Cameco v. The Queen: Prologue and Round II Preview

April 30, 2021 by Harold McClure
About the Author
Harold McClure
is an economist with over 25 years of transfer pricing and valuation experience.
Dr. McClure began his transfer pricing career at the IRS and went on to work at several Big 4 accounting firms before becoming the lead economist in Thomson Reuters’ transfer pricing practice. Dr. McClure received his Ph.D. in economics from Vanderbilt University in 1983.
Read more

More articles by this author

Find Harold McClure on LinkedIn

In February 2021, the Supreme Court of Canada declined to hear the Canadian Revenue Agency's (CRA) appeal in its case against Cameco Corporation. This marked the end to to Round I of CRA's unsuccessful challenge of the multinational's intercompany pricing with its Swiss affiliate for the years 1999-2014, but the decision addressed only the first 8 years, leaving the lucrative 2007-2014 period open for review. 

The Case: Cameco Corporation v. The Queen / Her Majesty the Queen v. Cameco Corporation 

CRA challenged the intercompany pricing between the Canadian producer of uranium and its Swiss distribution affiliate, which left almost all of the multinational’s profits over the 1999 to 2014 period with the Swiss affiliate.  In a lengthy decision, the Tax Court of Canada ruled in favor of the taxpayer in the September 26, 2018 decision in Cameco Corporation v. The Queen (2018 DTC 1138). 

The CRA appealed this decision only to lose in a June 26, 2020 decision by the Federal Court of Appeals in Her Majesty the Queen v. Cameco Corporation (2020 FCA 112). The Appeals Court noted that while the Tax Court devoted 197 pages of its decision to excerpts from various documents and the expert opinion reports, this information provided very little analysis of the evidence, describing as a “factual data dump.” The Appeals decision centered on the position of the CRA that no arm’s length party would have entered into the actual intercompany contract.

This contention was central to the CRA’s argument, which rested on an attempt to recharacterize the intercompany contract. The Tax Court and the Appeals Court rejected the CRA’s claim as well as the right of the CRA to recharacterize the intercompany transaction.

Our prologue presents two alternative analyses. The first analysis, which is the subject of this blog post, addresses the question that the CRA wished to pose under its recharacterization: what would represent an appropriate gross margin for a mere distributor of uranium?

The second analysis notes that third-party marketers such as Curzon Uranium Trading Limited enter into long-term offtake agreements, which would be an appropriate means for evaluating whether the actual intercompany agreement was arm’s length. The post describes the business and financials for Curzon, while a follow-up post will apply the logic of long-term upstream agreements to the taxpayer’s position in the Cameco litigation.

Analysis 1: A Co-Distribution Model for the Gross Profits for the Swiss and US sales Affiliates

While Cameco’s Swiss affiliate acted as the buy/sell distributor for uranium produced by its mining affiliate, a US affiliate performed many of the actual selling functions and the Swiss affiliate employed a modest number of employees. Dr. Anthony Barbera, as one of the CRA’s expert witnesses, asserted that the Swiss affiliate’s gross margin should be only 3.5% of sales, which represents a reasonable routine return to distribution since the marketing costs of the Swiss affiliate, including the efforts of the US sales affiliate, represented 2.5% of sales.

The CRA also hired Dr. Delores Wright to evaluate the appropriate gross margin for the Swiss affiliate under a variety of assumptions. One of her scenarios concluded that its operating margin should be a mere 0.1% if this entity performed very limited functions. Another scenario was based on the Swiss affiliate bearing all distribution functions, with her conclusion being that its gross margin should be 6%.

The IRS added to the confusion with its audit of the transfer pricing for the US sales affiliate for the years from 2007 to 2010. Initially the IRS expected an additional $122 million for the four-year period, which translates into an additional $87 million per year in US taxable income. Table 1 shows the Cameco revenues from 2007 to 2014 in terms of Canadian Dollars (C$), which we have translated into US Dollars (US$) using the average exchange rate for each year. If sales were approximately $2.2 billion per year, the IRS expectation would be to increase the US profit margin by 4%. The IRS, however, drastically reduced its expectations during the audit.

Table 1: Cameco Revenue from 2007 to 2014 (in Millions)

Year

Revenues (C$)

C$/US$

Revenues (US$)

 

2007

2309.741

1.0742

$2150.230

 

2008

2859.158

1.0668

$2680.209

 

2009

2314.985

1.1414

$2028.124

 

2010

2123.655

1.0301

$2061.551

 

2011

2384.404

0.9890

$2410.863

 

2012

2321.473

0.9994

$2322.847

 

2013

2438.723

1.0300

$2367.711

 

2014

2397.532

1.1047

$2170.334

 

 

Table 2 presents an illustration of the original IRS position as well as a more reasonable position. Our illustration assumes sales = $2 billion and US operating expenses represent 2% of sales or $40 million. If the US entity is a commission agent supporting the Swiss distribution affiliate, then a 2.2% commission rate would be reasonable under the arm’s length standard. Under this arm’s-length policy, US profits would be $4 million per year.

Table 2: An Illustration of the IRS Issue (in Millions)

 

IRS

Arm's length

 

Sales

$2000

$2000

 

I/C Price

$1880

$1956

 

Gross Profits

$120

$44

 

Gross Margin

6%

2.2%

 

Expenses

$40

$40

 

Profits

$80

$4

 

Operating Margin

4%

0.2%

 

 

The IRS position, however, was consistent with a 6% gross margin, which would suggest profits equal to $80 million per year. Note that this 6% gross margin was consistent with the position of Dr. Wright for a distribution affiliate performing all of the sales functions. We shall argue that this 6% gross margin was excessive even for her scenario. Meanwhile, 6% gross margin for a commission agent with very limited functions is clearly absurd.

The co-distribution model considers an intercompany structure where the distribution functions are divided between a distribution hub affiliate and other sales affiliate. The overall gross margin for the co-distributors is given by

GP/S = b(e1 + e2) + r(a1 + a2),

Where S = sales, GP = gross profits, e1 = expenses incurred by the distribution hub affiliate relative to sales, e2 = expenses incurred by the other sales affiliate relative to sales, a1 = assets owned by the distribution hub affiliate relative to sales, and a2 = assets owned by the other sales affiliate relative to sales. The model’s parameters are:

  1. The marginal Berry ratio (b), which will generally be less than the average Berry ratio (GP/E)
  2. The return to assets (r).

The markup over expenses can be seen as the marginal Berry ratio minus one. Let’s assume that this markup is 0.1 (b =1.1) and r = 5%.

We have assumed that the US affiliate’s expense to sales ratio (e2) is 2%, while the Swiss affiliate’s expense to sales ratio (e1) is 0.5%. Let’s also assume that the Swiss affiliate holds $300 million in assets (a1 = 15%), while the US affiliate holds no assets (a2 = 0).

The first scenario illustrates arm’s length pricing under these assumptions. If the Swiss (CH) affiliate pays the mining affiliate intercompany (I/C) price 1 equal to $1.93 billion, then the overall gross margin is 3.5%. The Swiss affiliate should also pay the US commission affiliate intercompany (I/C) price 2 equaling 2.2% of sales or $44 million. Overall operating profits for the co-distributors would be $20 million or 1% of sales, with 80% of those profits being retained by the Swiss distribution affiliate even though it incurs only 20% of the overall selling expenses. Under our assumptions, the return to assets represents the lion’s share of co-distribution profits under the arm’s length standard.

Colletively the co-distribution affilates receive a markup – defined as profits/expenses – equal to 40%. The US affiliate’s markup is 10 percent since it does not hold assets, while the markup for the Swiss affiliate is 160% reflecting the return to assets relative to its modest expenses.

Table 3: Two Transfer Pricing Policy for a Co-Distribution Structure

  TP Policy 1   TP Policy 2  

 

CH

US

 

CH

US

 

Sales

$2000

$0

 

$2000

$0

 

I/C price 1

 $1930
(Outbound)

$0

 

$1880

$0

 

I/C price 2

$44
(Outbound)

$44
(Inbound)

 

$44

$44

 

Gross Profits

$26

 

 

$76

 

 

Gross Margin

1.3%

 

 

3.8%

 

 

Expenses

$10

$40

 

$10

$40

 

Profits

$16

$4

 

$66

$4

 

Markup

160%

10%

 

660%

10%

 

 

The second scenario illustrates the implications of letting the gross margin be 6% as opposed to 3.%. In this case co-distributor operating profits total $70 million or 3.5% of sales even though expenses are only 2.5 percent of sales. This collective 140% return to expenses is implausibly high, as is the 660 percent markup for the Swiss affiliate.

Curzon Uranium Trading Limited

Curzon is a UK based trading company that purchases uranium from modest sized mining companies. Since its establishment in 2017, Curzon has become a well-known and respected name within the uranium industry. Curzon has a diversified portfolio of off-takes around the world to supply electrical utilities with nuclear fuel components on a long-term basis.

Over the two-year period ended April 30, 2020 its average annual sales have been approximately $50 million. Its gross profits have averaged $1.67 million per year, which represents a 3.34% gross margin. Its operating expenses have been $0.89 million each year so its typical operating expense to sales ratio has been 1.78% and its operating margin has been 1.56%. These financials support the position taken by Dr. Barbera that a limited function sales affiliate would receive a very modest operating margin under the arm’s length standard.

Further, it is curious that while Cameco argued that the long-term agreement between the Cameco mining affiliates and the Swiss distribution affiliate was consistent with market pricing, their various arguments never mentioned the type of off-take agreements Curazon engaged in, according to public statements. Each of these agreements granted the mining entity with economic rents with an agreed upon price that was well in excess of their operating costs:

  1. The most recent agreement was with Aura Energy Limited signed on January 29, 2019. The agreement covers production from 2020 to 2026 with the price being set at $44 per pound, which is well above the operating costs of Aura Energy’s Tiris Uranium Project. The announcement of this off-take agreement noted that the spot price at the time of the agreement was only $29 per pound.
  2. Curzon entered an off-take agreement in December 2017 with the Czech Republic based Uranium Industry A.S. for delivery of 4.75 million pounds of uranium over the 2020 to 2026 period at a price equal to $35 a pound. The spot price for uranium at the end of 2017 was less than $25 per pound.
  3. Interalloys Trading Limited, which became Curzon Uranium Trading entered to an agreement with Berkeley Energia Limited on November 28, 2016 for uranium produced from the Salamanca mine. The agreement was for a five-year period potentially covering 3 million pound of uranium with a price equal to $43.78 per pound. The spot price at the time of this agreement was less than $19 per pound.

Each of these agreements also were based on a fixed price above the spot price at the time of the agreement. Trading entities will agree to such high future prices if the price of uranium is expected to rise.

Another notable 3rd-party off-take agreement saw, on April 6, 2015, a uranium trading subsidiary of Areva sign an agreement with Texas Rare Earth Resources for delivery of 1.5 million pounds of uranium over the 2018 to 2022 period, according to company statements. The agreement stipulated a price equal to $44 per pound, which exceeded the $39 per pound spot price at the time.

Given the historical fact that uranium prices fell from 2015 to 2020, the Areva trading subsidiary incurred actual losses in its off-take agreement with Texas Rare Earth Resources.

This aspect of the litigation will be addressed in a follow-up discussion.

 

References

Her Majesty the Queen v. Cameco Corporation, 2020 FCA 112.

Cameco Corporation v. The Queen, 2018 DTC 1138.

Headquarters
EdgarStat LLC
6931 Arlington Road, Suite 580
Bethesda, Maryland, 20814-5284
USA
Customer Support
+1 (202) 558-2356
support@edgarstat.com