Dear Sir or Madam,
We thank you for the opportunity to provide comments on the Implementation Guidance on Hard to Value Intangibles released by the OECD on 23 May 2017, (the "Draft"). The comments we would like to offer for your consideration are stated below.
1. IP Life Cycle and Value It would be quite rational to expect that when IP in the early stages of development is valued in an arm's length setting, the value placed on such IP would be lower than the value of the same IP in the later stages of the development. The main reason for this is that in the early stages of development it would be natural to expect greater variability in the profit potential of the IP than in the later stages of development. Therefore, an arm's length investor would apply a higher discount rate to value IP in an early stage of development compared to the same IP in a later development stage. As an example, one can consider the value of the IP embedded in a pharmaceutical product that is just entering the Phase I of clinical trials vs. the same pharmaceutical product that has completed the Phase III trials and received marketing approval. Because failures of pharmaceutical products to pass successive phases of clinical trials are not uncommon, it would be quite rational for an arm's length investor to place a lower value on the IP of the product at the time when it was entering the Phase I trials compared to the same product a few years later after it has received marketing approval.
1 These comments represent views of the authors and do not necessarily reflect the views of NERA Economic Consulting.
Socit par Actions Simplifie R.C.S. Paris B 451 252 118
Therefore, tax authorities' position that the value of an early-stage IP has to be computed using the profitability realized in later years would be at odds with the arm's length evidence.
2. Contingent Payments and Adjustable Royalties
The Draft tends to view the valuation of Hard-to-Value Intangibles (HTVI) as a "one-shot deal." Although one-shot deals have their place among taxpayers in controlled relationships, such transactions may be quite rare, because most of the related-party transactions occur in the context of long-term, cooperative relationships. To understand the patterns of arm's length behavior in transactions that involve development of intangibles with uncertain value and continuous engagement of the parties, one can turn to transactions involving growth-stage companies funded by venture capital ("VC arrangements").
As mentioned in the preceding paragraph, in VC-backed deals, it is common to place a rather low value on an IP in the early rounds of investment when the profit-generating potential of the IP is associated with significant uncertainty. As the development of the IP progresses and some of the uncertainties get resolved, the value of the IP changes. Assuming a favorable resolution of the uncertainty (e.g., successful completion of a clinical study phase for a new drug, reaching of a development milestone by a software product), the value of the IP will increase. In a VC-backed transaction, this normally results in a higher valuation of the company shares in subsequent funding rounds. Often, VCs from the earlier rounds will participate in subsequent funding rounds, too, meaning that they may pay a higher price for shares of the same company. This is equivalent to paying a higher price for the same IP after some of the uncertainty about the profit generating potential of this IP has been removed.
Applying the same reasoning to transfers of HTVI in controlled transactions where the value of the intangibles increases after subsequent development, an arm's length behavior for the taxpayers would be to periodically re-value the IP associated with HTVI. In practice, this could be achieved via the mechanism of contingent payments and/or variable royalties. In view of the arm's length evidence discussed above, the authors believe that the "alternative pricing structures" mentioned in para. 12 of the Draft and para. 6.192 of the Guidelines should be discussed as important and, sometimes, necessary attributes of transactions involving HTVI.
3. How Significant Is the Information Asymmetry Between Taxpayers and Tax Authorities?
In the experience of the authors, for the intangibles transferred in the early stages of development, the information asymmetry between the taxpayer and the tax authorities may not be as large as is presumed by the Draft. The main reason for this is that taxpayers in the early stages of IP development rarely have insight into the true value of the intangibles they are developing. If the projections of cash flows from the intangibles are highly uncertain (which is the very definition of the hard-to-value intangibles proposed in para. 6.189), there may be a broad range of potential
NERA Economic Consulting
value outcomes for these intangibles once they are fully developed. For many types of intangibles under development, the resulting value depends on events outside of the taxpayer's control. As an example, such events may be (i) internal unanticipated events, such as delays or acceleration of the development schedule, deviation of the actual intangible development costs from the projections, and changes in the scope of the development work, (ii) external unanticipated events, such as state of the market at the time of the intangibles' rollout that may either be conducive or detrimental to the profits derived from the intangibles at the time or their commercialization; and competitive entry during the development period that could materially reduce profitability of the intangibles under development or even render them obsolete.
Thus, the Draft's presumption that the taxpayers always have more information about the true profitability of the intangibles transferred in an early stage of development leaves the taxpayer with a virtually unsurmountable burden of proof to argue that the events that led to a higher-thanexpected value of the resulting intangibles were unforeseeable. Even if the value of intangibles was estimated ex ante by a taxpayer using a probability-weighted forecast, what is to stop a tax administration from arguing that this valuation was deliberately skewed towards a low value?
Moreover, para. 6.193(i) seems to imply that a taxpayer must be able to provide contemporaneous documentation from the date of transfer in order to establish an ex post exception based on subsequent unforeseeable developments. This imposes an unrealistic burden and could be viewed as foreclosing entirely reasonable ex post explanations that could not reasonably have been foreseen at the time of the transfer. Rather, the ex post exception should be based on whether the taxpayer can reasonably establish ex post that the relevant developments were unforeseeable at the time of the transfer, including reference to documentation that may have been established at the time of the transfer.
4. False Positives
The same idea can be explained with the concept of "false positives" and "false negatives," which are concepts analogous to Type I and Type II errors in statistical hypothesis testing.
In a simplified manner, a "false positive" is a result that indicates a given condition has been fulfilled, when it has not--for instance, a medical test would return that a patient is infected when he is not. Conversely, a "false negative" is a result that indicates a given condition has not been fulfilled, when it has been--for instance, a medical test would return that a patient is not infected when he is actually.
As an illustration, let's consider a fictive population of 1,000 persons where 2% of people are actually infected by a given disease. A test exists to perform a diagnostic. This test has a false positive rate of 5% (0.05) and no false negative rate (all infected patients are tested positive by this particular test). The expected outcome of the 1,000 tests on population would be as follows:
NERA Economic Consulting
Table 1: Illustration of False Positive Paradox
Infection Actual Status
20 0 20
49* 69** 931 931 980 1000
* Calculated as 1,000 * 98% *5% ** The sum of truly infected patients and false positives
In the above situations, the test would result in identifying 69 "positive" (i.e., infected) out of which only 20 would be true positive and 49 would be false positive. This example illustrates that, in certain situations, false positive tests are more probable than true positive tests (a situation referred to as the "false positive paradox").
The authors believe that the concept of false positives may be insightful when applied to the HTVI Draft. As a matter of fact, it may be extremely difficult, ex post, to distinguish between "bad faith taxpayers" who abusively mispriced the value of a transferred intangible and "good faith taxpayers" who priced the intangible earnestly but for whom exceptional circumstances arose, resulting in a substantial deviation between the initial expectations and the actual business conditions.
Another way to put it would be to consider--as an illustration--a situation where all taxpayers act in good faith. Let's assume that in 80% of the cases, ex ante and ex post valuation of intangibles are aligned. Yet in 20% of the cases, there is a misalignment, with 10% of the cases resulting in an overvaluation of intangibles and an over-taxation and 10% of the cases where intangibles were undervalued resulting in an insufficient taxation.2 Under the approach recommended by the Draft, it is likely that the latter 10% of cases would result in a reassessment. Yet, assuming all of the taxpayers acted in good faith, these cases would just correspond to situations where the most extreme scenarios have occurred.
As such, as economists and transfer pricing practitioners, the authors are strongly concerned that the HTVI guidance would result in a substantial proportion of "false positives" as it would be very
2 The terms "overvaluation" and "undervaluation" here refer to the mismatch between the ex ante and ex post valuations of the transferred intangibles by more than 20%.
NERA Economic Consulting
difficult for taxpayers to establish that they were faithful but "lucky" versus a situation where they would have been "abusive." We can envision a situation where, tax administrations, using ex post results, would come to a conclusion that a group of taxpayers, on average, have underestimated the value of the transferred intangibles. Yet, at the same time, it is highly unlikely that all of these taxpayers would have acted in bad faith.
The process of intangibles development is often associated with significant risks. For the taxpayer, it may be extremely difficult to determine the full extent of reasonable scenarios that may exist for an intangible. Such a determination may require complex and extensive economic analysis that not all taxpayers in all situations may be able to afford. For instance, taxpayers may establish projections in "good faith" that turn out to be wrong for reasons outside of the taxpayers' control.
In this regard, an analysis of the evolution of comparable intangibles in the past, if possible, may be an appropriate option to determine the extent of reasonable evolution scenarios.
In particular, based on the authors' experience, the 20% threshold featured in the parag. 6.193 of the BEPS TP Report seems rather low. The authors find it disputable that a single threshold can be applied to all types of intangibles.
As such, the authors are concerned that the HTVI framework may be exploited by tax administrations to propose reassessments for taxpayers who acted in good faith.
5. Overpayment & Self-Adjustment The authors also think that the Draft should address the situations of overpayment for HTVI, when the ex post value of intangibles turns out to be lower than the expected ex ante value, and offer a mechanism of resolution for such cases, both from the perspective of the taxpayer and of the tax administration, particularly because para. 6.184 of the OECD Guidelines acknowledges the existence of such situations, at least in the arm's length setting.
It also would be a welcome addition to the Draft to acknowledge the possibility for the taxpayer to make self-initiated adjustments when either an underpayment or an overpayment for HTVI becomes apparent to the taxpayer.
6. Scenario Analysis
Notwithstanding the above, the author believes that the practice of preparing a scenario analysis for intangibles whose profit potential may vary greatly should be encouraged. In this regard, please refer to a publication by V. Starkov, "Applying Scenario Analysis for Computing Discount Rates in Cost Sharing Arrangements," Tax Management Transfer Pricing Report, Bloomberg BNA, Vol. 24 No. 4 (2015).
NERA Economic Consulting
Such a scenario analysis should be properly documented in terms of developing cash flow projections for different scenarios, assigning probabilities to each scenario, calculating discount rates, etc. so as to avoid suspicion by the tax administration that such analysis has been "skewed" by abusive taxpayers.
7. Comments on Examples in the Draft As a general comment, we think that examples should be used with caution. Ideally, the guidance should be clear and sufficiently logical so as not to require any illustration through examples. Yet, as a suggestion, the Draft should provide an example or examples where the valuation performed by a taxpayer at the time of the HTVI transfer, based on taxpayer's ex ante projections, does not get adjusted by the tax authorities, and an example when it does. The Draft should clearly spell the differences in the valuation techniques used ex ante by taxpayers that did not lead to an adjustment and those that did.
8. HTVIs and MAPs In the opinion of the authors, the HTVI framework may substantially increase the risk of double taxation. As such, we believe that any adjustment initiated by tax authorities under this framework should automatically be open to MAP proceedings.
Dr. Vladimir Starkov, Associate Director, Chicago http://www.nera.com/experts/dr-vladimir-starkov.html Guillaume Madelpuech, Principal, Paris http://www.nera.com/experts/guillaume-madelpuech.html
NERA Economic Consulting