Assessing the potential impact of differences in regional scope when valuing SEP licences

Courts and arbitrators are often faced with the question whether FRAND royalties for standard-essential patents (SEPs) should vary across regions. In principle, they might. But in practice, the force of the evidence commonly advanced is brittle.

In this article, Pekka Sääskilahti, Andrew Tuffin, Timo Autio, and Erkka Heikkinen [1] examine the most frequent methods proposed for calibrating regional adjustments: differences in regional product prices, and differences in regional SEP portfolio strength. Neither provides a compelling basis for regional adjustments. Price differences mainly reflect the variation in products sold, not the amounts that consumers pay for equivalent products. Further, adjustments based on regional patent strength would increase the transaction costs of licensing, something that willing parties would seek to avoid.

The views expressed in this article are the views of the authors only and do not necessarily represent the views of Compass Lexecon, its management, its subsidiaries, its affiliates, its employees, or its clients.

Introduction

Global technology standards sit uneasily on territorial patent rights. Standard Essential Patents, like all patents, are granted country by country, yet technology standards are created cooperatively and intended to be implemented globally. Courts have accepted this is a commercial reality for parties negotiating licences covering Standard Essential Patents. For global companies, negotiating and enforcing rights jurisdiction by jurisdiction would be “madness” – costly, slow, and often futile.[2] Willing parties with global operations would agree a licence with global scope.

However, it creates a practical challenge. When parties negotiate terms for a global licence,[3] would the reasonable royalty terms vary depending on the regional scope of licensed sales or licensed rights? The prevailing answer, in courts and arbitrations, is that they might. In public judgments[4] that consider the issue, the labels have varied, but the broad approach is the same: there may be “Major Markets”, where a particular royalty rate is reasonable; and “Other Markets”, where a lower rate is reasonable – determined by applying a regional “discount” to the rate in Major Markets.[5] The royalty for the global licence is therefore a weighted average, reflecting its specific mix of geographical markets.

This article considers whether the reasonable value of an SEP licence should vary by region and, if so, how that effect should be measured. The underlying rationale may appear plausible. However, in practice, the claim often lacks evidential support and the methodologies used to calibrate the discounts are flawed. So, even where accounting for regional variation may improve a valuation in principle, typically, it makes it worse.

Analysing differences in regional scope: in principle, and in practice

The reasonable royalty for the licence reflects the value of the technology for which it provides consent to use.[6] Should that reasonable royalty depend on where the licensed products are manufactured and sold? In principle, possibly. But even then, the scale of that impact needs to be identified in practice. In this section, we look at: the rationale for regional variation; an emerging framework in case law for analysing it in practice; and a practical test for assessing regional adjustments when comparing SEP licences.

Regional variation in principle: a proxy, not a cause

Two factors are typically claimed to motivate regional variation in royalties.

• Value and economic circumstances: variation in consumer demand for the technology itself. In principle, the value users place on the functionality enabled by a patented technology may vary by region. That may reflect differences in preferences, income, infrastructure, or usage patterns. If the licensed technology increases consumers’ willingness to pay for implementing products more in one region than another, the royalty for that technology may reasonably differ. A profit-maximising firm would seek to price regions differently based on the differences in willingness to pay, conditional on there being no resales between regions.
• Patent coverage: variation in the licensor’s ability to monetise the value of its technology. That ability varies with two factors: (a) where the SEP holder owns patents, and (b) where the implementer’s products are sold or manufactured. The claim is that, even where the technology adds value, the products manufactured or sold in some regions may be royalty free, if the licensor lacks patents there. Even where a licensor has patents, they are not self-executing. Their commercial force depends on the extent of infringement in a region, and on the relevant courts’, and competition authorities’ attitudes toward that infringement, patents in general, and for SEPs, the FRAND commitments required by standards organisations. Those attitudes can vary across jurisdictions and over time, which may affect the patent owner’s ability to monetise its intellectual property.

Each factor can, therefore, affect the royalty that parties would agree. However, neither applies automatically nor with equal effect. Whether these factors affect negotiations, and to what extent they affect them, are factual questions that should be tested in the evidence.

Regional variation in practice: an emerging framework for analysing regional scope

The extent to which an SEP licence's regional scope affects its FRAND rate has been an issue in FRAND determinations based on comparable licences. The issues are not limited to this context, but it provides a useful and practical way to explore them.

Comparable contracts can offer a guide on the rates that parties consider reasonable. Essentially, it is a benchmarking exercise. The basic premise is that, absent hold-up and hold-out (or its more neutral term, “non-FRAND” factors[7]), similar parties will agree similar rates for similar licences. However, exact parallels are rare for SEP licences. Contracts will differ in various ways that may affect the rates that parties agree to. In principle, the closer the parallel, the better the guide. But typically, even the closest comparables have differences that affect the value of the licence. Where possible, therefore, adjusting for the impact relevant differences have on royalty rates can improve the valuation.

Differences in regional scope is one of the potentially relevant differences. Over recent years, a broad framework has emerged for analysing the issue, based on four judgments: TCL v Ericsson (USA), Unwired Planet v Huawei (UK), Interdigital v Lenovo (UK), and Nokia v Oppo (China).[8] Their details differ, but the core elements are alike. The framework and specific approaches have then influenced the approaches put forward in subsequent negotiation, litigation, and arbitrations.

The judgments share three common elements that are worth generalising (Table 1):

• The rationale provided;
• The relevant regions identified; and
• The methodology used for calculating discount rates.

Table 1: Overview of rationale, regions, and evidence base in influential judgments

Element 1: Specifying the rationale for variation in the particular set of circumstances.

Each judgment accepts that, in principle, rates may differ across regions due to differences in the value of the technology itself, or patent coverage and enforcement. However, that is not a general truth. In a given case, both the fact of regional discounts and the rationale for them should be specified, investigated, and grounded in the relevant evidence.

Whether in fact regional variation has had any impact on the rates that parties have agreed to in practice, and particularly the extent to which it has had any impact, is a factual matter and may depend on the parties’ particular circumstances.

It also matters why regional variation has or has not affected the rates parties agreed to. If parties merely follow common commercial practice, the Court may still need to establish whether that commercial practice is reasonable, or it demonstrates a party’s acquiescence to a common non-FRAND factor that willing parties would not adopt.[9]

The validity of an approach depends on the specific set of circumstances. A rationale that applies in one scenario might not apply in others. For instance, in InterDigital v Lenovo, the Court explained that the approach taken in Unwired Planet v Huawei to assess differences in regional patent coverage applied well to licensors with very few SEPs, such as Unwired Planet, but not to licensors with a large portfolio, such as InterDigital. We explore the reasons for that below (see Section 4.1 of the full article).

Element 2: Specifying the relevant regions.

If it is considered that regional scope is relevant, then the next question is how to divide the world for the purpose of the analysis. Two considerations matter.

The first is granularity: how many regions should there be?

For purposes of the court’s valuation, the prominent court judgments divided global licences into only two or three regions: (a) one region where the “benchmark” FRAND rate would apply, and (b) one or two other regions, where another lower rate would apply. In theory, a more granular approach could be taken. Regional variation could be assessed continent by continent, country by country, or at an even more granular level – for instance, if regional variation is relevant within large and economically diverse sovereign states such as the U.S., or India, or China.

In practice, courts have acknowledged that the benefits of a more granular approach would quickly be outweighed by its costs. Common sense supports this. Practical and willing parties would limit the granularity of region analysis to relatively few groups that have material differences between them to economise on transaction, monitoring, and reporting costs related to geographic variations.[10]

The second consideration is how to allocate each regional market to a specific group.

In principle, that approach should align with the stated rationale for the adjustment. If a party claims that regional variation in patent coverage affects the rate, then the regions allocated to “Major Markets” should reflect that. If, however, the claim relates to differences in consumer demand, then the approach to allocation should reflect that.

However, here too, there may be a degree of reasonable pragmatism. The relationship between patent coverage and regions where technology adds most value is not independent. Filing and maintaining patents is expensive. If a technology creates substantial value in a market, a licensor has a stronger reason to seek protection there. If the value is limited, filing may not be worth the cost. The relationship is imperfect and may lag, but substantial overlap is unsurprising. So, in practice, a country such as the USA is likely to be a Major Market from both a patent coverage and consumer demand perspective, and Belize is unlikely to be.

Aggregation in broad categories is pragmatic, but it is not without risk and complications.

Groups should capture broadly similar regions and differentiate between dissimilar ones. But categories that seem intuitive to apply in principle can be difficult to impose satisfactorily in practice. For instance, Brazil is often grouped with Emerging Markets, even when its product prices exceed those in countries typically grouped in Developed Markets. The facts underpinning these considerations also change over time. Brazil is now an important and effective jurisdiction for SEP enforcement. China, once considered an emerging market in commercial and enforcement terms, is (at least) more ambiguous now.[11]

Element 3: Methodologies to estimate the scale of the adjustment to be applied.

This is where the approach to analysing regional variation often breaks down.

Each of the influential judgments applied substantial regional adjustments that resulted in “discounts” for implementing products sold (or manufactured) in some regions.

The evidence bases that have been used to calibrate or justify specific adjustment factors vary. Broadly, three sources have been relied upon:

• Commercial practice. In Unwired Planet v Huawei, a factor of 50% was applied to determine the rate to be applied in China. That was based on the practice observed in comparable licences, in which the rates agreed for China were often lower than for the rest of the world, although the specific factor varied.[12] That factor was also applied to “Other Markets” on the basis that all licensed products would be produced in China, so it provided a “manufacturing floor”. The adjustment factor was intended to capture the impact on rates of differences in regional economic circumstances, not patent coverage (which was addressed with a separate scaling factor).[13]
• Data on average selling prices for products. In InterDigital v Lenovo, evidence on the variation in regional product prices was presented as demonstrating the difference in regional economic circumstances. This was interpreted as “relevant data” that substantiated the commercial practice observed in comparable licences, and could “refine” subsequent negotiations.[14] This evidence base was subsequently presented and considered in Samsung v ZTE.[15]
• Data on regional patent coverage. TCL v Ericsson took a royalty rate it calculated for the USA as a baseline and applied a regional patent strength ratio to each generation of SEPs. When looking at 2G SEPs, for example, it applied a 27.8% adjustment to the rate for Europe and a 45.1% adjustment for the Rest of the World.[16] Nokia v Oppo took a similar approach, applying a 38.58% discount to the rate charged in Developed Nations based on differences in regional patent portfolios.[17] To avoid double counting, no adjustment for patent coverage was made in InterDigital v Lenovo.[18]

All three sources are problematic. Commercial practice alone can be relatively weak evidence. It may merely reflect the influence that non-FRAND factors can have on negotiations. It benefits from compelling empirical or theoretical support. However, neither the data on regional product prices nor patent portfolios support regional discounts. Both are flawed, failing to demonstrate that regional discounts are reasonable factors that willing parties would take into account.

The problem with adjustments based on regional average selling prices

Typically, the variation in average product prices between regions is assumed to show or approximate the scale of that variation in demand for the patented technology. It doesn’t. The same product with the same features sells for roughly the same price wherever it is sold. The implication being that the same technology used in the same way would have the same royalty wherever it is used.

In Section 3 of the full article, we discuss this issue in greater detail. Here we summarise the key points.

Firstly, prices differ regionally largely because consumers in different regions buy different phones from different brands and with different features that affect their prices. Cellular technology is only one feature among many that affects the phone prices consumers are willing to pay in each region. Table 2 shows, for example, that the average selling price of 5G smartphones in India is 60% lower than it is in the USA. However, it also reveals the prevalence of other features that the phones sold in each country have. Among other differences, more of the phones sold in the USA have premium features that increase their prices, including large high-resolution screens, more storage, a faster processor, and an iOS operating system. Variation in product prices, therefore, does not provide evidence that American consumers are willing to pay twice as much as Indian consumers for the same product, let alone a particular technological feature in that product. Rather, it shows they buy different products.

Table 2: Regional average selling prices and selected features for 5G smartphones: USA and India (2019Q2 – 2024Q3)

Instead of looking at average prices that ignore differences in phones’ product features, what happens if we compare the prices of phones with exactly the same features? Intuitively, a particular phone will sell for a similar price wherever it is sold. To illustrate: the iPhone 16 sold for $799 in the USA and $808 in India, not $400; the Samsung Galaxy S22 sold for $799 in the USA and $782 in India, not $400.

To see the whole picture, Figure 1 shows the prices in India and the USA of all phones that were sold in both regions over a five year period between 2019 and 2024. Any phone sold for the same price in the USA and India would lie on the black “equal regional prices” line. The red “regional discount factor in EWHC cases (50%)” line shows where prices would lie if phones were genuinely 50% cheaper in India on a like-for-like basis.

Figure 1: Comparison of phone prices in India and USA

Most phones cluster near the equal-price line, with variation concentrated at the premium end: high-end iPhones are more expensive in India, while Samsung Fold models are cheaper. Lower-end phones are often cheaper in India, but none approach the 50% discount line or the gap implied by simple comparisons of average prices. It is also important to note that relatively few phones are sold in both regions. Manufacturers tend to specialise, targeting each market with different products that have different features that command different prices.

Does this mean that 5G royalties should not vary at all by region? Not necessarily. But it does mean that a simple comparison of average phone prices cannot answer that question. Applying a 50% royalty discount based solely on differences in uncontrolled average prices cannot be supported. The results show that prices are broadly consistent across regions for phones with a given set of features. That should be the starting point when considering whether the royalty attributable to one of those specific features should likewise be consistent across regions, or whether it should vary.

One can argue that royalties for cellular technology should be higher when implemented in phones with premium complementary features than they should be when implemented in phones with basic complementary features. That may indirectly lead to regional variation in royalties, simply because there happens to be more “premium” users in some regions than in others.

Even so, that would still not motivate regional discounts. That variation, if valid, would be better captured directly, by using an ad valorem rate that scales with the product price (potentially subject to price and/or royalty caps and floors). Whether or not royalties should be ad valorem, or a fixed amount per phone is a complex and polarised debate. However, it is the right debate to have, and should be engaged with. That being said, whatever the actual outcome of that debate is, an ad valorem royalty should not be smuggled into contracts through an adjustment for “regional scope”

The problem with adjustments based on regional patent strength

The next major justification advanced by those advocating for adjusting for the regional scope of a licence is patent coverage. Typically, patents play two roles in the negotiation of global SEP licences: as security, that protects the patent holder of valuable technology from free riding; and as an indicator of the innovator’s contribution to the value of the standardised technology as a whole. Problems occur when the two become conflated. Rates may also vary in line with a region’s attitude to patent enforcement.

In Section 4 of the full article, we explore the importance of both roles. Here, we summarise the problem with using regional patent strength to approximate either of them.

Variation in regional patent portfolios has been used to adjust royalty rates in SEP contracts. However, comparing regional patent strengths assumes that a licensor would and should file a patent for each of its technologies in every jurisdiction. Although it may initially seem to both the licensor and licensee that this is a common-sense approach, it turns out to be irrational (other than on a strictly myopic assessment). Insisting on it makes both parties and consumers worse off, driving up costs for licensors in the long term, which would be passed on to licensees and consumers for no additional gain. From an economic perspective, rational willing parties negotiating a global licence would take a different approach to limit costs.

A practical starting point that parties may use is the licensor’s share of the global stack of patent families. Patent families group patents in different jurisdictions that protect the same invention. A licensor with 25% of global SEP families does not necessarily own patents in each jurisdiction for each of those families; it may do, but may only have a patent in one of the relevant regions for a given patent family.

To illustrate why, we contrast the composition of a hypothetical and simplified global patent portfolio from two perspectives: the “Patent Purist” and “Commercial Purist”. The Commercial Purist, modelled in Figure 2, seeks to ensure it has sufficient coverage in the main jurisdictions to enforce in each region: in each region it has patents that contribute to at least 5% of the global stack of patent families. However, it also seeks to avoid any duplication of costs across regions. It does this in two ways: (a) it only files a patent member for a particular family in one jurisdiction, and (b) it spreads its portfolio, so that it has sufficient patent members in each jurisdiction. So, although it holds 25% of the global stack of patent families, its holding in any given region never exceeds 15% of those families. Nonetheless, the parties use the global share as a starting point when assessing the licensor’s contribution to the standard, whether in a top-down analysis or when scaling a comparable licence from one portfolio to another.

Figure 2: Hypothetical composition of a global patent portfolio using the Commercial Purist structure

Conversely, the Patent Purist, modelled in Figure 3, illustrates the perverse incentive for defensive filing. If the simple difference in regional patent portfolio is used on the portfolio illustrated above, then the rate agreed for Region 2 and 3 will be a third of the rate agreed for Region 1. Even the rate in Region 1 may be 40% less than the licensor’s true global contribution. To avoid that, the licensor files a patent in each region for each patent family. That triples its costs, yet the technology offers no more value than before.

Figure 3: Hypothetical composition of a global patent portfolio using the Patent Purist structure

This does not mean the share of global patent families is a perfect guide. Real portfolios often fall between these two scenarios to balance the opposing risks. Single-member families can be viewed as “brittle”, increasing the legal and commercial risks involved. For instance, it exposes the licensor to the risk that simplistic regional portfolio comparisons are used when valuing its portfolio. There is also the risk that courts or licensees may view single-member families sceptically, particularly if they are concentrated in one specific region.[19] Further, a licensor’s share of the global stack, as measured by simple patent counting, is not determinative. Both Unwired Planet v Huawei and Samsung v ZTE are clear on that point; their reasoning applies across all jurisdictions.[20] The proportion of declared SEPs a licensor owns may not reflect its true contribution to the value of the technology standard. Validity, essentiality, and technical importance vary across patents. Although the average value per patent may vary less between different portfolios; it can still vary materially.[21]

However, it does show the problem with adjustment based on simplistic comparisons of regional patent portfolios.

A practical test for regional adjustments when comparing contracts: does the adjustment make the comparable contracts a better guide, or worse?

Finally, we illustrate a practical test for assessing regional discounts.

The most important test is whether regional discount has a robust rationale and reliable evidence base. In addition, a practical test to see if the adjustment helps to converge the rates in potentially comparable contracts, or makes the disparities worse.

Consider this simplified example of a comparables analysis.

Two parties negotiate a global licence for cellular SEPs for use in smartphones (“Focal Contract”). Table 3 compares this Focal Contract with other contracts. For the sake of illustration, assume that: (a) there are Major Markets, where a particular rate is charged, and Other Markets, where a “discounted” rate is charged; (b) that the regional discount is, in this hypothetical example, genuinely 50%; and (c) that there are no other differences between these contracts that affect their value.

Table 3: Stylised comparable contracts before and after a regional discount adjustment

Following these assumptions, two important dynamics are worth noting.

• First, the initial unadjusted royalty rates differ only because their regional scope differs. They are broadly comparable contracts, but not exact parallels. On a like-for-like basis, the contracts have agreed the same underlying prices: $4 per unit for all products sold in Major Markets and $2 per unit for all products sold in Other Markets.
• Second, in this idealised example where we can adjust perfectly and reliably for this difference, the adjustments converge on the same FRAND rate. For all contracts, we can estimate the adjusted rate that parties would have agreed in the Focal Contract’s circumstances. In this case, the adjusted rates converge to $3.50 per unit: a 75/25 blended rate of the Major Markets rate, and the discounted Other Markets rate. That they converge, demonstrates that the assumptions were valid, and we find that the adjusted rates provide the true reasonable rate for the Focal Contract.

In reality, adjusted rates rarely converge in this precise manner. In practice, we have seen some parties present regional adjustments that widen, rather than narrow, the gap between otherwise comparable contracts. That may be justified, but it is a warning sign that further investigation and explanation is required. When adjustments fail to narrow the range of rates, it can indicate that the adjustment itself is flawed, either in principle, or in how it has been applied in practice.

There are two general lessons here, that should be applied when working with comparable contracts.

• The closest comparable is the better starting point. The appeal of the comparable approach is that it provides evidence on what the market considers reasonable, not what an analyst estimates is reasonable. Where adjustments are reliable, they can improve a valuation. But adjustments introduce a risk of error and dilute the initial appeal of the approach: relying on observable market evidence. As a general rule, that risk is less severe for a contract less in need of adjustment.[22]
• Even if an adjustment is necessary in principle, it will not improve a valuation unless it is reliable in practice. The methodology used to adjust rates must be grounded in the economic and commercial rationale for the adjustment, and the facts of the case. If it is not, there is no reason to believe it helps improve the valuation, and may make it worse.[23]

In our experience, the methodologies that are typically used when applying regional adjustments fail the second of these general lessons. Even where they are right in principle, they are often too unreliable in practice.

---

References