ATO Prevails in Challenging the Procurement Transfer Pricing for Ampol

On February 20, 2023 Ampol Limited settled its transfer pricing dispute with the Australian Tax Office (ATO), which involved the transfer prices paid by its Australian affiliate to its Singapore affiliate for refined fuel products. This years at issue were from 2014 to 2022, and the agreement provides certainty for the multinational and the ATO through 2033. Procurement affiliates have been a point of emphasis for the ATO in recent years, and have been gaining the attention of other tax authorities, including the US IRS.

Ampol is an Australian petroleum retail distributor with annual revenues that have been approximately 20 billion Australian dollars (AU$). Ampol Singapore was established in 2013 as its procurement affiliate. The ATO has been reviewing the transfer pricing for foreign based procurement hubs, which I discussed in a recent piece in Journal of International Tax Management.

Tables 1 and 2 provide an illustrative model of this transfer pricing based on financial and tax provision data noted in the Annual Reports of Ampol Limited. We also use this illustrative model to discuss potential benchmarking issues. The model assumes that sales = AU$ 20 billion per year, cost of goods = 91% of sales, and operating expenses = 5.5% of sales. These assumptions are consistent with the financial data for Ampol and imply consolidated profits = 3.5% of sales. The Australian affiliate incurs AU$ 800 million in operating expenses (4% of sales), while Ampol Singapore incurs AU$ 300 in operating expenses (1.5% of sales).

Table 1: Three Transfer Pricing Models for Ampol (AU$ Millions)

Table 1 also considers three transfer pricing policies. Policy A considers what may be an aggressive pricing policy where Ampol Singapore charges Ampol Australia an intercompany (I/C) price = 94% of sales. Ampol Singapore’s gross profits represent 3% of sales. Since consolidated gross profits = 9% of sales, Ampol Australia’s gross profits represent 6% of sales. Under this policy, the Singapore affiliate’s profits represent 1.5% of sales or a 100% markup over its operating expenses. The Australian affiliate’s profits represent 2% of sales or a 50% markup over its operating expenses.

Policy C considers what may be a conservative pricing policy where Ampol Singapore charges Ampol Australia an intercompany (I/C) price = 95% of sales. Ampol Singapore’s gross profits represent 2% of sales. Ampol Australia’s gross profits represent 7% of sales. Under this policy, the Singapore affiliate’s profits represent 0.5% of sales or a 33.33% markup over its operating expenses. The Australian affiliate’s profits represent 3% of sales or a 75% markup over its operating expenses.

Policy B considers what may be an intermediate pricing policy where Ampol Singapore charges Ampol Australia an intercompany (I/C) price = 95.5% of sales. Under this policy both affiliates have a markup over operating expenses near the overall markup over consolidated operating expenses.

Table 2: Effective Tax Rate Under Policy A and Under Policy C (AU$ Millions)

Table 2 considers the income tax implications of both the aggressive policy and the conservative policy. Under policy A, Australian taxes = AU$ 120 million, since the tax rate = 30%, and Singapore taxes = AU$ 51 million if the tax rate = 17%. The implied worldwide effective tax rate =24.43%. Under policy C, Australian taxes = AU$ 180 million and Singapore taxes = AU$ 17 million. The implied worldwide effective tax rate =28.14%.

Note E1 of Ampol’s Annual Report discuss the taxation of the Singapore affiliate and this transfer pricing issue with the following statement appearing in its Annual Report for 2019:

… the Australian Taxation Office (ATO) had not finalised its position in relation to the extent to which earnings from the Group’s Singaporean entities would be subject to income tax in Australia. Due to the uncertainty over the ATO’s final position, the Group has estimated and recognised tax liabilities for 2014 to date based on the income tax rate of 30%, being the Australian corporate income tax rate … The cumulative tax expense for the differential between the Australian and Singapore tax rates recognised in the financial statements from 2014 to 31 December 2019 is $163 million … If the outcome of the ATO’s decision is in Caltex’s favour, an amount of income tax expense recognised to date could be written back in future periods. If the tax matter is resolved such that the ATO’s position is sustained, there would be no impact on the Caltex income statement or net assets.

The reported effective tax rate was approximately 28%, which implies Ampol took a tax reserve based on the conservative transfer pricing position noted in table 1. The ultimate settlement between Ampol and the ATO appears to be consistent with this conservative position.

Choice of Tested Party and Benchmarking

Any benchmarking analysis under the Transactional Net Margin Method (TNMM) would begin by deciding whether the tested party should be the procurement affiliate or the retail distribution affiliate. Before we explore this issue, the overall 9% consolidated gross margin for Ampol can be seen as a comment on some faulty analysis by the Panamanian tax authorities in a litigation I criticized in a February 23, 2021 blog post entitled “Abuse of the Resale Price Method for a LATAM Petroleum Distributor.”

The tax authorities examined the gross margins for four companies the Administrative Tribunal (Case No TAT-RF-062, September 2020) did not disclose. I noted that the tested party in this litigation was a wholesale distributor of petroleum, which is similar to the functions of Ampol Singapore. While the functions of this distribution affiliate were very limited, the functions of the alleged comparable companies were likely to be more significant. The 9% consolidated gross margin for Ampol Limited represented the combined compensation for both wholesale distribution and the retail operations of the Australian affiliate. Had Ampol Limited been used as a comparable company in this Panamanian case, its gross margin would severely overstate the appropriate gross margin for the wholesale distributor.

Similarly, a benchmarking exercise for the Ampol issue would require a reliable means for splitting the overall gross margin between the Singapore affiliate and the Australian affiliate. One approach would be to use the Singapore affiliate as the tested party and then estimate the appropriate markup over its operating expenses using third-party wholesale distributors of fuel as comparables. Our prior discussion only picked one such company - World Fuel Services. Its operating expense to sales ratio average 1.66% over the 2011 to 2019 period and its operating margin averaged 0.7%. Such an approach would likely yield a result somewhere between Policy B and Policy C.

Another approach would be to select the Australian retail distribution as the tested party in a TNMM analysis that examined the financials for third party retail distribution affiliates. In order to determine the appropriate relative profitability for Ampol, a capital adjusted return to value-added expense approach should be utilized. Whether the appropriate markup should be closer to 50% (policy A) or 75% (policy C) would depend in part on the level of operating assets held by the Australian retail affiliate relative to its modest level of operating expenses.

References

Harold McClure, “Centralized Procurement Hubs: A Co-Sourcing Model,” Journal of International Tax Management, April 10, 2023.

Harold McClure, “Abuse of the Resale Price Method for a LATAM Petroleum Distributor,” EdgarStat Blog, February 23, 2021.