Thoughts on IRS Web Page titled "How criminal investigations are initiated" (11/13/24)

The IRS has a web page titled “How criminal investigations are initiated, here” (last reviewed or updated 8/20/24 and visited 11/13/24). It is a “lite” introduction to the topic. I found the concluding “claim” interesting: 

Conviction

The ultimate goal of an IRS Criminal Investigation prosecution recommendation is to obtain a conviction - either by a guilty verdict or plea. Approximately 3,000 criminal prosecutions per year provide a deterrent effect and signals to our compliant taxpayers that fraud will not be tolerated.

Notice how the subject shifts from the first sentence to the second. The first sentence talks about conviction (as does the heading). The second sentence talks about prosecutions which is not the same as convictions unless the implication is that there is a 100% conviction rate. (The conviction rate is not 100% which I address later in this blog post.)

The IRS offers these statistics in its annual Data Book, Table 26 currently titled “SOI Tax Stats - Criminal Investigation Program, by Status or Disposition - IRS Data Book Table 26,” here (last reviewed or updated 8/20/24; viewed 11/13/24). Here is a screenshot of the 2023 Table 26:

Note that the indictments and informations were 1,676, the total for legal source and illegal source financial crimes. The 2022 numbers were 1,670 for legal and illegal source prosecutions. So, the claim of 3,000 prosecutions is incorrect based on the IRS numbers. However, the above statistics are for prosecutions from investigations by IRS CI. Tax prosecutions may come from other sources (e.g., government investigations other than from IRS CI in which tax crimes may be added to nontax crimes for prosecution reasons). So, such nontax prosecutions may account for the difference of about 1,300.

Third Circuit Denies Post-Loper Bright Petition for Rehearing in Case Applying Auer/Kisor Deference (11/6/24)

In United States v. Chandler, 104 F.4th 445 (3rd Cir. 6/11/24), CA3 here and GS here, the Court sustained a sentence based in part on the application of Auer/Kisor deference to the Guidelines Commentary. (See Slip Op. 7, 17-19.) I refer to his panel decision as Chandler I. Chandler I preceded Loper Bright Enters. v. Raimondo, ___ U.S. ___, 144 S. Ct. 2244 (2024), which rejected Chevron deference (as well as, any similar deference that preceded Chevron). But Loper Bright did not address a deference subclass for agency subregulatory interpretations of legislative regulations (such as Guidelines Commentary on Guidelines). See Fourth Circuit Applies Auer/Kisor Deference to Include in Guidelines "Loss" the Commentary Inclusion of "Intended Loss" (Federal Tax Crimes Blog 8/24/24), here; and More on United States v. Boler (Federal Tax Crimes Blog 8/25/24), here.

On petition for rehearing in Chandler, the Court entered a document titled “Sur Petition for Rehearing,” denying the panel rehearing and en banc rehearing but with dissents by Judges Bibas and Matey. United States v. Chandler, 114 F.4th 240 (3rd Cir. 9/22/24), CA3 here and GS here. I refer to this denial of rehearing as Chandler II. Judge Bibas argued that, even if Auer/Kisor deference were otherwise applicable to Guidelines Commentary, deference only applied when the interpretive toolkit was otherwise empty, but that lenity was in the toolkit and applied to preempt ambiguity for Auer/Kisor deference. Judge Bibas said that applying Auer/Kisor deference without first applying lenity, “put us on the wrong side of a circuit split. At least three circuits hold that lenity trumps deference.” (Slip Op. 2-3; note the page numbers are for the pdf because the pages are not numbered.) Judge Matey dissented because he felt that the ordinary meaning of the statutory term was discernible without deference (sort of a Chevron Footnote 9 approach). Neither dissenting Judge reasoned that the Auer/Kisor deference applied in Chandler I (the pre-Loper Bright panel opinion) did not survive Loper Bright.

So, as of now, at least so far as I am aware, we still do not have a definitive ruling on whether Auer/Kisor deference survives Loper Bright, but the courts seem to be deciding cases as if it does survive Loper Bright. Most immediately, that means that the Sentencing Guidelines Commentary interpreting the Guidelines may qualify for deference at least when lenity doesn’t apply. (That sets aside the issue of whether lenity might apply to avoid getting to Auer/Kisor deference for Guidelines Commentary; and conceptually the ambiguity invoking lenity is the same as the ambiguity required for Auer/Kisor deference, lenity might always apply.)