Non-profit organisations that manage public funds are more likely to keep good financial records of their spending if there is a possibility that they will be audited. They will also deliver better value for money (sometimes returning funds they do not need to use) and are more effective in promoting their activities through the media.

These are among the findings of research by Niklas Bengtsson and Per Engström, published in the June 2014 issue of the Economic Journal. They report the results of an unusual policy experiment in Sweden on the effectiveness of audits in the public sector.

In 2008, the Swedish national audit office demanded that their colleagues at the Swedish foreign aid agency (Sida) stepped up their efforts to monitor results and audit their partner non-governmental organisations (NGOs). Civil society opponents of this proposal warned that such audits could paralyse the efforts of NGO staff.

The idea of the researchers was to see if Sida''s partner organisations would behave differently if the threat of audit became more palpable. The researchers and Sida therefore worked out a clever experimental policy design. Together, they implemented a large-scale audit of the non-profit organisations that received tax money from Sida. A key aspect of the experiment was that half the organisations were warned about the upcoming audit.

Niklas Bengtsson says:

''That the warnings about the upcoming audit were randomised was a critical part of the field experiment. It means that we can conclude that any difference in subsequent performance can be interpreted as the impact of the threat of audit.''

At the end of the year, the researchers compared the financial records and outreach of the two groups. The results show that the warned organisations were significantly better at documenting their expenditures – the number of reported irregularities fell by 20%.

In addition, the notified organisations also returned a significant fraction of the taxpayer funds, while the organisations that had not been warned revealed a clear tendency to maximise expenditures. Despite the reduction in expenditure, the notified organisations improved their outreach: their projects achieved more media coverage and their staff''s self-assessments were more positive.

Niklas Bengtsson adds:

''That fewer irregularities were found in the warned group is not surprising. But that they decided to return funds, while improving outreach, is the more interesting result.

''It suggests that, absent control, non-profit organisations have a tendency to spend public funds without improving outreach.''

The study was motivated by recent criticisms of the so-called ''New Public Management'' doctrines. Previous research has suggested that financial audits and an increased focus on quantitative evaluations can backfire, particularly if those monitored are ''mission-oriented'' rather than focused on profit. The researchers wanted to conduct an empirical test of these hypotheses as directly as possible.

Since the mid-1980s, modern welfare states have increasingly relied on quantitative evaluations and audits to improve the performance of the public sector. But a common criticism is that they tend to shift focus away from what is really needed to what looks good on paper. As a consequence, public servants – who once entered the sector with the hopes of contributing to the public good – lose their motivation.

But previous research has had little to say about the net effects of audits and whether the positive effects are larger than the negative effects. There is considerable public debate about the impact of audits in Scandinavian countries, where the public sector constitutes nearly half of GDP.

''Replacing Trust with Control: A Field Test of Motivation Crowd Out Theory'' by Niklas Bengtsson and Per Engström is published in the June 2014 issue of the Economic Journal. The authors are at Uppsala University and the Uppsala Center for Labor Studies (UCLS).