Tax Credits That Bond a Partnership: Revisiting Cross Refined Coal

I. Introduction

In our monthly Blue J Predicts column, we use Blue J’s tax research software to analyze pending or recently decided federal income tax cases. This month we consider the judgment of the D.C. Circuit in Cross Refined Coal,¹ which addressed whether a business venture that was guaranteed to be unprofitable pretax could still be considered a bona fide partnership for federal income tax purposes.

Last year, we examined the Tax Court’s decision in this case and predicted with 90 percent confidence that the D.C. Circuit would find that the taxpayer, Cross Refined Coal LLC, operated as a partnership for federal tax purposes and that the IRS’s appeal would be dismissed. The D.C. Circuit has done just that with its August 5 decision.

This month, we unpack the D.C. Circuit’s de novo decision and take a deeper dive into the factors that influenced the outcome. The decision is favorable for taxpayers who are partners in a partnership trying to profit from the tax credits created by Congress. Although the IRS adopted the position that a partnership formed solely to turn a profit on tax credits is not a bona fide partnership, the D.C. Circuit held that “partnerships formed to conduct activity made profitable by tax credits engage in legitimate business activity for tax purposes.” This is consistent with Congress’s expressed intent to use tax incentives to encourage economic activity, as well as with limitations on judicial doctrines disallowing congressionally intended benefits, including section 45 tax credits.²

II. Background

The passthrough nature of partnerships often makes them useful devices for tax planning. A business that is itself a partner in a partnership may offset its own tax liability if it is a partner in an entity that generates a tax loss.³ For tax planning reasons, many business ventures have tried to appear and be treated as partnerships to capture the benefits of passing through tax losses without actually taking on the requisite degree of risk and profit sharing that is the underpinning of a true partnership. This is demonstrated by the extensive litigation over whether a business venture is a bona fide partnership for federal tax purposes.

It has largely been left to the common law to determine whether a venture is a bona fide partnership. The IRC does not supply a comprehensive definition of the term “partnership.”⁴ Leading authorities such as Luna⁵ have established an eight-part test to distinguish true partnerships from non-partnerships. To make matters more confusing, a partnership that is recognized as such for state law purposes is not necessarily recognized as a partnership for federal tax law purposes.

The dispute in Cross Refined Coal involves the denial of section 45 credits claimed by Cross, a producer and seller of refined coal. The IRS made adjustments to Cross’s partnership items for the 2011 and 2012 tax years on the basis that it lacked a nontax business purpose and was therefore not operating as a bona fide partnership for tax purposes.

Congress created a refined-coal tax credit to encourage the production of treated, cleaner-burning coal.⁶ Taxpayers that opened refined-coal production facilities before 2012 could claim a tax credit for each ton of refined coal sold over the following 10 years.⁷ Motivated by these incentives, Cross was formed by AJG Coal Inc. to operate a coal refinery. AJG secured two investors to invest in Cross: USA Refined Coal LLC (USARC), a subsidiary of FMR LLC (Fidelity); and Schneider Electric Investments 2 Inc. Through a series of purchase agreements, Cross was eventually owned by three entities: Fidelity (51 percent indirectly through USARC), AJG (24 percent), and Schneider (25 percent).

The Cross venture, like many of AJG’s ventures, operated at a pretax loss. Its business model contemplated purchasing raw coal from a utility plant, refining the coal, and then selling the refined coal back to the utility plant at $0.75 less per ton. The discounted rate would ensure that Cross lost money on each resale, and it would give the utility plant an incentive to purchase refined coal rather than use raw coal. This business model made economic sense only after accounting for the section 45(e)(8) tax credit, because the only opportunity to turn a profit was to claim a tax credit that exceeded costs.

Fidelity and Schneider made initial contributions of $4 million and $1.18 million, respectively, for the purchase of their interests in Cross. They made additional contributions for operating expenses throughout 2010 to 2013 (Fidelity’s $22 million and Schneider’s $10.5 million). Fidelity’s agreement with AJG contained a liquidated damages clause that limited Fidelity’s risk of loss. If specified trigger events occurred, Fidelity had the right to exit from Cross and receive from AJG a pro rata portion of its initial $4 million contribution. Fidelity was not entitled to a return of its due diligence costs or its monthly contribution to the operating costs.

For the 2011 and 2012 tax years, Cross claimed more than $25.8 million in refined-coal tax credits and $25.7 million in ordinary business losses. It distributed the credits and losses among its three members proportionally. However, the IRS determined that Cross was not a partnership for federal tax purposes “because it was not formed to carry on a business or for the sharing of profits and losses,” but instead “to facilitate the prohibited transaction of monetizing ‘refined coal’ tax credits.”⁸

The Tax Court found that Cross was a bona fide partnership for federal tax purposes. It further found that each of the three partners had contributed to the partnership, shared in the profits, and agreed to bear the risk of loss.

III. The IRS on Appeal

On appeal, the IRS advanced two main arguments. First, it maintained that a partnership is bona fide only if each partner expects a pretax profit “at some point in time.”⁹ Under that reasoning, because Cross’s business model built in a guaranteed loss but for the tax credits, it could not be considered a true partnership.

Second, the IRS argued that the lion’s share of the capital contributions were made by AJG and that the Tax Court grossly overstated the actual risk borne by the other partners (Fidelity and Schneider). The IRS noted that the Cross facilities were already operational before Fidelity and Schneider invested, and that the months in which the facility could not operate were ones with fewer expenses such that there was minimal risk of loss for Fidelity and Schneider. Ultimately, the IRS argued that the risk of loss and initial capital contributions were minor and disproportionate to the anticipated high returns associated with the tax credits.

IV. The D.C. Circuit’s Decision

In response to the IRS’s first objection — that Cross had no pretax profit and thus could not be considered a partnership — the D.C. Circuit found that a partnership’s pursuit of after-tax profit can be legitimate business activity for partners to carry on together. Particularly for activities that are socially beneficial, tax incentives exist precisely to encourage activity that would not otherwise be profitable. Cross engaged in business with a “practical economic effect”: the production of cleaner-burning fuel that “Congress specifically sought to encourage.” Thus, Cross did not simply engage in “wasteful activity that is typical of sham partnerships which exist merely to manufacture tax losses.”¹⁰

The D.C. Circuit quoted the Ninth Circuit in Sacks,¹¹ by stating that if “the government treats tax-advantaged transactions as shams unless they make economic sense on a pre-tax basis, then it takes away with the executive hand what it gives with the legislative.”¹² Because Congress legislated particular tax credits to encourage specific, unprofitable business activities, it does not make sense to preclude the realization of those credits by finding a partnership organized around pursuing those activities and credits to be a sham.

In response to the IRS’s second objection — that Fidelity and Schneider did not jointly take on risk — the court found that although AJG took on more risk and much of the heavy lifting to launch Cross, Fidelity and Schneider also made monthly contributions to Cross “commensurate with their status as partners.”¹³ In total, Fidelity contributed $26 million, and Schneider contributed $12.3 million. Both remained members of Cross for several years, even during unprofitable shutdowns.

V. Applying Machine Learning

In our June 2021 installment of Blue J Predicts,¹⁴ we forecast, with 90 percent confidence, that the D.C. Circuit would find that Cross is a valid partnership.

Our prediction was based on a data-driven analysis using a machine-learning model trained on the facts and results of 169 tax decisions between 1949 and 2021 that directly address whether a true partnership exists. The Blue J model generates predictions about hypothesized new scenarios based on an analysis of the outcomes of previously decided cases. Each factor in a hypothesized scenario is weighed according to the detailed factual context. The model can detect when a given factor should be given more significance based on the facts at hand, and it provides an overall forecast of the most likely result, coupled with a confidence level.

Our experience suggests that a supervised machine-learning modeling approach can be quite accurate. Our diagnostics reveal that the model for partnership determinations is 91.4 percent accurate at predicting the outcomes in federal partnership disputes since 1949.

VI. Machine-Learning Insights

In our earlier analysis of Cross Refined Coal, we applied machine learning and identified the specific arguments that the IRS could not afford to lose for its appeal to succeed. Our machine-learning model also pinpointed the importance of whether the tax credits are considered profits in determining the partnership question. We predicted that if the IRS lost on this point alone, it would more likely than not lose on the ultimate issue of the existence of a bona fide partnership.

In this installment of Blue J Predicts, we go further by assessing the other part of the IRS’s argument: the risks borne by the parties. The underpinning of a valid partnership is the sharing of profits or risk, or both.¹⁵ By applying our machine-learning model, we can uncover how the degree of risk undertaken by the parties affects the ultimate result.

A. Minimizing Risk Had Little Effect

Our analysis shows that for hypothetical situations similar to the facts of Cross Refined Coal, the fact that members like Fidelity could minimize their risk with a liquidated damages provision was practically insignificant in situations in which there is also clear and convincing evidence of profit sharing between the parties. In other words, when there is clear evidence in writing that the parties have agreed to share in the profits, and that they in fact divided the profits proportionally to their contributions, the fact that some members may have significantly limited their risk exposure has relatively little influence on the outcome.

To demonstrate, we tested our machine-learning model on facts nearly identical to those in Cross Refined Coal. In this hypothetical, we assume that it is uncontested that (1) the parties explicitly agreed to share the profits of the business venture; and (2) they then did share the profits proportionally to their contributions to the venture. By holding those factors constant, we can test the effect of the degree of risk undertaken by the members.

Machine learning assessed the impact of risk on the outcome when there is clear evidence that the members share profits.

Table 1. Impact of Risk

Whether Any of the Members Could Restrict Their Loss	Whether Any of the Members in Fact Actually Restricted Their Loss	Whether the Capital Contribution of Any Member Constitutes Debt	Predicted Outcome
No	No	No	88% partnership
Yes	No	No	86% partnership
Yes	Yes	No	82% partnership
Yes	Yes	Yes	78% partnership

As reflected in Table 1, we tested all the iterations of just these three factors relevant for risk-minimization clauses. Although Blue J’s machine-learning model takes into account 27 other factors, the three listed factors are most relevant to a member’s ability to control risk. We can see that toggling those factors to represent different hypothetical situations had a negligible impact on the predicted outcome. All test scenarios resulted in a prediction that a partnership exists, and changing each factor made an incremental difference (2 to 4 percent) on the confidence of the prediction. Again, remember that these hypothetical scenarios assume that there is clear evidence of an intent to share profits and that there is actual sharing of profits proportionally among the members. In this type of scenario, the effect of the ability to control and reduce risk is relatively insignificant.

This insight aligns with previous case law. In Hunt,¹⁶ for example, the Tax Court found that a member could have a 98 percent guaranteed return of its capital contribution and still qualify as a bona fide partner.¹⁷ The Tax Court in that case found that the guaranteed return was a financial incentive for the partner to enter into the partnership arrangement with the limited partners. The court also noted that the distribution scheme could have yielded healthy profits for all the partners and would have returned to them their capital contributions.

Hunt and Cross Refined Coal are notable for their emphasis on profits. What emerges from the case law — something that our machine-learning model has also identified — is that when there will likely be healthy profits for the partners and there was a clear intent to share those profits in accordance with the intent of the parties, the fact that one partner has been able to minimize or even neutralize its risk usually is not a powerful enough factor to materially affect the outcome that a partnership exists. It seems that the courts will avoid heavy-handed approaches to establishing an appropriate degree of risk if the evidence indicates a clear agreement and a practice of sharing profits.

Applying this insight to Cross Refined Coal, unless the case can be decided entirely on substance-over-form principles such as the economic substance doctrine, sham doctrine, or business purpose test, determining whether there was a valid partnership using the Luna factors means that if the sharing of profits is clearly established, the more forgiving a court may be of guarantees of a member’s investment. Thus, in the overall factual matrix of Cross Refined Coal, the liquidated damages clause that minimized risk for some members is not determinative of the issue, particularly when the intent to share profits was clear.

B. Importance of Contributions Made to Cross

As a corollary to that insight, we tested hypothetical situations nearly identical to those above. However, we explore the importance of whether the contributions of each member are necessary and valuable to the operations of the business.

Table 2 shows the same findings as in Table 1, but with a new variable: the importance of the capital contribution.

Table 2. Impact of Risk and Capital Contribution

Whether Any of the Members Could Restrict Their Loss	Whether Any of the Members in Fact Actually Restricted Their Loss	Whether the Capital Contribution of Any Member Constitutes Debt	Contribution Was Necessary and Valuable	Predicted Outcome
No	No	No	Yes	88% partnership
Yes	No	No	Yes	86% partnership
Yes	Yes	No	Yes	82% partnership
Yes	Yes	Yes	Yes	78% partnership
Yes	Yes	Yes	No	60% partnership

It was when we introduced the fourth factor — that the capital contribution of one of the parties was not necessary and valuable to the partnership — that we see a material effect on the outcome. There was a change of -18 percent — down to 60 percent likelihood that there will be a partnership for tax purposes. To be clear, the Blue J model continues to predict that a court would conclude that there is a bona fide partnership, but it is a less confident prediction.

Building on the previous finding that minimizing risk is not a strong factor against finding a partnership for tax purposes, this finding suggests there is a material impact if neutralized risk is paired with the fact that a purported partner’s financial contribution was neither valuable nor necessary to the partnership. This makes intuitive sense. A partner may be able to use its leverage, such as the case with Fidelity, to insulate itself from a significant amount of risk because it can contribute key services and loans to the partnership. Those contributions may be crucial to the operation of the partnership. Further, these limitation clauses may provide a necessary incentive to entice some partners to join the business. However, when a contribution is neither necessary nor valuable to the partnership, this could indicate that the investment is likely so trivial as to make would-be partners indifferent to the success or failure of the operations.

Readers may be surprised to see that all of the above test scenarios resulted in a prediction that a partnership exists, even though some hypotheticals involved members who insulated themselves from risk and whose financial contributions were considered unnecessary. It is important to note that the Luna test is multifaceted and involves eight significant factors. Our machine-learning model assesses 30 different types of facts tied to the Luna factors.

Other facts of Cross Refined Coal that were not discussed here also weighed in favor of a partnership finding, such as the fact that all three members were involved in the day-to-day management and decision-making to operate the facility, and that the members did in fact split the losses when the facility was nonoperational. Further, it is important to note that we are here testing the financial risk of the day-to-day business operations. As the D.C. Circuit noted, there were other risks borne by each member; no party was able to insulate itself from potential regulatory or litigation risk.

VII. Conclusion

The passthrough benefits of a partnership and the decision of the D.C. Circuit in Cross Refined Coal make it tempting for taxpayers to organize a partnership to take advantage of tax credits and pass through its tax burdens and benefits to its partners. Although the decision is favorable for taxpayers using a partnership to monetize tax credits, there is still a fine line between tax minimization as a primary consideration — which the courts have allowed — and a sham.¹⁸

The question of whether a partnership exists and whether the purported partners will be respected as such (and not simply as tax credit buyers) has been reinvigorated by the Inflation Reduction Act (P.L. 117-169). New section 6418 states that if an election is made, payments from a sale of specified tax credits are not includable in gross income, and in some instances they will be treated as tax-exempt income under section 705(a)(1)(B). While distinct, the question of who is a partner is driven in part by the partnership analysis discussed above.¹⁹

As the Supreme Court held in Tower,²⁰ a valid partnership requires the partners to intend to “share in the profits or losses or both.” As our analysis using machine-learning algorithms demonstrates, the clear intent to share in profits and an actual sharing of the profits goes a long way toward a determination of a valid partnership. When the profits-sharing side of the equation is satisfied, the significance of the intent to share in losses and the actual sharing in losses has a minor impact on the outcome. Thus, taxpayers who have significantly contained their risk exposure through liquidated damages clauses can still qualify as partners of a bona fide partnership by ensuring that the intent to share in profits and the actual sharing of profits is clear.

However, the analysis is even more nuanced than that. As our machine-learning model demonstrates, courts are understanding of partners who entice others to join the partnership by allowing the new partner to minimize its risk because the new partner can provide something important for the business. When this is no longer true, and the financial contribution of the partner is neither necessary nor valuable, this can significantly reduce the likelihood of a finding that a valid partnership exists to a mere coin toss.