Determining the SGI – the process, organisation and governance

In recent years, the Swedish Social Insurance Agency has made SGI management more efficient through increased digitalisation and automation. Furthermore, at a general level, SGI decisions reflect an individual’s actual income circumstances when a person has submitted new income details. However, there are systematic discrepancies between an individual’s actual income and the SGI determined by the Swedish Social Insurance Agency in cases of part-time work and when a long time has elapsed since the last income check. The audit also shows that the Swedish Social Insurance Agency takes limited account of risks specifically linked to the quality of individual-level income data from the Swedish Tax Agency, even though such income data may be misleading or incorrect, for example as a result of criminal schemes.

The audit shows that many SGI decisions are based on out-of-date information, which often leads to an SGI that is too low. The majority of the Swedish Social Insurance Agency’s SGI decisions – around 7.5 out of 9 million per year – are based on previously determined SGI figures and are subject to limited checks. The audit shows that the accuracy of the decisions deteriorates significantly the older the most recently determined SGI figure is. One reason why the SGI is set too low is that people often confirm that previous information remains valid, even when their income circumstances have changed. This contributes to out-of-date information forming the basis for the level of benefits and suggests that the Swedish Social Insurance Agency needs to strengthen its preventive measures at the application stage.

The audit shows that the design of the SGI regulatory framework can lead to inefficient case processing. For example, the Swedish Social Insurance Agency’s investigations into how far back in time the SGI can be secured (SGI security) can be very demanding in terms of resources, as the regulations can be complicated to apply. The audit also shows that there are shortcomings in SGI security processing and that the risk of errors is high.

Another problem area is annual working hours, which are often difficult for individuals to specify as few people understand what the term means. At the same time, the Swedish Social Insurance Agency checks the information on annual working hours only to a very limited extent. This leads both to resource-heavy follow-up checks and the risk of incorrect annual working hours being determined. In addition, for technical reasons relating to the system, the Swedish Social Insurance Agency investigates and determines annual working hours in situations where it is not really needed, which contributes to further inefficiency.

The audit shows that the Swedish Social Insurance Agency has not designated or clarified which department has overall responsibility for SGI governance and monitoring. While case processing is spread across several departments, the division of responsibility is incomplete and not clearly set out in governing documents. Productivity and quality monitoring is fragmented and does not cover all SGI processing, which means that the Swedish Social Insurance Agency lacks an overall picture of how SGI is managed. This, in turn, increases the risk of inconsistency, inefficiency and incorrect SGI decisions.

The audit shows that the SGI management is particularly vulnerable when there are high volumes of cases and resource constraints. During the COVID-19 pandemic and in the years that followed, the Swedish Social Insurance Agency made extensive deviations from the SGI process in order to cope with the workload. These shortcuts have increased the risk of SGI decisions being based on incorrect information and have contributed to a rise in the number of reassessments.

The Swedish National Audit Office considers that the design of the regulatory framework and the substantial caseload are the underlying reasons why the administration of SGI is resource-intensive. At the same time, there is scope for efficiency improvements within the framework of the current system, for example by further developing automated checks. However, more far-reaching efficiency improvements also require action from the Government in the form of initiatives to amend the regulations.