Germany: Statistical approaches to TP adjustments for profit split systems
Yves Hervé and Philip de Homont of NERA Economic Consulting discuss the application of transfer pricing ranges in profit split applications.
One key development in the OECD transfer pricing guidelines (OECD TP Guidelines) – and consequently in many national tax rules – has been the strengthening of the profit split method. In many cases, this will be the most reliable method to assess TP, especially between group entities, which each make unique and valuable contributions.
The application of the profit split method is not necessarily based on actuals. In many cases, the profit split analysis is conducted on an ex-ante basis and translated into specific prices or non-routine mark-up throughout the year.
In such cases, one challenge in the application of the profit split method is the missing guidance of the OECD on how ranges of acceptable outcomes can be established and how deviations between target profit allocations and actual outcomes should be handled. In the absence of such dedicated guidance, taxpayers should look to tested economic and statistical approaches as demonstrated herein.
In this article, the case of a multinational group with several manufacturing and several distribution companies is shown. Based on an ex-ante profit split analysis, transfer prices are set at the beginning of the financial year and normally kept constant throughout the year.
Ex-post adjustments in TP
In most business circumstances, ex-post actual outcomes at year-end will differ from ex-ante budgets. In TP, such deviations can be handled in different ways. The extremes would be to either make a full adjustment, (i.e., to make an ex-post determination of appropriate transfer prices and book a correction payment) or to make no adjustment at all (i.e., to simply rely on the transfer prices set ex-ante).
In practice, a frequently applied approach is a middle-ground solution: transfer prices are tested ex-post and, if they fall into a certain range, are accepted without further adjustment; only larger deviations are corrected. This corresponds to observed third-party behaviour, as independent parties rarely adjust prices ex-post, at least for ‘normal’ deviations.
This is also applied in our case example, in which the group would generally prefer to not conduct any adjustment payment due to administrative burdens and to not distort their management incentive structure. However, in order to decrease tax risks (which may more example may arise if an expected profit for one party turns into a loss while overall the business remains profitable overall), adjustment payments would be acceptable for ‘large’ deviations that fall outside of acceptable ranges.
The challenge of ranges in profit splits
For standard, plain-vanilla TNMM TP, the arm’s length range is usually determined by the interquartile range of benchmarked comparables’ results. However, for profit split solutions, no such natural benchmarking method to compute ranges exists, since the complementary intangible contributions of the transacting parties are unique.
In principle, a range could be established by applying different allocation keys, but ultimately this runs the risk of having tax authorities pick and choose the allocation key they prefer. This would also be in conflict with traditional pricing thinking where in which taxpayers need to justify which key they consider most appropriate. Thus, in practice, a different approach is needed.
Comparison between of ex-post outcomes and ex-ante expectations
From an economic perspective, we propose differentiating between, on the one hand, developments that reflect commercial uncertainties as they ordinarily occur throughout a normal financial year and, on the other hand, developments that would be considered extraordinary such that they may trigger renegotiations between unrelated parties. For this purpose, the natural uncertainties in the budgeting process are evaluated by considering three variables:
1. The ex-ante target margin that is determined on budget profit split analysis. For a specific supply relationship, the overall expected profit might equate to 20% of manufacturing costs, which we assume is split 50/50. Thus, the profit split analysis might, for example, translate into a manufacturing mark-up of 10% that will be used for ex-ante price setting. We call this number the ‘target TP budget’.
2. The ex-post target margin that is determined on actual profit split analysis. For a specific supply relationship, this might, in a simplified example, be equivalent to a mark-up of e.g. 15% (e.g., because the business turned out to be more profitable than expected). We call this number the ‘target TP actual’.
3. The actually realised margin. Since all financials (distribution costs, productions costs, sales, etc.) are uncertain, this number can deviate from both target numbers (ex-post and ex-ante). In a simplified example, the realised mark-up might be 17%. We call this number the ‘realised TP’.
The comparison between target margins (budget and actual) shows the uncertainties inherent in TP. In our example, a deviation of five percentage points already occurred between the target TP budget (10%) and the target TP actual (15%). This could be an indication that the actual realised margin is not too far off the ex-post target. Indeed, the two percentage point difference (17%–15%) is smaller than the uncertainties due to the budgeting process.
On a pure bilateral transactional basis, gaps between ex-ante and ex-post profit margins for a single year provide no great insight regarding the uncertainties of related to the budgeting process. However, a statistical analysis is possible in the assumed set-up of multiple manufacturing and distribution entities. In such cases, a distribution of the deviations of profit share outcomes from profit share targets considered during budgeting can be assessed statistically. If this procedure can be done, for example, for for e.g. 3–5 years and 10 intercompany supply relationships, this would generate 30–50 data points, which can then be subject to statistical analysis.
In particular, it is possible to identify typical deviations and deviations that can be considered as an outlier. Typical statistical metrics to apply to this data set would include either the interquartile range, or the standard deviation.
Deviations within the interquartile range or within the standard deviation from the target value would be considered related to acceptable normal business fluctuations, and would not require TP adjustments. Deviations out of this range would qualify to be related to exceptional business circumstances. Ex-post TP adjustments to the acceptable range would then be accepted.
As shown, in certain business set-ups, it is possible to design a TP system based on the profit split method that identifies not just point value targets, but also allows for certain well-defined deviations.
In our practical case, we could identify a threshold of 3.5 percentage points to a target mark-up derived from an ex-ante profit split analysis – if the realised transfer prices deviate by less than this amount from the target transfer price mark-up, no adjustment is booked. In practice, this solution allowed the client to implement a profit split solution with very few year-end adjustments, together with a consistent documentation of how and when adjustments are made.
Managing Director, NERA Economic Consulting
Philip de Homont
Director, NERA Economic Consulting