Page 6 - Data-driven Tax Administration

Advanced Analytics for Tax Administration: New
Opportunities, Old Challenges

By Daniel Sinnott, Chief Analytics Officer, Office of the Revenue Commissioners, Ireland

Introduction
Advanced analytics creates major opportunities for tax administrations to extract insight and
practical value from their data. In discussions of the topic, the focus tends to be on technical
challenges: What database management system should we deploy? What analytics software
package should we use? Which machine-learning algorithms should we apply to our data?

Despite the interest that such questions generate among data engineers and analytics
experts, they are in many ways secondary issues. Ireland’s experience, and that of other
administrations in the Advanced Analytics network of the OECD Forum on Tax
Administration (FTA), suggests that the key to making analytics work is to concentrate on
establishing the organisational fundamentals. Once these are in place, the technical
challenges become substantially easier to address.

This article will summarise the findings of a research project carried out by the FTA’s
Advanced Analytics network in 2015 and early 2016. The findings are based on insight
gathered through a survey of 18 FTA-member administrations, and a series of follow-up
interviews with analytics managers in different administrations.

Overall, the research of the FTA Advanced Analytics Network indicates that, while analytics
may be a technically complex topic, the principal challenges to be addressed are of a
practical nature. Administrations will greatly improve their chances of establishing effective
advanced analytics functions if they can:

1. Carefully match the right analytical technique to the exact problem and dataset at
hand;

2. Build strong working relationships between analytics and operational units;
3. Learn to manage the uncertainty associated with analytics projects;
4. Begin to manage and develop their data as a key resource in its own right;
5. Invest in the change-management efforts needed to bring analytics out of the lab

and into the field.

Overview of activities
Advanced analytics is the practice of using statistical techniques to make predictions and
draw inferences about cause and effect. From its initial use in the selection of cases for audit,
the scope of advanced analytics applications has broadened, to the extent that analytic
techniques are now used to optimise debt-management processes, secure filing and

6