Divorce brings all kinds of consequences. Today’s lesson is about one of the tax consequences. In last week's case of Brigette Ogden v. Commissioner, T.C. Memo. 2019-88 (July 15, 2019) (Judge Halpern), the divorced taxpayer sought to be relieved of her obligation to pay tax reported on two prior joint returns. She sought relief based in part on a claim of spousal abuse and a claim that a state court’s divorce decree absolved her of responsibility. Judge Halpern’s opinion teaches lessons about: (1) the relationship of spousal abuse to spousal relief in §6015, (2) the relationship of state courts to federal tax law, and (3) the relationship of sub-regulatory IRS guidance---here a Revenue Procedure---to Tax Court review of an IRS decision about spousal relief. Details below the fold. It's all about relationships.
A Sketch of Spousal Relief Law
Section 6013 provides that spouses are jointly and severally liable for the taxes owed on a jointly filed return. But sometimes there are unpaid tax liabilities after a divorce and one spouse might not want to be jointly liable for those liabilities.
Section 6015 allows divorced individuals the ability to be relieved of the obligation created by §6103. These relief provisions are a mixture of legal concepts and equitable concepts. For the gory details, feel free to download my article “The Unhappy Marriage of Law And Equity in Joint Return Liability," 108 Tax Notes 1307 (September 12, 2005). I also blogged about this legal/equitable mix back in 2018 in the post “The Role of Equity.”
Today’s post looks at one part of §6015 in a little more detail. Basically, §6015 gives taxpayers three pathways to obtain relief from unpaid tax liabilities associated with a joint return.
First up is §6015(b). This path is usable when there was a deficiency on the original return. The statute allows subsection (b) relief only when the requesting spouse shows: (1) they filed a joint return; (2) it had an understatement of tax attributable to the other, non-requesting, spouse; (3) the requesting spouse both did not know and did not have any reason to know about the underlying cause of the understatement; and (4) it would be inequitable to hold the requesting spouse liable for the liability (that is, they must show they did not benefit from the cause of the understatement).
Subsection (b) is the traditional form of spousal relief that is typically called “innocent spouse” relief. I teach students to remember this by remembering the requesting spouse must be innocent of knowledge and innocent of benefit, because those are often the biggest hurdles to relief.
Second comes §6015(c). This path is equally usable when the unpaid tax liability is due to an understatement of tax on the return. It is different than subsection (b) relief, however, because this statutory provision allows one spouse to simply “unwind” a joint return if (1) the requesting spouse had no actual knowledge that there were erroneous items on the return; and (2) the requesting spouse is now divorced or separated under the applicable rules. The chief difference between subsection (b) relief and subsection (c) relief is the burden of proof. Subsection (c) presumes innocence. It does not play the "should have known" card against the taxpayer but instead shifts the burden to the IRS to show actual knowledge of item giving rise to the understatement.
A taxpayer seeking relief under either of these first two paths may not act before the IRS starts collecting against them and then must act within 2 years of the first collection action. §6015(c)(B)(3). This is a tricky requirement. Not all IRS collection action triggers the 2-year period. Worse, it is not intuitive what actions trigger the clock. For example, the Notice of Intent to Levy and Notice of Your Right to a Hearing, which is sent before any actual levy action takes place, counts as a collection action. It triggers both the 2-year period to seek relief and it also triggers CDP rights. Treas. Reg. 1.6015-5(b)(2). In contrast, the Letter 3172 (“Notice of Federal Tax Lien Filing and Your Right to a Hearing under IRC §6320”), which is issued after the NFTL is filed and starts hurting the taxpayer’s credit, is not sufficient to trigger the 2-year period even though it does trigger CDP rights. Id. Go figure.
The third path to relief is §6015(f). It is very different from subsections (b) and (c). First, it is available as long as the collection statute of limitations is open. There is no 2-year period. Second, it applies to both understatements and underpayments. That is, it can apply even if there is no deficiency but simply an unpaid tax liability. Third, the statute is based on equity. It allows a spouse who cannot get relief under either §6015(b) or (c) to be relieved of responsibility for an unpaid liability when “taking into account all the facts and circumstances, it is inequitable to hold” the requesting spouse liable.
The IRS has published guidance on what “facts and circumstances” it will consider in Rev. Proc. 2013-34. While the Tax Court repeatedly says it is not bound by the Rev. Proc., as a practical matter it always follows the Rev. Proc.’s structure and approach to evaluating the suitability of relief. This case illustrates how that works.
Rev. Proc. 2013-34 puts a taxpayer’s request through three equity screens, each of which asks slightly different questions, all variants of “is it fair to allow relief?” Section 4.01 is the first screen; it is called the “threshold” conditions. Most of them are basically a checklist of procedural hoops. If a taxpayer does not meet the threshold conditions, the IRS rejects the request without further analysis.
The key equitable concept in these threshold conditions is an attribution rule: the unpaid tax liability must be due to the other spouses’ behavior. §4.01(7). If it is attributable to the actions of the requesting spouse, relief is not appropriate. Since this is an equitable determination, however, the attribution rule has exceptions, all grounded in various reasons why it may be unfair to enforce that rule. As relevant to today’s lesson §4.01(7)(d) provides this exception for abuse:
“If the requesting spouse establishes that he or she was the victim of abuse prior to the time the return was filed, and that, as a result of the prior abuse, the requesting spouse was not able to challenge the treatment of any items on the return, or was not able to question the payment of any balance due reported on the return, for fear of the nonrequesting spouse’s retaliation, the Service will consider granting equitable relief even though the deficiency or underpayment may be attributable in part or in full to an item of the requesting spouse.”
Notice that the requesting spouse must “establish” the abuse. Yep. It’s that old bugaboo about substantiation. The Tax Court really prefers some solid evidence, like a police report or doctor’s testimony. For example, in Deihl v. Commissioner, T.C. Memo. 2012-176 (Judge Marvel), Ms. Deihl claimed she was abused by her husband such that she could not question the positions taken on the return. She brought in her son to testify. That’s it. Judge Marvel was not convinced, writing:
“Abuse is a genuine reason to grant relief from joint and several liability, and we are sensitive to the legal and emotional issues related thereto. However, we cannot conclude on this record that petitioner did not question the treatment of items reported on the joint returns for fear of Mr. Deihl's retaliation. We did not find petitioner's testimony credible or convincing. Petitioner provided no substantiation of the alleged abuse, such as a police incident report or a medical report. As corroborating evidence, petitioner introduced only the testimony of [her son], which we do not find credible.”
Notice too that the requesting spouse not only must show abuse, but must also connect the abuse to the return preparation by showing the abuse prevented the requesting spouse from doing anything about the return. That proves important in today’s lesson.
When a taxpayer satisfies the threshold requirements the IRS applies a second screen, found in §4.02. This second screen applies just three additional conditions. If all three apply, then the IRS will grant relief without the need for further analysis. So this second screen is called the “streamlined determination.” Basically, the three conditions are: (1) the requesting spouse is properly divorced or separated; (2) paying the tax would cause the requesting spouse suffer economic hardship; and (3) (only in underpayment situations) the requesting spouse reasonably believed the other spouse was supposed to have paid the tax.
If a taxpayer passes the threshold requirements but does not qualify for the streamlined determination, then §4.03 lays out seven non-exclusive factors to evaluate to see if it would be fair to grant relief. Some of them repeat the above equitable considerations, such as the idea of who bears the legal obligation under state law to pay any unpaid tax: the requesting spouse, the non-requesting spouse, or both in some fashion.
This case involves two tax years: 2008 and 2010. Ms. Ogden married Mr. Ogden in 1991. In 2003 she became ill and applied for Social Security Disability (SSDI). In 2007, while the SSDI process was pending, the Ogden’s separated. In 2008, the SSA finally agreed that she was disabled and sent her a lump sum payment of $36,000. Apparently she then began receiving a yearly benefit. In 2010 she received a yearly benefit of $10,000 from Social Security.
The Ogdens filed jointly for both 2008 and 2010 even though they were separated. Both returns were prepared by Mr. Ogden. On their 2008 return they totally omitted the $36,000 lump sum payment. On their 2010 return the properly included the $10,000 yearly benefit, but underpaid their reported tax by over $5,000.
In 2011 the Ogdens divorced. The state court divorce decree provided that Mr. Ogden was to assume all Federal and State tax liabilities of the parties for all taxable years through 2010.
In 2014 Ms. Ogden asked for spousal relief from both the deficiency in the 2008 return and underpayment of the 2010 return. The IRS denied her any relief for 2008 because the $6,300 deficiency was attributable entirely to the omission of the $36,000 lump sum payment and that income was attributable to her. The IRS gave her almost total relief for 2010, except for $327 of the unpaid tax liability that was attributable to her Social Security payments.
Ms. Ogden appealed both determinations.
Lesson 1: The Role of Abuse
Ms. Ogden claimed that her husband abused her such that she was unable to object to or contest the way he prepared the returns for both 2008 and 2010.
Ms. Ogden’s claim of abuse failed for two reasons. First, she had a problem proving the abuse occurred. Remember, the Tax Court really likes to have some independent third-party proof, such as police reports, medical evaluations, neutral witness testimony, etc. Ms. Ogden had none of that. Further, Judge Halpern noted, she and Mr. Ogden separated in 2007, well over a year before the 2008 returns would be prepared. That only increased her burden of proof. Yet all she had was her testimony, which Judge Halpern called “vague and generalized.” That was not enough to overcome the income attribution rule for either 2008 or 2010.
Her second problem is the lesson: she was unable to connect the alleged abuse with the 2008 return preparation. To succeed with an abuse claim, the requesting spouse must show how the abuse prevented the requesting spouse from contesting or reviewing the errors on the return. Here, Judge Halpern found that Ms. Ogden was “apparently unaware that the SSA payment was taxable.” Well folks, if she did not know that the $36,000 lump sum was reportable income in the first place, how could she contest its omission? The alleged abuse simply does not connect to the return preparation if she had no reason to try and contest the erroneous item.
Ms. Ogden argued that the state court’s decree making Mr. Ogden liable for all unpaid tax liabilities through 2010 meant the IRS should relieve her of the obligation. Judge Halpern disposes of this with a footnote that “Enforcement of the terms of petitioner's divorce is a civil matter which has no impact on her tax liability to the IRS.” I do not think that is quite accurate and the point is worth a bit more explication because it gives us two additional lessons.
Lesson 2: The Effect of State Courts On Spousal Relief
State courts certainly cannot countermand federal law, because state law cannot prevail over federal law. U.S. Constitution: Article VI, Clause 2. In that sense that Judge Halpern’s statement is true. A state court cannot determine a federal tax liability. But state court divorce decrees can certainly "impact" federal tax liabilities (or "influence" them, if you like that word better). For example, while a state court divorce decree cannot establish who is the “custodial” parent for dependency purposes, it can certainly order that a child spend the majority of nights with one or the other parent. See Treas. Reg. 1.152-4(d)(1) (defining custodial parent as "the parent with whom the child resides for the greater number of nights during the calendar year"). That puts a heavy state thumb on the scale in favor of one parent over the other.
As relevant to spousal relief, one way a state court decree can affect tax liability is by allocating responsibility for payment of any tax obligations arising from joint returns. That is because one of the equitable factors in both §4.02 and §4.03 of Rev. Proc. 2013-34 is the factor of who bears the legal obligation to pay. If a state court order puts responsibility for payment on the non-requesting spouse, then that favors relief under both equity screens. So to say that the terms of Ms. Ogden’s divorce decree has “no impact” on her federal tax liability is not quite true. Again, it's a state court thumb on the scale of federally determined equity factors.
Lesson 3: The Effect of Rev. Proc. 2013-34 on Spousal Relief
Here is where the structure of Rev. Proc. 2013-34 is important to understand. You do not even get to the legal obligation factor under the Rev. Proc. until you pass the first screen, which includes the threshold attribution requirement.
Notice how this worked both to help and hurt Ms. Ogden here. The 2010 joint return underpaid the self-reported tax by over $5,000. Most of that underpayment, however, was attributable to Mr. Ogden. So for that portion of the underpayment Ms. Ogden was able to get through the first equity screen (threshold requirements) and into either the second equity screen (the three-factor streamlined determination) or the third equity screen (the seven factors). I don’t know which the IRS applied because the opinion is silent on Ms. Ogden's economic situation, but in either screen the legal obligation factor favored Ms. Ogden as to all but $327 of the underpayment.
What Ms. Ogden wanted Judge Halpern to do was to ignore the Rev. Proc. structure and skip right to the “obligation to pay” factor. In effect, she wanted the state court's assignment of payment responsibility to Mr. Ogden to become another exception to the attribution rule rather than one factor that would come into play only for liabilities not attributed to her.
Judge Halpern refused to do that, sub silentio. Judge Halpern did not explicitly say so, but what he did was hew to the structure of the Rev. Proc. by limiting his analysis of the attribution rule to consideration of the abuse exception, and not even for a moment did he consider departing from the structure of the Rev. Proc. to create a new exception to the attribution rule for state court orders that put the entire obligation to pay on the non-requesting spouse.
I am here making only a descriptive claim, not a normative claim. That is, I am just pointing out what I see going on in this case: the Rev. Proc. cabins Tax Court review. I am not saying whether this is a good thing or a bad thing. It's just a thing. FWIW, I tend to think it good. After all, equity is not a free-floating concept of "do the right thing." It has structure and the IRS invented the wheel of the Rev. Proc. to structure equity in a way that a bureaucracy can apply it to a high volume of claims. The Tax Court has no need to re-invent that wheel in this case (and most cases), even if it has the power. Full exploration of that normative question, however, is the stuff of a longer article than a blog post.
Coda: We marry, thinking it’s going to last. We divorce, often wondering why it didn’t. While the divorce rate is falling, it still likely approaches 1 million per year. I say “likely” because it appears that no-one collects that data anymore. The best data I found was this webpage from the Center for Disease Control website, but it was based on reports from only 45 states (plus DC). Five states no longer report that data, including California. So my guess of about a million just uses the CDC data from 2017 and adds the divorce rate (2.9 divorces per 1,000 population) to the population of the five non-reporting states.
Bryan Camp is the George H. Mahon Professor of Law at Texas Tech University School of Law