The Mirror Code Concept; Some Thoughts and Ruminations (5/27/13)

The United States has a "mirror code" system with certain of its territories -- including U.S. Virgin Islands, Guam and the Commonwealth of the Northern Mariana Islands (CNMI).  The mirror code concept treats the U.S. tax code as the tax law of each jurisdiction -- U.S.,on the one hand, and the other jurisdiction,on the other.  In effect, the two jurisdictions are treated as separate countries for purposes of the mirrored Code applied by each.  Double taxation is generally avoided by requiring a single filing to the jurisdiction in which the taxpayer resides.  If the U.S. citizen is a "bona fide resident." of the U.S. V.I., the U.S. resident reports and pays tax to the U.S. V.I. and not to the U.S.; If the U.S. taxpayer is resident anywhere else, he reports and pays his tax to the U.S.  By contrast, if the U.S. V.I. citizen is resident in the U.S., he reports and pays tax to the U.S.; if he is resident anywhere else,he reports and pays tax to the U.S. V.I.  In each of these cases, if the citizen pays tax in the noncitizenship country of residence -- i.e., U.S. citizens reports and pays tax in U.S. V.I. or U.S. V.I. citizen reports and pays tax in the U.S. -- the country receiving the tax will remit -- a process called "cover" -- the portion tax received that relates to income in the other country.

A pure mirror code system will result in the same tax regardless of where the return is filed and the tax paid.  The territories are, however, allowed to give tax breaks with respect to taxes paid on income source in the territories under the system.  Thus, if the reporting of U.S. V.I. sourced income is to U.S. V.I., the U.S. V.I. is permitted to give a tax break with respect to that tax.  If the reporting of U.S. V.I. sourced income is to the U.S., then the U.S. should cover that portion of the tax to the U.S. V.I., whereupon, from the tax thus remitted, it can given any break it otherwise allows.  Thus, at least in theory, wherever the return is filed, the same ultimate result should obtain.

The problem comes when, after the taxpayer has filed in good faith with the mirror code territory (e.g., the U.S. V.I.), the U.S. attempts to force the U.S. taxpayer to file in the U.S., despite being required under the treaty to make the single filing in U.S. V.I. The circumstances for double taxation are present if the U.S., through making a different sourcing determination, has no plan to cover the amount to the other country.

This battle was fought out in Appleton v. Commissioner, 140 T.C. ___, No. 14 (2013), here.  The IRS took the position that the taxpayer was required to file in the U.S., he had filed in U.S. V.I. rather than the U.S., and that, as a result, the IRS had an unlimited statute of limitations to send a notice of deficiency with respect to the tax.  The Court held that the single filing with the U.S. V.I. required under the mirror code scheme was a return filed under the Code and therefore the unlimited statute of limitations did not apply.

I think that is the right result and refer readers to the opinion for the analysis supporting it.  I thought I would use the balance of this blog to deal with a nuance that I dealt with in Preece v. Commissioner, 95 T.C. 594 (1990), here.  This nuance is a policy argument for how to make Congress' clear purpose to have a single filing provision work where the two jurisdictions are feuding -- insisting that the taxpayer file in each jurisdiction -- as they were in Preece.

Preece involved the CNMI's mirror code (which was the same as Guam's and somewhat like the U.S. V.I.).  That mirror Code had a similar single filing requirement based on residence (without the "bona fide" limitation).  The question in the cited decision was whether residence was determined by the objective substantial presence test of Section 7701(b)(3)(A), as mirrored into CNMI under the mirror code.  The Preeces clearly met the substantial presence test.  The IRS argued, however, that the more subjective test of residence applied which was the law prior to enactment of the substantial presence test.  Section 935, the single filing requirement turning on residence, was enacted before the substantial presence test.  Judge Nims effectively held that, based on his reading of the tea leaves, the prior subjective test applied to the Section 935 determination.

Despite an inference in the holding that it would be tough for the taxpayers to establish residence, the case then settled on terms that, I think, were favorable to the taxpayers.  Thereafter I wrote the piece which I include below to discuss a key aspect of how these mirror code treaties should work.  Basically, I conclude that, rather than permitting the two jurisdictions to fight over the residence issue, the jurisdiction of noncitizenship (in Preece, CNMI as to a U.S. citizen living in CNMI; and the U.S. as to the CNMI citizen living in the U.S.) has the right to make the residency determination and, if it makes the residence determination as to the noncitizen, that determination should control the single filing requirement.  This requires simply a clear focus on the clear congressional purpose to have a single filing requirement and how that has to be effected when the two jurisdictions cannot agree (as was the case in Preece and Appleton).

With that introduction, here is the piece I wrote (which I do not indent, since the balance is the piece itself and is my work):

The Preece litigation was hard fought.  I believe, however, that the correct holding is that the substantial presence test (rather than the facts and circumstances test) should apply to the determination of residency, for all of the reasons that the substantial presence test was adopted in the first place.  United States citizens and CNMI citizens physically present in the other "country" (for purposes of the treaty and most particularly residency determination in the Code, I think they are treated as separate countries) and the respective tax administrations have the right to certainty on the issue of residency.

There is a threshold conceptual problem.  Who makes the residency determination and from what perspective?  The Code residency test was not designed to determine the U.S. residency of a U.S. citizen.  Rather, it was designed to determine the U.S. residency of an alien.  U.S. residency is ordinarily irrelevant to the determination of whether the U.S. citizen should file a U.S. return and pay U.S. tax.  Applying this analysis to the Code as “mirrored” into CNMI, the Code test of residency determines whether the U.S. citizen is a resident of CNMI (the U.S. citizen being an “alien” to CNMI under the mirrored code).  If, under the mirrored Code, the U.S. citizen is a resident of CNMI (regardless of which test applies), then that is where the U.S. citizen files his or her tax return and pays tax (and, of course, qualifies for any rebate that CNMI offers on the tax thus paid).  CNMI clearly made the determination that the Preeces were CNMI residents and was fully justified in doing so under either the substantial presence test or the facts and circumstances test.  The conceptual problem is that, if the residency test were applied to determine the U.S. residency of the U.S. citizen, then it is possible that the taxpayer could be residents of both the U.S. and CNMI and then the treaty would totally fail in its clear purpose to have one filing and one payment of tax, for each jurisdiction could then apply its Code’s residency test to require filing in that jurisdiction.

What this suggests is that, if the one-filing concept is to work in any rational way, one country’s determination should prevail over the other’s.  As I suggest above, that is inherent in the adoption of the Code residency test -- i.e., it was not designed to determine the residency of a citizen of the Code country making the determination.  Specifically, in this context, the residency test in the U.S. Code is irrelevant to a U.S. citizen, and is relevant only to aliens (including for this purpose citizens of CNMI); by reverse token, the residency test in the CNMI Code (the “mirrored” code) is irrelevant to CNMI citizens, and is relevant only to aliens (including for this purpose U.S. citizens).  The residency test thus “traffic-cops” potential conflicts between the two jurisdictions under the one filing concept, ceding primary tax jurisdiction to the country of non-citizenship.  That does not mean, of course, that the U.S. does not get tax, for as you know, the CNMI collects tax on U.S. source income and covers (i.e., pays) that tax to the U.S. Treasury.

This system is a rational system and is not dissimilar to the foreign tax credit system, except in the one-filing concept.  I illustrate this by assuming that country X has a code just like the U.S. and U.S. citizen A resides in country X, so that under the Country X Code Citizen A must pay tax on his or her worldwide income to country X and in exactly the same amount as Citizen A will have to report to the U.S. on his or her U.S. return.  The key to avoiding duplicative and onerous taxation is the foreign tax credit (“FTC”) system that the U.S. gives the U.S. citizen in filing its U.S. return, so that the real tax jurisdiction (up to the amount of U.S. tax) is ceded to the foreign country based upon its determination of nexus to tax -- i.e., residency of the U.S. citizen in the foreign country.  The U.S., of course, retains the right to tax U.S. source income and Country X would allow its U.S. resident a foreign tax credit on the U.S. tax paid.

Of course, foreign countries do not have a U.S.-like Code, but the mirror code concept applying to certain jurisdictions does in effect give the “foreign” jurisdiction (specifically CNMI) exactly that phenomenon.  In the absence of a one-filing concept, the primary taxing jurisdiction would be determined under the FTC regime by CNMI as to U.S. citizens and by the U.S. as to CNMI citizens.  A dual filing would be essentially duplicative and informational only in the country of citizenship.   The U.S. tax imperative to at least collect tax on U.S. sourced income is protected because CNMI is required to cover to the U.S. the tax it receives on U.S. sourced income (just as, as to CNMI citizens filing with the U.S., the U.S. must cover to the CNMI the tax on CNMI sourced income).  Thus, the system works essentially like the FTC system, except that it avoids double filings and, assuming each jurisdiction covers to the other jurisdiction the tax on the income sourced in the other jurisdiction, there can be no revenue hemorrhage.

The problem, of course, is in the rebate which introduces discontinuities in this one-filing model.  If the two-filing system were in place, the rebate would mean that the filing with the country of citizenship would not be just informational.  The U.S. FTC would be reduced by the amount of the rebate.  Thus, the U.S. citizen resident in CNMI would, under a two-filing concept, owe the U.S. net tax to the extent of the rebate.  The U.S. would thus collect tax on the U.S. sourced income and a tax on the CNMI sourced income, with a credit only for the net taxes paid.  Under the one-filing system, however, the foreign tax credit is replaced by a system that permits CNMI to tax the CNMI sourced income (either directly or by a U.S. cover of the tax) and the U.S. to tax the U.S. sourced income (either directly or by a U.S. cover of the tax).  But, in authorizing the mirror code for CNMI, the framers of the treaty expressly allowed CNMI to rebate the taxes it collected and CNMI chose to do that.

The problem therefore in Preece was not really residence, for the Preeces clearly were residents of CNMI under the Code test applied from CNMI’s perspective.  But, regardless of where they were resident and assuming that the income was properly sourced in CNMI (as it clearly was applying the Code as it then existed from CNMI’s perspective), the covering mechanism would get the Preeces to the same point -- i.e., the CNMI Treasury would get the tax whether it was collected in the U.S. or in CNMI or both and would rebate the appropriate amount to the Preeces.  CNMI had already made the determination that the income was sourced in CNMI under the then Code-applicable sourcing rules (that really was a bullet-proof determination) and was not about to cover to the U.S. any amount of the tax on that income.  The IRS, unhappy that the Preeces and others like them had discovered the ultimate tax shelter in the form of the CNMI rebate, wanted to solve the problem simply by forcing the Preeces to file a U.S. return (requiring a residency determination) and then refusing to cover the tax to the CNMI regardless of CNMI’s good faith and bullet-proof determination under the mirror code that the income was sourced in CNMI.  The key to relief for the Preeces (that is, getting the benefit of the rebate to which they were entitled) would be for the U.S. to cover to CNMI.

The problem in the Preece case was that the taxpayers arguably could not litigate the sourcing battle and in any event, regardless of the treaty’s clear intent that CNMI sourced income be covered, had no way to force the U.S. to cover.  Although we noised about in Preece that the sourcing issue was before the Court for resolution (figuring that, should the Tax Court sustain CNMI sourcing for the income, the U.S. would be morally compelled to cover the tax to CNMI), the truth is that it may not have been before the Court except in an expansive sense of the Tax Court’s jurisdiction.  So, the U.S. might be able to achieve certain self-help for this genre of tax shelter simply by refusing to cover, period.

The one filing concept would place the onus upon CNMI to insist that it be covered the tax and, ordinarily, CNMI would have self interest to do so because it would still have net revenue even after the rebate.  But, CNMI had already collected its tax directly from the Preeces, so that it had no net revenue interest in insisting that the U.S. cover the tax to CNMI and probably figured that if the U.S. wanted to hammer its citizens that way, that just is too bad.  Of course, that would not take away the immorality of the U.S. refusal to cover if indeed the income were sourced in the CNMI.  But it would leave the Preeces without any practical remedy, except to fight the only battle that they could -- i.e., the residence battle to avoid having to file a U.S. return ab initio.  But, as noted above, that is a battle that neither U.S. nor CNMI citizens should have to face if the treaty is properly applied and the two jurisdictions deal in good faith to determine sourcing in the same manner it would be determined for purposes of the FTC.

Despite the loss on the test of residency and uncertainties as to whether we could resolve the sourcing issue, we achieved a settlement in Preece that was far more favorable than the loss would have indicated.  I think most people reading the case would think that the Preeces would have a difficult time establishing CNMI residency because of the inherently fuzzy facts and circumstances test and the fact that the U.S. Tax Court, like the IRS, would be influenced by the U.S. tax hemorrhage if CNMI residency were too easily obtained.  The IRS nevertheless settled on quite favorable grounds because, I believe, it finally realized that we were right on the main issues -- (1) despite Judge Nims’ holding,  the substantial presence test should apply and (2) even if it did not, the critical test for residency and sourcing should be made by CNMI as to U.S. citizens, in which case the facts and circumstances test would still dictate CNMI residency and sourcing as determined by CNMI and thus no U.S. return.

Although that was a tough sell to Judge Nims, I had notified the IRS attorney that I would call the key IRS players in the application of the treaty for trial, and hoped through their testimony to get this position adequately presented before the court (rather than through just my logical argumentation on brief).  We settled just short of trial for a quite acceptable result.

By the way, for the local CNMI BGRT, we used Nancy Beckner from Los Angeles with whom you may have had some contact.  She was (and I presume still is) very good.

And, I spoke on several occasions with Stephen Cohen who worked with the Attorney General or Department of Revenue and Taxation.  He was quite knowledgeable and helpful, except it was clear that he (and the CNMI tax officials) did not want to enter the brawl to support my clients’ position even though I perceived that they thought it was the right one.  (An essential part of my case at trial would be proof that they had made a good faith determination of CNMI residency.)  As I understand it, the relations between the IRS and the CNMI tax authorities was not the best and, I believe, the cover mechanism which, as noted above, should have gotten to the same result had even broken down to some extent.  I viewed the CNMI tax authorities’ desire to stay uninvolved as being more political than anything else. I know that the IRS was not happy about the rebate which had clearly been an essential part of the treaty.