The Benefits Agency and local councils are striving to reduce benefit fraud by tightening procedures and introducing new technology. So far there is little to show for their efforts. Meanwhile the Government Statistical Service has cut the annual fraud bill by some 50m pounds after revisiting earlier calculations. The importance of this revised base line figure for fraud is that this is one of the targets in the Public Service Agreement between the Benefits Agency and the Treasury.What has emerged from the Government Statistical Service statement is that it is not possible to place the same reliance on figures of fraud as it is, for example, on the size of hospital waiting lists. Earlier in the year the Audit Commission published a figure for benefit fraud some 100m pounds below the latest estimate.

Benefit fraud figures have much in common with crime statistics, they both use indirect methods for getting a result. Fraud figures rely on sampling and crime statistics on police records. Sampling is not a precise science and many victims do not report crimes. As public services become more target oriented and budgets linked to results, ‘soft’ baselines are likely to lead to disputes.