Purpose – This paper aims to develop a classification scheme of different types of value for money(VFM) audits with different degrees of compliance audit, and to classify the performance audits carriedout by the Swedish National Audit Office (SNAO) during its first six years as an independent stateaudit organization reporting to parliament.

Design/methodology/approach – The empirical data were gathered from all of the 150 auditreports published by the SNAO from its establishment in 2003 to the end of 2008. Seminars werearranged to discuss the classifications for validation.

Findings – The focus on traditional VFM audits (the “Three Es”) is unusual. Most audits carried outby the SNAO combine different types of extended VFM audits with compliance audit. On the onehand, they audit how the government and/or central agencies fulfil their mandates (from good to bad).On the other, they audit how the government and/or central agencies adhere to legislation, rules andpolicies (right or wrong). In some cases, the SNAO equates compliance audit with performance audit.

Practical implications – The authors have verbalised and visualised performance audit activities of interest not only to state auditors, but also to external stakeholders. One practical implication is thatthe Swedish national audit committee has conducted an evaluation of the SNAO that is partly based onthe national report of this study, and has proposed a stronger focus on the Three Es.

Originality/value – The study addresses a new approach in terms of a classification scheme forperformance audits, consisting of eight types.

This paper reports from a study of how performance is stabilized by the Swedish National Audit Office (SNAO) in its audit reports. The paper builds on all the 206 reports published by the SNAO from its inauguration in 2003 until 2010 and analyzes the (28) audit of substance reports (Pollitt, Girre et al. 1999). By investigating the ways in which the SNAO makes it possible to answer the question of what constitutes good performance in the public sector, the paper suggests that the audit institution operates in three phases. In phase one, the SNAO must address the controversial task of defining performance. In phase two, this definition is tested by obtaining for creating evidence. In the third phase, the auditors mobilize benchmarks as a means to argue for good (or poor) performance. Unlike the arguments posed by the International Organization of Supreme Audit Institutions about using common sense to evaluate good performance, SNAO audit reports include arguments about what is “good” that emanate from other actors. The analysis of the paper builds on some analytical tools from Actor Network Theory. Specifically, the idea of a linguistic actor is developed and used as a means to show how the concept of “independence” helps the audit institution to become Mr. Somebody but impedes the possibilities of producing an answer to the question to what constitutes good performance. 

This paper builds upon a study of the Swedish Energy Agency’s (SEA) process of defining and accounting for output. New regulations have been implemented regarding the Swedish central agencies’ performance reporting. According to these, the agencies should define their performance independently and decide how to represent their work in their annual report. The single restrictive requirement is that the agencies should focus on their “output” in terms of volume and cost. Based on the study of SEA’s process of defining and accounting for output, this paper suggests that performance reporting can be a challenging task for an organization when no clear performance standards are set by the organization’s superior. Whereas researchers have often written about the challenges of accounting for “outcome”, this paper shows that even the concept of “output” can be problematic and that it cannot be taken for granted. Contrary to many other accounting studies, this paper also suggests that the idea of “becoming visible” through accounting is not necessarily met with resistance among individuals in organizations. On the contrary, it can be regarded as highly desirable. 

This paper builds upon a case study of the way organizational activities are represented with numbers in a performance report. A common notion in accounting literature is that accounting numbers enable “government at a distance” (Miller & Rose, 1990, p. 9) by making organizational activities visible, calculable and comparable. The paper shows that if organizational activities are to be represented with numbers that can be calculated and compared, these numbers must be made calculable. For this to happen, the represented activities must be framed in a certain way. Drawing on Goffman’s (1974) frame analysis, this paper suggests that organizational activities have to be keyed, i.e., transformed, into quanta and then re-keyed into generic concepts. Only then can the numbers be open for calculation by others and used as tool for governing. The paper also shows we should not overestimate the capacity of accounting numbers to enable government from a distance.