Innovation and Dynamic Efficiencies in Merger Review


April 9, 2007

Prepared By:
Andrew Tepperman and Margaret Sanderson
CRA International

Date: April 9, 2007
CRA Project No. D09208‑00

This report provides the views of the authors.  It does not necessarily reflect the views or opinions of other staff or consultants with CRA International.  As a result, CRA International makes no representation or warranty as to the accuracy or completeness of the material contained in this document and shall have, and accepts, no liability for any statements, opinions, information or matters (expressed or implied) arising out of, contained in or derived from this document or any omissions from this document, or any other written or oral communication transmitted or made available to any other party in relation to the subject matter of this document.

Table of contents

  1. Introduction
  2. Innovation and dynamic competition
  3. Innovation in a competitive effects analysis: Threshold issues
  4. Evaluating innovation concerns in merger review
  5. Dynamic efficiency considerations
  6. Conclusion

Executive summary

This report addresses issues that arise when incorporating innovation issues into the merger review process.  Competition based on innovation, whereby firms attempt to gain market share through the introduction of new or improved products or services‑dynamic competition‑is at the heart of many modern industries.  Accordingly, an understanding of dynamic competition is relevant for merger review.  When incorporating innovation issues into merger review several considerations arise.  First, there is no settled economic model that relates the extent of market concentration to the extent of innovation and as a result, we do not know how concentration today affects firms’ levels of innovative activity, which differs from the clear link that exists between concentration and pricing.  Second, innovation is highly uncertain, making it much more difficult to measure and quantify than price and output.  Third, these measurement problems make it difficult to quantify a merger’s likely effect on the rate or outcome of innovation.  Finally, innovative activity is a form of up‑front investment, and prices must be expected to exceed short‑run marginal cost to justify the investment, on average.  As a result, static measures of economic efficiency that fail to account for the flow of surplus from the introduction of new products cannot tell the whole story when innovation is important to the merger review.  An appropriate treatment of efficiency must recognize these dynamic gains, without ignoring the importance of competitive rivalry among firms within markets at a point in time.

With these considerations in mind, in this report we propose a framework that allows for the effect of merger transactions on innovation to be incorporated into merger review, where the current approach found in the Merger Enforcement Guidelines is not sufficient to fully capture dynamic competition.  The current approach found in the Guidelines works well for addressing when mergers are likely to lead to a reduction in either actual or potential competition in an existing goods market.  The framework we propose is aimed at addressing future goods markets, while making use of information available today.  It should be of assistance to competition authorities faced with mergers that may raise competition issues in future goods markets, or that may involve future innovations which obviate any competition concerns in existing goods markets.

The framework proceeds by making the following inquiries:

  1. Is innovation important in the industry in question?
  2. Can affected future products and firms be identified?
  3. Would the merging firms compete against each other in those future markets but for the merger?
  4. Would the merger reduce the existing level of innovation?
  5. Would the merger result in an increase in prices in the future market above what they would be without the merger?

In light of the special considerations associated with innovation, application of this framework is likely to be highly fact specific. A number of factors that can be taken into account under each of these five inquiries are discussed.

The merger review process also calls for gains resulting from improved innovative conditions to be considered as a potential offset to static efficiency losses resulting from price and output changes.  Such dynamic efficiency considerations may be difficult for parties to prove, but could be expected to arise from a number of plausible sources.  In particular, merging firms may be able to eliminate duplicative research and development programs, or may be able to realize economies of scale or scope in research and development.  Again, a dynamic efficiency analysis must be case‑specific in nature and would require a careful consideration of factual material.

1. Introduction

We have been retained by the Competition Bureau (“Bureau”) to consider how the Bureau might assess innovation and dynamic efficiency when undertaking merger reviews.  While the nature and extent of change and innovation in a relevant market is an identified factor in the Competition Act to be considered in order to determine if a merger is likely to substantially prevent or lessen competition,Footnote 1 to our knowledge it has not been addressed in a broad Bureau policy report.  In particular, a practical framework to address issues of innovation and dynamic efficiency in merger review has not been previously advanced.  We hope this report will fill that gap.

Having said this, we do not wish to imply that innovation is a foreign concept in current merger review.  The 2004 Merger Enforcement Guidelines (“MEGs”) discuss change and innovation in the context of determining when mergers may have anti‑competitive consequences and when discussing the types of efficiencies that the Bureau will consider under section 96.  The MEGs indicate the Bureau will consider the competitive impact of technological developments in products and processes arising from many areas of the merging parties’ operations.Footnote 2  Change and innovation may reduce barriers to entry, for example, thereby making it difficult for the merging firms to sustain a material price increase post‑merger.Footnote 3  Alternatively, a merger may facilitate the exercise of market power if it eliminates an innovative firm from the market that posed a serious threat to incumbent firms.  In such circumstances the merger may substantially lessen or prevent competition by delaying the introduction of new products or processes.Footnote 4

Gains in dynamic efficiency arising from the introduction of new or improved products and processes are also recognized by the Bureau as important considerations in merger review.Footnote 5  Such gains are described in the MEGs as “crucial to both the general evolution of competition and the international competitiveness of Canadian industries.”Footnote 6  They are also regarded as inherently difficult to measure or substantiate.  Improved dynamic efficiency as a public policy goal is sufficiently important in the context of merger review that it warranted a chapter of discussion in the Report of the Advisory Panel on Efficiencies.Footnote 7  Nonetheless the Advisory Panel Report did not provide (nor was it asked to provide) practical advice on how to consider dynamic efficiencies in merger review.

In this report, we explore these and related issues.  We accept as a general proposition that Canadian competition law is flexible enough to allow for a rigorous analysis of innovation and its potential competitive consequences in merger review.  Indeed, competition in a market with differentiated products will often take the form of product innovation as well as price and cost reduction.  In such markets, a successful product innovation by one firm will stimulate imitation of the innovation by others, and may spur competitors to conduct innovation aimed at leapfrogging one another’s technological advancements.  As a result, innovative rivalry can be viewed as one form of competition that the Competition Act seeks to protect by preventing mergers that substantially lessen or prevent competition along this dimension.  Similarly, the long‑term gains that accrue to consumers (and firms) as a result of innovative behaviour can be encompassed within the efficiencies provisions of Section 96 of the Act.

Rivalry in the form of innovation can be distinguished from our usual focus in merger review on pricing in the following ways:

  1. Innovation involves future products or services making it intrinsically uncertain;
  2. Innovation is a form of investment, requiring (at least) a normal rate of return on average, which is typically obtained through above‑cost pricing in the short term, leading to a potential tension between short‑term and long‑term efficiency concepts;
  3. There is no agreed‑upon economic theory of innovation‑based competition, and as a result a detailed analysis of the facts surrounding each potential transaction is critical; and,
  4. Measurement issues make it difficult to say what constitutes “more” or “less” innovation in relation to a pre‑merger benchmark.

In light of these special characteristics associated with innovation, we do not presume that the merger review process described in the MEGs can necessarily be applied without modification to cases where innovation concerns are paramount.  While the framework embodied in the MEGs can be successfully applied to most mergers, it may not provide the correct conclusion in situations where the merger involves two firms that do not currently compete, but that would be significant rivals in respect of a future goods market (which may include existing goods), or if the merger raises substantial competition concerns but future innovation is likely to reduce these concerns.  Thus, we depart from the MEGs on certain specific issues, as we will describe in more detail below.

Prior to setting out the framework for analyzing innovation in merger review, it is useful to define key terms.  We do this in the next section (Section 2), following which we elaborate on some of the inherent properties of innovation‑based competition in Section 3.  In Section 4, we outline a pragmatic analytical framework that we believe allows for due consideration to be given to innovation and dynamic efficiencies in considering a merger’s potential competitive effects, while not ignoring the traditional merger concerns related to price and output.  This is followed in Section 5 with a discussion of how merger review should address efficiency claims in respect of innovation.  Concluding remarks are found in Section 6.

2. Innovation and dynamic competition

2.1. Key concepts

We begin by defining several of the terms we use extensively throughout this report.  Innovation can be thought of as the creation and development of new or improved products or processes.  A product innovation is a new or improved product (or service); a process innovation is a means by which existing production technology is made more efficient, with the consequence that costs of producing the product are decreased.  Quality improvements in existing products or the development of new features for existing products are captured in our notion of product innovation.  New or improved processes need not relate solely to production, but may also include cost‑saving means of distribution or marketing.

Innovation is inherently dynamic, which distinguishes it from more static price and output decisions.Footnote 8  This is so because it is generally necessary to incur costs at some point in time (or over some extended interval of time) in order to enable an innovation to be realized later.  In this sense innovative activity—research and development (“R&D”)—is a form of sunk and risky investment, in which a cost (in dollars or hours of effort) precedes a potentially uncertain outcome.  Both product and process innovations contribute to economic efficiency.  To see this, consider only the outcome of innovation, ignoring the costs of research and development, which are assumed to have taken place in the past.  Product innovations contribute to economic efficiency by increasing consumer surplus—the total amount by which consumers’ willingness to pay exceeds the price of the product for those consumers purchasing the product.  For existing products, an innovation leading to a product improvement increases the value that consumers place on using the product, assuming price is held constant.  For example, the introduction of the minivan generated large gains for consumers who valued its attributes as compared to existing cars.Footnote 9  In the case of innovations that result in entirely new products, such as the introduction of the cellular telephone, consumer surplus increases from a level of zero pre‑innovation to some positive and potentially substantial amount post‑innovation.Footnote 10  The potential gains to consumers from the introduction of new goods are typically far larger than those available from optimizing the pricing of existing goods.Footnote 11

Process innovations also promote economic efficiency by allowing existing products to be produced at lower costs.  This may not immediately translate into lower prices (and thus increased consumer surplus).  The innovating firm may instead hold a competitive advantage over its rivals for which it earns additional profits compared to the pre‑innovation period.  As a result producer surplus is increased and efficiency improved even in the short term.  Then, as rival firms seek to imitate the innovation or develop their own cost‑saving techniques, the cost‑reducing technology is diffused across firms in the market leading to lower prices than would have existed without the process improvement over time.Footnote 12  Consumer surplus may then be increased.

When considering competition through innovation, it is important not to confuse the improvement in earnings by the innovating firm with monopoly profits.  Instead, the firm is earning a return on its investments, which economists refer to as a quasi‑rent.Footnote 13  The existence of quasi‑rents is not analogous to market power, yet the two are often confused, particularly if analysts look to firm profits at a single point in time as a means of measuring the extent of market power.

Having established that innovations, once realized, lead to gains for consumers and producers, and that the act of innovation is not itself typically costless, it is clear that society overall is faced with a tradeoff.  A greater expenditure devoted to R&D can be expected, on average, to result in a greater number of innovations being realized.  Some positive level of expenditure on R&D would then seem to be desirable from a social perspective.  At the same time, there will eventually be diminishing returns to R&D expenditure.  The optimal level of innovative effort is that which economists call the dynamically efficient level.  Dynamic efficiency implies that the flow of surplus realized through the introduction of new products or processes over time, net of the cost of researching and developing these new products and processes, is at a maximum.

2.2. Dynamic competition

Competition based on the successive introduction of new or better products over time is called dynamic competition.  Dynamic competition based on investment in R&D may be thought of as a form of “competition for the market”Footnote 14 in contrast to price competition which is “competition in the market.”  This characterization is overly simplistic, however.  There are certainly many situations in which both forms of competition operate—firms may compete for customers’ business by reducing price and improving quality for existing goods, and by pursuing innovation in an effort to introduce new goods to market.Footnote 15  Nonetheless, this way of dichotomizing competitive rivalry serves to emphasize an important contrast.  Static views of competition take the existing set of products and market participants as given, describing the outcome of competitive behaviour among those market participants using strategic instruments such as pricing or advertising that can be applied and varied in the “short term.”  Dynamic competition involves the creation of new products and potentially also new markets, along with the replacement or obsolescence of older products.  It also implicitly or explicitly involves entry and exit by firms—there is no guarantee that today’s successful firms will be able to offer the product attributes demanded by tomorrow’s consumers.

Anecdotal examples of dynamic competition are not hard to come by.  Consider an example from the pharmaceutical area.  SmithKline’s breakthrough anti‑ulcer medication Tagamet was largely free from direct competitors in the late 1970s.  Newer products with desirable attributes soon arrived on the market, beginning with Glaxo’s Zantac in the early 1980s, and Tagamet’s share dropped off as several competing firms offering different chemical entities to treat the same set of symptoms emerged.Footnote 16  In the microprocessor industry, Intel continues to undertake a tremendous amount of innovative effort in order to stay ahead of its rivals, introducing generation after generation of chip since its first products were launched in the 1970s.  Only recently have other microprocessor firms such as Advanced Micro Devices been able to make inroads into Intel’s share of the chip market by offering high‑quality and innovative products with characteristics desired by consumers.Footnote 17

As manufacturers bring new technical features forward, they can create demand for their particular product from both new and established customers.  Consider for example the increased demand generated over time for computing processing power.  Throughout the 1990s, IBM made a series of investments in its mainframes that significantly enhanced mainframe functionality.  A fundamental innovation during this time was the introduction of technology to alleviate the need for large water‑cooling systems that prior processor technology required, thereby allowing mainframes to be much smaller in physical size and dramatically lowering their cost.  Competing mainframe manufacturers Amdahl and Hitachi had difficulty replicating the significant technology advances of IBM and ceased producing mainframes.  Yet IBM’s mainframes still faced competition from an alternative technology known as Unix.  Sun Microsystems and Hewlett‑Packard in particular advanced Unix technology to the point that by 2001 Unix‑based server systems had mainframe‑like power and comparable performance characteristics.  Technology has since advanced further still, allowing firms to link a number of Intel‑based servers together to have the computing processing power and performance features of Unix supercomputers.  With each new development, IBM invests more in its mainframe technology to improve upon it, while Sun and Hewlett‑Packard invest further in Unix technology.  Intel and Microsoft, meanwhile, invest in their technologies.  New technologies appear, disrupting old technologies and providing the motivation for incumbent firms to continue investing and developing new products.

2.3. Basic principles

From this high‑level overview of the relevant concepts, we present two principles to inform the remaining discussion.

2.3.1. Tension may exist between static and dynamic efficiency

To sustain innovative efforts, and thus support dynamic efficiency, firms do not expect to price at short‑run marginal cost at every point in time and as a result some degree of allocative inefficiency may be inevitable.  Motivating firms to make costly investments in R&D requires some prospect of “profit,” which as noted above is in the form of quasi‑rents.  In the absence of this positive return per unit of output sold, a firm would never be able to recoup its up‑front investment in R&D, and would therefore have no incentive to undertake this investment.  In other words, innovating firms anticipate a period of “incumbency” during which they are able to sell a product at a price exceeding not only the short‑run marginal cost of production, but potentially also the price of existing products (if any) that do not incorporate the innovation.  Consumers are willing to pay the higher price because they value the additional attributes embodied in the new or improved product sufficiently to pay a premium for it over other firms’ products.

An implication is that textbook “perfect competition” may be inconsistent with dynamic efficiency, a point recognized by Joseph Schumpeter decades ago, and developed in subsequent macroeconomic and microeconomic literature.Footnote 18  Once we realize that the basic assumptions necessary for perfect competition to hold—including, inter alia, homogeneous products, infinitesimally small firms, perfect information, and small fixed costs per firm—cannot possibly apply in the vast majority of real world markets, it is clear that competition is often “imperfect,” and prices are expected to exceed short‑run marginal cost.  As a result, a price above marginal cost does not suggest market power exists, in and of itself.Footnote 19

The existence of above‑marginal cost pricing in any particular market may be explained by a wide range of factors.  One common explanation is that products may be differentiated from the perspective of consumers (for example, breakfast cereal), which tends to reduce the intensity of price competition.  Alternatively, production may require firms to incur substantial and ongoing fixed costs in order to set up or to stay in business (as is the case with infrastructure industries such as transportation or telecommunications service provision).  From a purely static perspective, products with similar (but not identical) attributes may coexist, exerting competitive pressure on the innovator’s pricing decision.  From a dynamic perspective, any discussion of whether returns are competitive or supra‑competitive must confront the problem of how to define market power in the context of a dynamically competitive market; we defer discussion of that issue to a later section.

Incumbency status enabling above‑marginal cost pricing must be expected to persist for long enough to make the investment worthwhile.  It is a firm’s expectation that quasi‑rents will exist that drives the investment decision.  Should a different state of the world materialize in the future such that ex post returns on the product are insufficient to justify the investment, this will not alter the product’s current existence.  Instead, it would be expected to factor into the firm’s next investment decision, or those of similar firms.  Legal instruments such as patents may serve to protect a firm’s investment to a certain extent.Footnote 20

In many instances, the process of dynamic competition prevents any one firm from enjoying a position of leadership for very long, unless that firm continues to innovate.  “Follower” firms will perceive the benefits of incumbency status and will compete by engaging in R&D, or by imitating.  This process normally imposes limits on the length of time any given incumbent can expect to remain ahead of its rivals before its existing product is replaced in a new round of innovation.  Recently, economists studying the determinants of economic growth have sought to understand and fully model these forces.Footnote 21

2.3.2. More R&D is good, but more innovation is better

The private incentives to innovate that firms face may not give rise to the optimal level of innovation from a social perspective.  As discussed above, the socially optimal level of innovation is that which maximizes dynamic efficiency, or the flow of surplus net of R&D expenditure.  The incentives faced by private firms, on the other hand, are determined by a number of different factors, which can give rise to either too much or too little innovation.Footnote 22  Two effects in particular tend to promote underinvestment in innovation because: (i) innovating firms fail to capture the benefits that their innovations provide to future researchers by moving the technological frontier ahead they invest less than the socially optimal level;Footnote 23 and (ii) firms cannot typically receive all of the increase in consumer surplus that is created by a successful new product.Footnote 24  Working in the opposite direction toward overinvestment in R&D is the “business stealing effect.”  This effect is driven by firms failing to take into account the fact that while each firm benefits when it replaces another as incumbent, no such gain accrues to society; so firms may have an excessive incentive (relative to that which is socially optimal) to seek to replace other firms.

On balance, the evidence suggests that firms face insufficient incentives to innovate relative to socially optimal levels.  Numerous studies have attempted to empirically measure the spillover benefits generated by firms’ R&D efforts in order to test this theory.Footnote 25  These spillovers seem to be large, and as a result the social rate of return from R&D exceeds the private rate of return.  The implication is that underinvestment in R&D is real and significant.Footnote 26

This is not to suggest that a discussion of innovation should begin and end with R&D.  The considerable attention R&D expenditure receives in practice is warranted to a certain extent because worthwhile innovations do not typically spring fully‑formed from inventors’ minds, and thus R&D is a necessary condition for innovation.  In addition, R&D is easy to measure, so it is used extensively in empirical studies; for example, public firms report R&D spending in their financial statements, and most empirical studies of innovation rely on publicly available data.  Yet too close a focus on R&D risks obscuring the important fact that it is not R&D per se that is of interest—it is innovative output.  Consumers do not derive benefits from an additional dollar of R&D spending unless that dollar results in an increased likelihood of either a new product being developed or an existing product being made available for a lower price.  As a result, the ultimate focus of any investigation into the impact on innovation of a particular transaction must be on the output of the innovative process.  Transactions that would reduce total R&D expenditures but leave the level of innovation constant (perhaps because of the existence of duplicative R&D efforts across firms) should not be subject to challenge.

3. Innovation in a competitive effects analysis: Threshold issues

To usefully incorporate innovation and dynamic efficiency into merger review we need to recognize several distinguishing features of innovative competition.  First, there is no robust model of innovation that links the rate of innovation to industry concentration.  Second, there is far more uncertainty surrounding innovation than is normally encountered in a typical merger review focused solely on price and output.  Third, associated with innovation are complicating measurement problems that do not ordinarily exist with more conventional merger analysis.

These issues do not prevent us from taking innovation and dynamic efficiency into account when assessing mergers.  But they do mean that we cannot rely on simple measures to quantify levels of innovative activity or to confidently predict how mergers will impact on innovative activity.  Instead, as we discuss in the subsequent section, a more fact‑intensive, case‑by‑case approach is required.

3.1. No predictive model relating innovation to concentration

When considering the potential price effects of a merger in the short run, economists typically make reference to a number of accepted models of competitive firm interaction.Footnote 27  Such models provide guidance on the relationship between firm concentration and price, once other variables such as the demand elasticity and ease of entry are defined.  We would like to have a similar model that relates innovation to industry concentration.  In particular, we would like to know whether the rate at which firms pursue a particular innovation (and accordingly their likelihood of success) increases with the number of firms involved in the pursuit.Footnote 28  Unfortunately, a robust framework that predicts the amount by which innovation will increase for a given increase in the number of firms has proven elusive.

One model that is frequently invoked (to varying degrees of explicitness) to address this issue is the “patent race” model.  Firms in patent race models interact under the following specific set of assumptions: (i) all firms are assumed to be pursuing the same (known) prize; (ii) the number of firms is fixed at the outset; (iii) the time needed to innovate successfully is uncertain; (iv) only the first firm to attain the innovation will be able to profit from it; and (v) firms that spend more are more likely to innovate first.  This model leads to some interesting conclusions, a fundamental one being that R&D efforts are “strategic complements”: the optimal strategy in response to an increase in innovative effort by a rival is to increase one’s own R&D.Footnote 29  The rate of R&D spending by each firm, and thus the speed at which the innovation is realized, depend positively on the number of participants.  As an obvious application, a merger between two participants in the patent race will decelerate the pace of innovation and so there might be some justification for challenging the merger.

But patent race models, and the conclusions reached from these models, are highly dependent on the underlying assumptions.  Consider, for example, the repercussions of changing the “winner take all” assumption.  Instead, make the assumption that an R&D success by one firm leads to a phase of product market competition against other participating firms.  This might arise if success by one firm provides important benefits to other R&D competitors, such that they are then able to complete their own projects and sell products in competition with the successful innovator.Footnote 30  Models in which research projects are assumed to be substitutes and in which there can be only one “winner” will therefore describe this structure very poorly.Footnote 31  Here, it is theoretically possible for R&D effort to be negatively related to the number of competitors, so that a decrease in the number of firms involved in a race will tend to increase each remaining firm’s R&D expenditure.Footnote 32  A merger between R&D competitors would then have an ambiguous effect, as there would be fewer participants each performing more R&D.  This innocuous‑seeming change in assumptions completely eliminates the “classic” result.  Careful consideration of market conditions is generally necessary before the patent race model can be used.

Empirical studies of the broader relationship between concentration and innovation have also failed to arrive at robust results.  Cohen and Levin’s comprehensive review of the early empirical literature finds little compelling evidence of a systematic relationship between concentration and innovation in cross‑industry studies, with much of the variation in R&D intensity explained by industry‑specific effects and technological opportunities.Footnote 33  A more promising model has recently been advanced that relates innovation to competition in an “inverted U” pattern: innovation increases with competition up to some critical point after which the relationship is reversed.Footnote 34  Empirical research seems to bear this out.  If industrial sectors are ranked according to a measure of competitive intensity based on profit margins, an increase in competition seems to promote innovation in less competitive sectors and stifle it in sectors that are already highly competitive.Footnote 35

The results of economic research on competition, firm size and innovation generally are sufficiently mixed that the Advisory Panel on Efficiencies advised against either a deliberate strategy of increasing competition, or a deliberate strategy of encouraging market concentration as a means by which Canada’s innovative capacity could be predictably and reliably improved upon.Footnote 36

3.2. Uncertainty

In the typical merger analysis, we can observe what products the merging parties are selling and where they are selling them.  Customers’ choices are also observable, as is the existence of rival firms.Footnote 37  In many cases, potential competitors can also be identified by observing which firms have the necessary production capacity to sell the relevant product in the relevant geographic market.

Matters are different where mergers among innovating firms are concerned.  In the case of product innovations, we are dealing with products that likely do not yet exist.  Two subsidiary issues flow from this.

First, innovation is highly uncertain.  When firms set out to do R&D, they often do not know whether their investment will lead to a product that works from a technical perspective and that would be desired by consumers.Footnote 38  In many innovative industries, failure is more common than success for any particular inventive path.  As a result, firms will pursue numerous possibilities, particularly at the early stages of R&D in order to increase the probability that one of these will be successful.  This form of R&D is tremendously costly and exposes companies to a great deal of risk, since even if a product is approved for sale it may not ultimately be successful in the market.

Second, innovation takes time.  In some cases there may be years between the R&D stage and the point at which a product can be brought to market.  During this intervening time period, innovating firms receive information about the technical characteristics of an innovation and its likely market prospects, and make decisions based on this new information as it arises.  As a result, a product may be quite different by the time the end of the innovation process is reached than was envisioned earlier on.  The time it takes to complete R&D may also make it difficult to identify other firms that would be positioned to undertake similar types of innovation.  Suppose a firm estimates that an R&D project is likely to take five years to complete, and that at the end of that time period the firm expects to have a successful product.  It is conceivable that two years into this time period, as a result of general technological progress or specialized knowledge, another firm may be able to complete a similar project within three years, such that both resulting products arrive on the market simultaneously.  The entry of this second firm would have been completely unanticipated at the outset.  Carlton and Gertner argue that since many innovations arrive from unexpected sources, such situations are relatively common, and as a result it is difficult to anticipate the firms that will offer competition in the future to firms that are innovating today.Footnote 39

3.3. Measurement problems

Associated with uncertainty are measurement issues.  First, consider price in the context of innovation.  It would be desirable to know the price consumers would be willing to pay for an innovation, since this would be the first step toward being able to determine whether a proposed transaction would likely increase this price.  For innovations that have not yet been realized, price is unknown.  The value that consumers are expected to place on an innovative product, net of the value placed on other non‑innovative aspects of the product, is unobservable prior to the product appearing on the market.  Even once a product is sold it may be difficult to disentangle the innovation itself from other aspects of the product contributing to its value.  Economists have estimated what are referred to as “hedonic” pricing models in an effort to measure the effect of individual product characteristics on value.  These empirical models allow for the measurement of quality‑adjusted prices, facilitating a comparison across products that would be uninformative were only nominal prices to be used.  Where an innovation relates to one identifiable product characteristic, and where sufficient data exists to be able to isolate the effect of the innovation from everything else, it may be possible to estimate the price effectively paid by consumers for that innovation using these methods.Footnote 40  Data requirements normally make this very difficult.  In addition, except for perhaps minor incremental innovations, it will be risky to attempt to use a model based on existing data to predict the value placed by con sumers on future improvements.

This brings up a related complication.  Many products are made up of numerous components.  Each of these components may be a distinct product which the firm “bundles” with other components before making it available to consumers.  Consider as an example the personal computer, or PC.  The PC can be crudely described as a collection of components: microprocessor, memory, storage devices such as hard drives, interface devices, and so on.  Each one of these components is available on a standalone “unbundled” basis, the price of which can be observed on the market.  If an innovation improves one of these components, its value could (in theory) be inferred by observing the price of the component before and after the new technology was introduced.  Yet many innovations occur at much finer levels.  A microprocessor firm might develop a new means by which logical operations could be performed more efficiently on a chip.  This innovation might be incorporated into the next generation of chips sold on the market, along with numerous other improvements and established technologies.  In such cases, which are very common, the value of that one logic improvement cannot be isolated because it is not sold on its own, and many other factors contribute to chip performance and thus to the value placed on the product by consumers.

Similar reservations apply to measurement of “quantity” in the context of innovation.Footnote 41  In general, no completely adequate measures of quantity exist in the context of innovation, although the level of R&D expenditure, and indicators of innovation output such as patents, have sometimes been used. 
R&D spending is a useful indicator of a firm’s overall involvement in innovation at a point in time, but it is problematic to use to represent quantity on at least two counts.  First, it is a measure of input rather than output.  While greater R&D expenditure toward a particular end may represent a greater chance of eventual success, it is unknown precisely how this relationship works at the individual firm level.  If one firm spends twice as much on R&D as another in pursuit of an innovation, it is probably inappropriate to say that the first firm has twice the competitive significance as the second firm.  If we were to suggest the first firm is twice as big as the second based on this information we do not even know the extent to which we have over‑ or underestimated the firms’ relative position.  Second, being a measure of the flow of dollars, R&D spending may not provide a sufficiently precise indicator of innovative rivalry with respect to any one particular target.  R&D may be shared across numerous projects if a company’s scientists or equipment tend to be engaged in various different pursuits.  Economics does not help in determining how shared R&D should be allocated across various projects.

Patents have been suggested as a potential alternative to R&D as a measure of innovation quantity.  In contrast to R&D spending, patents have the benefit that they are a unit of R&D “output” that is correlated with success, at least as measured along some technical dimension.  Moreover, firms often patent relatively early in the R&D process, long before they are ready to sell related products, so patenting might be observed early enough for the information to be useful in an analysis of a merger between innovating firms.  Finally, enough information is contained in patent documents to be able to associate each patent with a particular R&D program, at least in principle.

Unfortunately, there are also serious problems with the use of patents as a measure of innovation quantity.  The main problem is that not all innovations are patented.  Patents appear to be much more useful as a means of protecting returns to innovation in some industries than in others.Footnote 42  In many industries, patents do not appear to be used for this purpose to any great degree.  Therefore, in industries like semiconductors, attempting to infer innovation quantity by observing the number of patents would be misleading.  Even in industries where firms actively use patents to protect their innovations, it is widely recognized that simply counting patents gives a relatively “noisy” indication of research output.Footnote 43  If patents are to be used, it is preferable to find some way of indexing them for their importance.  Weighting patents by the number of citations each has subsequently received from later patents has been used with some success.Footnote 44
We are therefore left with no single good measure of innovation quantity.  Instead, several imperfect indicators, such as R&D spending and patents, must typically be relied upon.

4. Evaluating innovation concerns in merger review

4.1. Motivation

We now turn to how best to account for innovation concerns in merger review.  We begin by identifying the types of mergers in which innovation and dynamic efficiencies could play a significant role when reviewing the competitive effects of the transaction.  A useful classification scheme for mergers is provided by Richard Gilbert and Willard Tom, which they apply to U.S. mergers evaluated by the Federal Trade Commission (“FTC”) and the Antitrust Division of the Department of Justice (“DOJ”).Footnote 45  Gilbert and Tom group challenged mergers, including those involving innovation, into four categories:

  • Mergers in which actual competition in an existing goods market would be reduced;
  • Mergers in which potential competition in an existing goods market would be reduced;
  • Mergers in which actual competition in an “innovation market” would be reduced; and,
  • Mergers in which actual competition in a future goods market would be reduced. 

The first two types of mergers are adequately addressed using conventional tools, provided attention is paid to dynamic efficiency issues (which we discuss in the next section).  The third category is a special type of analysis that has been proposed by the FTC and DOJ for use in some cases.Footnote 46  The concept of the “innovation market” does not exist in Canadian jurisprudence, and we do not advocate it.  Instead, we believe the competitive effects related to innovation can most practically be addressed under the fourth merger category.Footnote 47  In speaking about a “future goods market” here, we do not intend to restrict ourselves to markets consisting of new goods that do not currently exist.  We have in mind a more general definition, in which a future goods market is simply a market as it exists at some future date; as a result, it may include existing goods as well as products that will come into existence at a later date.

It is useful to additionally motivate the discussion by considering two examples of mergers in which innovation effects are important, and for which a conventional merger analysis may not lead to the correct answer.  Any practical framework should be able to deal correctly with examples such as the following:

  • Example 1: Two firms, A and B, currently sell products that are commonly understood to be complements from a demand perspective (i.e., an increase in the price of one of the goods decreases demand for the other).  However, B is currently engaged in innovation such that in the future, there is a strong likelihood that B’s product will be valued by consumers as a replacement for A’s (instead of used in connection with A’s), and there is likely to be little other competition for A’s product.  Firm A proposes to acquire firm B.

  • Example 2: Firms C and D currently compete with each other in the sale of some product.  No other firms currently sell similar products.  However, some set of firms E, F, and G are engaged in R&D that is likely to lead to future products that will be competitive with those offered by C and D.  Firms C and D propose to merge.

In Example 1, it is clear that if innovation concerns are not taken into account a merger might be permitted that would decrease economic efficiency if further entry in competition with a merged A/B is unlikely, despite the fact that at the time of the merger, the parties are not producing close substitutes for each other (and consequently their products are not considered to be part of the same relevant antitrust market at the time of the merger).  For instance, if Microsoft had proposed to acquire Netscape in the mid‑1990s, and accepting the U.S. government’s theory of anticompetitive harm (i.e., that Netscape’s browser would eventually represent a strong threat to Microsoft’s dominance in the market for operating systems), a strong case for blocking the merger should be made.Footnote 48  Example 2 illustrates that accounting for innovation concerns can also have the effect of demonstrating that some mergers thought to be undesirable may actually not lead to a lessening of competition—the state of current competition, as represented (for example) by market shares, need not indicate the “true” long run level of competition after accounting for the likelihood of innovative entry.

4.2. A framework

Given the discussion in the previous section,Footnote 49 we suggest that potential competitive effects of innovation are best dealt with using an approach that proceeds by asking five questions.

  1. Does innovation matter in the industry?

    This is a critical threshold issue, since if innovation is not a key attribute of competitive rivalry between firms, it is unlikely that innovation concerns are truly at the heart of any negative competitive effects feared from the proposed transaction.  On the other hand, if the merging firms operate in a highly innovative industry then an exclusive focus on existing goods markets may overlook competitive issues relating to future goods markets.
  2. Is it possible to identify the type (or identity) of firms that may participate in the future goods markets?

    The claim that a merger is likely to result in a lessening of competition in a future goods market (as in Example 1) can only be compelling if it is possible to identify the firms that may plausibly compete in that future goods market and the innovative products that these firms would likely be selling, and if the relevance of the innovation to competitive success in that market is transparent.  Similarly, it is only reasonable to allow a merger that currently seems problematic on the grounds that innovative new products will be competitive in the future (as in Example 2) if the firms that may develop and sell those products—as well as the nature of the potentially competing products themselves—can be readily discerned.
  3. Are the merging firms likely to compete against each other in an identifiable future market, but for the merger?

    Assuming we are able to determine with some degree of confidence the future products and firms that will be of interest, this step involves determining if the existing or future products of the merging firms will compete in that future market.  At this stage, the question allows for an appropriate consideration of uncertainty.  Firms may be engaged in R&D which would result in competing products if both firms are successful, but if success by both occurs with only a low probability, then there is little reason to fear a merger among the two firms will substantially lessen competition in the future market, and the case for intervention is weak.
  4. Would the merger result in a decrease in the resources devoted to one or more R&D programs, or the diversity of R&D programs such that the rate of innovation is likely to change following the merger?

    Provided (3) is answered in the affirmative, we ask whether evidence exists that the parties would change the manner in which they are pursuing innovation if the merger is permitted to proceed.  An affirmative answer at this stage would support a conclusion that the merger is likely to affect competition in future markets by reducing the level of innovation today.  This consideration is absent from merger analysis under the MEGs, but it is potentially important in scenarios involving innovation.  For instance, it may be the case that a merger will not be found to have any substantial effect on the prices of existing or future products, but may negatively impact on the speed at which the merging parties introduce new products.  If this effect is substantial enough, it might be desirable to challenge the merger despite the absence of price effects in either the existing or future market.  Step (4) therefore provides the opportunity to do so.
  5. Would the merger result in the merged firm being able to raise prices in the identified future market, relative to prices but for the merger?

    This is analogous to the traditional question asked under merger analysis, and discussed in the MEGs.  Under the facts assumed in Example 1, we would expect the merger to allow the parties to raise the price of the future product, since the parties would otherwise be in competition but for the merger.  In Example 2, we would expect competition from innovators entering in the future to be able to constrain pricing by the merging parties.  Relative to the conventional case, another additional nuance is that there may be no pre‑merger price to use as the base price; this complication arises in Example 1, for instance. 

    In setting out this framework, we do not advocate an exclusive focus on innovation at the expense of conventional price concerns.  A merger may lead to higher prices in a future market than would otherwise exist if the question posed in (3) is answered in the affirmative.  Nonetheless, a substantial price effect is not an inevitable consequence of a merger between likely competitors in a future market, just as it is not for a merger between actual competitors in an existing market.  Non‑merging parties selling other products—including, as in Example 2, products that do not yet exist—may discipline the merged firm’s pricing.  To take another example, merging firms may be conducting R&D toward the introduction of new drug products that, if successful, would be sold in competition with a number of therapeutic alternatives offered by other firms, so while the merger may eliminate competition between the merging firms, this is not expected to have a substantial effect on prices in the future market given the competition expected from additional rivals.

4.3. Practical issues in applying the framework

In this subsection, we focus on practical issues that are involved in using this framework to evaluate specific mergers.  Each of the five questions used to frame the analysis is discussed below.

4.3.1. Is innovation important in the industry?

This threshold question is usually relatively easy to address.  Research into the industry should be able to establish certain basic facts, including whether firms attempt to gain competitive advantage through innovation in any form, and whether consumers derive benefits from the introduction of innovative products.  Analyst reports covering the merging firms or the sector in which they operate may provide valuable information on the importance of innovation.  Studies of the industry in general or of specific markets within that industry may also exist.  These are often useful in understanding whether innovative product attributes drive consumer demand.  Any evidence that consumers tend to switch purchases among different producers based on product attributes—especially those attributes that have recently been introduced—is indicative of active innovation.

Qualitative information of this type should be supplemented by more quantitative inquiries.  One approach is to look at innovative inputs that are known to be correlated with innovative output.  R&D spending by industry participants is relevant, since industries in which participants make substantial and ongoing R&D commitments (normally measured as a share of sales) tend to be more innovative.Footnote 50  Another metric of innovative input that has been used with some success is scientific research activity, as indicated by publications in scientific journals.  Lim has shown that both pharmaceutical and semiconductor innovation (as represented by firm patenting) are strongly correlated with the number of publications by firm personnel in applied scientific journals.Footnote 51

Another approach would be to look for evidence of innovation in market outcomes.  An observation of frequent changes in market share is one such indicator.  Fluctuations in market share from quarter to quarter or from year to year tend to indicate the presence of dynamic competition, as firms are constantly forced to reevaluate the effectiveness and attractiveness of their product offerings in light of their competitors’ performance in the marketplace.  The cellular handset industry provides an example of dynamic competition and changing market shares.  Leaders such as Nokia are unable to rest on past market successes, finding that if they do so other innovative cellular handset producers such as Motorola or Samsung will rapidly introduce products with attractive characteristics and gain market share.Footnote 52  Note that market share turnover is not a necessary condition to identify the importance of innovation to the sector, as is witnessed by Intel for example.

4.3.2. Can firms and products in future goods markets be identified?

This question is posed to determine whether there is likely to be an identifiable future market for which the transaction raises competitive concerns.  To identify any such market, we must be able to describe the firms that could plausibly be participants in the future market along with the product(s) that could potentially be at issue.  As we have discussed above, these products could include those that already exist and are expected to remain available, as well as new products that will result from innovation.  Analyst reports will again be of use here—stock analysts are particularly attuned to factors that may influence the future development of markets in which the companies they cover operate.  In telecommunications, for example, analyst reports offer various opinions on the likely take‑up of voice‑over‑Internet Protocol telephony, and what this will mean for incumbent telephone providers.

The parties’ internal documents will also offer insight, particularly in sectors where innovation is central.  In any “racing” situation, for example, we would expect some discussion in the firms’ internal documents comparing its research advances to those of important rival firms, thereby allowing the Bureau to identify the most relevant rivals to the merging firms.  Thus, for instance, IBM will comment on Sun’s and HP’s latest Unix advances when discussing funding for new mainframe models.  Documents that discuss research setbacks or that make requests for further funding may contrast the firm’s innovation success with important rivals. 
In addition to reviewing available descriptive material on the industry, certain economic and technological considerations must be taken into account.  Consider first how information on the form of innovation could be useful in identifying possibly competing firms.  Innovation in some industries may arrive unpredictably from outside sources, or from firms or individuals that are only peripherally involved in the industry at present.  By looking at firms that currently operate in the industry in such cases, we would be unable to determine the firms that might be active in particular markets in the future, and the sources of innovation for those firms.  An example of this type of industry might be computer software.  Many software innovations arrive in the form of computer code written by programmers who may not be affiliated with established firms.  Software firms can spring up almost overnight to take advantage of perceived market opportunities, facilitated by the fact that entry into software markets requires programmer effort and time, and relatively little capital investment.Footnote 53  Moreover, programmers are ingenious at devising methods to achieve the same functional goals by means that are different from those currently practiced.Footnote 54  Predicting which software firms might be competing in some future market years hence would then be extremely difficult, if not impossible.  In cases such as this we would recommend against challenging a merger based on innovation concerns.

At the other extreme, innovations in some industries may arrive in a relatively systematic fashion as a result of R&D efforts conducted within established firms.  Drugs or medical devices are examples of industries in which many innovations result from diligent, systematic internal efforts.  Only a select handful of firms may have the necessary physical and human capital assets to pursue particular innovations and commercialize the resulting products.  Industry participants and informed observers will normally be well aware of the identities of these firms.  Pharmaceutical and medical device products must also pass through a lengthy governmental review and approval process before they can be sold on the market.  A wealth of regulatory documentation, together with patents filed by innovating firms, is publicly accessible.  These facts ensure that it is relatively straightforward to identify well in advance firms that may ultimately compete in a future market.  It is probably for these reasons that the majority of merger challenges brought by the U.S. agencies based on innovation issues have involved pharmaceutical or medical device firms.

Equally important is isolating the contribution of the innovation to the value and functioning of the resulting product.  For the reasons discussed previously, it will normally only be possible to do this in a rough and qualitative way.  For illustrative purposes we offer two extreme cases for discussion.

At one extreme, the typical innovation in an industry might resemble the incremental improvement to logical processing noted previously in the microchip example.  Even if it is somehow known that a set of firms are pursuing similar innovations, it is difficult to predict how a merger between two firms pursuing this research would likely have a negative impact on future product markets, given the great disparity between the scale on which innovation takes place and the ultimate scale of the product.  A great deal would have to be known or assumed about the technical relationship between the innovation and the additional complementary components necessary to build up from the innovation level to the product level—for example, would innovation failure imply that no product would be forthcoming?  Or would the product simply be different somehow?  If innovation succeeds, would different firms develop different products using alternative sets of complementary components?  Generally, the more oriented the innovative activity is toward “basic research” (or the further the product is from commercialization) the more difficult it is to evaluate reliably the potential competitive effects resulting from a merger.  In such circumstances, a merger between two firms, even when both are pursuing similar research, is not likely to result in an identifiable lessening of innovation or increase in prices in future goods markets.
At the other extreme, some products are effectively made up of a single innovation.  Pharmaceutical products are a good example.  Successful R&D in the pharmaceutical industry results in an “active pharmaceutical ingredient”—essentially a chemical compound with attractive clinical attributes—that can be incorporated into various dosage forms for sale to consumers.  Here, it may be relatively easy to predict the extent of competition in a future market, given the observed relationship between R&D along a particular technological trajectory and competition in existing (and past) product markets.  Two firms engaged in clinical trials for drug products that aim to treat the same condition in these circumstances will very likely be product market competitors should their development programs succeed.

With this in mind, it is apparent why many of the recent mergers challenged by the U.S. agencies on the basis of innovation issues have involved firms in the pharmaceutical industry.Footnote 55  It is not because these were the only innovation‑intensive industries in which firms proposed to merge.  Rather, we believe it is because problems of identifying the likely firms and products in future goods markets, which tend to be highly speculative in many other industries, can more frequently be resolved in the pharmaceutical industry.Footnote 56  As a result the potential competitive repercussions from mergers involving firms that are likely to compete in these future markets can be more readily predicted.

4.3.3. Would the merging firms compete against each other in an identifiable future market, but for the merger?

Suppose we are able to adequately resolve the previous set of issues, and can identify a class of innovation‑embodying products (perhaps together with existing products) that would be considered close substitutes in a future market.  Additionally, assume we can determine the firms that are engaged in innovation with the goal of selling products in the future goods market, and also those that are expected to be able to sell such products through the use of alternative technologies.Footnote 57  One might be tempted to conclude (erroneously) that since these firms would compete in the same market following the conclusion of the innovation stage, a merger involving firms that account for a substantial part of this market (as later defined) will be found to be anticompetitive.  However, it does not follow from these assumptions that there is an appreciable likelihood that the merging firms will actually compete against each other in the resulting market, viewed from today’s perspective.  Several specific areas of concern—the uncertainty associated with innovation, the potential for patent holdings to interfere with the ability to bring a product to market, and the need for complementary assets for commercialization—may substantially decrease the likelihood that innovating firms meet in the future product market following the innovation stage.  We advocate a great degree of sensitivity to these issues to prevent challenges based on an unrealistic view of the post‑innovation market.Footnote 58

Consider first an example addressing the issue of uncertainty.  Suppose it is known from an analysis of similar projects that the R&D undertaken by two particular firms is highly risky, such that each firm will succeed with an estimated 25% probability.  (In other words, three out of four similar R&D projects have failed to lead to a viable product.)   Then in the absence of a merger, these firms would only compete against each other with a 6% probability.Footnote 59  This would suggest that the competitive effect of concern, whether it is in the form of reduced innovation or increased pricing, would have to be quite large to warrant intervention in the presence of any offsetting efficiencies.

One might wonder whether it is reasonable to expect data on the probability of R&D success to be readily available.  In fact, where R&D involves a closely monitored regulatory procedure, as in pharmaceuticals and medical devices, data are available.  A recent study of R&D success factors in drug development used data collected on over 1,900 drug compounds developed in the U.S. between 1988 and 2000.  The authors calculated various m easures of R&D “success” based on whether or not the Food and Drug Administration (“FDA”) approved the compound for treatment of particular symptoms at various stages of clinical trials. Footnote 60  Success rates conditional on a treatment starting Phase 1 clinical trials (initial human testing for safety) vary across therapeutic area, ranging from 30% for respiratory drugs to 78% for hormone preparations.  Drugs targeted toward a larger market seem to have a lower probability of success.  Thus, at least in the pharmaceutical area, good data are available on R&D success rates.

A second factor may pose a barrier to direct competition between innovators even if both firms’ innovations successfully proceed to the commercialization stage: the existence of intellectual property rights.  Firms in many industries vigorously patent their discoveries.  To the extent that a firm’s patent rights for a particular discovery are valid, in that the discovery is novel, useful, and non‑obvious,Footnote 61 the rights conferred allow the patentee to exclude others from practising the invention.  Products sold by firms that do not have a licence to the patent and that embody technologies that fall within the patent’s claims are said to infringe.  The patentee has a right to sue the infringer to prevent infringement.  Of course, the patentee may choose to license to the infringing firm in return for a royalty, but it need not do so.  In some cases, where the technologies contained in products are complex and involve multiple discrete intellectual property rights, patentees may find themselves in a mutually infringing situation, where neither firm can sell its product without access to the patent rights held by the other.  Patents in this situation are said to be “blocking.”  Blocking patent positions are commonly resolved through licensing negotiations, whereby each firm grants the other the right to produce and sell (referred to as a “cross‑licence”), possibly with the exchange of a fee or royalty.  Mutual infringement of this type is especially common in high‑tech industries such as semiconductors, where any given product may draw on literally thousands of disparate inventions, each potentially covered by some firm’s patent rights.

If the expected result from the innovation is mutual blocking or simply one‑way infringement, the existence of valid patent rights may pose a barrier to direct and unfettered competition between innovating merging firms.  Where this is likely to be the case, it would be inaccurate to compare the post‑merger outcome to a but‑for environment that entailing direct competition between the two firms.Footnote 62  In a world without the merger, competition would simply not arise unless the firms chose to license.  Even if a licence were agreed upon, its terms might call for running royalties that would affect the prices charged in the future product market.  For example, holders of blocking patent rights may seek to resolve their dispute by licensing each other with a running royalty that results in each choosing to sell at the price that maximizes their joint profit, just as though they were part of the same firm.  Such a licence would be permissible under the Bureau’s Intellectual Property Enforcement Guidelines, since it is consistent with a “mere exercise” of valid intellectual property rights.  Alternatively, the parties may resolve their dispute by licensing in a lump‑sum form that does not impact on market prices.  Where patent threats are roughly symmetric, blocking positions are often cleared up by royalty‑free cross‑licensing.  These possibilities highlight the difficulty in speculating ex ante upon the eventual outcome in markets characterized by broad patent rights.  That being said, in industries where licensing is common there may be past examples upon which the Bureau may rely to determine the most likely outcome should the merger not proceed.

A final issue to be considered is whether the innovating firms possess the assets necessary to commercialize the products resulting from the innovation stage.  The farther away from commercialization is the state of technology under development, the greater the need for inventor cooperation to bring the innovation to the commercialization stage.Footnote 63  Some innovating firms may be vertically integrated into the production and distribution stages and may therefore have access to all the necessary assets, including perhaps specialized production equipment and highly trained marketing and sales forces.  Others may be innovation “specialists” that do not have access to these assets.  Such firms are faced with the choice of licensing their successful innovations to others, or attempting to develop the complementary assets needed for commercialization themselves.  Licensing is often the preferred alternative.Footnote 64

There are likely to be differences in the ex ante competitive considerations depending on whether innovating firms are “integrated” or “specialized.”  Our analysis thus far implicitly assumes that innovating firms are integrated; but what if one or both innovators are specialized and have no intention of producing and selling the resulting product?  Although the innovating firms, if specialized, may not participate directly in the resulting product market, they would participate indirectly in the absence of the merger by offering licences to firms with the requisite complementary assets.  Should the complementary assets be scarce, however, there may be a limited number of attractive licensees, and these licensees would have bargaining power over licence terms.  In the extreme case, if there is only a single viable licensee, the licensee would capture the gains from competition among the licensing firms, and may or may not pass these on to final consumers, depending on the structure of the future product market.  More generally, the extent to which competition among licensors translates into gains for consumers in the future product market would depend on the structure of the “market” for licensees, and these licensees’ competitive strengths in the product market.  While such considerations complicate the analysis, an evaluation of the distribution of complementary assets is essential if the merging firms are not integrated.

4.3.4. Would the merger result in decreased R&D resources being allocated to innovation?

If the innovating firms are expected to compete against each other in a future product market, one competitive concern is that the merger would result in a decrease in the combined firm’s level of innovation.  For example, where the separate firms are pursuing different R&D paths that, if successful, would result in two differentiated products being sold, the combined company might choose to cancel the least promising program.  Consumers would be deprived of the expected benefits of price competition offered by the forgone product, as well as any special benefits offered to particular consumers by the cancelled product’s specific attributes.  As an alternative example, both R&D programs might be maintained, but the merged firm, perceiving less urgency than two separate firms, might decide to scale back each program.  Innovation might be expected to take longer to yield results in this case, so gains to consumers would arrive at a more distant date than would be the case but for the merger.

Whether a merger would impact on the resources devoted to innovation is fundamentally a context‑specific issue.  This follows from the earlier discussion of the general inadequacy of theoretical or empirical results concerning the effect of concentration on innovation, and the lack of a clear predictive model relating concentration to innovation activity.  A fact‑specific analysis might take into account the following factors:

  • Information available from parties on resources devoted to R&D, and planned expenditures going forward.  Basic planning documents maintained by the parties in the ordinary course of business would help to characterize the level and type of R&D that would likely exist in the absence of the merger.

  • Information that might have been revealed since R&D began on the size of the eventual market or the expected costs of completing the R&D phase.  There is nothing wrong with planning to reduce or eliminate R&D if that R&D is no longer deemed to be warranted in light of market circumstances.  Should this information come to light at around the time the merger is announced, it might be erroneously linked to the merger.  We must be careful to distinguish actions caused or made possible by the merger from those that the parties would have engaged in regardless, as is done now with any merger investigation.  Studies performed or collected by the parties on the possible market prospects for the expected products, as well as internal studies updating estimates of the likely R&D resources necessary to get there, would be helpful for this analysis.

  • Indications that the first firm to innovate would garner a disproportionate share of the rewards.  Recall that the patent race model assumes that only the first innovating firm will earn returns from innovation.  This gives the model its “racing” property.  The winner‑take‑all assumption is unlikely to hold generally, but it is worth investigating whether the market in question has any characteristics that would allow some of the insights of the patent race model to apply.  For example, if the first innovating firm is able to lock consumers in to a proprietary standard, so that later innovators would have great difficulty surmounting this first‑mover advantage, there would be a strong incentive for participants to bring innovations to market as soon as possible.

  • Information on the degree of overlap between participating firms’ R&D programs.  A transaction that has the effect of reducing duplicative R&D, while leaving the rate of innovation effectively unchanged, should not be subject to challenge on innovation grounds.  The obvious difficulty is in determining whether R&D is truly duplicative or whether firms are instead pursuing different paths toward a similar goal.Footnote 65  In this latter case, elimination of an R&D program would be undesirable.  Technical expertise will normally be necessary to make this assessment.

The recent FTC Genzyme decision is an interesting example of the use of case‑specific information in practice.  It also highlights the conflict between those that would adhere to a more formalistic view of innovative rivalry and those, such as then‑Chairman Muris, who advocate a more fact‑intensive inquiry.  The FTC’s initial challenge was brought on an “innovation markets” theory: since Genzyme and Novazyme were the only two firms conducting R&D toward a particular product, the transaction amounted to a merger to monopoly in the innovation market consisting of firms performing such R&D.  Ultimately the Commission voted not to challenge the transaction.  Chairman Muris summarized the factors that contributed toward his decision to vote with the majority against a challenge.Footnote 66  He commented that “[a]ssessing the effects of a merger on the pace of innovation is especially fact‑dependent,” and pointed toward the “lack of any clear theoretical or empirical link between increased concentration and reduced innovation.”Footnote 67  Muris also argued that the merging firms would not have been “racing” to market in the absence of the merger, since there was some evidence that Genzyme’s product was viewed as a short‑term treatment while Novazyme’s was more in the nature of an “improved, second‑generation therapy.”Footnote 68  The majority appears to have concluded that the facts supported an inference that the firms’ R&D programs were separate and independent, rather than competing and thus interrelated.  A dissenting statement was offered by Commissioner Thompson who argued that Genzyme would face reduced pressure to innovate with Novazyme removed as a rival, essentially under a formal patent race theory: “[Competition from Novazyme] was important because it created a race between Genzyme and Novazyme to develop Pompe ERTs [enzyme replacement therapies], thus increasing the pace of innovation.”Footnote 69  As under the patent race model, Commissioner Thompson viewed the removal of an innovator as likely to lead to a decrease in the overall incentive to innovate.Footnote 70

4.3.5. Would the merger lead to increased prices?

In the final step of the analysis, the traditional merger competitive effects assessment is undertaken, but with the complication that we are dealing with a future post‑merger environment compared to a future market without a merger.  As a result, there is likely to be no pre‑merger price or output level for the future product to use as a comparison.  This is obviously true when the innovative product is in some sense pioneering and would face no close competition from existing products.  Less obviously, it is true even if the innovation leads to an improvement in an existing class of products.  If firms are pursuing product improvements, quality‑adjusted prices would be expected to change pre‑ and post‑innovation.  This effect is distinct from the purely nominal change in price that would result from a firm having increased market power in a given product market.  The critical question is then whether we can identify the undesirable price impact in an environment of changing quality‑adjusted prices, where the latter are often unobservable.

Data and measurement problems will normally prevent us from being able to answer this question directly—we will rarely have the information necessary to predict post‑innovation prices with and without the merger.  Accordingly, we propose a practical indirect solution.  One way of proceeding is by abandoning the (somewhat artificial) distinction between the merger’s effects on innovation and its effects on pricing, and treating these as a single inquiry.  The value of this approach can be seen by examining a simplified framework in which there are four possible cases of interest, depending on whether the merger materially reduces the level of innovation, and whether the combined company has a high share of the post‑innovation market.

  • No impact on innovation, large share. Here, the traditional conclusion would apply; prices (both quality‑adjusted and nominal) might be higher than without the merger, depending on other characteristics of the post‑merger market.

  • No impact on innovation, small share.  Here there should be no cause for concern. 

  • Reduced innovation, small share.  If only the merging firms’ level of innovation is expected to change, while other non‑merging participants are expected to be equally active in the post‑innovation market regardless of the merger, the reduction in innovation by the merging parties is likely to have little effect on consumer welfare.  Product characteristics and quality‑adjusted prices are expected to be virtually unchanged with the merger compared to a world without the merger, and therefore a challenge is unnecessary. 

  • Reduced innovation, large share.  The firm’s large share in the future market suggests that for a given set of products, it would be able to sustain a nominal price increase relative to the price level that would exist in the absence of the merger, holding all else (including the level of innovation) constant.  Combined with the negative impact on innovation, we would conclude that quality‑adjusted prices are likely to be higher compared to a world without the merger.

This set of results indicates a compelling case may be made for a merger challenge if (and only if) the merger would result in the combined company having a large share of the future market.  This result is intuitive, and it also has the attractive property that it “nests,” as a special case, mergers that do not have innovation effects.  We caution that these results are obtained from a highly simplified example; in any practical application, there are more possibilities than just “high” or “low” shares and “negative” or “neutral” impacts on innovation.  As a specific example, it is conceivable that a merger may result in an increased level of innovation, an efficiency effect that would tend to offset a large combined company presence in the future product market.  It is to such cases that we turn in the next section.

5. Dynamic efficiency considerations

Parties may claim that a merger that gives rise to the exercise of market power, and consequently reduces allocative inefficiency, may be beneficial overall due to a positive impact on dynamic efficiency.  In this section we discuss generally how such issues may be incorporated into the framework proposed in the prior section.  We provide examples of the types of dynamic efficiencies that might be reasonable for parties to articulate and for the Bureau to seriously consider.

5.1. Impact and quantification

Suppose a merger is expected to result in a significant increase in market power in an existing market.  The MEGs recognize that parties may attempt to prove offsetting efficiencies arising from a variety of possible sources, including dynamic efficiencies.Footnote 71  The types of efficiencies that may be considered in the conventional case, and the process by which cost savings are found to be true savings in resources, are well understood.  In this respect, we note that the Competition Bureau considers fixed cost savings, as part of any efficiency claim: “Both variable and fixed cost savings are relevant to the analysis because both generate producer surplus (even though it is recognized that generally only variable (i.e. marginal) cost savings lead to price reductions).”Footnote 72  The list of cost savings that are either variable or fixed and that the Bureau will consider include the traditional product, plant‑level and multi‑plant level savings related to economies of scale, economies of scope, and economies of density and specialization, as well as savings in respect of distribution, advertising and raising capital.   This differs from U.S. practice, although the Tentative Recommendations of the US Antitrust Modernization Commission include a recommendation by some Commissioners that increased weight should be placed on fixed‑cost efficiencies, particularly for dynamic, innovation‑drive industries where marginal costs are low relative to prices.Footnote 73

Demonstrating causation and magnitude where dynamic efficiencies are relevant will normally be much less clear‑cut than for these traditional fixed and variable cost savings.  This is largely due to the uncertainty and measurement problems associated with innovation, as described above. 
In general, it may be very difficult for parties to show that a merger will appreciably increase innovation, given that innovation is subject to a high degree of uncertainty to begin with.  For example, pre‑merger the parties may expect to introduce new products within a certain time frame, say two years, although given uncertainty it could take longer or shorter; they may claim that by merging, they will be able to reduce the expected time for the next innovation to be commercialized to one year.  Proving that this is likely may be difficult for the party bearing the burden of proof, in this case the merging firms.Footnote 74  In addition, the parties would have to demonstrate that the proposed dynamic efficiencies are merger‑specific, and not available through an alternative organizational form such as a strategic alliance devoted to R&D.

Even after addressing such causation issues, merging firms will still face the problem of estimating the incremental surplus created by the additional innovative effort, most or all of which may be tied to future products.  The attributes associated with those products may not be known with certainty at present, which will complicate the quantification of the surplus created.  We would expect that merging firms will normally have difficulty meeting the requisite level of proof required for causation and quantification.  However, given the ease with which claims of dynamic efficiencies can be made, and the fact that the merging firms will have much more information than the Bureau regarding their innovation prospects, the Bureau is right to approach such claims rigorously and with some skepticism. 
Clear results exist only for certain limiting cases that, while unlikely in practice, provide the bounds within which cases will generally fall.  Suppose that there are only two firms with the knowledge and expertise necessary to develop a new line of products, which will be the only line of products sold by the firms, and that for technical reasons R&D will only succeed if the firms combine their efforts by merging.  Post‑innovation, the merged firm is expected to price higher than the prices that would be obtained if there was no merger and the two firms were competing.  Here, static allocative inefficiency and dynamic efficiencies arise from the same source: the introduction of the innovative products.  Consumers would receive no surplus without the merger, as arguably the product would not be introduced.  In such a case, dynamic efficiencies outweigh static inefficiencies.  At the other extreme, suppose the same merger would have little or no impact on the pace or likelihood of innovation.  Then static inefficiencies would be found to outweigh dynamic efficiencies because the negative price impact would remain post‑merger without any positive improvement in consumer surplus from the introduction of new goods.

The problem, of course, is that the direct link between changes in allocative inefficiency and dynamic efficiency that is present in these hypothetical illustrative examples will normally not exist.  If, for instance, the parties also sell an existing line of products in competition with each other, the allocative losses associated with the pricing of these products would have to be compared to the gains from the introduction of the future products.  Alternatively, suppose innovation was possible but uncertain pre‑merger, but the merger would render it a certainty.  Then the surplus created under a no‑merger scenario, which would arise with some probability less than one, would have to be compared to the smaller amount of surplus created by merging, which would exist with certainty. 
In general, we are unable to escape the need to make at least rough comparisons of the sizes of static and dynamic efficiencies.  While this may seem unpalatable, at least two arguments suggest that if the merging firms can successfully demonstrate plausible and likely dynamic efficiencies, these should be weighted heavily against suspected inefficiencies arising from price effects.

First, as we discuss in the next subsection, there are many sources of dynamic efficiencies for innovating firms but there is just one source of allocative inefficiency.  Innovating firms may not even be aware of all the ways in which their actions may enhance the flow of consumer gains over time.  For example, there may be important spillover effects.  Parties do not take spillovers to other industry participants or society into account when making investment decisions—hence the bias toward too little innovation relative to the socially optimal level that we described earlier.Footnote 75  Yet such spillovers are a source of real gains to society, and may be orders of magnitude more important than the losses incurred by select consumers as a result of increased prices.

Second, when dynamic efficiencies can be successfully demonstrated, any possible price effects, if they exist, may tend to be transitory, given the dynamically competitive nature of any industry where demonstration of important dynamic efficiencies is possible.  In such cases, higher initial prices are likely to bring forward a new round of innovation intended to replace the merged entity, assuming it attains market leadership status following the merger.  This effect would be absent, or mitigated, if the merger allowed the parties to gain control of scarce resources that would be necessary for others to innovate, or if barriers to entry by other potential innovators were otherwise high.

5.2. Sources of dynamic efficiencies

Given the variety of different types of innovation and forms of dynamic competition across firms and industries, merging parties may be expected to make a wide range of arguments in support of dynamic efficiencies.  Provided parties are able to demonstrate the necessity of the merger to realize these efficiencies, and quantify their magnitude against any inefficiencies associated with the merger, they should be free to make such arguments.  We provide in this subsection a brief discussion of sources of dynamic efficiencies that can be given economic support.

5.2.1. Elimination of duplicative R&D

The most obvious efficiency that may arise from a combination of R&D programs is elimination of redundant R&D.  Firms seeking to introduce a new product may need to complete certain stages of basic or applied research in order to make product development possible.  As a result, when multiple firms are seeking to innovate in the same area, the same research stages may be performed multiple times.  Efficiency would be improved if some of this duplication were eliminated, provided there is little or no effect on the pace of innovation.

The difficulty here is judging “whether a company’s decision to shut down one of two research tracks, to focus its resources on just one track, would likely be procompetitive or anticompetitive.”Footnote 76  If there is a strong positive relationship between the number of competitors engaged in R&D and the likely pace of innovation, a case can be made that R&D programs are not likely to be duplicative; the existence of multiple programs results in consumers enjoying the benefits of new products at a faster rate.  This is clearly a factual question.  In addition, the Bureau must be sensitive to the possibility that often R&D programs are not substitutes for each other but rather complements: “reductions of so‑called ‘duplicative’ R&D may actually represent the elimination of diverse research paths that could lead to different results[,] and … even if research paths were identical, different R&D researchers in different companies might draw different inferences from them, and hence achieve different results from the same discovery.”Footnote 77  Given the uncertainties associated with R&D, we urge caution in accepting claims that R&D programs are duplicative.Footnote 78  At the same time, it is not procedurally efficient for the Bureau to conduct large‑scale investigations of the nature of merging parties’ R&D operations.  Our view is that if this argument has merit, it should be evident in the parties’ contemporaneous documents (e.g., strategic planning memoranda or presentations), and/or the assessments of independent market observers.

5.2.2. Economies of scale or scope in R&D

Another possibility is that a merger may allow the resulting firm to take advantage of economies of scale or scope in R&D activities.  Economies of scale in R&D arise when an R&D program of some size, say S, is more productive (in terms of the generation of innovative output) than two separate programs of equal size ½S.  Economies of scope in R&D exist when a firm engaged in a range of different R&D activities is more productive at each activity than a firm performing some smaller number of activities.  It has long been speculated that “larger” firms, measured according to some metric, are more successful innovators than “smaller” firms.  Indeed, this is one way of articulating the classic Schumpeterian hypothesis that posits a relationship between monopoly situations and innovation.  As Tirole summarizes:

[Schumpeter] suggests that large firms are better qualified or more eager to undertake R&D than smaller firms because increasing returns are prevalent in R&D; because R&D activity involves a high level of risk that is difficult to eliminate with insurance (for reasons of moral hazard), and large firms are more diversified and therefore more willing to take risks; because innovation, once generated, is implemented more rapidly in a large firm because there is an appropriate production structure; and because a monopolist does not have competitors ready to imitate his innovation or to circumvent an existing patent on this innovation.Footnote 79

Empirical studies focusing on innovation at the industry level have produced mixed results on whether there are increasing returns to R&D activity by virtue of economies of scale and scope.  More recent studies have focused on data related to individual R&D programs, which is the level of disaggregation necessary to evaluate the issue.  The key references are Henderson and Cockburn’s studies of pharmaceutical firm R&D.  In a 1996 paper, these authors presented evidence from the R&D programs of ten large pharmaceutical firms observed over an average of 20 years per firm.Footnote 80  They test whether research output, as measured by the output of “important” patents (those filed in several different countries) is responsive to the firm’s total R&D effort across all programs, and whether research productivity increases with the number of programs within a firm.  They find evidence supporting each of these hypotheses, with a mean research elasticity of about 0.3 in each case.  In other words, a 10% increase in total research spending results in about a 3% increase in research output for a given project, and a 10% increase in the number of large R&D projects also leads to a 3% increase in output.  In a later paper, Henderson and Cockburn ask whether scale and scope confer similar advantages at the development (as opposed to the discovery) stage.Footnote 81  They test whether scale or scope affect the likelihood of receiving FDA approval for a new drug product and find a significant impact for scope, but not scale.  From this research, they conclude: “The performance advantage of large firms appears to lie in economies of scope rather than economies of scale: all else equal, a development program initiated within a more diverse development effort is significantly more likely to result in an [approved drug] than one initiated within a more narrowly focused effort.”Footnote 82

As a result, some good evidence exists that mergers enhancing the scope of R&D efforts may have salutary and measurable effects on innovation, at least in the pharmaceutical area.  This is a developing research area in economics, and few results are available for other industries.  One exception is a recent paper by Helfat studying R&D on coal conversion technologies in the petroleum industry.Footnote 83  She finds a positive association between R&D and a measure of R&D capital stock, implying that firms find additional impetus toward innovation from their direct experience with past R&D efforts.

5.2.3. Improved intellectual property enforcement

Many firms seek to protect the returns from their innovations using intellectual property rights.  Indeed, a key principle underlying the intellectual property system is that in the absence of additional incentives for innovation provided by the formalized right to exclude, firms will under‑invest in R&D.  As we have argued above, generating a new idea in the industrial context typically requires incurring sunk, up‑front R&D expenditures.  Once these costs have been incurred, the incremental cost of employing the innovation, or creating copies of it, is generally quite low.  Moreover, since any number of people can use the innovation at the same time, the inventor or author is unlikely to be able to recoup his or her sunk investments if others can exploit the innovation by incurring only the incremental cost after the fact.  Intellectual property rights exist to give innovators the incentives to incur the up‑front costs of innovation. They do so by giving the inventor or author the temporary right to prevent others from “free riding” on his/her efforts. 

Recent research has shown, however, that in some cases, firms are unable to access the intellectual property system in the manner it was intended, and as a result these firms may not be able to protect the returns on their R&D expenditures to an efficient degree.  Lanjouw and Schankerman have studied the determinants of patent suits and settlements across a range of technology areas over 1978‑1999, focusing on the ability of small firms to protect their innovations.Footnote 84  They find that the probability of filing a suit on any particular patent owned by a firm is negatively related to the number of patents that firm holds in its portfolio, and that this effect is stronger for smaller firms (as measured by employment).  In other words, small firms are less likely to be able to settle disputes on advantageous terms and more likely to find themselves embroiled in costly, risky patent litigation, unless they have unusually strong patent portfolios.  Lanjouw and Schankerman conclude that as a result, “the enforcement process undermines the R&D incentives of small firms.”Footnote 85

Intellectual property issues may also negatively affect smaller firms in specific high‑tech sectors.  Josh Lerner, studying patenting behaviour in the biotechnology industry, has found that firms—especially those that are relatively inexperienced or financially weak—tend to avoid research areas in which others have patented intensively, even if these might be fruitful avenues for research.Footnote 86  With reference to the semiconductor and communications industries, the widespread practice of cross‑licensing of entire patent portfolios for relatively long (typically five‑year) periods has developed to resolve patent issues.  Through this means firms are able to settle potential infringement battles before they occur and avoid costly overlapping licence payments.Footnote 87  Those firms that are especially reliant on technologies that are covered by a diverse and fragmented set of ownership rights find it less costly to negotiate all the necessary licences ex ante than to develop strong patent portfolios of their own to use as a reciprocal threat if necessary.Footnote 88  Smaller firms without strong portfolios may find it necessary to enter into complex and resource‑consuming negotiations in order to obtain access to the necessary intellectual property if they wish to compete in such technology areas.

In sum, intellectual property rights may justify a set of dynamic efficiency considerations, provided firms are able to document these appropriately.  Mergers to consolidate resources may allow firms to better protect their intellectual property rights, thus enhancing returns to their R&D efforts; and, depending on the industry in which the firms operate, mergers may also help firms to clear a path through the thorny intellectual property positions of other firms, freeing up resources for other value‑enhancing activities.

5.2.4. Increased financial resources for R&D

Tirole’s synopsis of the Schumpeterian hypothesis, excerpted above, alludes to a possible beneficial effect of increased internal financial resources on innovation.  The argument is that it may be excessively costly for firms to obtain external financing (from equity or bond markets) for R&D activities and, as a result, firms with larger internal financial resources are better positioned to engage in R&D.  If this is true, then a merger that combines one firm’s ideas with another’s cash would tend to lead to more innovation than if the creative firm attempted to exploit its ideas on its own.

Hall summarizes the empirical work that has been done in this area to date.Footnote 89  She observes that R&D is quite sensitive to a firm’s internal cash flow, and that as a result, “any problems associated with financing investments in new technology will be most apparent for new entrants and startup firms.”Footnote 90  The availability of venture capital funding does not appear to offer a complete solution to these problems, despite the fact that venture capital financiers are able to much more closely monitor the use of funds than are ordinary stock market investors.  Lerner, Shane, and Tsai study a related issue in biotechnology, which is a very R&D‑intensive sector.  Innovative biotechnology firms often turn to alliances with larger, better‑funded partners to overcome financial constraints.Footnote 91  They observe that during periods of unfavourable equity market activity, innovative biotechnology firms in need of financing are more likely to enter into alliance agreements which in the longer term are less successful, as measured by subsequent progression through clinical trial and approval phases.  This would suggest that hybrid organizational forms such as alliances are an imperfect solution to the funding problem, and that in some cases allowing firms to merge may be the best approach.

6. Conclusion

We have sought in this report to provide a sound general grounding for the incorporation of innovation and dynamic efficiency effects in merger review.  Recognizing that innovation and dynamic competition are critical drivers of economic growth, our discussion has relied in large part on two fundamental principles.  First, promotion of dynamic efficiency, and thus productivity growth over the long term, normally requires tolerating some degree of static allocative efficiency over the short term.  It is this effect that will lead to a real trade‑off in many merger situations: should a merger that is expected to raise prices be allowed to proceed, provided that it supports increased innovation?  Second, we have suggested that on balance, the incentives created for innovation in a market economy are likely to be insufficient, and that as a result increased innovation—and not just increased R&D—should generally be regarded as desirable.

Analyzing whether a given transaction is likely to lead to increased innovation is difficult from a practical perspective.  We pointed to a number of issues that should remain in the foreground during any actual analysis.  With these in mind, we proposed a method for incorporating innovation issues into merger analysis that focuses on effects in future product markets.  This method is appealing to us because it does not rely on somewhat hazy concepts such as “innovation markets.”  Nonetheless, applying it to any real situation would not necessarily be easy, and in particular, the set of factual information on the markets and firms involved would require careful analysis.

Finally, we examined incorporation of sources of dynamic efficiency into the merger review process.  In our view, based on the existing economic literature, there are quite a number of plausible, defensible dynamic efficiency claims that can be made.  Of course, firms face a daunting challenge substantiating these in any given case.  It may be especially difficult to quantify the gains that can be realized from future surplus as against any expected anticompetitive effects from increased prices, beyond providing rough orders of magnitude.  Nonetheless, we do not believe the burden of proof should be shifted.  It rightfully should remain with the merging firms, as they have the greatest information available to substantiate any claims of dynamic efficiency.  A more relaxed quantitative burden might be provided to those merging firms that are able to adequately demonstrate plausible and likely qualitative improvements in dynamic efficiencies resulting from the merger.

Date modified: