Transfer pricing continues to generate friction globally. It remains vulnerable to manipulation as economies digitize and multinationals seek tax savings. In this climate, regulators are trying to stem the shift of profits from their origins into affiliated entities in low-tax locales.
The shifting tide of transfer pricing regulations and compliance is ripe for disruption. Artificial intelligence and machine learning tools have potential to satisfy demand for consistent valuation and requirements. An open-source and public-facing AI model that provides tax transparency, and a safe harbor for taxpayers who abide by its valuations, is key for compliance.
Where AI Shines
The field of AI is fertile ground for solutions to transfer pricing oversight. Multinational enterprises are being asked to place a value on transfers between controlled entities—often “pricing” things such as intangible assets that, absent transfer pricing concerns, would never be assigned a valuation.
There is no repository of transfers in the marketplace to consult, so an alternative is needed. AI, which can simulate market conditions, could assess and predict the value a given transfer would have in an imagined arm’s-length transaction.
More broadly, AI and machine learning have potential to resolve inconsistencies. An open-source, publicly accessible AI model could revolutionize the practice by providing accurate, reproducible, and consistent valuations.
Transfer pricing’s efficacy hinges on precise valuations—even of difficult-to-valuate intangibles. Perhaps more important, taxpayers and regulators must have a mutual and clear understanding of the factors in that valuation. There must be a known and agreed-on set of underlying valuation factors to have sufficient overlap from which to engage in a productive back-and-forth.
AI’s ability to analyze massive data sets and consistently apply sophisticated algorithms would produce more precise and consistent valuations, with substantially lower administrative overhead for both taxpayers and regulators.
Companies and governments could harness this technology to sift through vast troves of financial data, adjust to market conditions, and calibrate valuations based on transactions across jurisdictions. A sophisticated AI model might generate an arm’s-length valuation by executing the underlying transaction in a model of the market writ large—in more of a prediction than a calculation.
AI could also make iterative adjustments as it learns and adapts to the market. This way, it ensures transfer pricing valuations remain relevant and accurate with a fidelity to time that auditors or compliance experts could only dream of.
This AI-driven approach would ensure tax transparency, making data and models readily available for public review. In exchange for adoption, transparency, and adherence to the model’s valuations, taxpayers would gain a safe harbor.
This approach would boost adoption by mitigating risk of non-compliance and fines. With this opportunity for increased corporate transparency, regulators must seize the moment to develop and provide a carrot along with the sticks of fines and penalties.
Practice of Formulas
Following the OECD’s base erosion and profit shifting initiative, international transfer pricing regulations have only become more complex. Documentation requirements were substantially expanded to ensure valuation mechanisms aren’t being used to facilitate profit shifting.
The writing is on the wall: Transfer pricing has been identified as a key way to combat base erosion. This means increased regulatory scrutiny of an area of law that was already a compliance challenge.
Aligning income with the economic activities that produce that income is one main target of reform—curtailing the ability of multinationals to separate income and shift it to lower-taxed subsidiaries.
Transfer pricing was replete with formulas and algorithms even before the Organization for Economic Cooperation and Development’s BEPS and war on international tax avoidance. In the US, if a company changes the price of an import after it has been imported, the company can adjust their transfer prices and potentially get a refund for any overpaid import duties.
However, this refund is only available if the original price was set based on a clear and predetermined policy for valuing these imports. Put differently, refunds of excessive duties paid are only available if the initial valuation of an import is based on a documented and previously established objective valuation policy.
The valuation policy must comply with the US Customs and Border Protection’s objective formula standard and be established and documented—pricing can’t have been at the whims of the taxpayer. It should demonstrate that the transaction value declared to customs reflects an arm’s-length price—the price that would apply had the related-entity exchange been transacted under normal market conditions.
Ultimately, the valuation must be made through a formal policy that aligns with Section 482 of the Internal Revenue Code. Even if the entire process is handed over to an AI model, accountants will unquestionably be, and probably already are, using models in those internal policies.
Looking Ahead
Tax disputes involving transfer pricing are the shark attacks of the tax world—you may be more likely to get clipped by a boat motor, but nothing strikes fear into the heart of a swimmer quite like the sight of that dorsal fin.
So too with transfer pricing disputes, a multinational may be more likely to get hit with unpaid value-added tax in an obscure jurisdiction, but few accounting missteps take a bite out of a corporation’s bottom line quite like a transfer pricing tax claim.
AI inevitably will be adopted and integrated into every facet of the tax field. What remains to be seen is which side of the compliance and regulatory dichotomy will own the models.
Properly envisioned and funded, AI models could be developed and shared through intergovernmental organizations and made available without cost—offsetting the cost of compliance for developing countries. A model, once developed, would have de minimis operational costs that could be absorbed by those entities.
In the long term, deployment of such an AI tool in the realm of international taxation—particularly with regard to transfer pricing, would generate fiscal benefits far exceeding the initial or ongoing costs.
Transparency has always been a critical element to tax justice and ensuring corporations are paying their fair share. The complex transfer pricing regulatory landscape, coupled with AI’s rapid development, can further the interests of transparency.
Let’s use AI to lure corporations into conspicuous compliance. An AI tool at the border would be an immediate boon to world trade.
Andrew Leahey is a tax and technology attorney, principal at Hunter Creek Consulting, and adjunct professor at Drexel Kline School of Law. Follow him on Mastodon at @andrew@esq.social
Read More Technically Speaking
To contact the editors responsible for this story: