Yesterday at the Tax Policy Colloquium, Leandra Lederman presented her paper, Of Risks and Remedies: Best Practices in Tax Rulings Transparency. (I'm not including the link because we discussed an early draft.)
The paper looks at the issue of confidentiality versus disclosure with respect to private letter rulings (PLRs), which a tax authority issues to a particular taxpayer, definitively setting forth how a particular transaction will be treated under its tax laws, but generally without precedential effect on other cases.
The paper responds in particular to two distinct episodes in the history of PLRs. The first is American, while the second relates to Luxembourg.
In the United States, PLRs were long treated by the IRS as confidential. Taxpayers and their advisors weren’t barred from disseminating them as they liked, but the IRS itself would not do so. The IRS rationale was that there were administrative advantages, to both taxpayers and the IRS itself, to settling particular questions in advance so that everyone could go forward. But because doing this in any helpful volume required allowing lower-tier IRS personnel to greenlight particular rulings, the agency was anxious not to allow for any appearance that these were legal precedents that taxpayers could follow against the IRS. To this day, for this reason, PLRs state on their faces that they do not have precedential value, cannot be relied upon other than by the taxpayer to whom they were issued, etc. But the IRS felt better still if there was no publication of them. (Taxpayers often had no reason to share their rulings with others.)
This led, however, to insidious practices. For example, insider law and accounting firms that did a lot of rulings business with the IRS had their own libraries of PLRs, which perhaps they swapped a bit when both sides had enough to make it worthwhile. The tax shelter in the infamous Knetsch v. U.S. case was based on a PLR that the promoter had evidently extracted from a naïve lower-level IRS decisionmaker, and then used as a marketing tool to sell more of the same.
Eventually Tax Analysts, the publisher of Tax Notes, sued under the Freedom of Information Act, and ever since PLRs have been published, albeit with redactions that aim to preserve the anonymity of the taxpayer. But this policy has not, despite another Tax Analysts lawsuit, been extended to Advance Pricing Agreements (APAs), which are basically PLRs that address transfer pricing issues. These ostensibly are too fact-intensive and particularistic to have the same dissemination value – a conclusion with which the paper does not agree (nor do I).
So that’s the American thread of the paper: looking at this history, evaluating the pros and cons, offering a typology re. how to think about the issues, and proposing at the end that APAs be publishable like PLRs, with suitable redactions.
Now to the Luxembourg story. As Omni Marian has helpfully figured out (cited in Leandra’s paper), Luxembourg was for some years issuing private letter rulings with next to no effort to do a bona fide tax analysis, as a way of helping EU companies avoid EU taxes. Here is a representative example (again, taken from Omri’s work):
Say that Deutsch-Co (DC), a German company, wants to invest €100 in a French subsidiary, DC-France, in the expectation of earning €10/year that will annually be paid over to DC. If DC invests using equity, DC-France’s €10 payout will be a dividend, nondeductible in France but exempt in Germany, so its €10 of income will be taxed just in France. If it invests using debt (and let’s assume, but ignore for arithmetic simplicity, that DC uses enough equity to avoid thin capitalization rules), the €10 of interest zeroes out DC-France’s net income but is taxable in Germany.
The trick to avoiding taxation in either country – cross-border tax arbitrage, aka hybridity – is to come up with an instrument that France deems debt and Germany deems equity. Not necessarily impossible, at least in years past (before OECD-BEPS), but also perhaps not so easy. So Luxembourg would insert itself as an accommodation party. DC would invest €100 of equity in DC-Lux, DC-Lux would invest €100 of debt in DC-France, and Luxembourg would issue a PLR misclassifying DC’s stake in DC-Lux as debt (for Lux tax purposes). So taxable income is zeroed out by interest deductions in both France & Luxembourg, but Germany, “rightly” classifying the stake in DC-Lux as equity, would still classify the €10 payment as equity, so DC ends up paying tax nowhere.
What’s in it for Luxembourg? Rather than the numbers matching exactly as in my simplified example, the DC-Lux conduit would net a slight spread between the cash flow entering and that going out again – say, less than 1% of the total amount of income that is being sheltered. That would be taxed at Luxembourg’s 29% rate. So they’d get a nice little fee, tied to the total amount of income that was being sheltered, for their accommodation services via the PLR.
This is quite a different type of example from the US-only scenarios described above. In particular, there’s a strong case that Luxembourg benefits from doing this due to the negative revenue spillover to Germany and/or France, for which it is compensated. If so, then the secrecy may help Luxembourg as well (even if other countries know in general terms what’s going on). The exposure of such previously secret rulings in the Lux Leaks scandal were therefore at least arguably bad news for Luxembourg, whereas for Lux rulings covering purely internal matters there would be no reason for presuming ex ante that secrecy was nationally beneficial.
For this reason, I see the cross-border and same-country examples as importantly different. This is probably the direction in which the paper draft is going – I gather that Lederman has a lot of interesting research into the Lux situation that was already excised here (partly for length reasons) and that may end up in a separate paper.
With all this as background, the article has two main aspects. The first is to establish a very useful typology of the types of issues raised by PLRs et al in terms of the societal interests (and the benefits or costs for particular types of taxpayers and professionals) associated with greater versus lesser secrecy. It’s off to an excellent start in that regard, although I’m urging greater use within the categories of insiderism as a key analytical issue. Second, it argues very persuasively for greater transparency and less secrecy, albeit with redactions insofar as they are needed to protect valid taxpayer concerns about confidentiality. Two especially significant aspects of the paper’s recommendations are its urging:
1) that APAs issued by the IRS be disseminated publicly along similar lines to those governing PLRs, and
2) that the adoption of generally US-style publication approaches for PLRs be at least seriously considered in other countries.
It’s nice, at a time like this, to be reminded of something that the US might currently be doing better than peer countries.