Supporting and enabling tax authorities: the role of advanced analytics


Tax authorities may not be everyone’s favourite organisations. But around the world, they have a key role in collecting revenues that enable governments to spend money on essential public services. You only need to read the media coverage of the tax avoidance of some of the big tech companies, and the efforts by tax authorities to ensure they pay tax, to understand that this is an important and emotive issue.

It is, therefore, good news to find from this brand new publication that OECD countries are applying advanced analytics to improve tax administration. Tax work is a natural area for analytics, as it enables decisions supported by evidence. While analytics does not fundamentally change what is done, it improves the way it can be done, and hopefully improves both efficiency and effectiveness of operations.

According to this newly published OECD report we at SAS can proudly call ourselves the primary analytical partner to the OECD tax authorities. And it is not just simple statistics we provide the Tax Authorities with; The analytical disciplines involved are multi-faceted.

A developing picture

It is probably no surprise to find that the picture of analytics used in tax authorities is developing and changing over time. The first area where analytics was used was in selecting cases to be audited. This use has now broadened significantly, for example into making decisions on how to manage debt, improve payment compliance and reduce non-payment. It has also been used in strategic decisions such as designing communications with taxpayers to improve compliance and payment, and understanding the impact of policy changes on taxpayers.

Alongside this expansion in applications, tax organisations have also broadened the type of analytics techniques they use. Initial applications were all about general modelling. Applications now include detailed models to assess individual risk types, mining correspondence for insights into particular issues, and use of predictive modelling.

As organisations’ way of using analytics matures, they encounter different problems and challenges. In the early stages of analytics use, the main issues tend to be organisational and structural, with the key problem being to balance centralisation with decentralisation. Centralisation supports the development of a strong, cohesive analytics function, which in turn enables the tax authority to build skills and control the quality of its analytical work. Moving forward with analytics, it needs to become more integrated into the whole organisation: more ‘the way we do things’. As a result, it becomes more decentralised, as analytics teams establish stronger relationships with business teams.

Tax authorities also face an interesting challenge in integrating analytics into their way of working: the uncertainty of most advanced analytics projects. The OECD report likens analytics projects to R&D work, and suggests that experimentation is vital. Most organisations use an iterative approach to analytics, with ‘test-and-learn’ being a common approach that enables ongoing improvement in response to feedback. This places a burden on analytics teams to follow up projects and ensure that they continue to provide value.

Challenges and competence

From the SAS point of view, one of the most interesting sections of the report concerns the software and tools used for analytics. Commercial analytics software is the most common, particularly in the early stages of analytics use. However, open source options are likely to become more popular as organisations’ use of analytics matures. The report notes that open source technology is significantly cheaper, but does bring its own challenges. These include, in particular, ensuring that the organisation has the necessary skills.

The report also noted that data quality and governance is a key issue for tax authorities in increasing their use of analytics. They need to move from seeing data as a side-effect of their activities to an important input into analytical processes, and therefore an asset to the organisation. This means seeing data in new ways, for example, building representative databases, and actively seeking out new sources of data that may improve insights.

A place for ongoing collaboration

It will come as no surprise that the OECD report concludes by recommending ongoing cooperation between tax authorities to improve analytics practice. The report suggests the importance of sharing information about three key areas: best practice, knowledge about specific applications of analytics that have proved effective, and finally, the sources and types of data that have been found to be most useful for analytics. In other words, the entire process from data load to engagement of decision support in the field operations is covered in one coherent solution.

As one of the key providers of commercial software, it is very satisfactory for us working at SAS to be part of these collaborative conversations, and help to share good practice around the world.


About Author

Mads Krogh

Business Advisor

Datamining and explorative analytics thought leader. Targeting the dramatic change in the use of data to explore, detect and prevent fraud, non-compliance and waste in transaction driven organizations as we know it today. Firmly believe that this is one domain where analytics really can do a significant difference. I find it rewarding to support the anti-fraud community’s knowledge through my white papers and blogs, conference speaking engagements and participating in public debates.

Leave A Reply

Back to Top