An inside look at the innovation impairment myth of patent reform
Legislative reform of the US patent system may alleviate the concerns of companies that face assertions, but they could end up with a far more difficult situation on their hands as a result
I began writing this as I returned from a trip to Vietnam on behalf of the International Finance Group and its parent organisation, the World Bank. The topic of interest was IP-backed lending in Vietnam, which I found to be remarkable given the relative lack of interest by US financial institutions for financing patent-backed assets in the United States for any but the largest patent owners in the US system. Vietnamese companies and institutions were keen to learn from experts in the banking, legal and valuation professions about best practices for IP-backed lending. Much to my surprise, an unexpected number of recommendations started with a common phrase: do not follow the United States in current legal developments. Much of the subsequent discussion focused on the dizzying and continual changes to the legal certainty of US-based assets in which companies invest tens of billions of dollars every year. Complicating the situation further is the fact that some judges admit to a core lack of understanding of many of the unique technical and economic attributes that drive patent value and innovation – a particular problem given that they sit on the bench of the United States’ highest court. Naturally, current patent reform efforts in the United States, occurring after Congress passed the first substantial patent legislation only three short years ago, was a major topic of discussion.
The question of US patent reform, like many issues pending before Congress, is heavy on hyperbole and rhetoric, and light on fact. Proponents of patent reform start with good intentions, but often degenerate into personal attacks and hasty generalisations, with the figure of $29 billion as the cost to innovation – a number arrived at by Boston University – frequently used. While various authors have discredited this figure and the underlying methodology, it still continues to find an audience. When comedian John Oliver hosts a television programme on so-called ‘patent trolls’, one must question the veracity of the underlying story.
Innovation impairment myth
Proponents of patent reform argue that reform is necessary because patent litigation represents a tax on innovation and stifles innovation. However, a reasonable search into product developments from proponents of patent reform yields little tangible support for this presumption. For example, what actual products did Cisco, Google, Amazon, Dell, Intel, Micron, Comcast or others refuse to launch because of fears about patent litigation? If these companies did indeed scuttle projects for such reasons, they should easily be able to present those projects – as well as the patents blocking them – for public inspection and debate. Further, a basic analysis into the heuristics associated with the patent data completely refutes any such arguments. This article presents some such heuristics.
First, there is empirical evidence to suggest the link today between patent assets and the market value of companies. A 2010 estimate of the US economy concluded that IP assets comprised 55% to 62.5% of gross domestic product. If patent lawsuits were indeed stunting the pace of innovation, we would expect to see a retraction of this ratio at a macro level (while one could argue that the economy could be growing faster, such analysis is remarkably specific and requires direct micro-economic measurement). Second, there is a much more direct measurement of the rate of patenting activity and innovation. Analysis of the patent data shows clearly that patent grant activity increased at a compound annual growth rate (CAGR) of 2.13% between 1976 and 2014, as the data in Figure 1 demonstrates.
Figure 1. US patent grant activity 1976-2014
Between 2004 and 2014 – the alleged period of innovation impairment, if one believes the reform narrative – patent activity actually increased at a faster CAGR of 2.68%. However, studying patent grants tells only part of the story, because of the natural time lag inherent in the grant process. An analysis of patents published between 2004 and 2014 indicates that patent activity increased at a CAGR rate of 13.29%. Such a pattern hardly indicates that litigation is stunting the rate of innovation. (It is not as helpful to study the patent application dates at this time because the last three years in the analysis tend to bias down the patent application activity due to the 18-month lag between patent application and publication.)
An additional metric to demonstrate the increasing pace in IP growth – besides cumulative patent filings – is the interval between successive patent filings. Figure 2 demonstrates the rate of change in time between each set of 250,000 patents that the United States has issued since 1976.
Figure 2. US patents granted by date interval 1976-2014
As the data indicates, the date interval between every 250,000 patents filed since 1976 shrank at an astounding rate. In 1976 the US Patent and Trademark Office (USPTO) issued Patent 4 million – it issued Patent 4.25 million on February 10 1981 or 1,505 days later. By 2015 the interval between every 250,000 patents issued was just 301 days. The USPTO published over 380,000 patent applications in 2014, compressing the timeline even further. In other words, the United States is creating an increasing number of patents and is issuing them at five times the rate it did in the late 1970s and early 1980s, contradicting the notion that patent litigation stifles innovation – at least from the perspective of companies building patent portfolios to protect their innovations.
Next, when one examines patent litigation activity and patent prosecution activity, the correlations are largely positive. Consider Microsoft: the software giant has done a remarkable job building its patent portfolio and then licensing it to others in the market. Its core entity received fewer than 100 patent grants in 2000, yet today it is one of the most prolific prosecutors of patents in the US system. A quick data analysis reveals why. Between 2004 and 2009, Microsoft was a defendant in many patent cases. When one cross-references Microsoft’s patent litigation history as a defendant against its patent application activity over the same period, the data is telling, as Figure 3 demonstrates.
Figure 3. Microsoft patent applications filed versus patent suits filed against
Figure 4. Microsoft correlation increased patenting activity and increased litigation activity
For example, there is a strong correlation between Microsoft’s increased litigation activity and its increased patent activity. As Figure 4 shows, Microsoft’s patent litigation record explains some 91.24% of the increase in its patent activity over the observed period. Microsoft appears in some ways to have adopted an aggressive patent position for defensive reasons. There is remarkable validation in the procurement of intellectual property as a defensive measure. In 2011, a consortium including Microsoft, Research In Motion, Apple, Ericsson, Sony and EMC outbid Google for a collection of over 6,000 patent and patent applications owned by the bankrupt Nortel Networks. To help bolster its own defensive posturing, Google bought more than 1,000 patents from IBM a short while later to build out its patent portfolio.
However, while Microsoft owns a robust patent portfolio comprising tens of thousands of assets which it uses to protect its own products and deter litigation, the company also licenses its assets to competitors in the market. Microsoft initiated a licensing programme for its patent portfolio in 2003 and has since executed more than 1,100 licences to other companies in the market for its patents. For example, Microsoft garners significant revenues from the Android operating system because it originated many of the key inventions that make up the system. As a result, Microsoft has entered into licensing agreements with many manufacturers that sell Android devices (arguably, one could label Microsoft a ‘patent troll’ as it relates to Android because it produces no Android product that it sells in the market). Samsung, the largest seller of products that use Android, pays Microsoft over $1 billion a year for access to patents which cover technologies captured within Android. Apple and Google both implemented similar strategies, as did other major rights holders in the US system, which one can verify by searching the US patent application, grant and assignment databases.
What this data tells us is the exact opposite of what patent reform advocates claim. Contrary to popular belief that patent litigation hampers innovation, it in fact appears to bolster it. Patent monetisation efforts through licensing or assertion have increased the liquidity of IP-related assets and attracted capital and inventors alike. Empirical evidence indicates that the actual liquidity for patents in the United States may have increased, in part because of the growth in litigation and alternative monetisation pathways. This is apparent if one studies the rate of assignment of patents and the rate of patent grants. When the rate of patent assignments tends to exceed the rate of patent grants, that data provides evidence of a more liquid patent marketplace because assets are changing hands among willing buyers and sellers at a pace that is greater than self-development efforts. Figure 5 illustrates this phenomenon.
For most years since 2000, the rate of assignments has outpaced the rate of patent grants, as indicated by the shaded region. The pace of change took a dramatic turn in 2014, which coordinates neatly with many of the recent key changes in patent law and the number of headlines touting the contraction of the patent marketplace. What this data suggests rather concisely is that recent changes have depressed the assignment of patents. While grants continue to increase, further analysis reveals that companies tend to benefit from the increases as opposed to individual inventors, as the data in Figure 6 suggests.
This shows that patents with no known assignee at the time of grant (ie, an individual) have dropped to just 6% of all patents granted in any given year, down from nearly 20% in the late 1970s and early 1980s and about 10% as recently as 2000. Given that the top 300 patent holders generate more than half of all patents in the US system, the data suggests that individual inventors need greater assistance through structures that enable the liquidity and monetisation of ideas from the innovation economy, not greater barriers courtesy of litigation reform.
Defendants in patent litigation often complain about the cost of patent litigation, yet their actions are often the reason for this protracted and expensive litigation. The balance of delays can result from plaintiffs having unrealistic value expectations for their inventions, which removes the defendant’s motivation to settle.
Patent litigation is the sport of kings, given its complexity, cost and labour intensity. Patent assertion entities (PAEs) that sue larger companies maintain extraordinary amounts of cash to demonstrate their capacity to pursue litigation to whatever means necessary. As an indication of this, the net cash component of PAEs’ stock shares that trade in the public markets often exceeds 30% to 50% of their stock price. Despite their claims and ability to exit cases quickly, large companies routinely overspend on litigation in an economic war of attrition. Empirical evidence supports this trend of overspending. An analysis of 800 patent cases in DocketNavigator between December 29 2010 and May 7 2015 where there was a finding of damages demonstrates this concept clearly. Figure 7 captures a frequency histogram of damages by amount.
Figure 5. Patent assignment rate versus patent grant rate in the United States
Figure 6. Individual versus company ownership of US patents at grant
As the data in the chart suggests, if there is a finding by a judge or jury, the amount of damages will generally be quite small (ie, well below the headline-grabbing range). In fact, there is less than a 50% chance that any patent litigation will generate a damages award greater than $1 million and less than a 30% chance that damages will go above $5 million. According to 2013 American Intellectual Property Law Association statistics, companies are willing to spend an average of $3 million to defend claims of damages that range between $1 million and $25 million. This basic economic disparity between spend and economic risk highlights the level of overspending by defendants and the economic war of attrition going on, and deserves further inspection.
Presume that the average patent litigation lasts for 26 months and costs $3 million to get through trial. Presume further that the expenses amortise evenly over the 26 months (ie, $115,385 spent per month on average). With a corporate discount rate of 15%, the present value of the defence spend is $2,573,308 (determined by calculating the present value of a 26-month annuity payment of $115,385 using a geometric monthly discount rate that equals 15% on an annual basis). To justify the litigation defence strategy, the value at risk must exceed $2,573,308 on average; otherwise, company management destroys shareholder value.
To test this, one must also study the value at risk. As discussed, empirical evidence suggests that the median damages award is about $1 million, payable about 26 months after the initiation of a complaint. With this data, the present value at risk can be calculated; only a discount rate is needed to finish the calculation. To calculate that discount rate, a peer-reviewed formula can be used to calculate a risk-appropriate discount rate. The formula considers the following items:
- the target rate of return for the investment;
- the holding period for the IP investment; and
- the systemic success rate of the litigation strategy
The formula is as follows:
Recall that the corporate hurdle rate (ie, the target rate in the formula) is 15% and the holding period is about two years. If the defendant had a 50% chance of prevailing in court, then the indicated discount rate using the formula would be about 58%. With a discount rate of 58%, the present value of the median damages award is $369,367 (determined by calculating the present value of a 26-month delay on a payment of $1 million using a geometric monthly discount rate that equals 58% on an annual basis). Based on this example, the defendant spends $2,573,308 on a present value basis to neutralise a present value at risk of about $369,367. The level of overspend is about seven times the value at risk and the defendant overspends by about $2.2 million for each case. By any reasonable measure, this strategy is an economic loser and destroys shareholder value. In one historical example, the Recording Industry Association of America (RIAA) pursued a similar litigation strategy from an economic perspective, initiating over 35,000 lawsuits from 2003 through 2008 in an attempt to combat rampant sharing of music files on peer-to-peer networks. In what became an utter PR disaster (it wound up suing the elderly, children and even a dead person), the RIAA ultimately abandoned the strategy in December 2008 – but not before incurring costs of over $100 million.
Further, the losses scale linearly (ie, the more lawsuits, the worse the shareholder value destruction). Thus, if the company is a defendant in 10 lawsuits, it effectively spends about $25.7 million defending a value at risk of about $3.6 million. Who rationally spends their own money in this way?
The argument proffered by large defendants is that a strong defence acts as a deterrent to future litigation activity. Spending more and being difficult or burdensome in the process reduces the incentive for plaintiffs to bring new cases. However, empirical analysis of the litigation data refutes this fact, as the litigation rate is fairly stable over time. Has the litigation rate dropped off for any of the pro-reform parties because of a vigorous defence strategy? The answer is no, according to data in DocketNavigator. Apple is one of the most sued companies as it relates to patent infringement in the US system, yet Apple’s reputation for being difficult in litigation has done little to reduce its litigation risk. We see similar strategy outcomes in personal injury lawsuits and securities class action lawsuits in the United States, which certainly have not declined due to vigorous defences.
The argument proffered by large defendants is that a strong defence acts as a deterrent to future litigation activity. Spending more and being difficult or burdensome in the process reduces the incentive for plaintiffs to bring new cases
Admittedly, one may reasonably argue that a company’s vigorous litigation defence strategy tempers the growth rate of litigation (ie, a company’s litigation rate would have been even worse were it not for a vigorous defence). However, the side effect of a disproportionate defence strategy is that case quality should rise, because the economic risks of failure increase with each change in law (eg, fee shifting). The net effect of this outcome is that defendant behaviour would erase any incentive that plaintiffs might have to settle. This is because the interim costs to get to a judgment force the plaintiff to go further in the litigation process in order to generate economic returns from the assets. The longer and more complex the litigation, the greater the recoveries that the plaintiff must seek. Economically, a plaintiff is thus better off avoiding settlement and proceeding to trial if there is any material finding in its favour, because interest accruals and the elimination of any settlement discounts offset the risk-adjusted time value of the payment delay. Further, as case quality improves, the defendant has lower chances of success on the merits in court.
Next, consider the fact that the inventor typically does not receive all of the economic benefits of pursuing litigation, as litigation costs consume a significant portion of the proceeds. For example, in a recent case where Rembrandt Wireless Technologies sued Samsung and won a $15.7 million judgment, the inventor received just 2.5% of that amount, or about $392,500. If the total licence value was only about $400,000 to the inventor, then Samsung would have been much better off to short-circuit the litigation process and pay an equitable licence fee upfront. In fact, after considering the costs of paying attorneys and expert witnesses along the way, which can easily consume 40% to 50% of most settlements or judgments, the licensee may easily be better off with an upfront licence as well.
Figure 7. Frequency histogram of damages occurrence by amount
Many small patent owners have no real understanding of the litigation process and the associated costs, and are surprised at how little inventors will actually receive. Rights holders are frequently advised that the patents that they spent tens of thousands of dollars developing are effectively worthless, either because of a changing legal standard or because the costs to generate economic returns are so high.
On the other side, we find that large players have a well-earned reputation for being difficult to work with. For example, Google bid $4.4 billion for the Nortel patent portfolio because it wanted access to inventions contained in it, but lost the bid to Rockstar Bidco. Yet Google later rebuffed Rockstar Bidco when it requested a licence for Google to access the patented inventions within the portfolio, forcing Rockstar Bidco to sue it. While both parties settled the case within a year, Google’s reputation for vigorous defence and the year-long litigation certainly did not dissuade Rockstar from pursuing its licence, any more than Cisco’s litigation defence reputation and strategy had a noticeable impact on its dealings with Rockstar (where Cisco ultimately settled for about $188 million).
There is an interesting side effect to defensive posturing and behaviour: precedent. Interestingly, had Google or Cisco dealt privately and equitably with Rockstar upfront (ie, not forcing Rockstar to sue in a public forum), the mere existence of a licence and the associated amounts would have remained in the shadows, open only to speculation. However, the aggressive positioning of defendants has created a rich database of heuristics and behaviour. PAEs today use that data to learn from those cases and adapt based on experiences from previous trials. It stands to reason that some portion of these heuristics and behaviour is unflattering to the defendant position (eg, sanctions awards, inequitable conduct determinations, trial transcripts demonstrating bad behaviour or damages theories), which may be helpful to later litigations. None of those facts would have ever seen the light of day except for patent litigation
The effect of overzealous defence behaviour, unintended as it may have been, has been to force inventors to pursue patent monetisation in other forms – hence the rise of the non-practising entity. The result is that many larger companies, because of an unreasonable posture in licensing discussions, actually created the patent assertion problem that they now deride and seek to reform. We have a direct proxy for this outcome in other areas of the law. For example, personal injury law and securities class action law operate mainly on a contingency basis (as do many patent litigation firms). Their strategy is to sue first and ask questions later, which is why we see personal injury claims and securities class actions filed within hours of major events. That activity is a direct result of large companies and insurance companies forcing injured parties to find equitable remedies in whatever form is economically available.
As Congress continues to debate patent reform, it is imperative to do so from a reasonable and factual basis. An objective analysis of the heuristics demonstrates that patent litigation has had little impact on the rate of innovation in the United States. However, changes in the legal certainty of patent assets have clearly done a remarkable job in stifling liquidity in the patent marketplace, which is unfortunate. Whatever reform Congress pursues, it is clear that this will only change the rules of the game – Congress and the small number of companies lobbying for patent reform will not change the basic economic principles that govern IP investment. The next generation of patent monetisation against large accused infringers is likely to morph into something much more burdensome than what any of the reform proponents have likely contemplated to date.
Many of the perceived problems with today’s US patent litigation system have their roots in unrealistic expectations, posturing or just a plain lack of factual knowledge on the part of both plaintiffs and defendants. Here is some advice for both parties.
- have realistic expectations – do not be unreasonable or greedy;
- few litigated standalone inventions generate values of over $5 million, so do not go into negotiations asking for $50 million;
- understand the true, risk-adjusted economics of the litigation outcome;
- you have a 50% chance that your nominal damages will be less than $1 million;
- on a risk-adjusted, present value basis, your actual recovery will be much less;
- even if successful, a significant portion of the value will go to intermediaries; and
- using that true, risk-adjusted understanding of the litigation outcome, set reasonable expectations for patent value before going into a licensing discussion or mediation.
- stop with the hyperbole;
- except perhaps in the pharmaceutical space, few companies have cancelled any material innovation programmes because of patent litigation risk;
- reconsider the true value at risk and alter your thoughts on inbound licence requests;
- offering a fair deal to an inbound request on an important patent asset is not a sign of weakness, but of economic prudence, practicality and equity;
- remove emotion from decisions regarding litigation defence – such decisions should be based on fundamental economics, as anything less destroys shareholder value; and
- dealing pragmatically with inbound licensees in private masks defendant behaviour, suppresses associated heuristics analysis and reduces the risk of dirty laundry making its way into the public view.