organisation’. However, we should nuance this statement. The interviews with the auditees of performance audit 2 brought to the fore that the changes made to the policy process did not only result from the performance audit itself, but also from a wider range of public management reforms that were already going on for a longer period in the municipal administration. Consequently, the auditees looked upon the performance audit report as an external source of legitimacy for the new line of policy that they were already following of their own accord. Furthermore it is important to mention that the auditees spontaneously defined the concept of impact very widely. Making use of the questionnaire we developed, we examined whether in the auditees’ perceptions the performance audit made their organisation more cost efficient, whether it helped their organisation to reach her goals, whether it helped them to work more result‐oriented, whether it contributed to the computerization within their organisation, or whether it made their policy initiatives more evidence‐based. Although some auditees had indicated that the performance audit did have an impact on their organisation, they assigned only low scores to the questions dealing with the topics outlined above. During the semi‐structured interviews we of course discussed this issue, which seemed rather contradictory to us at first sight, with the auditees. The greater part explained their seemingly contradicting behaviour by referring to the positive impact the performance audit had had on their awareness of the new public management reforms. Although they acknowledged that these reform processes were already going on for a longer period, they pointed out that the performance audit for the first time really confronted them in a direct way with its implications. Finally, as table 6, which also can be found in appendix B, strikingly demonstrates, wide differences existed between the responses of the audited organisations of performance audit 1. This was not the case for the two other performance audits we studied. Table 7, also added to appendix B of this paper, for his part presents the scores the auditees attributed to the different success factors put forth by Morin (2001). A first remark should be made with regard to the relatively low scores the auditees assigned to the success factor ‘credibility of the auditors’. The semi‐structured interviews revealed that the auditees did not at all couple this success factor to the auditors’ professional competences. When assessing auditors’ credibility they rather referred to other qualities, like for example auditors’ empathy with the specific features of the audited organisations. The main criticisms directed against the auditors were that they acted too much according to the letter of the law, instead of according to its spirit. In the opinion of other auditees auditors could have presented certain findings in their end report in a more positive way, if they had taken into account the context in
Katrien Weets
16
How effective are performance audits? A multiple case study within the local audit office of Rotterdam
which the audit took place. Finally, some auditees felt that the auditors did not pay enough attention to changes the audited organisation had already made on its own initiative. Next to this, the rather low scores for the success factor ‘auditees’ perception of power relations between auditors and auditees (absence of struggle for power)’ immediately catch the eye. Although they create the impression that the greater part of the auditees were convinced that there was a struggle for power between themselves and the auditors, this conclusion should be softened somehow. After all, the scores for this success factor did not only vary widely between the different cases, but also between the different auditees of a same performance audit. Furthermore all performance audits received very high scores for the success factor ‘auditees’ level of commitment’. In order to substantiate their point of view, the auditors referred to the openness and the kind cooperation of the auditees and the fact that these actively thought along with them. Regarding one of the cases they especially pointed to the time and energy the audited organisations invested in the audit, and to the conversations between auditors and auditees, which in fact sometimes became quite emotional. Other striking points of similarity between the three selected performance audits are the high scores they all gained on the success factors ‘timing of the performance audit’ and ‘place of the activity audited and of the recommendations within the priority scale of the audited organisation’s management’. Regarding the latter, all the auditees shared the same opinion, namely that the performance audit touched to the core activities of their organisations. With respect to the former, the greater part of the auditees viewed the timing of the three performance audits as contributing to their success. The argument most often used by the respondents to motivate their statements that the performance audits came at the right moment, was the fact that a new policy plan had to be drawn up shortly after the publishing of the performance audit report, so that the results of the audit could serve as feedback for this new plan. With regard to the other success factors, one can contend that their scores differ widely along the three cases. Table 8 presents these results by category of success factors. Table 8. Global overview of the results by category of success factors, advanced by Morin (2001) Category of success factors PA 1 PA 2 PA 3 Factors linked to the performance audit process 25.25/40.00 22.75/40.00 21.17/40.00 Factors linked to the existence of environmental 13.01/20.00 17.75/20.00 19.00/30.00 conditions Total score 38.26/60.00 40.50/60.00 40.17/70.00 Based on these findings, we can identify for each of the three selected performance audits the factors that fostered their success. Table 7 and 8 show that these varied strongly across the three performance audits. In addition to this table 9 in appendix B shows that also with regard to their scorings of the success factors wide differences existed between the different audited organisations of performance audit 1. In general one can state, in accordance with Morins’ approach, that the factors linked to the performance audit process and the factors linked to the existence of environmental conditions quite equally contributed to the effectiveness of performance audit 1. With regard to performance audit 2 and 3 the environmental factors seemed to play a more important role. Especially in the case of performance audit 2 their contribution to its effectiveness is quite remarkable. Although performance audit 2 did not do very well on the factors linked to the performance audit process itself, he came out as most effective thanks to the favourable environmental conditions. The same can be said for performance audit 3, but in this case the factors linked to the existence of environmental conditions only compensated for the moderate scores on the factors linked to the performance audit process to a lesser extent.
Katrien Weets
17
How effective are performance audits? A multiple case study within the local audit office of Rotterdam
By way of recapitalising our results, table 10 shows the total scores of the three selected performance audits in accordance with Morins’ approach. Table 10. Total scores of the performance audit in accordance with Morins’ approach Total score on the performance Total score on the success Performance audit factors (in %) indicators (in %) Performance audit 1 61.99 63.77 Performance audit 2 71.79 67.50 Performance audit 3 56.56 57.39 Finally, as was already outlined, independently of Morins’ framework, we also registered the scores for five other variables, of which is stated in literature that they impact upon the degree of effectiveness of a performance audit. Table 11 in appendix B presents the scores auditees and / or auditors attributed to these variables. This table shows clearly that while performance audit 1 & 2 where deemed relevant by the auditees, this was not at all the case for performance audit 3. The auditees of performance 3 motivated their answer by stating that although they and the people within their organisation considered the audit very important, for example because it touched to their core activities, they did not view the conclusions and recommendations of the auditors as good starting points to improve their organisation and its policy initiatives. According to the auditees this was a consequence of the fact that auditees and auditors did not agree about the way in which an audit should be conducted and the way in which policy making and policy implementation should be organised. They particularly criticised the fact that the auditors focused on the letter of the law, instead of safeguarding its spirit. Next to this they felt that the auditors laid too much emphasis on quantifying goals, what according to them was simply not always possible for the policy areas in which their organisation was active. As table 11 demonstrates, these confrontational views between the auditees and auditors of performance audit 3 also manifested themselves in their scorings of the degree of cognitive coupling between auditors and auditees. The technical quality on the other hand was quite positively assessed by the auditees of all of the three selected performance audits. However it is remarkable that, while performance audit 2 came out as the most effective one according to Morins’ approach, he did not get the highest score on this feature. This could be an indication that the technical quality of a performance audit is indeed not an important determinant for its effectiveness. This view was confirmed by the statements of the auditees during the interviews. An example can illustrate this. Although the auditees of performance audit 2 were very positive about the conclusions of the auditors, they were not completely satisfied with the way in which the audit was conducted. In addition to this they indicated that the performance audit only partly covert the policy area in which their organisation was active, so that he could not grasp fully its complexity. The auditees of performance audit 3 for their part acknowledged that the auditors were very competent and very professional. In their opinion however, the auditors’ lack of empathy did do harm to the technical quality of the audit and to its conclusions. Furthermore table 11 shows that the three performance audits are also quite similar in their scorings on the degree to which they were seen as a challenge to the status quo by the auditees. The greater part of the auditees of performance audit 1 and 2 reported that these audits had had a distinguishing effect on the existing conceptions and assumptions within their organisation. In order to substantiate these statements, they mainly referred to the public management body of thought that as a consequence of the audit would have been incorporated more solidly within their organisations.
Katrien Weets
18
How effective are performance audits? A multiple case study within the local audit office of Rotterdam
Finally, we examined to what extent certain potential side effects (ossification, tunnel vision, short termism, reduction of auditees’ motivation and disruption of the daily work scheme of the audited organisation) manifested themselves in case of each of the three performance audits in the auditees’ perception. The results in table 11 indicate that side effects were especially present in the case of performance audit 3. According to the auditees of this performance audit three side effects appeared: disruption of the daily work scheme of the audited organisation, ossification and tunnel vision. Quite remarkable was their answer to our question regarding short‐termism. These auditees stated that thanks to the performance audit they actually felt more inclined to pursue their long term goals and that thus the opposite of short termism had occurred. According to them, the fact that they had had to defend themselves and their organisation as a consequence of the negative criticisms of the auditors, had stimulated them to formulate their goals in the longer term.
5. Limitations to the research method and to research dealing with performance audit effectiveness in general – directions for further research The results of our study seem to corroborate Morin’s (2001) findings. Of course this statement should be put into perspective, for within the limited scope of our research we could only examine three cases. While conducting this research, we bumped into some limitations to our research approach and to research dealing with performance audit effectiveness in general. These are briefly outlined below, because they could be useful as a starting point for further performance audit effectiveness measurement attempts. Firstly, readers should be aware that by using the research approach we proposed in this paper one does not measure performance audit effectiveness in terms of a better policy and a better public management, but rather in terms of the success of auditors’ influence attempts. It should be stated clear that our aim was to examine the effectiveness of performance audits, and not the effectiveness of the policy measures subject to these performance audits. As such, by using this research method, one cannot guarantee that the implementing of the auditors’ recommendations truly contributed to the putting up of more efficient, effective and transparent policy and management processes. A second remark that should be made is that based on the research method we applied, one cannot state for sure that, when auditors in a performance audit end report had recommended to make certain alterations, and the audited organisation had implemented these changes in the meantime, the audited organisation had displayed this behaviour solely as a result of the performance audit. During the semi‐ structured interviews, some auditees for example indicated that organisational reforms had been carried through rather as a consequence of the new public management body of thought that was permeating the municipal administration, than as an immediate result of one of the performance audits. Regarding public management reforms one of course cannot create a laboratory environment and in that way eliminate all external influences that distort the degree of effectiveness of a performance audit. Nevertheless, it is important to take into account such methodological issues in the future if one wants to make progress in studying and measuring performance audit effectiveness. The following comments deal with some potential hiatuses in the research method we used. Firstly, our approach fully disregarded institutional factors. Yet several authors contend that the institutional context of an audit office at least partly determines the way in which it conducts performance audits, which for his part can be supposed to have an impact upon performance audits effectiveness. In the case of our research, this lack of attention for institutional factors did not cause any problems, for we chose three cases that took place within the same institutional environment, namely that of the ‘Rekenkamer
Katrien Weets
19
How effective are performance audits? A multiple case study within the local audit office of Rotterdam
Rotterdam’. We thus already neutralised the institutional context in advance by means of our research design. Conversely, scholars that aim at comparing different audit bodies should include these institutional factors in their research framework, for we can reasonably assume that they do exert influence upon performance audit effectiveness. In addition to this, the research method we used does not rule out the possibility that next to the effect of a certain single performance audit, a cumulative effect of several performance audits occurred. This possibility was precluded in case of our research, for the interviewed auditees did not have any previous experiences with the ‘Rekenkamer Rotterdam’. However, in our opinion, further research should take into account this possibility of performance audits producing cumulative effects, in order to paint an as correct as possible picture of performance audit effectiveness. Furthermore one should be careful in dealing with causality. Often literature about performance audit impacts and effects assumes linear causal relationships as a given. Consequently other possible causal relationships are already precluded in advance. However, as the following examples strikingly demonstrate, by using such a unilateral approach towards the causality concept, one bears the risk to misrepresent the extent to which performance audits are truly effective. By using the indicators ‘the amount of attention a performance audit gained in parliament (here: municipal council)’ and ‘the amount of press coverage a performance audit received’ one for instance implicitly assumes that the more effective a performance audit is, the more attention he will gain in parliament and the more press coverage he will get. However, this reasoning can equally well be reversed. One could just as well state that the more attention a performance audit gains in parliament and the more press coverage a performance audit gets, the more effective he will be. Next to this also circular causality could arise. We speak of circular causality when a factor A codetermines the condition of a factor B, that for his part codetermines the condition of factor A. We can illustrate this somehow abstract statement with an example. If the media and the parliament, for instance based on their previous experiences, expect that a performance audit will have a big impact, it is plausible that they devote more attention to this particular audit than if they wouldn’t have these expectations. The attention gained from the press and from the media could then for their part foster the performance audit’s success, and as such enlarge its’ impact. Moreover the possibility that third factors have an impact upon the relations between certain independent variables and the dependent variable, in this context performance audit effectiveness, cannot be precluded. It is for instance possible that media attention and/or attention from the parliament on itself did not have an impact upon performance audit effectiveness, but that these factors for their part stimulated the audited organisation in following up a performance audit closely, which in the end led to a more effective audit. The attention a performance audit gained from parliament and/or media should then be seen as an intermediary variable, that did not have an impact on performance audit effectiveness in a direct way, but rather did exert influence in an indirect manner. Besides all this, the nature of a causal relation can be quite complex in itself. An independent variable can for instance either be a necessary and sufficient condition, either a necessary but insufficient condition, either a sufficient but not necessary condition, or a contributing cause for an event to occur. As was demonstrated extensively above, performance audit effectiveness research involves dealing with complex causal relationships. Therefore in our opinion, next to a quantitative research component, a qualitative research component is essential to studies attempting to examine and/or to measure performance audit effectiveness. After all, qualitative data permit us to disentangle and explore the different factors that impact upon performance audit effectiveness, to see precisely which events led to which consequences, and to derive some fruitful understandings. They also help scholars to get beyond
Katrien Weets
20
How effective are performance audits? A multiple case study within the local audit office of Rotterdam
their initial conceptions and to generate new conceptual models (Miles & Huberman, 1994). These are, in our opinion, necessary for making progress in studying performance audit effectiveness. Finally, with regard to Morin’s (2001) approach, we are fully convinced about the surplus value it offers. Based on our experiences we would however recommend to other scholars who are interested in using it, to examine a larger number of cases. In that way, statistical techniques can be applied and the number of success factors can be reduced by means of factor analysis, for, while conducting our research, we felt that there existed some overlap between these factors. As such, the research method used in this paper can be refined, and serve as a comprehensive model to measure performance audit effectiveness.
6. Conclusion Our explorative research had three main ambitions. Firstly, we sought to explore the extent to which performance audits are truly effective. Secondly, we aimed at uncovering the various factors that impact upon and provide an explanation for the degree of effectiveness of a performance audit. Thirdly, and no less important, we wanted to reopen the debate on performance audit effectiveness on the one hand and on the way in which it should be measured on the other. With these ends in mind, we undertook an extensive review of the existing literature on the use, the impacts and the effects of performance audits on the one hand and policy evaluations on the other hand. This review showed that there is a dearth of research on performance audit effectiveness. Moreover the few contributions that do exist tend to be very fragmented. Typically they only focus on particular aspects of performance audit effectiveness and they often lack empirical evidence. In addition to this, using the literature review as a starting point for our research, we identified various factors that, according to previous research, foster performance audits’ success. Furthermore we argued against the ‘popular’ indicators that are nowadays most commonly used by Supreme Audit Institutions in order to determine the impact of their performance audit work, for they only measure performance audit effectiveness in a limited way and they cannot fully grasp the complexity of the causal relations involved. Mainly inspired by Morin’s (2001) approach to performance audit success, we selected three performance audits, all conducted by the Rekenkamer Rotterdam, and applied the more comprehensive framework proposed by her to them. Our research results appeared to corroborate Morin’s (2001) findings, in that the total scores for the performance indicators and those for the success factors seemed to correlate. Besides the empirical testing of Morin’s more comprehensive model, we identified for each of the three selected performance audits the factors that fostered their success. The research showed that these varied strongly across the three performance audits. With regard to one of the selected audits, these factors even varied strongly across the different audited organisations. Furthermore, especially in one case, so called environmental conditions (like for example the timing of the performance audit, the importance of the audited subject for the audited organization, etc.) seemed to affect the degree of effectiveness of a performance audit to a large extent. Based on the results of our research and on our own experiences we brought to the fore a number of limitations to the research method we used and to research dealing with performance audit effectiveness in general. As such, directions for further research could be identified.
Katrien Weets
21
How effective are performance audits? A multiple case study within the local audit office of Rotterdam
References Alon, G. (2007). State audit and the media as the watchdogs of democracy – a comparative view. Journal of the Office of the State Comptroller and Ombudsman of Israel, 61, 55‐100. Blume, L. & Voigt, S. (2007). Supreme Audit Institutions: Supremely Superfluous? ‐ A Cross Country Assessment. 25 p. (http://papers.ssrn.com/sol3/papers.cfm?abstract_id=965434 or http://www.icer.it/menu/f_papers.html) Bowerman, M., Raby, H. & Humphrey, C. (2000). In Search of the Audit Society: Some Evidence from Health Care, Police and Schools. International Journal of Auditing, 4, 71‐100. Cummings, R. (2002). Rethinking Evaluation Use. Paper presented at the 2002 Australasian Evaluation Society International Conference October/November 2002 – Wollongong Australia (www.aes.asn.au).. Dolmans, L.J.F. & Leeuw, F.L. (1997). Performance auditing. Vraagtekens bij de effectiviteit van een onderzoekstraditie. NIVRA, Naar een doelmatiger overheid. Goed georganiseerd mensenwerk, 81‐88. Eichenberger, R. & Schelker M.. (2007). Independent and competing agencies: An effective way to control government. Public Choice, 130(1), 79‐98. Elliott, J. (2002). The impact of intensive ‘value for money’ performance auditing in educational systems. Educational Action Research, 10(3), 499‐506. Gunvaldsen, J.A. & Karlsen, R. (1999). The auditor as an evaluator. How to remain an influential force in the political landscape. Evaluation, 5(4), 458‐ 467. Jackson, A. (2005). Falling from a Great Height: Principles of Good Practice in Performance Measurement and the Perils of Top Down Determination of Performance Indicators. Local Government Studies, 31(1), 21‐38. Johnsen, A., Meklin P. & Vakkuri, J. (2001). Performance audit in local government: an exploratory study of perceived efficiency of municipal value for money auditing in Finland and Norway. The European Accounting Review, 10(3), 583‐599. Johnston, W.P. (1988). Increasing Evaluation Use: Some Observations Based on the Results at the U.S. GAO. New Directions for Program Evaluation. 39, 75‐84. Kirkhart, K.E. (2000). Reconceptualizing Evaluation Use: An Integrated Theory of Influence. New Directions for Evaluation, 88, 5‐23. Korsten, A.F.A. (1983) Wat is goed genoeg? Benutting van onderzoek in overheidsbeleid, Amsterdam, Kobra, 68 p. Lapsley, I. & Pong, C.K.M. (2000). Modernization versus problematization: value‐formoney audit in public services. The European Accounting Review, 9(4), 541‐567. Leeuw, F.L.. (2006). Over impact en neveneffecten van performance monitoring en auditing. Vlaams Tijdschrift voor Overheidsmanagement, 11(4), 37‐44. Leviton, L.C. & Hughes, E.F.X. (1981). Research on the Utilization of Evaluations ‐ A Review and Synthesis. Evaluation Review, 5(4), 525‐548. Miles, M.B. & Huberman, A.M. (1994). Qualitative Data‐Analysis: An Expanded Sourcebook, 2nd ed., Thousand Oaks, Sage, 338 p. Morin, D., Influence of Value for Money Audit on Public Administrations: Looking beyond Appearances. Financial Accountability & Management, 17(2), 99‐118. Morin, D. (2004). Measuring the impact of value‐for‐money audits: a model for surveying audited managers. Canadian Public Administration, 47(2), 141‐164. Neely, A. (1999). The performance measurement revolution: why now and what next? International journal of operations & production management, 19(2), 205‐228. Pollitt, C. & Summa, H. (1996). ‘Performance Audit and Evaluation: Similar Tools Different Relationships?’, in C. Wisler (ed.) Evaluation and Auditing: Prospects for Convergence (special issue of New Directions for Evaluation 71(Fall)). San Francisco, CA: Jossey‐Bass. Pollitt, C. & Summa, H. (1997). Reflexive Watchdogs? How Supreme Audit Institutions Account for Themselves. Public Administration, 75(2), 313‐36. Pollitt, C., Girre, X., Lonsdale, J., Mul, R., Summa, H. & Waerness, M. (1999). Performance or compliance? Performance Audit and Public Management in Five Countries, New York, Oxford University Press, 248 p.
Katrien Weets
22
How effective are performance audits? A multiple case study within the local audit office of Rotterdam
Pollitt, C. (2003). Performance Audit in Western Europe: Trends and Choices. Critical perspectives on accounting, 14, 157‐170. Pollitt, C. (2006). Performance Information for Democracy. The Missing Link? Evaluation, 12(1), 38‐55. Power, M. (2005). The Theory of the Audit Explosion, in E. Ferlie, L.E. Lynn & C. Pollitt (eds.), The Oxford Handbook of Public Management, Oxford, Oxford University Press, 2005, 326‐344. Put, V. (2005). Normen in performance audits van rekenkamers. Een casestudie bij de Algemene Rekenkamer en het National Audit Office, Leuven, K.U.Leuven, 243 p. (Diss. Doc.). Shore, C. & Wright, S. (1999). Audit Culture and Anthropology: Neo‐Liberalism in British Higher Education. Journal of the Royal Anthropological Institute, 5(4), 557‐575. Smith, P. (1995). On the unintended consequences of publishing performance data in the public sector. International Journal of Public Administration, 18(2), 277‐305. Torgler, B. (2005). A knight without a sword or a toothless tiger? The effects of audit courts on tax morale in Switzerland. Working Paper No. 2004 – 06. Center for Research in Economics, Management and the Arts, 19 p. Torgler, B. & Schaltegger, C.A. (2006). The Determinants of Political Discussion: How important are audit courts and local autonomy?” in Berkeley Program in Law & Economics, Working Paper Series, 198, Berkeley, University of California, 26 p. Van der Meer, F.B. (1999). Evaluation and the Social Construction of Impacts. Evaluation, 5(4), 387‐406. Vanlandingham, G.R. (2006a). A Voice Crying in the Wilderness – Legislative Oversight Agencies’ Efforts to Achieve Utilization, Florida, The Florida State University, 2006, 165 p. (Diss. Doc.) Vanlandingham, G.R. (2006b). A Voice Crying in the Wilderness – Legislative Oversight Agencies’ Efforts to Achieve Utilization. New Directions for Evaluation, 112, 25‐39. Vries, G.J. de (2000). Beleidsdynamica als sociale constructie. Een onderzoek naar de doorwerking van beleidsevaluatie en beleidsadvisering. Eburon Delft, 264 p. Weiss, C.H. (1979). The many meanings of research utilization. Public Administration Review, 39(5), 426‐431. Weiss, C.H. & Bucuvalas, M.J. (1980). Social science research and decision‐making, New York, Columbia University Press, 332 p. Yin, R.K. (1994). Case Study Research. Design and Methods. 2nd ed, Thousand Oaks, Sage, 171 p.
Katrien Weets
23
How effective are performance audits? A multiple case study within the local audit office of Rotterdam
Appendix A Table 3. Factors that have an impact upon the degree of effectiveness of a performance audit Author(s) Factors
Weiss & Bucuvalas (1980)
• receptivity and openness of decision makers for social science research • frames of references of the decision‐makers relevance to issues their office dealt with research quality: the technical quality, objectivity and cogency of a study conformity with User Expectations: its plausibility given their prior knowledge, values and experience action orientation: the explicit guidance it provides for feasible implementation challenge to the status quo: its challenge to existing assumptions, practice and arrangements
Leviton & Hughes (1981)
• the relevance of an evaluation to the needs of potential users: with respect to the content as well as with regard to the timing of the evaluation • direct and flexible communication between the potential users and the producer(s) of the evaluation • clear and accessible communication of the research results to policy makers (translation of evaluations into their implications for policy and programs) • a positive perception of the credibility of the evaluator by policy makers • commitment and advocacy by certain individual users in the policy making process
Johnston (1988)
• the types of change put forward in the recommendations • role as an external evaluator • activities set up in order to promote audit work
Lonsdale (1999)
• factors regarding the organisation and the functioning of a SAI: introduction of new methods and techniques the recruitment of staff with new skills the development of better planned studies longer term audit strategies efforts to work more closely with auditees the development of greater publicity for reports the move from consolidated reports to single, stand‐alone publications efforts to obtain broader and deeper rights of access to bodies spending public money • environmental factors: the political context in which auditors work the means a SAI has at its disposal the way in which auditees react on a performance audit report the interests of the auditees the reaction of the public opinion on a performance audit report the reaction of the parliament on a performance audit report luck (for example with regard to the timing of publication)
Van der Meer (1999)
• the degree to which and the manner in which the repertoires of the auditors and auditees match • context (for example the repertoires of other actors)
Katrien Weets
24
How effective are performance audits? A multiple case study within the local audit office of Rotterdam
de Vries (2000)
Morin (2001)
Morin (2004) Vanlandingham (2006)
Katrien Weets
• • • •
degree of cognitive coupling between auditors and auditees the degree of social coupling the degree of embedment the degree of multiplicity: regarding the internal differences within an organisation regarding a third actor which is considered important by the audited organisation regarding the policy networks in which the audited organisation participates • factors linked to the performance audit process auditees’ perception of a participating leadership style in auditors auditees’ perception of a preference for collaboration on part of auditors auditees’ perception of power relations between auditors and auditees credibility of auditors in eyes of auditees auditees’ perception of connotation of auditors’ modes of influence and type of message auditees’ degree of influenceability auditees’ level of commitment auditees’ level of tolerance to criticism degree of fluidity in communications between auditors and auditees • factors linked to the existence of environmental conditions will at staff level and in the central authority of the organisation being audited political will timing of the performance audit major reorganisation in the body being audited reform at the government level place of the activity audited and of the recommendations within the priority scale of the audited organisation’s management • • • •
relations between auditors and auditees environmental factors press coverage attention from parliament
• location of the audit institution: within a legislative unit • research standards that stress utility to stakeholders • active participation in networks with decision makers
25
How effective are performance audits? A multiple case study within the local audit office of Rotterdam
Appendix B Table 4. Average scores given by auditors and auditees to the performance indicators advanced by Morin (2001) by performance audit5 Performance indicator auditees feeling towards auditors auditees’ satisfaction with regard to the performance audit auditors’ work 3. cooperation offered by auditees to auditors (in auditors’ opinion) 4. absence of doubts about the legitimacy of auditors’ influence attempt 5. auditees’ perception of added value of the performance audit 6. auditees’ evaluation of auditors’ findings 7. willingness of auditees to follow‐up on auditors’ recommendations 8. auditees’ perception of the usefulness of the performance audit 9. changes made by auditees to management practices 10. auditees perceptions of the overall effect of the performance audit 11. attention from the audited organisation 12. coverage by the press (in auditees’ perception) 13. attention from the municipal council 14. attention from other institutions Maximum score Total score (in absolute numbers) Total score (in %) 1. 2.
PA 1 3.25
PA 2 2.50
PA 3 1.67
3.25
3.00
2.33
3.00 4.25 3.00 4.00 2.50 3.00 2.00 2.63 3.50 3.00 3.63 2.38 70.00 43.39 61.99
4.00 4.00 4.00 4.00 4.25 4.00 3.75 3.50 3.75 3.25 4.00 2.25 70.00 50.25 71.79
3.50 4.67 1.67 2.00 2.83 2.33 2.25 3.00 3.92 2.92 3.83 2.67 70.00 39.59 56.56
5
Average scores on a scale ranging from 1 to 5, whereby 5 represented the best result and 1 the poorest. The performance indicators evaluated by the auditees were 1, 2, 4, 5, 6, 8 & 10. Performance indicator 3 was scored by the auditors. Both the perceptions of the auditees and the perceptions of the auditors were taken into account for the following performance indicators: 7, 9, 11, 12, 13 & 14.
Katrien Weets
26
How effective are performance audits? A multiple case study within the local audit office of Rotterdam
Table 6. Average performance indicator scores of the two most divergent audited organisations of performance audit 1 by audited organisation Performance indicator 1. auditees feeling towards auditors 2. auditees’ satisfaction with regard to the performance audit auditors’ work 3. cooperation offered by auditees to auditors (in auditors’ opinion) 4. absence of doubts about the legitimacy of auditors’ influence attempt 5. auditees’ perception of added value of the performance audit 6. auditees’ evaluation of auditors’ findings 7. willingness of auditees to follow‐up on auditors’ recommendations 8. auditees’ perception of the usefulness of the performance audit 9. changes made by auditees to management practices 10. auditees perceptions of the overall effect of the performance audit 11. attention from the audited organisation 12. coverage by the press (in auditees’ perception) 13. attention from the municipal council 14. attention from other institutions Maximum score Total score (in absolute numbers) Total score (in %)
Katrien Weets
Audited organisation X
Audited organisation Y
1.00 1.00 3.00 5.00 1.00 5.00 1.50 1.00 2.00 1.00 3.75 3.25 3.25 1.75 70.00 33.50 47.86
4.00 4.50 3.00 4.50 4.00 4.00 3.00 4.00 2.25 3.75 3.00 3.00 3.75 2.50 70.00 49.25 70.36
27
How effective are performance audits? A multiple case study within the local audit office of Rotterdam
Table 7. Average scores given by auditors and auditees to the success factors advanced by Morin (2001) by performance audit6 Success factor
PA 1
PA 2
PA 3
auditees’ perception of a preference for collaboration on part of auditors auditees’ perception of power relations between auditors and auditees (absence of struggle for power) 3. credibility of auditors in eyes of auditees 4. the role and function adopted by the auditors 5. auditees’ degree of influenceability 6. auditees’ level of commitment 7. auditees’ level of tolerance to criticism 8. degree of fluidity in communications between auditors and auditees 9. the will at staff level and in the central authority of the organisation being audited 10. political will 11. timing of the performance audit 12. major reorganisation in the body being audited 13. reform at the government level 14. place of the activity audited and of the recommendations within the priority scale of the audited organisation’s management Maximum score Total score (in absolute numbers) Total score (in %)
3.00
1.50
3.00
2.25
2.50
1.33
3.50 3.25 4.00 4.00 2.00 3.25
2.50 1.50 3.50 4.00 4.00 3.25
3.67 1.00 3.50 4.00 1.50 3.17
2.75
4.50
3.00
2.38 4.00 / /
4.50 4.00 / /
2.42 3.17 3.08 3.00
3.88
4.75
4.33
60.00 38.26 63.77
60.00 40.50 67.50
70.00 40.17 57.39
1. 2.
6
Average scores on a scale ranging from 1 to 5, whereby 5 represented the best result and 1 the poorest. The success factors evaluated by the auditees were 1, 2, 3, 4 & 9 Success factors 5, 6 & 7 were scored by the auditors. Both the perceptions of the auditees and the perceptions of the auditors were taken into account for the following success factors: 8, 10, 11, 12, 13 & 14.
Katrien Weets
28
How effective are performance audits? A multiple case study within the local audit office of Rotterdam
Table 9. Average success factor scores of the two most divergent audited organisations of performance audit 1 by audited organisation Success factor
Audited organisation X
Audited organisation Y
1.00
5.00
1.00
3.50
3.00 1.00 4.00 4.00 2.00 2.00
4.50 5.00 4.00 4.00 2.00 3.75
1.00
4.00
1.50 3.75 / /
2.75 4.25 / /
3.75
3.75
60.00 28.00 46.67
60.00 46.50 77.50
1. 2.
auditees’ perception of a preference for collaboration on part of auditors auditees’ perception of power relations between auditors and auditees (absence of struggle for power) 3. credibility of auditors in eyes of auditees 4. the role and function adopted by the auditors 5. auditees’ degree of influenceability 6. auditees’ level of commitment 7. auditees’ level of tolerance to criticism 8. degree of fluidity in communications between auditors and auditees 9. the will at staff level and in the central authority of the organisation being audited 10. political will 11. timing of the performance audit 12. major reorganisation in the body being audited 13. reform at the government level 14. place of the activity audited and of the recommendations within the priority scale of the audited organisation’s management Maximum score Total score (in absolute numbers) Total score (in %)
Table 11. Average scores for other variables put forth in literature as having an impact upon the degree of effectiveness of performance audits Variable 1. 2. 3. 4. 5.
auditees’ perception of the relevance of a performance audit to issues the audited organisation deals with degree of cognitive coupling between auditors and auditees auditees’ perception of the technical quality of a performance audit challenge to the status quo: its challenge to existing assumptions, practice and arrangements presence of potential side effects (ossification, tunnel vision, short termism, reduction of auditees’ motivation and disruption of the daily work scheme of the audited organisation) according to the auditees
Average score PA 1 (in %)
Average score PA 2 (in %)
Average score PA 3 (in %)
70.00
80.00
43.40
60.00 85.00
55.00 67.50
20.00 63.33
54.38
60.00
41.67
59.00
54.00
70.67
Katrien Weets
29