Battling Uphill Against the Assignment of Income Doctrine

Abstract

We examine a Tax Court case that our machine-learning model suggests was correctly decided (with more than 95 percent confidence). Ernest S. Ryder & Associates Inc. v. Commissioner, T.C. Memo. 2021-88, has received significant attention from the tax community. It involved tax avoidance schemes marketed by the law firm Ernest S. Ryder & Associates Inc. (R&A) that produced more than $31 million in revenue between 2003 and 2011 and for which the firm reported zero taxable income. The IRS unmasked more than 1,000 corporate entities that R&A’s owner, Ernest S. Ryder, had created and into which he funnelled the money. By exposing the functions that these entities performed, the IRS played the most difficult role in the case. Yet, there are deeper lessons that can be drawn from the litigation by subjecting it to analysis using machine learning. We shine an algorithmic spotlight on the legal factors that determine the outcomes of assignment of income cases such as Ryder. For Ryder, the time for filing an appeal has elapsed and the matter is settled. Thus, we use it to examine the various factors that courts look to in this area and to show the effect those factors have in assignment of income cases. Equipped with our machine-learning module, we are able to highlight the fine line between legitimate tax planning and illegitimate tax avoidance in the context of the assignment of income doctrine.

Keywords: tax avoidance, assignment of income, machine learning

JEL Classification: H25