The long and the short of it

Matt Stack, XLP Capital

Matt Stack – a co-founder and managing director of investment firm XLP Capital – has profited throughout his career from both backing IP-rich businesses and from turmoil in the US patent system. His career track record makes Stack part of the small group of investors who marry their investment acumen with a deep knowledge of IP monetisation and someone who has seen both the positives of IP value creation and how changes in the legal and regulatory climate have sown uncertainty and fear.

At Boston Consulting Group, where he started his career, Stack quickly developed an appreciation that IP-rich companies were among the best-performing investments. He took that experience with him to his next roles at a venture fund and a technology licensing private equity business, spending much of his time investing in IP-rich university spinouts from the likes of Harvard, MIT and Princeton. He then moved into private equity with a fund called 3LP where he started investing in standalone patent portfolios across the United States and parts of Asia.

It was there that he started to appreciate the patent market bubble that developed through the early 2000s and which reached its height with the 2011 auction of Nortel’s assets, which were ultimately acquired by the Rockstar consortium. “We saw prices, peaking roughly around the time of the Nortel transaction and the Micron Round Rock deal, which represented both the high point and the low point of the IP market, in close proximity,” Stack recalls. “The Nortel deal was the largest IP deal done to date and the Round Rock deal was effectively an off-balance sheet, unreported event valued at nothing. So, we thought that was really interesting: intellectual property was both hugely valuable and devalued at the same time.”

Together with a few colleagues Stack left 3LP in 2015 to form XLP Capital, an equity-focused short fund where he started to profit from the turmoil hitting the public IP company space. “We noticed that one of the most compelling arguments was that any publicly traded company holding patents would be a fantastic short opportunity,” he explains. “That’s what we’ve been doing for some time and have since expanded into long-short credit, volatility and data ventures.”

Stack’s experience has given him an excellent view of how changes in US patent law – particularly around patent-eligible subject matter and the introduction of inter partes reviews – have affected the investment proposition in IP monetisation. Those changes helped to pierce the patent bubble and Stark paints a fairly dismal picture of the long-term value creation opportunities currently available for parts of the patent market.

However, he insists that there are still intangible assets (eg, the monetisation of companies’ non-core data) that remain attractive to investors. “I anticipate, and I believe that we will see, the meteoric rise of data itself as an asset over the next 10 to 15 years, in a similar way to how the patent bubble rose and fell,” he insists. Here he talks to IAM’s North America Editor Richard Lloyd to throw some more light on the IP investment climate and why IP leaders should be focusing on large datasets as the next big value creation opportunity.

RL: From your perspective what do you pinpoint as being particularly destructive to those looking to monetise their intellectual property?

MS: As with many assets, in capital equipment assets there is a lifecycle. So, in the early stages of a new piece of equipment or a new model of a car there are front-end buyers that will buy and pay retail margins. For a brand-new edition of a car, there is usually somebody who will pay a premium for an early prototype. Then the asset goes mainstream and prices go down as margins are tightened. However, there is always a tail end of the value chain – for instance, not all of the cars that were driven in the United States in the 1980s ended up in landfills or dumps. Instead they were sold and are driven in other countries and regions (eg, India, South America and Africa). So, cars as assets have this elaborate lifecycle, where even though the average US consumer sees a car for only a few years, it can remain an asset for as much as 15 to 20 years.

Patents have the same lifecycle stages. So, in the early days patents are traditionally valued very highly. In brand-new technologies (eg, three-dimensional printing or solar or wearable electronics), early-stage patents can be extremely valuable as the technology is just starting out – even one or two patents can be worth tens of millions of dollars.

As the lifecycle of that patent evolves, what you see is the patent becoming more of a commodity asset. Today a lot of different companies have these types of patents, which they cross-license with each other. However, when it comes to the final stage at the very end stage of the patent lifecycle, many of the companies that had attempted to build their own products no longer exist. So, that tail end of the asset value has become critical. And just as the terminal value of a car is whatever you can sell it to somebody, the terminal value of a patent is what it can be licensed for.

What happened in the patent marketplace is that when licensing – specifically litigation – became more difficult with the passing of the America Invents Act, the Alice decision and the introduction of inter partes reviews, the entire tail end of intellectual property effectively became worthless. And when you take the terminal value out of the equation of an asset’s longevity, this has a profound impact on its value.

If you look at private equity, when you buy and sell companies it is often true that the terminal value of a company in a discounted cashflow is as much as half (if not more) of its enterprise value. For a car, the terminal value can be up to one-third of its value. The terminal value of a patent, given the asset’s overall lifespan, applies for between 10 and 15 years and can be worth anywhere from one-half to two-thirds of the entire value, sometimes more.

Figure 1. The core mechanism behind tech investing at a macro level

Introducing the inter partes review system effectively destroyed the secondary market for intellectual property, almost overnight. The patent acquisition market effectively deteriorated and disappeared within between 12 and 24 months and we saw licensing volumes decline within three years. The inter partes reviews system took between one-half and two-thirds of the value out of patent valuations by destroying the terminal value.

RL: There are some who would argue that inter partes reviews are doing a good job, the early statistics around patent invalidations have come down and they are not the death squad that they were initially made out to be. So is there a risk that you overstate the impact of inter partes reviews or can you just give me a little bit more detail around how you see this destruction of value from an investor’s point of view?

MS: A lot of the stats around inter partes reviews are all just window dressing for investors. These are measurements of something that does not actually change the prospective valuation. As an investor, you do not want any kind of system under which you could spontaneously lose an asset. If you look at the credit markets, there is a general and widespread fear of bankruptcy – even though bankruptcy rates are traditionally very low. However, bankruptcy is such a traumatic event to a bond holder or a credit holder that there is an entire ecosystem of service providers which specialise in liability insurance or default insurance and which are designed to protect bond holders from exposure, risk or loss of income due to a bankruptcy. You can also typically see bankruptcies coming, signalled by unhealthy industries and cyclical movements across sectors. In addition, one develops a good sense for when companies might be approaching bankruptcy, so you can hedge or rotate your exposure.

Figure 2. Number of patent transaction deals over $100 million / year: 2012-2014 marked the peak of the IP transaction marketplace

In the patent world, you have very little ability to predict when inter partes reviews will happen. An inter partes review is to an IP investor what bankruptcy is to a credit investor, with the difference that it is unpredictable and unforecastable. If you look at the capital market economy, unpredictable bankruptcies at the company level are catastrophic events. You see this with some of the large tech bankruptcies – there was little warning so people had no chance to insure, to put in hedges or to diversify their exposure to an event capable of eliminating fortunes. From a wealth management perspective, you never want to be holding debt or credit during a bankruptcy that you did not anticipate and where you had no time to manipulate or manoeuvre your exposure.

Figure 3. Fall of centralised APIs and the rise of decentralised developers transforms how data is managed

An inter partes review is simply the threat of a spontaneous bankruptcy-like event for an IP rights holder. If it does occur, it is usually at the stage that the patent’s value is finally ready to be asserted. You do not need to have large numbers of inter partes reviews and you do not need to have a large number of patents killed for the mere threat to completely wipe out and eliminate the secondary market for intellectual property – and that is what we have seen.

RL: So, if inter partes reviews are to be found unconstitutional, will a lot of investment money come flooding back into the space?

MS: Define ‘flooding’. I think enough investors in intellectual property have been burned in recent memory so there will need to be new money and this will likely be less of a flood, more of a slow trickle. Also, over the last 10 years many of the professionals who were skilled at licensing and who struck some of the early licensing deals have moved on to other fields. If licensing and the value of intellectual property are to be restored, I would expect it to take between three and five years to ramp up and, perhaps, another decade to reach levels that we saw five years ago.

RL: If you listen to one side of the patent debate in the United States, you hear comments that inter partes reviews and various cases concerning patent-eligible subject matter are improving the system because they are weeding out a lot of poor-quality patents and, ultimately, we will be left with a much stronger pool of intellectual property. As an investor, do you not think that is a good thing, which will strengthen the asset base?

MS: Well, I think this is where we have to focus the answer around software patents. We have a small number of large tech companies that did not like software patents and therefore were eager to see them invalidated. To some extent, the elimination of the IP threat has driven high degrees of consolidation and scale in the software and web service world and thus it is no surprise that in the top 10 companies in the world by market cap there is a stronger representation of software tech companies today.

Figure 4. Data deals are the new IP transactions: exponential rise in data deals over $10 million in the past few years

I think the pendulum is beginning to swing the other way. Many of these software companies are beginning to regret their decision to be so anti-intellectual property, specifically as they develop intellectual property in machine learning. I look at many of the novel machine-learning algorithms out there and I think that the arguments about how machine learning might replace jobs might actually be discarded if you could patent machine-learning algorithms and then license them to other people. There are novel approaches in machine learning – specifically in forecasting methodologies, image detection and recognition – that bear all of the hallmarks of being industrial pieces of technology, which is the very thing that should be patentable. So, instead of throwing these into the public domain immediately, I could imagine them being patentable and enforceable, with new fortunes and new companies being built around this machine-learning software.

RL: Let us move away from the United States. Are there other geographies where you are long on the IP system and long on straight licensing or IP-rich businesses?

MS: Well, I think when you look at the technology industry and the difference between the tech industries in Japan and China, it is fascinating to compare. I would say that the Japanese have a very strong patent system and therefore it thrives in sectors such as semiconductors and automotive.

You would predict that a Japanese patent system that is strong in this way would continue to do well, to the extent that those industries do well. On the other hand, you look at China where the patent system encourages copycat and public innovation in a different way. Physical innovation and physical part innovation are highly leaky and patents are not necessarily useful for parts. However, they are useful for integrated systems and for outcomes, where patents are somewhat stronger. In fact, as a technology integrator, I think that the Chinese IP system may actually scale up better over the next decade. It may turn out that the subtleties in the Chinese system are better at protecting highly complex, integrated systems than the US and Japanese systems. While the Japanese IP system encourages subcontracting, have-made rights and therefore strong patent licensing, the Chinese market encourages tear down, reverse engineering and process trade secret disclosure contracts.

You have to evaluate the strengths of any IP regime in the context of the industries that are doing well under its purview. There is an open question as to whether different IP systems globally will incentivise certain industries to perform well. Which is the cart and which is the horse?

RL: You mentioned the opportunities that you are seeing in the data space. Can you give us a bit of insight about what you are seeing there? What are the kinds of assets that are being traded and where is the value being created?

MS: In the last decade, particularly when I was at 3LP, I executed dozens of patent sale agreements and saw hundreds of licensing agreements thereafter. That was at the peak of the patent licensing and patent sale economy. Today at XLP Capital, I see that same volume of deals – not in patents but in data licence agreements. Today, for instance, I am involved with about a dozen large data licensing deals where I am brokering exclusive datasets from companies that are in one business, but realise that they are generating data in completely orthogonal industries in the same way that people holding patents 10 years ago would happily license them out to businesses that were not directly competing with them (so-called ‘non-core’ businesses).

Today we ask CEOs and investors the same question: what is your core data and what is your non-core data? If as a company you have a non-core data asset, we at XLP are eager to buy, license or resell it. In fact, we are now almost exclusively in the business of buying, selling, auctioning, brokering and relicensing non-core but valuable datasets. Data is the new IP asset and every company has it.

I will say that only the most progressive companies today have regarded their datasets in this way and are beginning to make IP managers responsible for non-core data management. I see this trend accelerating and I would anticipate that the next intellectual asset of significance that IP managers and IP executives will manage is non-core data. First, by identifying what data there is. Second, by identifying what is strategic, what is core and what is non-core data. And, finally, coming up with valuation techniques for all three of those groups and determining whether it is a good idea to sell or to monetise non-core data as an asset.

RL: Do you think that this market might be subject to the same kinds of wild swing that we have seen in the patent market over the last seven or eight years?

MS: Swings or not, I think it is clearly the next frontier of value and asset value realisation. And so, it falls to the intellectual asset manager or the IP manager to evolve with the state of the market and with the supply and demand of intangible and intellectual property. We have assessed the IP bubble to have lasted somewhere between 20 and 30 years.

In the same way, I think data licensing agreements are the next frontier. Almost every company I have encountered has non-core data and yet does not know how or struggles to manage it. If data licensing agreement transactions turn out to be a bubble, then it could be one that lasts for between 20 and 30 years. But that does not mean that it should not be on the radar of every IP manager.

RL: Do you think there is a role for the US Patent and Trademark Office and other IP offices to police this process?

MS: I think there is a dual role. There is an opportunity realisation role, which includes valuation and identification. It may be time to create a classification system for types of data, in the same way that technologies are classified under US, international or cooperation patent classifications. I also think there is a policing role. Many of the same roles that were needed in the IP patent creation and management process will be necessary here. In the same way that on the patent side there was a licensing or at times a litigation role for patent managers, I think that those same people will have an important part to play in detection and litigation when it comes to dataset leaks.