« The question of the major challenges in the digitalization of taxation has a technological and legal dimension. »

Interview with David Hadwick, Doctoral researcher (FWO Aspirant/FWO PhD Fellow), Faculty of Law, Anvers University

( (Navigate freely through the interview by clicking on the menus below)

1. David Hadwick
2. Opinion about EDHEC Tax digitalization study
3. Difference between the evolution of practices on the administration side and the company side
4. Next steps or major challenges in digitalization of tax
5. New key skills for tax professionnals
6. Download EDHEC Study on tax digitalization

Could you please introduce yourself, David?

My name is David Hadwick, PhD researcher at the Centre of Excellence DigiTax of the University of Antwerp and fellow of the Research Foundation for Flanders (FWO) on an individual grant. My research relate to the use of AI by tax administrations and their impact on the fundamental rights of taxpayers.
Prior to Academia, I was working as tax and financial advisor in Luxembourg (Arendt & Medernach), advising large businesses, involved particularly in the technology sector.
I hold a master’s degree in corporate law from the University of Maastricht and another master degree in international business taxation from the University of Tilburg.

——————————————————————————————————————————————————————————————

David Hadwick is deeply involved in the Tax digitalization research, both from an academic and from a professional perspective. He is involved in several research groups and projects on the use of AI for tax management purpose, mainly from the tax authorities perspective whereas in parallel the EDHEC Augmented Law research produced in partnership with Algonomia and Fidal focuses on the companies practices.

Very early June 2024, he will be a panelist about “ The impact of digitalisation on the right to fair treatment of the taxpayer” during the 9^th International conference entitled Toward a Digital Taxpayer Bill of Rights that will be hosted by the DigiTax Centre at the University of Antwerp early June. (https://taxpayer-rights.org/international-conference/)

Among his most recent publications:

Hadwick, D. (2022). EUcrim :‘Error 404 – Match not found : Tax enforcement and law enforcement in the EU AI Act’

Hadwick, D. (2022). Peer Reviewed Articles:‘Behind the One-Way Mirror: Reviewing the Legality of EU Tax Algorithmic Governance’. EC Tax Review, 31(4).

Hadwick, D. (2023). Breaking the fiscal omerta: the roadmap to transparency in EU tax algorithmic governance. In Proceedings from the First Annual International FIRE Conference: 10th–11th of November 2022, Örebro University, Sweden/Kristoffersson, Magnus [edit.] (pp. 107-121).

Find out more

——————————————————————————————————————————————————————————————

What elements surprised you the most when reading EDHEC study on tax digitalization in partnership with Algonomia and Fidal? Conversely, which were the most predictable to you?

Because I previously worked with large MNEs, I was aware of the fact that large groups had already integrated the vital necessity of adopting specifically dedicated technological tools to deal with tax externalities. A dynamic regulatory landscape can be a blessing for a company, but for a large group it is first and foremost a curse, to which one must be able to adapt. Tax-Technology can help espouse the dynamic from the get-go. The indicators chosen and the answer provided by respondents generally reflect the pre-conceptions lawyers with a tech-background would entertain. These indicators empirically confirm a number of these pre-conceived notions, e.g.: access to budget for digitalization projects, dependance on other departments and different barriers to digitalization. Having observed these same phenomena on the market, these points empirically confirmed foreseeable observations.

Nonetheless some elements of the study were surprising, particularly: the proportion of teams declaring having a member specifically dedicated to managing technological tools. In my professional experience, the development and deployment of technological tools was operated by hybrid teams formed ex officio from tax and IT teams. The development of ‘tax-tech’ roles within companies highlights the importance of fiscal governance tools on the market. The study demonstrates how ‘tax-tech’ is becoming part of the company’s lexicon, and the high saliency of that factor within large enterprises. Certain examples cited, such as the high proportion of respondents having deployed their own risk management systems, show the importance of that change. It is surprising to see these developments so soon, a couple of years after leaving the industry.

What is the difference between the evolution of practices on the administration side and the company side?

Based on this study, it is clear that technology is primarily a mean for corporations to automate tasks that are recurrent, voluminous, with low added-value.
In many aspects, the experience of the tax administration is different because technology was first and foremost used to operate the selection of taxpayers for audit, which is a high added-value priority for the organization. The tax administration’s core business is forensics accounting and the selection of taxpayers for audit, the control of fiscal risks and the guarantee of compliance. Accordingly, the technological tools used to operate that task were of the highest importance for the administration. Yes, the tax administration also uses technology to automate low added-value tasks such as answering taxpayer queries through chatbots, but this does not constitute the core of their use of technology.

Because automation related to a high priority process, tax administrations invested more heavily in the use of technology, when compared to companies. It is important to realize that the use of data analytics and machine learning ushered in the most profound reform of tax administrations, the move towards a centralized statistical selection of taxpayers. Such a move would not have been possible without technology. Hence, the use of fiscal governance tools entirely changed the structure and functioning of the tax administration: reducing its manpower, reducing discretion for field agents, selecting the majority of taxpayers statistically. For corporations, technology is used mainly for low added-value tasks, in turn this has an impact on the degree of digitalization, maturity and importance of tax-tech roles within a company. In my opinion, companies cannot automate their core business in the same way as the tax administration, because for companies technology is synonymous with low added-value. Thus a company that would automate their high-priority task would send the message that their service is low-value. For the tax administration, being low in value is also of high priority because it is synonymous with more efficiency and less cost for the taxpayer.

Nonetheless, using AI for risk-management whether as a company or a tax administration, is a tentacular objective composed of thousands of sub-tasks. Thus, for companies who face a large population of customers and offer (semi-)personalized pricing and assistance services, e.g. insurance, many core tasks will be identical. Tax administrations did not invent the formula used today to select taxpayers. The formula was long used by banks and financial institutions to calculate our car insurance premiums, our medical costs, our default risks. Tax administrations adopted the formula and sometimes perfected it, by virtue of access to more powerful tools and enlarged access to taxpayer data. Hence, it makes sense that a number of processes will be the same when comparing tax administrations with these kind of market players.

How do you anticipate the next steps or major challenges in the digitalization of taxation?

This depends on what is meant by the digitalization of taxation. In my opinion, this question has a technological and a legal dimension.
From the perspective of the adoption of tax-tech, the integration of technology to our previous/current fiscal paradigm creates a new phenomenon formed from the conjunction of the human and the technological tool. The next steps of the digital transformation of taxation is to understand how to interact with that new phenomenon. For instance, these dimensions include:

– The degree of discretion of the tool and the degree of human control over the tool – to what extent should the tool carry its task with or without human involvement and where will the human control be inserted. Many are speaking of ‘human-in-the-loop’ forms of governance, not many are capable of explaining how it will look in practice.

– Exogeneity – to what extent can external users exercise some control over the use of the tool through transparent information, direct interaction, right to refuse processing, etc.

– Capability – to what extent do the tool increase human capability and to what extent this transformation is desirable.
Through the integration of technology, particularly machine learning, organizations are including an agent. Similarly to any human colleague, organizations must understand how to communicate with that colleague, its limitations, capabilities, externalities, and how the new agent should interact in consumer-facing relations.

From a legal perspective, the digital transformation of taxation equally transforms the constitutional balance of powers between the different players involved, i.e. taxpayers and tax administration, and even service providers to a certain extent. This means that the adoption of new tools sometimes disrupts the regulatory landscape and renders certain norms obsolete. Legally, market players must understand the impact of the adoption of certain tool on their own compliance prerogatives.

For tax administrations, this means understanding the impact of their tools on the rights of taxpayers, fundamental, procedural or otherwise. For instance, to what extent legality, transparency and fairness are guaranteed when adoption a new technological fiscal governance tool. What values are being upheld through the adoption of that tool, whether efficient tax collection or equitable collection of tax revenue. It is important to realize that from a statistical perspective, these values are often opposed. Hence, understanding the impact of a tool is the first step in ensuring trust, and consensus over their use.

For companies, the perspective is slightly different, but mostly the same. Companies must ensure that their tools seamlessly fit within the regulatory landscape and limit its disruptive effect across the board. Sometimes, this means that internal processes must be adapted to the technology, not the opposite. Understanding how to fit technology, through individual increments within a legacy system composed of a myriad of layers, that is the next challenge.

In your opinion, what are the new key skills that tax professionals, whether in companies, consulting firms, or tax administrations, must quickly acquire?

In my opinion, tax experts whether as consultant, lawyer or tax official are increasingly developing a side expertise in data science and statistics. This constitute a key competence for anyone who wishes to evolve within the fiscal landscape. The goal is not for tax experts to become software engineer or develop machine learning models from scratch, these careers are by all means separate and will remain so. However, the tax regulatory landscape creates an environment where any fiscal expert is relying on a flow of data that is too large to be processed by hand.

When operating internal controls, a VAT consultant cannot process all the invoices sent within the year. A small selection will be operated on the basis of statistics, preferably leveraging some data from previous years. A diligent internal control requires thus a dual process where a selection is first informed by data science and statistics, and then by taxation rules. True expertise in taxation requires understanding both aspects of the process. The same way as a painter understands its brushes, a tax expert must understand its tools, which go beyond the taxation rules only, but includes a myriad of tools to operate data analytics, ERP and statistics.

The tax expert must not become the engineer or data scientist, but must understand enough to communicate the legal imperatives into computable requirements. The tax expert must be able to become the bridge between the different fields of expertise involved.