Strengthening Intercompany Financing and Treasury Compliance Through Data Analytics
This study develops a traceable control framework for intercompany financing and treasury compliance by validating Shell International Finance B.V.'s 2022–2024 reports and SOFR data, demonstrating that while aggregate totals remain consistent, specific maturity and currency details can change year-over-year, and highlighting how interest calculation conventions significantly impact financial outcomes.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
Multinational corporations are vast networks of companies that operate across borders, moving money between their own branches to fund operations, manage risk, and invest in new opportunities. To keep this internal economy running smoothly, these companies rely on intercompany financing, which is essentially a system of loans and cash transfers between different parts of the same organization. Because these transactions happen between related entities rather than independent strangers, tax authorities and regulators require them to be priced fairly, a concept known as the arm's-length principle. This means the interest rates and terms must look like what the companies would have agreed upon if they were dealing with the open market. Ensuring this fairness is a complex task for treasury departments, who must verify that every dollar of debt is classified correctly, calculated using the right calendar rules, and supported by solid evidence. If the numbers add up but the underlying details are wrong, a company could face significant compliance risks, even if its total balance sheet looks perfect.
A recent study by researcher Emmanuel Brefo explores how data analytics can help untangle these financial knots by looking closely at the public reports of a major global finance company, Shell International Finance B.V. The research focuses on a specific challenge: how to ensure that the numbers reported in annual documents are not just mathematically correct, but also semantically accurate. In other words, the study asks whether the labels attached to the money—such as the currency it is denominated in or the specific year it refers to—remain consistent over time. The researcher treated the company's published reports from 2022, 2023, and 2024 as a set of documents to be audited, extracting hundreds of rows of data regarding loans, interest rates, and maturity dates. By comparing the original reports with the comparative figures that appear in later years, the study created a traceable chain of evidence to see if the data held together under scrutiny.
The investigation began by checking the basic arithmetic of the financial tables. The researcher verified that the sums of smaller time periods added up correctly to the total for each currency and that the grand totals across all currencies matched the company's reported figures. These checks passed with flying colors; the numbers reconciled exactly, with only the tiniest rounding differences of one thousand dollars, which is well within the acceptable margin of error for such large financial statements. This confirmed that the company's internal math was sound. However, the study then moved beyond simple addition to look for deeper inconsistencies. It compared the 2023 report as it was first published against the 2023 figures that appeared in the 2024 report. This is a crucial step because companies often update their historical data in new reports, and those updates can sometimes introduce subtle changes.
What the researcher found was a hidden discrepancy that the total numbers completely masked. In the original 2023 report, a specific amount of debt was labeled as being in British pounds, while the same amount in the 2024 report was labeled as being in US dollars. The reverse was true for another portion of the debt. While the total value of the debt remained exactly the same in both versions, the currency classification for twenty specific data points had flipped. This discovery highlights a critical weakness in traditional auditing: if you only check the final sum, you will miss the fact that the underlying details have changed. The study did not determine which version was the correct one or why the change happened, but it proved that public data alone could reveal that a disagreement existed between two official versions of the same financial story.
To further test the reliability of financial calculations, the study also examined how interest rates are applied over time. The researcher used official market data for the Secured Overnight Financing Rate, a standard benchmark for short-term borrowing costs, to calculate what the interest would be on a hypothetical loan of one hundred million dollars. The calculation followed strict rules about how many days are in a year and how interest compounds. The study showed that changing a single rule, such as using a 365-day year instead of a 360-day year, would alter the final interest charge by nearly seventy-two thousand dollars. This demonstrated that even when the market rate is known, the final cost of borrowing depends entirely on the specific conventions used to count the days. The researcher also successfully recreated a published average interest rate from a three-month period, proving that the mathematical logic behind these calculations could be reproduced independently with high precision.
The core contribution of this work is a new way of thinking about financial compliance. Instead of treating a balanced ledger as the final proof of correctness, the study argues for a system that keeps the source of the data, the specific labels attached to it, and the calculation rules all linked together. The researcher proposes a framework where every number is tied to its original document version and its specific meaning, allowing auditors to spot issues like the currency swap that would otherwise go unnoticed. This approach does not claim to solve all tax problems or to prove that a company is avoiding taxes. Rather, it offers a reproducible method to distinguish between a simple math error, a change in how data is presented, and a genuine issue with how a financial transaction is structured. By preserving the history of how data was reported and how it was calculated, companies can build a more robust defense against compliance failures, ensuring that their financial stories are not just numerically accurate, but also factually consistent.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.