4 The bottom-up approach

Topics covered in Part 4:

• An overview of the bottom-up approach

• Step 1: The economic principles of modeling a hypothetical negotiation

• Step 2: An empirical assessment of technical contributions and their economic value

4.1 An overview of the bottom-up approach

In license negotiations, a “bottom-up” approach estimates a reasonable royalty for a specific patented contribution. It anchors the rate, or a range of rates, in the technology’s incremental value. Incremental value refers to the additional benefit the technology delivers compared with the next-best alternative way of performing the same function.

In SEP licensing, practical alternatives are constrained – or eliminated – once a standard has become widely adopted. In this setting, the bottom-up approach therefore seeks to anchor a FRAND rate in the SEP’s incremental value ex ante: that is, the royalty that the parties would have regarded as reasonable immediately before the standard was set, taking into account the incremental benefit that the chosen technology offers over the alternatives available at the time.

4.1.1 The rationale

The motivation for this approach to SEP valuation is to address the main concern with ex post negotiations – lock-in. Once a patented technology has become essential to a widely adopted standard, implementers’ competitive alternatives narrow, and the royalty that an SEP holder seeks may reflect the benefit of complying with the standard, rather than the use of the patented technology itself. This concern is most acute where compliance is necessary for the implementer, but the patented technology contributes only limited incremental value to the product – for example, because it relates to a minor feature, or because similarly effective alternative technologies were available when the standard was being developed.

The ex ante benchmark therefore asks what price the parties would have agreed when implementers still had credible outside options, such as rival technical solutions that could have been incorporated into the standard, or the option to delay, redesign or forgo the relevant functionality. These outside options limit what an implementer would rationally be willing to pay and, correspondingly, constrain what a licensor could demand in a competitive setting.

In such a setting, license terms would fall within a bargaining range bounded by two constraints:

• The licensor’s minimum willingness to accept (WTA). The lower bound reflects the value of the licensor’s best alternative to the agreement – its opportunity cost. A patent owner’s alternatives to an agreement may include profits from using the technology itself, licensing the technology to other implementers or deploying the invention in other markets.

• The implementer’s maximum willingness to pay (WTP). This is determined by the incremental value of the licensed technology over the next-best alternative, taking account of any costs of adopting that alternative. If the royalty demanded exceeds that incremental value, the implementer would rationally prefer the alternative. In the limiting case, the “next-best alternative” may be to forgo the functionality and to incur neither the costs nor the benefits associated with it.

In some settings – particularly where rivalry for selection into the standard is strong and the leading technical alternatives are close substitutes – the ex ante benchmark can imply a low royalty, in the limiting case approaching zero, even where the chosen solution will provide substantial value to implementers. This outcome occurs where two conditions coincide: (iI) the licensor’s WTA is low due to the limited availability of outside options even before the standard is set, ⁽¹⁾See Section 5.2 explaining possible reasons. and (ii) the incremental value of the patented technology over the next-best alternative is small.

Whether it is reasonable, as a matter of economic theory and commercial practice, for a valuable technology to command a zero, or near zero, royalty is contested. The debate turns on issues such as the realism of “perfect substitutes,” the appropriate timing of the ex ante benchmark, the need to preserve long-term innovation incentives and how to treat alternatives that may have existed in theory but not in practice. These issues are discussed in Section 4.2 below.

In principle, focusing on the incremental value that a technology provides over its alternatives provides incentives to develop beneficial technologies. First, if a technology’s advantage over its alternatives is small, a low royalty can be economically consistent with competitive markets. Second, innovators are incentivized to develop technologies that offer high incremental value over the value that existing alternatives can already provide.

Box 4.1 sets out the dynamics of two otherwise identical license negotiations: one conducted ex post and the other ex ante.

4.1.2 The approach

Typically, a bottom-up valuation is conducted ex post, after the relevant SEPs are already essential to a successful standard. To estimate the royalty range that would have been available under competitive constraints, it therefore seeks to establish the outcome of a hypothetical negotiation.

In practice, the approach combines two linked components:

• Step 1: An economic model of the hypothetical ex ante negotiation. This sets out the principles and assumptions used to identify a bargaining range for the license.
  

• Step 2: An empirical assessment of technical contributions and their economic value. This provides the evidence needed to apply (or “calibrate”) the model to the particular SEP(s) or portfolio at issue and its credible ex ante alternatives.

4.1.2.1 Step 1: Modeling a hypothetical negotiation

Modeling a bargaining range is a standard tool in economic analysis and is used relatively frequently when economists assess reasonable royalties in US patent-infringement cases. ⁽²⁾See Meyer, C. and B. Ray (2005). A critique of noneconomic methods of reasonable royalties. In Leonard, G.K. and L.J. Stiroh (eds), Economic Approaches to Intellectual Property. NERA Economic Consulting. Swanson and Baumol first formalized the approach for SEPs. ⁽³⁾The ex ante benchmark was developed in Swanson D.G. and W.J. Baumol (2005). Reasonable and non-discriminatory (RAND) royalties, standards selection, and control of market power. Antitrust Law Journal, 73(1), 1–58. Also see: Layne-Farrar, A., A. Jorge Padilla and R. Schmalensee (2007). Pricing patents for licensing in standard-setting organisations: Making sense of FRAND commitments. Antitrust Law Journal, 74(3), 671–706; and Lemley, M. and C. Shapiro (2007). Patent holdup and royalty stacking. 85 Texas Law Review, 1991 (2007), 1 Stanford Law and Economics Olin Working Paper No. 324, 1990-2049, available at SSRN.

The result of any hypothetical negotiation depends on the economic assumptions used to represent bargaining dynamics and competitive constraints. Therefore, a key task is to ensure that the model reflects economically plausible behavior and a realistic decision point for standard selection.

In particular, the timing of the hypothetical negotiation will affect the result. This approach typically frames the negotiation as occurring just before the standard specifies the preferred technology, when manufacturers – through the standard-setting process – could hypothetically still have chosen between rival technical solutions based on the net value they would provide, based on the benefits of their functionality minus any costs, including royalties. ⁽⁴⁾See the discussion on timing below, in “practical considerations.” In the United States, in reasonable royalty cases, the hypothetical negotiation takes place on the “eve of infringement,” which is determined by the implementer’s first use. However, in the context of FRAND, in principle this would reflect the ex ante principle – i.e., the parties at the negotiating table would be focused on the FRAND obligation, which would price the value of the technology prior to standardization. Further considerations about selecting this timing as a reference point are addressed in Section 4.2.2.

The structure of the bargaining process will also matter. In principle, the negotiation can be characterized as either:

• bilateral negotiation between a patent holder and a prospective implementer, in which rival technologies act as outside options that constrain the royalty that the patent holder can achieve; or

• auction for inclusion in the standard, in which rival technologies effectively compete for selection based on the functionality they offer and the licensing terms on which they are made available.

The bilateral negotiation is typically the default structure used for the ex ante approach and produces a bargaining range, rather than a winning “bid.”

4.1.2.2 Step 2: Assessing technical contributions and their value

To apply the approach to a particular SEP or SEP portfolio, the analysis then requires evidence on (i) the value of the patented solution’s technical contribution to the implementing product and (ii) the value of any credible ex ante alternatives that could have been adopted instead.

In practice, this assessment can be structured in two stages:

1. a technical assessment, comparing the performance and trade-offs of the patented solution against contemporaneous alternatives;

2. an economic assessment, translating incremental technical improvements into incremental economic value, recognizing that technical improvements and economic value may not be proportional.

The main considerations for each step are set out below.

4.2 Step 1: The economic principles of modeling a hypothetical negotiation

A bottom-up valuation depends on the economic logic that underpins the hypothetical ex ante negotiation.

The central idea is that a license determines how the parties divide the gains from trade, i.e., the economic value created when patented technology is combined with an implementing product.

4.2.1 Understanding the economic principles that determine the bargaining range

In a hypothetical negotiation, the bargaining range is shaped by the gains from trade created when patented technology is combined with an implementing product and by the outside options available to the parties.

4.2.1.1 Gains from trade: the economic value of implementing a technology

The starting point for the hypothetical negotiation is the economic value that the licensed technology unlocks in use. Standard essential technologies are valuable because they provide technical functionality that benefits users and can reduce costs, or improve performance and reliability, for implementers. Ultimately, the amount that users are willing to pay for the resulting product improvements, together with any cost savings, determines the technology’s economic value. For an implementer, the value attributable to the technology is the incremental profit it can earn when the product implements the technology, compared with the position it would have faced without it. That incremental value may be driven by higher demand – reflected in prices or volumes, or both – and by lower production or operating costs (see Part 1).

That technical value is conceptually distinct from the value that arises when a market coordinates around a common standard. Interoperability, scale economies (lower per-unit costs as production volume increases) and network effects (product value increases as more users adopt it) can be important sources of value, but they play a different role in ex post and ex ante negotiations.

• In an ex post negotiation, coordination benefits have already formed around the chosen technology, weakening competitive constraints and insulating it from price competition.

• In an ex ante setting, by contrast, the benefits of coordination will form around whichever technology is selected. However, the scale of those coordination benefits may differ depending on which technology is selected; a higher-quality solution may attract more implementers and consumers and thereby generate larger network effects and economies of scale than the inferior alternative would have done. But that difference does not distort the comparison between technologies. It simply affects the size of the gains from trade that each ex ante alternative will generate.

Finally, as discussed in Section 1.1.2 above, the economic value created by implementation is complementary: it depends on both the patented technology and the implementer’s product. The technology generates value only when implemented in products; products capture additional value only when they can use the technology.

The resulting value is therefore a gain from trade – joint value that depends on an agreement to combine the patent holder’s technology with the implementer’s product.

Box 4.2 illustrates the gains from trade in the negotiation between Gamma and Alpha described in the hypothetical ex post and ex ante negotiations discussed above.

4.2.1.2 Negotiating a price that divides the gains from trade

The parties negotiate terms of a license agreement. They do not negotiate a specific division of the gains from trade; they negotiate a price. However, in principle, each party will reject terms that leave it worse off than it would be if no agreement is reached: therefore that price will be mutually beneficial.

In an ex ante negotiation, before parties are locked into implementing a particular technology, the price they agree will divide the gains of trade: the royalty is the patent holder’s share, the remainder is the implementer’s share – its additional profit after royalties.

The range of prices that the parties would mutually agree on depends on the net value provided by their outside options at the time of the negotiation. The implementer will choose the option that provides it with the greatest net value: the difference between the benefit and the cost of choosing that option.

In the typical setting for an ex ante hypothetical negotiation, for the purposes of the bottom-up approach assumed to take place immediately before the standard is agreed, the minimum royalty that innovators competing to be part of the standard are willing to accept may be low. In certain scenarios, this value may be zero. Under the assumptions typically used for this method, two reasons explain this phenomenon.

First, for the patent holder, the alternatives to an agreement may include profits from using the technology itself, licensing it to other implementers or deploying it in other markets. However, in the context of standards, these outside options may be constrained or eliminated by the process of developing a successful industry standard – even at the ex ante stage, immediately before the standard is set. Where a market is expected to coordinate around a single standard, a patent holder whose technology is not selected may anticipate that it has a limited prospect of competing against the eventual standard and its associated network effects. If the technology is selected, standard adoption and a FRAND undertaking commonly imply broad access on non-exclusive terms, which can limit reliance on exclusivity, or on playing implementers off against one another, as a source of bargaining pressure on royalties.

Second, whether a patent is essential to a standard or not, the patent holder’s development or acquisition costs will not affect the minimum royalty it is willing to accept. At the point of negotiation at the selected point in time, those costs are sunk: they have been incurred and cannot be recovered whether an agreement is reached or not. So, it may accept little to no royalty and incur a loss, even if its technology offers value. The prospect of that loss may have deterred it from its earlier decision to invest, but by the time the standard is being set, its sunk costs do not make a given royalty offer less attractive than the option of walking away.

The significance is that, in this situation, the patent holder of the next-best alternative will be willing to accept no royalty, even if its technology offers value to implementers. In that case, the net value offered by an implementer’s outside option will be the same as the economic value of that option. This constrains the maximum amount that the implementer is willing to pay for the preferred technology to the incremental value that technology offers over the next best alternative– as illustrated in Box 4.3.

4.2.1.3 Negotiations in three cases of ex ante competition

Here, we illustrate the dynamics of the hypothetical ex ante negotiation between an implementer and an innovator. As in Box 4.1 above, the implementer, Gamma, sells consumer devices for $500, earning $100 profit on each one. It implements a wireless communication standard, to which the innovator, Alpha, contributes. Alpha’s patented technology, covering a data-packaging technique, increases Gamma’s profits by $1.00 per unit (compared with not having that functionality).

The outcome of ex ante negotiation between the parties depends on the strength of Gamma’s outside option. There are three possible situations:

• No ex ante alternative. Where no realistic alternative exists for the relevant function, the incremental value may approach the technology’s full economic value, subject to the implementer’s ability and incentives to adopt. Even then, an implementer would not rationally pay more than the economic value that the technology enables; beyond that point, it would prefer to forgo the functionality.

• An imperfect alternative. Alternatives may be inferior in quality or differentiated in ways that affect the implementer’s preference. Where the next-best alternative would have delivered less value, the incremental value reflects that difference – which may be substantial or modest, and may vary across implementers.

• A perfect (or near-perfect) alternative. Where the next-best alternative would have delivered substantially the same value, incremental value is small. That does not imply that the total value of the functionality is small; it means that the best option and the next-best option benefit the implementer to a similar extent, which may or may not be large. In either case, the royalty is tightly constrained and may tend toward the licensor’s minimum WTA.

Ex ante negotiation with no alternative

If the parties agree terms, the gain from trade is $1.00 per unit. Without an agreement, neither party benefits.

The minimum amount that the licensor, Alpha, is willing to accept is nothing. It has no outside option: its past development or acquisition costs are sunk at the time of negotiation, therefore any positive payment improves its current position. In practice, if anticipated, such an outcome would have discouraged Alpha from conducting R&D, and it may dissuade it from developing technology for inclusion in subsequent standards. However, it does not affect the current negotiation in this hypothetical scenario.

The maximum amount that the implementer, Gamma, is willing to pay is $1.00 per unit. Above that price, the cost of licensing exceeds the benefit it provides, and so Gamma would rather do without.

In a bilateral negotiation, the bargaining range lies between these two points. The parties determine the royalty based on their relative bargaining power. In an auction setting, which differs from a standard setting, the royalty may simply be the implementer’s willingness to pay, if the patent holder can make take-it-or-leave-it offers. Box 4.4 illustrates the bilateral negotiation.

Ex ante negotiation with an imperfect alternative

This section considers the same negotiation, at the ex ante stage, where the relevant part of the standard could instead have adopted a credible alternative technical solution that provides the same or similar functionality – although not as well. The next-best alternative is an imperfect substitute that offers less value.

As before, if the parties agree on terms, the gain from trade is $1.00 per unit. However, this time, the implementer, Gamma, can opt for an alternative agreement, which provides $0.80 per unit. Both licensors have no outside options, so the minimum each is willing to accept for the benefit that their respective technical solutions provide is zero. On that basis, the net value offered by Gamma’s outside option is $0.80 per unit: its entire value for no royalty.

The bargaining range for Alpha’s superior technology is $0.00 to $0.20 per unit – the incremental value it offers over the next-best alternative ($1.00 – $0.80). The royalty they determine depends on their relative bargaining powers; the rest of the gains of trade is Gamma’s share. This is illustrated in Box 4.5.

Ex ante negotiation with a perfect alternative

This section considers the same negotiation but in this instance the next-best alternative at the ex ante stage is a perfect substitute that offers the same value.

As before, if the parties agree terms, the gain from trade is $1.00 per unit. Both technical solutions offer the implementer Gamma exactly the same benefit, and both licensors are willing to accept nothing as they lack outside options. Both offer the same net value to Gamma, but no incremental value. The royalty that the parties agree is zero, as illustrated in Box 4.6.

4.2.2 Contentious issues when establishing a reasonable hypothetical negotiation

The outcome of the hypothetical negotiation depends on how it is set up. In the context of FRAND valuation for SEPs, three issues in particular require careful consideration.

1. Timing. At what point in time should the hypothetical negotiation be placed?

2. Information. Should the model use only information available ex ante, or may it rely on ex post evidence about value?

3. Role of the bargaining range. Does the model identify a single FRAND rate, a ceiling or simply a bargaining range that must then be applied with judgment?

4.2.2.1 The timing of the negotiation

The issue here is whether modeling a hypothetical negotiation immediately before the standard is set provides reasonable terms.

The issue

In non-SEP reasonable royalty analysis, the hypothetical negotiation is often placed at the “eve of infringement” – typically linked to the implementer’s first use. In the SEP context, that timing is less informative. By the time infringement occurs, the standard has already been specified and the relevant technologies are essential to it; in many cases, viable technical alternatives may already have been eliminated. Placing the negotiation at this point in time can therefore undermine the motivation for an ex ante framework, as the implementer already lacks effective outside options to constrain the royalty demands.

For this reason, the conventional ex ante framing for SEPs places the hypothetical negotiation earlier: immediately before the relevant technical solution is selected into the standard (the “eve of standardization”). At that stage, implementers can still choose between competing technical proposals – through the standard-setting process. This timing is attractive because it captures the central ex ante intuition: competition between alternatives constrains the royalty that a successful technical proposal can command.

A key criticism, however, is that this timing can be asymmetric. It occurs before implementers of the standard are locked into the standard, but after innovators of the standard are locked in. Implementers retain outside options, leveraging competition between patent holders. Innovators do not: their R&D costs are sunk, and the network effects that will form around the winner mean that the losing technical proposals cannot viably compete against the proposal that becomes the standard. Furthermore, the selected technical proposal cannot leverage competition between the implementers for access to the benefits it provides, as the innovator cannot implement or license their technical solution exclusively – the implementers form a de facto monopsony, particularly where the prospect of standardization drives strong coordination and interoperability benefits.

That criticism has been used to dismiss the approach entirely. However, analyzing the dynamics of the bargaining process can help a valuation. So, in principle, one potential solution proposed to address this criticism is to further modify the timing of the hypothetical negotiation to a point when neither implementers nor innovators are irreversibly committed to the standard. ⁽⁵⁾Sääskilahti P. and A. Tuffin (2023). What the ex-ante benchmark reveals about the reasonable price for SEP licences . Compass Lexecon. In practice, that means setting the hypothetical negotiation at a point in time before the innovator has incurred the cost and risk of development. At that point, the costs of development are incremental, and so inform the innovator’s minimum WTA. Competitive pressure will still encourage it to reduce its demands, but not to the point where it would make a loss from developing the technology. Below that “break even” point, or “hurdle rate,” it would be better off not to compete at all. The principle of the ex ante framework therefore remains, but an innovator is not irreversibly committed to a loss-making agreement. Of course, establishing the minimum amount that an innovator would hypothetically have been willing to accept before it was committed is difficult in practice.

Reasons to keep the conventional timing

Some economists do not consider the outcome of the conventional timing to be necessarily paradoxical. Innovation is inherently risky: firms invest ex ante without certainty of selection or commercial success, and a successful technical solution may still earn less than originally hoped. Low royalties may simply reflect strong ex ante competition and the presence of close substitutes. Essentially, even a successful innovation might not make the money that was initially expected. ⁽⁶⁾Melamed, D. and C. Shapiro (2018). How antitrust law can make FRAND commitments more effective . Yale Law Journal, 127(7), 2110–2141. This is not a problem that is unique to standards, per se. Standards and SEP portfolios often consist of many patents and not every patent will be a “winner.” As noted above, it is the long-term expected value of a potential portfolio of patents that draws firms to participate in innovation markets. Furthermore, if the sales volume of products implementing the standardized technology are expected to be very large, inclusion of a technical solution in the standard and the ability to earn even marginal royalties over close substitutes may provide sufficient incentive to engage in R&D. This perspective also notes that many non-SEP patents earn little or no royalty because they do not offer meaningful incremental advantages over competitive alternatives.

The issue may, however, be mitigated. In many real-world settings, the “zero-royalty” result is an edge case associated with perfect (or near-perfect) substitutes. Where alternatives are imperfect, incremental value is positive, and the benchmark can yield royalties that plausibly reward successful innovators. Even where the criticism is accepted in principle, it need not undermine the usefulness of the ex ante framework. A common practical approach is to retain the ex ante incremental-value logic as the core discipline, while recognizing that extreme outcomes may require careful scrutiny – particularly where they would imply implausible incentives given the scale of investment and the nature of competition.

Reasons to adjust the framing to reflect innovators’ ex ante incentives

The alternative view is that competition should not systematically imply that there is no expected return for efficient innovators where the technology creates value for users. The concern is that, if selection into a standard is expected to eliminate innovators’ effective outside options and push royalties toward zero even when the technology is valuable, participation in and intensity of competition to develop valuable technologies may weaken – particularly for firms that rely on licensing to fund R&D.

In this view, innovation may skew toward vertically integrated firms, which can recoup returns through downstream product sales rather than licensing, reducing the pool of potential innovators that could compete to contribute to standards. It also makes it more likely that standards rely on spillover innovation, where technologies are developed primarily for other markets and only incidentally contribute to standards. Finally, innovators would have incentives to contribute in settings where competition is expected to be weak, where they anticipate earning returns closer to the full value of their technology.

4.2.2.2 Information in the hypothetical negotiation

The issue here is whether a hypothetical ex ante negotiation may use information that emerges ex post.

The issue

In practice, many key inputs are uncertain at the time of standard setting: how widely the standard will be adopted; what products and use-cases will emerge; and how much value a given feature will create in the market. If the negotiation takes place before the innovator has incurred any development costs, then all information about future costs and benefits may be particularly uncertain at that point in time.

A central modeling question is therefore whether the hypothetical negotiation must be restricted to what parties could have known in practice when that negotiation took place, or whether the negotiation may appeal to foresight – informed by evidence that only emerged subsequently.

A literal ex ante approach

Under a literal approach, the hypothetical negotiation is modeled using only information that was available at the time. This has two consequences:

• Uncertainty – the model must incorporate uncertainty about adoption, demand and costs, and about the performance and commercial significance of alternatives.

• No benefit of hindsight for either party – the implied bargaining range must reflect the value expected at the time, not value realized later. If realized value turns out to be higher than expected value, the model does not retroactively increase the royalty (in the licensor’s favor); if realized value is lower than expected, it does not retroactively reduce it (in the licensee’s favor).

This approach resembles how parties would actually contract a fixed price under uncertainty and avoids the bias of hindsight.

An ex ante negotiation with ex post evidence

This approach allows the parties to have “perfect foresight” during the negotiation. They negotiate with full knowledge of the evidence that becomes available ex post. This allows them to more accurately estimate the value that the parties are bargaining over. The core rationale is that the ex ante benchmark is meant to remove distortions from lock-in and ex post bargaining power, not to freeze the valuation in the face of uncertainty that is unrelated to opportunism.

On this approach, ex post evidence can help measure value more accurately. For instance, it will be clearer what performance improvements were achieved, what use-cases emerged and how consumers responded. That evidence does not re-create hold-up, as it does not affect the licensor’s bargaining leverage in the hypothetical negotiation – rather, it refines the estimate of the surplus to be shared.

However, it may also introduce error. In particular, parties make decisions without perfect foresight, so that risk affects their incentives. If risk is not adequately factored into the negotiation, then it might not truly represent the prices that the parties would have been willing to pay or accept. In that case, use of ex post evidence must be as an input must be viewed with caution.

4.2.2.3 The role of the bargaining range

The issue here is whether the ex ante bargaining range is determinative. There are two aspects to this discussion.

What does the hypothetical negotiation yield: a FRAND rate, a ceiling or a range?

Within the bilateral negotiation model, the implementer’s WTP sets the ceiling and the licensor’s WTA sets the floor; any price between these could, in principle, be FRAND. Any price lower than the WTP makes an implementer better off than it would have been with the next-best alternative. Any price above the WTA makes the licensor better off than it would otherwise have been without the agreement.

Most applications would stop at identifying this bargaining range and not identify a single price. The exact price would depend on other frictions that the model typically abstracts – in particular, the factors that affect the parties’ relative patience and negotiation skills, which may be hard to parameterize in an ex ante setup immediately before the standard is specified. ⁽⁷⁾Thus far, the discussion has focused on one patented technology, but there is also the consideration of complementary technologies, which is discussed further below.

If the hypothetical negotiation is modeled as an auction among rival technologies for inclusion in the standard, then WTP may effectively pinpoint the price rather than set the ceiling of a range.

How determinative should the model be in real cases?

The framework shows a close link between outside options, the bargaining range described by parties’ willingness to pay and to accept, and the agreed price.

Like all models, that relationship is a deliberate simplification. It is designed to isolate how lock-in changes pricing negotiations relative to a world in which implementers could still choose among alternatives. It is not intended to be a literal reconstruction of all ex ante bargaining dynamics. ⁽⁸⁾The starting point for hypothetical bilateral negotiation is Nash (1953). Here, the approach does not model the firms’ literal behaviors but finds the solution assuming certain axioms – of which, the main ones are: rational players, binding agreements, Pareto efficiency, independence of irrelevant alternatives and availability of complete information. Sutton (1986) provides an overview of the bargaining literature, including: (i) Rubinstein model, (ii) some alternative games with outside options, (iii) the role of incomplete information and (iv) negotiations with more than two players. Nash, J. (1953). Two-person cooperative games. Econometrica: Journal of the Econometric Society, 21(1), 128–140; Sutton, J. (1986). Non-cooperative bargaining theory: An introduction. The Review of Economic Studies, 53(5), 709–724.

In real negotiations, additional factors can shift outcomes relative to the model’s central prediction: asymmetric information, portfolio effects, risk allocation, repeat interactions and institutional features of standard setting. For example, where there is asymmetric information about the value of each other’s outside options, there may not be a direct causal link between the actual value of outside options and the price that the parties agree; it would need to be demonstrated. ⁽⁹⁾For instance, in Nash (1953), an outside option is just an exogenous “threat” that is relevant even if it is worse than the expected outcome. This simplification is unintuitive. Sutton (1986) notes “only threats which are credible will have an effect on outcomes.” Nash, J. (1953). Two-person cooperative games. Econometrica: Journal of the Econometric Society, 21(1), 128–140; Sutton, J. (1986). Non-cooperative bargaining theory: An introduction. The Review of Economic Studies, 53(5), 709–724.

In some settings, a party may be able to make a credible take-it-or-leave-it offer, in which case changes in outside options may not translate into proportional price changes. For these reasons, the ex ante bargaining range is often best treated as an organizing framework and a constraint on plausibility, rather than a mechanically determinative rule.

4.3 Step 2: An empirical assessment of technical contributions and their economic value

A bottom-up valuation requires information on the value of both the patent technology and its next-best ex ante alternative to calibrate the hypothetical negotiation. In practice, this involves an empirical assessment, with two component steps.

• Technical assessment. This step identifies the incremental technical contribution; that is, what the patented solution contributes relative to the alternatives that could realistically have been adopted at the time of standard setting.

• Economic assessment. This step seeks to value that incremental technical contribution, which can then inform both the implementer’s WTP and a plausible royalty range.

4.3.1 Assessing the technical contribution of patented technology

The objective of this step is to assess the patented solution’s incremental technical performance relative to contemporaneous alternatives – the options that the SDO considered, or could reasonably have considered, for the same part of the standard.

The analysis is typically led by technical experts, drawing on the evidence submitted during the standard-development process, which typically scrutinizes the relative technical merits of rival proposals in expert working groups. ⁽¹⁰⁾See the discussion of the ETSI/3GPP process in Section 1.2 above. See Gupta, K. (2017). How SSOs work: Unpacking the mobile industry’s 3GPP standards . SSRN. The aim is to identify what area of the standard’s functionality the patented solution improves – for example, in wireless standards, a particular technical solution may contribute to the performance of data throughput, reliability, latency, power consumption or implementation costs – and how that specific functionality compares to the contribution that rival proposals, if any, could have contributed in their place. The next-best alternative proposal provides the outside option in the hypothetical negotiation. The assessment should also note that some improvements may involve trade-offs – for instance, higher peak speeds may come at the expense of greater power consumption.

A practical limitation in such an assessment is scale. In many standards, an SEP holder may declare hundreds or thousands of SEPs. In principle, the bottom-up approach would value them all individually. In practice, that patent-by-patent assessment may not be feasible. ⁽¹¹⁾For example, in TCL v. Ericsson, Ericsson owned over 200 SEP families essential to the 2G, 3G and 4G cellular standards. The court set portfolio FRAND rates based on top-down and comparable license analyses rather than patent-by-patent assessments. TCL Communication Technology Holdings Ltd. v. Telefonaktiebolaget LM Ericsson, No. 8:14-cv-00341, (C.D. Cal. Dec. 21, 2017) (link) , amended and superseded (C.D. Cal. Sept. 14, 2018) pp. 33, 98 –104, rev’d in part, vacated in part 943 F.3d 1360 (Fed. Cir. 2019) , cert. denied 141 S.Ct. 239 (2020). In addition, the technical contribution of a single patent can be difficult to identify in isolation, because it may operate in conjunction with related patents. For these reasons, courts and parties often conduct the analysis at the portfolio level, covering all related patents for a specific innovation area rather than individual patents.

Box 4.7 gives two simplified examples of technical contribution assessments for hypothetical SEP portfolios relating to Wi-Fi and cellular standards.

4.3.2 Assessing the value of the technological contribution

This second step is the economic component of the assessment. It seeks to translate the incremental technical improvements into some measure or indicator of the incremental value – for example, the financial benefits to the implementer that the incremental technical contribution provides through greater sales volumes, higher prices or lower costs.

The valuation of the incremental technical contribution should be quantitative – relying on, for example, demand estimations, conjoint studies, consumer surveys, cost models or engineering studies. Where no such evidence is available, conclusions may be qualitative – for example, whether there is evidence that a technological contribution enabled a meaningful capability or use-case change, or if it was a minor incremental refinement – but the results of the valuation may only be indicative.

The relationship between technical improvements and their value may be non-linear. There are several reasons for this, in particular the following two.

• Improvements in use-case performance that have diminishing marginal utility. Once a use-case has been satisfied, additional performance may not add to consumer value to the same degree. There may be diminishing marginal utility. For example, moving from “minutes” to “seconds” to download a small file may be highly valuable initially, but further speed increases may matter far less for that particular use case. In principle, there may also be cases of increasing marginal utility.

• Changes in primary use-cases, increasing marginal utility. Performance improvements can unlock entirely new uses, which result in step changes in the value of a technical improvement, rather than smooth linear gains. For instance, the increase in cellular wireless data speeds and usage over the past 10 years has not simply resulted in people performing the same tasks they performed 10 years ago, but faster. It has changed what they use their phones and that data for. The majority of cellular data usage is consumed by high-definition video, which was not possible using previous generations of cellular technology. Enabling a new use-case may result in a step change in the incremental value that an improvement in technological functionality provides.

In practice, the valuation typically takes one of two forms.

• Direct valuation of the technical contribution. In the experience of the contributors to this Report, this is the preferred approach of competition authorities and US courts. The techniques applied use the method referred to above. Unfortunately, many of the decisions are not publicly available, and experts’ reports are not typically publicly disclosed.

• Relative valuation. This second application assesses how technical contributions protected by a particular portfolio of SEPs compare to those of SEPs that are essential to the relevant standard on average. The core idea is that when comparing patent portfolios to benchmark their values, patented technologies with perfect substitutes would receive a lower royalty than those with imperfect substitutes or no good substitutes. This variation in patent values is also consistent with economic studies. ⁽¹²⁾Vigil, R. and X. Zhang (2022). Determining the price for one when all you have is the price for many . Analysis Group Forum 2022.

The second approach was adopted by the court in Innovatio, for example, when it conducted an analysis of “The Importance of Innovatio’s Patents to the 802.11 Standard,” including “Alternative Technologies that Could Have Been Adopted Into the Standard” and the specific contributions of Innovatio’s related SEPs.  ⁽¹³⁾ In re Innovatio IP Ventures, LLC, Patent Litigation, 921 F. Supp. 2d 903 (N.D. Ill. 2013) (link) , pp. 34-49, specifically, for the quoted text, pp. 34 and 36. The court then took these bottom-up considerations into account when deciding where to place Innovatio’s patents in the distribution of all patents.

4.3.3 Practical considerations

A bottom-up valuation raises practical challenges that can materially affect results. Three issues are particularly important for applying the approach reliably:

• granularity: whether to assess individual patented technologies or to value coherent patent groups;

• complementarity: how to treat technologies whose value depends on other technologies, either within the standard or outside the standard but implemented in the product; and

• aggregation of implementers: how to treat differences in willingness to pay across products, markets and implementers.

4.3.3.1 Considering granularity

Individual SEPs often protect narrow contributions within complex standards comprising many thousands of specifications. This creates two opposing issues.

The first issue is that this can discourage parties and courts from examining technical contributions at all. This lack of examination, however, may raise a backward induction problem. In licensing negotiations, the parties will work backward from the end of litigation; if courts only ever evaluate whether SEPs are valid and essential, and not their technical and economic contributions, then negotiations will not focus on these issues. In turn, licenses signed under these conditions will also not reflect the economic value, but other considerations.

The second issue is the aggregation (or heap) fallacy – the observation that although each patented contribution may be small enough that its contribution seems insignificant in isolation, that does not entail that the aggregation of those parts is also insignificant. Each minor contribution may appear irrelevant to the overall functionality and value of the standard, and yet the value of those contributions in aggregate may be substantial.

In practice, this approach seeks to assess the value of patents at the most granular level that is feasible and efficient. Valuing at the level of individual patents is conceptually closest to the bottom-up principle, because it attempts to measure the contributions of the individual patent rights actually being licensed. This approach can be informative when the number of relevant patents is limited, when each patented contribution can be clearly identified and accurately measured, and when the evidentiary record supports a meaningful comparison with alternatives. In some cases, while it may be difficult to assess all or even a large number of patents, it may be possible to evaluate a subset of patented technologies to assess whether the licensor has made significant contributions to the standard.

However, the valuation often requires greater pragmatism: for example, grouping patents into manageable groups, such that economic value to users of the incremental technical contribution can be meaningfully assessed. In many technical standards, patent-by-patent assessment is not feasible and can create false precision, especially where individual patents cannot be understood independently of related filings. A common and sensible alternative may be to assess coherent groups of patents linked to discrete, consumer-relevant features. This preserves the spirit of a bottom-up assessment while reducing cost and improving technical interpretability. Courts and experts have sometimes taken this approach in the case of complex standards. ⁽¹⁴⁾For instance, In re Innovatio IP Ventures, LLC, Patent Litigation, 921 F. Supp. 2d 903 (N.D. Ill. 2013) (link) , pp. 40–58 and 84–86. First, Judge Holderman assessed patents in three related groups: The Channel Sharing Patent Family; Multi-Transceiver Family; and Sleep Family. Second, Judge Holderman determined that Innovatio’s patents were among the top 10 percent of Wi-Fi SEPs and, using research submitted by the manufacturers that found that the top 10 percent of the patents accounted for the top 84 percent of the value, he attributed 84 percent of the profit margin on the chip to the top 10 percent of patents. He then calculated the royalty due to Innovatio based on its numerical share of the top 10 percent of Wi-Fi SEPs. Also, on the pragmatic need for some degree of aggregation when assessing SEP portfolios, see Unwired Planet v. Huawei [2017] EWHC 711 (link) (Pat), para. 182: “There was ample evidence before me that apart from Ericsson (see below), parties negotiating SEP licences in fact use methods which are based on patent counting. That is evidence which supports a finding that a FRAND approach to assessing a royalty rate is to engage in some kind of patent counting. Indeed when one thinks about it some sort of patent counting is the only practical approach at least for a portfolio of any size. Trying to evaluate the importance of individual inventions becomes disproportionate very quickly.”

4.3.3.2 Considering complementarity

Many patented contributions to a standard are complements: they work together, so that the value attributable to one depends partly on the presence of others. If each patent is valued in isolation – “holding everything else fixed” – the sum of the individually assessed contributions can exceed the value generated by the technology collectively. This happens if each patent receives the full marginal “technical contribution” which is created jointly by the standard. ⁽¹⁵⁾For a full discussion on applying the ex ante incremental value approach to complementary SEPs, see Layne-Farrar, A., A. Jorge Padilla and R. Schmalensee (2007). Pricing patents for licensing in standard-setting organizations: Making sense of FRAND commitments. Antitrust Law Journal, 74(3), 671–706.

For example, in a cellular standard, patents that improve power efficiency and patents that increase data throughput both increase the value to consumers of an implementing product – but the value of each depends partly on the other. Higher data speeds raise power demand, which increases the benefit of superior power-saving technology; conversely, better power management makes high-speed use-cases more attractive. Valuing each separately without accounting for the joint (overlapping) value can therefore overstate their combined value.

Complementarity can be analyzed, and there are theoretical methods suggesting how to allocate joint value – for example, the “Shapley value” and related cooperative-game approaches. ⁽¹⁶⁾Shapley, L.S. (1953). A value for n-person games. In Kuhn, H.W. and A.W. Tucker (eds), Contributions to the Theory of Games (Vol. II), Annals of Mathematics Studies, 28. Princeton University Press, 307–317.

In practice, however, addressing complementarity pushes the analysis toward a top-down approach (see Part 5), rather than a pure bottom-up one. A middle ground may suffice, that avoids excessive aggregation, but which permits grouping of related and complementary patents where separate valuation would materially distort the results.

Box 4.8 illustrates the potential challenges in separately assessing the contribution of complementary patented technologies.

4.3.3.3 Considering aggregation of implementers

Conceptually, an implementer’s WTP for a patented feature will reflect how that feature affects its own profits – by increasing demand, enabling higher prices and/or reducing costs. In practice, therefore, WTP will not be uniform across implementers. Those that stand to gain greater benefit will be willing to pay more than those that stand to earn less.

In practice, the benefits may vary across:

• product categories: such as the value that Wi-Fi technology provides devices with high-resolution screens and IoT devices with different data needs;

• market segments: such as the value that a particular technology provides to a device maker with premium features versus a competitor with budget features; and

• implementers’ business models: depending on their position in the value chain, business strategy and portfolio of complementary products.

This raises a practical question when establishing the value of a technical contribution. To some extent, its value is in the eye of the beholder. If the hypothetical negotiation is framed as a bilateral negotiation between one licensor and a specific implementer, then technically the value is defined solely by that implementer’s WTP. In contrast, the negotiation may treat implementers collectively – sometimes implicitly as all “similarly situated” implementers – as though they act as a collective single buyer in an auction or tender framework.

Where the negotiation concerns a specific implementer and there is reliable evidence on the value of the technology in that implementer’s products, in principle it may be preferable to estimate that specific implementer’s WTP. This aligns the valuation with the economic value actually realized in the relevant products and markets. However, in practice that can be challenging, as it is subject to uncertainty and could make the assessment of non-discrimination within a product market difficult.

Where implementer-specific WTP cannot be reliably established (particularly for an individual patent or a small subset of a portfolio), aggregation can be a practical simplification. However, aggregation should be treated with caution. Averaging – lumping together implementers or product markets with materially different WTPs – can produce an “average” that is too high for low-value applications and too low for high-value ones.

Pragmatically, it is often preferable to analyze implementers in coherent groups of similarly situated parties, where their WTPs should be broadly similar, based on the specific benefit that the technical contribution adds to the implementers’ products. Where the products compete directly with each other, it is more likely that the value which the technical contribution adds to each one is broadly similar, and their manufacturers would be willing to pay similar rates for those benefits. Where products do not directly compete with one another, the technical functionalities might not provide the same value to each, and so the royalties for each product may be better assessed separately from each other.

4.4 Summary

The bottom-up framework offers a structured way to estimate FRAND terms by anchoring valuation in the ex ante incremental value benchmark: the royalty (or range of royalties) that the parties would have regarded as reasonable before the standard was set, when implementers could still choose between competing technical alternatives.

The core intuition is that ex ante competition constrains royalties to reflect the patented technology’s incremental contribution over the next-best alternative, rather than the value that arises solely from ex post lock-in to the standard and exclusion of competition.

As this part has shown, the approach may be appealing in principle insofar as it seeks to reward the patented technological contribution while excluding additional value arising solely from the inclusion of that technology in the standard. However, it depends heavily on modeling choices, evidence and practical judgments.

The hypothetical negotiation must be specified in a way that is economically plausible, including careful consideration of the timing of the negotiation, what information about value may be used (ex ante expectations versus ex post evidence), and whether the model yields a bargaining range rather than a single price estimate.

Calibrating the model requires an empirical assessment of:

• the patented technology’s incremental technical contribution relative to contemporaneous alternatives considered (or reasonably considered) during standard setting; and

• the economic value of that incremental technical contribution – recognizing that technical improvements and economic value may be nonlinear and may need to be assessed quantitatively or, where data is limited, qualitatively.

The principles of the bottom-up approach or ex ante incremental value benchmark may be used in combination with other valuation approaches. First, to benchmark the potential value of one SEP portfolio against the value of a license for a potentially comparable SEP portfolio. Second, in combination with the top-down framework, where the complementarities between patented technologies mean that they need to be valued at a higher collective level of aggregation, to avoid the risk of royalty stacking.

Part 5 explores top-down valuation methodologies.