Disguised Distributions and Management Fees: Aspro Revisited

Abstract

Our study suggests that there is a significant overlap between the factors that drive the ordinary and necessary analysis and the reasonable compensation analysis in determining the deductibility of business expenses. Our review demonstrates that the taxpayer in Aspro would have benefited from considering all the factors identified by Blue J’s algorithm at the tax-planning stage. Although we did not render a prediction in our previous article on the deductibility of the expenses in dispute, we did examine how machine learning could be used to assess the likelihood of whether the payments in question were ordinary and necessary expenses. Recall that a payment must not only be (1) ordinary and necessary to be deductible, but must also be (2) reasonable and purely for services (that is, the payment cannot be found to be a disguised distribution of profits). The Tax Court in Aspro denied the deduction of all the management fee payments — holding that the payments made to corporate shareholders were not ordinary, necessary, and reasonable — while payments made to the individual shareholder satisfied the ordinary and necessary test, but were not reasonable. As it happens, the Eighth Circuit affirmed the Tax Court’s denial of the management fees Aspro paid to its shareholders, finding no error in the lower court’s determination that Aspro failed to demonstrate that the management fees were reasonable and failed to present evidence showing what like enterprises under like circumstances would ordinarily pay for similar management services. But the appellate court did not engage in any significant substantive discussion on whether the expenses were ordinary and necessary and decided the issue on the narrower basis of the Tax Court’s finding that the taxpayer failed to establish the fees were reasonable and for services performed. At first glance, the absence of any major discussion of the ordinary and necessary nature of the expenses would appear to render our original analysis using machine learning as moot. However, our analysis reveals that the taxpayer would have been in a much stronger position on both fronts by addressing the factors identified by Blue J’s algorithm at the tax-planning stage.