Bunge Ibérica S.A. v Administración General del Estado

Distilled

The facts

The arrangement was a physical, zero-balancing pool with a daily sweep. Spanish accounts closed each day at nil, euro balances moving through a Madrid bank and dollar balances through London, with positions transferring to current accounts held with the Dutch pool leader. The Court found the leader's role purely administrative — a clearing function. It did not decide where funds went, could not refuse contributions, did not own the liquidity, and assumed no default risk, a matter neither its accounts nor the current account contract reflected.

The taxpayer had priced the two sides differently. Contributions to the pool were benchmarked against short-term bank deposit rates. Borrowings were benchmarked against Bloomberg yield curves using Bunge Ibérica's own credit rating at a five-year tenor. The authority rejected both. Contributions are not deposits, it said, but short-term loans between non-financial entities; and the curve selection was wrong on rating, which should be the group's, and on tenor, which should be the shortest available. It rebuilt the comparables accordingly and imposed symmetry between the two sides. The adjustment for 2010 to 2013 came to EUR 4,726,666.21.

The decision

The taxpayer lost at every stage — the tax tribunal, the Audiencia Nacional, and now the Supreme Court. The Court dismissed the appeal, confirmed the liquidation, and made no order as to costs.

On the substance it was explicit about what it was doing. The matter coincided substantially with one already resolved, and so, to preserve unity of doctrine and legal certainty, it would reiterate what had been fixed there — reproducing that judgment in practically its entirety. The doctrine stands: in cash pooling within a multinational group, the arm's length method requires that the interest rate on amounts contributed and amounts received by participants be symmetric, and that the credit rating applied to the lending transactions be the group's rather than the borrower's.

ajiho commentary

Confirmation is the point

This ruling adds precedential weight rather than content, and that is precisely why it matters. A single Supreme Court decision on a novel cash pooling question can be read as turning on its own facts. A second, for different years and on a separate appeal, reproducing the first almost word for word, is what converts a holding into settled jurisprudence — which is why the Court invoked unity of doctrine. For any group with a Spanish pool participant, the question has moved from whether the doctrine will hold to what to do about it.

But the Court fenced it in

Read the operative words carefully. The doctrine is prefaced "in the specific circumstances of the present appeal". The Court then observed that the tax authority's report on which the assessment largely rested, and the judgment under appeal, had no vocation of generality. And the State itself — the winning party — had argued that the method does not require this or that of cash pooling in general, citing authority that no doctrine can be projected onto every contract of this nature. A rule stated in universal form, hedged three times against generalisation. The Court leaves that tension standing. A pool whose leader genuinely performs functions, bears risk and owns the liquidity is not this case, and the judgment does not say otherwise.

Two things the Court did not do

It applied the 2022 Guidelines to 2010–2013 on a dynamic interpretation, and neither party objected — worth noting wherever the retroactive reach of Chapter X is still contested. And it never answered the suggestion that a pool denaturing the comparable uncontrolled price method calls for a profit split instead. But note why: the taxpayer itself accepted that neither had a profit split been applied nor the applicability of the comparable uncontrolled price method been challenged, so the dispute fell to be resolved under that method. The argument was conceded away rather than rejected.

What actually sank the file

Three choices, each arguable alone and fatal together: pricing contributions as bank deposits, using the participant's standalone rating, and taking a five-year tenor for balances swept nightly. The tenor point is the most easily avoided — a position cleared every night is not five-year money, and no rating analysis repairs a curve selected at the wrong maturity. Worth adding that the loss was substantially evidential. The taxpayer's own expert report was rejected as generic and unconnected to the facts of the case, and a cassation court cannot revisit the trial court's assessment of evidence.

The full reasoning on symmetry and the group rating is analysed in the FT Case on sentencia 985/2025, covered separately in this library. ajiho will continue to monitor Spanish cash pooling decisions.

Case reference

Bunge Ibérica S.A. v Administración General del Estado, Tribunal Supremo, Sala de lo Contencioso-Administrativo, Sección Segunda, sentencia núm. 489/2026, recurso de casación 2786/2024, 22 April 2026. ECLI:ES:TS:2026:1746. On appeal from Audiencia Nacional, Sección Segunda, 13 December 2023, recurso 120/2020; TEAC resolution of 8 October 2019. Ponente: Isaac Merino Jara. Corporate income tax 2010 to 2013. Primary source: official judgment.

The published text carries several internal inconsistencies: the earlier Supreme Court ruling is dated both 15 July and 15 April 2025, and the Audiencia Nacional judgment appears under two different dates and recurso numbers. The operative doctrine sentence is also grammatically defective, so the reasoning — that the method requires symmetry — rather than the headline sentence should be quoted.
This note confirms the earlier ruling covered separately in this library: Bunge Ibérica S.A. — Tribunal Supremo, 15 July 2025 →