The nascent patent market in the automotive sector

It is hard to turn on the television today without seeing commercials for automobile manufacturers touting technological features in their vehicles, such as in-car built-in WiFi (Buick), heads-up texting (Volkswagen’s 2016 Passat), an active collision-avoidance system (Audi) and a mobile office (Chevy’s Silverado pickup truck). While some of these features require new in-vehicle hardware, many are merely software features enabled in one or more of the dozens of embedded control units already sitting inside the vehicles. These embedded control units manage various systems within the vehicle, including:

• engine;
• transmission;
• anti-lock braking systems;
• cruise control systems;
• airbag safety systems;
• entertainment systems;
• climate control systems; and
• instrument clusters.

Some estimates claim that today’s vehicles have up to 100 million lines of installed software source code. Coincident with this growth in technology integration is the interest in new patents in the automotive space. This pattern also applies to heavy equipment manufacturers. For example, Caterpillar received more patents in software-related art units between 2011 and 2016 than it did in art units relating to internal combustion engines or power plants. While auto manufacturers integrate inventions from many scientific disciplines, the direct volume of patent activity in the automotive space generally tends to follow other sectors such as pharmaceuticals, semiconductors and software – although patents in the automobile space tend to be worth less.

This is not to say that automobile patents are worthless. Perhaps one of the most interesting automobile patent stories (aside from Robert Kearns’ fight over the intermittent windshield wiper with Ford and Chrysler) dates back to 1969, when General Motors paid $50 million for a licence to the Wankel rotary engine design. To put this into perspective, that $50 million initial licensing fee today would be worth about $311 million – a magnificent amount for an initial licence fee and quite unusual outside the pharmaceutical, medical device and certain semiconductor sectors. So what creates the draw to the automotive space and do the economics warrant what might be a new gold rush for patents in this area?

Making a car

To answer the question of where patent value lies in a car, one must first look at what constitutes a car from top to bottom. The major cost drivers include raw materials, labour, administration, advertising and R&D (see Figure 1).

Figure 1. Total composition of all costs in automobile production and marketing

Given that raw materials (eg, steel, iron, aluminium and plastic) comprise 47% of the average car, there is little opportunity for apportioning IP value to those items. The same could be said for direct labour, logistics, administration, depreciation and advertising. Patent rights typically manifest in the R&D component of the vehicle, which comprises about 6% of the total. However, this hardly accounts for the specific IP content that exists at the system level for automobile original equipment manufacturers (OEMs). So while the cost apportionment directly affects what OEMs can pay for patent rights, the raw cost apportionment does not highlight where high-value opportunities are actually situated in the vehicle.

For years, Merrill Lynch has tracked and issued research reports on the source of value within cars in a report called “Who Makes the Car” (last published in 2012), which provides a detailed breakdown of the cost centres in a car by system type. Table 1 summarises the details of this.

As the data in the table clearly demonstrates, the value of the intellectual property at issue depends on where it is located. Many high-tech features that appear in vehicle commercials today tend to be concentrated in the audio and telematics systems (while a subset of certain advertised features also appears in the electronics and electrical systems, that grouping also includes items such as light switches, headlights and wiring harnesses). So if the audio and telematics systems represented by 250,000 patents or more (eg, mobile communications units, GPS locators and speech recognition systems) tend to reside inside the audio and telematics units, it becomes clear that the total amount of economic apportionment to patented inventions within those systems is low. However, this is not to say that the costs are low. Rather, OEMs may be underpaying for access to such inventions. One explanation for this relates to the cost leverage that OEMs exert on Tier 1 suppliers, which cannot capture economic value attributed to the rights associated with such patents through pricing premiums. Put more succinctly, GM may require its Tier 1 supplier Delphi to swallow the cost of any royalties due to a rights holder suing Delphi or GM for patent infringement. Delphi cannot raise prices on the parts that it sells to GM to account for the new royalty costs to an external rights holder. In the end, both the rights holder and Delphi lose in this business arrangement: the rights holder fails to capture the true market value potential from its innovations and Delphi pays the patent infringement royalties and defence out of its own profits. Ultimately, the rights holder’s value tends to transfer to GM and consumers who buy GM’s cars without it receiving equitable compensation.

Automotive value chain problem

The inventor/rights holder equity problem in the automotive industry begins with the value chain. This is ruthless about squeezing costs out of the entire manufacturing system and transferring risk. The hyper-competitive nature of the market forces this relentless cost pressure onto OEMs. If one couples the cost pressure with a supplier’s willingness to take on litigation risks for patent infringement actions through indemnity requirements, it becomes clear where the economic limitations are. The unfortunate impact on a rights holder is that sharing industry profits can be difficult unless the rights holder can make it to the Tier 1 or OEM level. Consider a hypothetical company which develops a new thermoelectric chemistry which is more efficient than others currently on the market. The contemporary applications of the new chemistry may arise in the heated and cooled passenger car seats that OEMs build into medium and high-end vehicles today. The problem for the rights holder is that there are so many companies in the value chain, from the raw materials producer to the OEM. Figure 2 gives one example which captures all of the players in an automotive industry value chain.

Figure 2. Climate-controlled seat value chain

For example, suppose that Gentherm sells a climate-controlled seat (CCS) unit for about $70. It reports that direct costs comprise 70.21% of revenues, or about $49.14 (this and other cost examples are based on figures reported in corporate reports and regulatory filings). Each CCS unit contains several components, including thermoelectric modules, a heat exchanger, a blower motor, an electronic control module, assembly labour and ductwork. Suppose further that the allocation of the costs in the thermoelectric module follows the schedule in Table 2.

Figure 3. Climate-controlled seat value chain – Gentherm to rights holder

Under this allocation, the thermal module costs about $10.44. Gentherm purchases its thermal modules from Ferrotec. Ferrotec’s direct costs are 73.8% of the product value, or about $7.70. A report from the University of California Berkley and Lawrence Berkeley National Laboratory identifies the typical cost structure of a thermoelectric module. According to the report, the material portion of the breakdown refers to the cost of thermoelectric material, which is 17%. Of the module’s $7.7 cost, about 17% of the costs relate to materials. Thus, materials cost represents about $1.30 of the module’s costs. Ferrotec buys those materials from a chemical company. The chemical company’s direct costs are about 72% on average, leaving 28% as gross profits – about $0.36. If the licensee were to receive 33.33% of those gross profits for a licence fee, it would earn about $0.12, or about 0.17% of the total value chain up to Gentherm. Figure 3 shows how the value allocates from Gentherm down to the rights holder.

Gentherm would have to sell more than 8 million units for the rights holder to earn $1 million in licensing fees – and that assumes that the licensor (ie, the chemical company) would want to part with 33% of the gross profits.

As the example demonstrates, these economics are not compelling, especially if the rights holder invested millions of dollars in R&D for the new thermoelectric chemistry to begin with. Moreover, even if the market embraces the technology, the rights holder is so far down the value chain that it would reap little proportional reward for this. The companies that make the most in this deal are the OEMs (which charge a significant premium for heated/cooled seats), Johnson Controls (which sells the integrated seats) and Gentherm (which makes the CCS units). Everyone else in the value chain is a small player.

Tesla patent conundrum

In June 2014, Tesla CEO Elon Musk penned a blog post announcing that Tesla will not initiate offensive litigation against anyone that uses its patents in good faith. Wall Street types pulled out their hair, while the Harvard Business Review even wrote an article about it. In economic terms, a patent, when wielded appropriately, allows its owner to restrict the supply of its patented invention to the market. With sufficient demand, the patent owner can then raise prices for access to the patented product, enabling it to price skim. This basic economic property is the foundation of innovation in the for-profit context and is why consumers must pay more than the true cost of manufacturing the product when the utility of the invention provides compelling value to the market. Absent typical market distortions (eg, government intervention), this process works remarkably efficiently.

The economic value that Tesla can generate from its portfolio generally falls into one of four categories:

• Direct exploitation – Tesla generates direct value from its portfolio in the marketplace by selling products that embody the patents in the market.
• Licensing – Tesla generates value from royalties paid by authorised users of its portfolio.
• Cross-licensing lever – Tesla generates value by using its patents as a cross-licensing lever to avoid paying royalties to other parties that might choose to assert their patents against it. This value source tends to provide Tesla with a certain degree of design freedom and cost savings.
• Blocking value – Tesla generates value blocking its competitors from selling competing products, thereby driving demand for its products instead.

As far as patent portfolios go, Tesla is relatively young and provides the company with a reasonably long period of protection, as the mortality schedule in Figure 4 shows.

Figure 4. Tesla Motors’ patent mortality schedule

As is clear from the mortality analysis, Tesla’s patents do not start to expire until 2026 or so. Moreover, it appears that Tesla is a good steward of its existing patents. Despite not enforcing its patents, Tesla continues to maintain them, which if carried to full term costs the company hundreds of thousands of dollars per year on average, as data in Figure 5 demonstrates. So why did Musk take the approach that he did?

Figure 5. Tesla Motors’ projected patent maintenance budget

Despite the initial negative reactions to Tesla’s actions, a closer look shows that Musk may not have damaged his company as much as some initially believed. In fact, basic economics suggest that in the automotive context, Tesla’s patent portfolio may have a relatively low value. While Musk has said that his actions were designed to spur growth in the market, it is also true that defending its patent portfolio would only have lost Tesla money.

Tesla’s enforcement economics

A review of the enforcement economics of Tesla’s patent portfolio highlights why it may have little value in an automotive context. The fruits are not worth the trouble of enforcement and the positive economic benefits are tenuous at best. First, consider some simplifying assumptions. Tesla is the defendant. Quarterly vehicle sales for the Model S are as laid out in Table 3 based on its reports.

As the data in Table 3 demonstrates, the total applicable market on a historical basis is about 90,000 vehicles. To demonstrate the diseconomy of a Tesla lawsuit, consider the following parameters:

• The intellectual property of interest resides in the drive unit – the smallest saleable patent practising unit.
• The drive unit costs about $15,000 (see www.greencarreports.com/news/1093713_tesla-model-s-drive-unit-replacements-how-big-a-problem/page-2).
• The apportioned value of the hypothetical patent in suit to the drive unit is 10% of the $15,000 cost or about $1,500 (recall that a good chunk of the value in an auto part is raw materials, so the amortised value of the intellectual property in the unit may be proportionally smaller than in other markets, such as pharmaceuticals).
• The royalty rate is 3% of the value of the drive unit.
• The total royalty is $45 per drive unit.
• The total nominal damages potential, not including interest, is $4,043,925 (based on 89,865 units at $45 per unit).

Cross-checking this outcome against empirical results suggests that the economics are worse even than this optimistic model. The median damages award at trial is about $1.3 million over the last several years – yet the average cost to get to trial at this level is about $2.8 million (see my previous article “An Inside Look at the Innovation Impairment Myth of Patent Reform”, IAM, Issue 74, November/December 2015). If a rights holder has to sue Tesla, the outcome may indeed be worse than not filing an enforcement action in the first place.

Further, even if a rights holder succeeds in obtaining a district court judgment, Tesla can appeal this to the Federal Circuit. The appeals process adds time and cost to the case, and incidences of district court judgments being changed or modified on appeal are remarkably high. PwC’s 2015 Patent Litigation Study found that, on appeal, the court changes or modifies a decision in district court judgments 52% of the time, modifies a judgment in whole or in part approximately 67% of the time and reverses an entire judgment 19% of the time. Overall, only 18% of appellate court decisions substantively address damages. Of those that do, 80% reverse, vacate or remand the damages in some way. Thus, even if the damages component survives, it could be years before the prevailing party receives payment. The numbers do not improve if the patented feature is in a lower-value part of the car. In fact, according to a review of cases against Tesla in DocketNavigator, the material cases all tend to relate to telematics applications, which drive a lower value in the car, as opposed to the drive unit and other higher-value systems.

In summary, aside from the design freedom that Tesla receives, Musk deliberately decided not to enforce the company’s patents because they are largely worthless in the context of manufacturing cars. However, Tesla’s decision may have created consequential damage for the company as it removes any lever for cross-licensing in the event of mass litigation. For instance, if Ford sues Tesla for patent infringement, Tesla has no recourse to compel Ford into a cross-licensing arrangement because it has deliberately chosen not to enforce its patents. The result is that Tesla could experience margin compression because of having to pay royalties to others for infringing their patents.

Optimal opportunities

The casual reader may surmise that inventions in the automotive industry come with low relative economic power. OEMs may read this article and use the economics to justify holding out on small rights holders that seek licences for inventions captured in vehicles. As is the case with many items in the patent space, the economics depend on the context. Anyone that truly wants to generate significant revenues from patents on a massive scale (ie, tens or hundreds of millions of dollars) must create a product that integrates the patented features on a massive scale (ie, tens of millions of units or more per year) and the market must demand this patented feature. Examples include industrial designs for iPhones, pharmaceutical pill compositions of matter and semiconductors.

In the context of the automotive space, the heuristics associated with valuable patents may include systems that have an associated government mandate, require high-tech components, have low raw materials concentration and pricing volatility, have high manufacturing margins and have replacement potential. Out of the many systems in a vehicle today, direct tyre pressure monitor system (TPMS) sensors serve as an excellent example of an opportunity ripe for monetisation. TPMS sensors are attractive as they exhibit the following attributes:

• Government mandate – in the late 1990s, tyre failure caused over 250 deaths and more than 3,000 injuries. In order to reduce such incidents, Congress passed the Transportation Recall Enhancement, Accountability and Documentation Act in 2000. This called for all passenger vehicles weighing less than 10,000 pounds to be equipped with TPMS sensors capable of warning the driver about underinflated tyres. By model year 2008, all passenger vehicles and light duty trucks weighing less than 10,000 pounds were shipping new from the factory with TPMS sensors. Some 17.5 million vehicles were sold in the United States in 2015. With four sensors per vehicle at an average OEM cost of $20 per sensor, the total market potential because of the mandate is $1.4 billion.
• High-tech components – the operational parameters for TPMS sensors are extreme, with the devices having to operate soundly and reliably in harsh physical environments. Such devices depend on high-tech components, long-life batteries and wireless technology, among other factors, to operate properly. Such areas are ripe for patent activity.
• Low raw materials concentration – TPMS sensors are small and do not require significant amounts of heavy or bulky materials (eg, steel, aluminium or rare metals). Raw materials comprise a relatively low part of the development cost.
• High manufacturing margin potential – given the relative simplicity of TPMS sensors (ie, pressure transducer, analogue-to-digital converters, microcontrollers and wireless communication modules), the sensors have high manufacturing margin potential. The retail prices for sensors are as low as $20 apiece, indicating manufacturing costs well below that range.
• Replacement potential – TPMS sensors are long-life wear items that include batteries. These batteries are not user-serviceable and will die over time, requiring replacements, usually between seven and 10 years after initial use. Thus, the market potential and demand for additional TPMS sensors coming online is somewhat disconnected from new vehicle sales and provides additional value capture potential for innovators. Average costs for aftermarket replacement TPMS sensors range from between $20 and $40 apiece on Amazon.com. Given that a car needs four sensors, the replacement cost for vehicles may range from $80 to $160 every seven to 10 years, with a spike in aftermarket systems in 2016 and 2017 (ie, about seven to 10 years after the 2008 government mandate). To put that amount into perspective, market research company IHS reported 2008 vehicle sales at 13.2 million units. If 50% of those vehicles are still on the road, then 6.6 million vehicles will require 26.4 million replacement sensors. With total vehicle replacement sensor costs of between $80 and $160, the revenue potential in the aftermarket alone would range from $528 million to $1.05 billion. That amount stacks on top of the units sold for new car sales.

At current unit sales levels, including aftermarket replacement parts, patent investment and enforcement makes sense as the economics allow reasonable recapture of the investments and enforcement costs even if apportioned per-sensor royalties are as low as $0.05 to $0.20 per unit.

Next, automotive patents that drive greater value will apply to a large set of the vehicle population. The maths is simple. The more vehicles that adopt the technology, the greater the market reach and amortisation of R&D costs. Consider Table 4, which outlines 2015 vehicle production statistics for the US market.

As the data in the table suggests, GM and Ford are better prospects for owning or licensing valuable patents because they ship a greater number of vehicles relative to Jaguar, Tesla or other small-volume brands. Next, the value may also vary by model. Table 5 shows 2015 production data for the top 20 vehicle models in the US market.

As the data in the table demonstrates, the unit volumes can matter as well. If one is making $5 per vehicle model selling widgets, it is better to create value by selling products that integrate into a Ford F150 pickup truck than into a Jeep Grand Cherokee because the volume is nearly four times greater. As the data suggests, the right product with the right attributes can indeed be valuable in the automotive context.

Reappraising patent value

There is significant emphasis on technology in the automotive sector which has permeated the popular press and headlines and served as a point of differentiation among various OEMs. While there is much interest in providing intellectual property to the automotive sector, the current value chain makes it difficult for most rights holders to generate positive returns on investment. That Tesla wilfully decided to open its patent portfolio reinforces this fact and the ultimate result is unsurprising when cross-referencing the enforcement value of its patent portfolio against the broader automotive market.

Valuable patents in the automotive sector will not centralise on high-tech, expensive, low-volume makes. Rather, the most valuable assets will apply to the broadest set of automobile manufacturers and models, often paired with a low raw materials component and a government mandate. Rights holders ultimately will likely maximise value by dealing directly with OEMs or Tier 1/Tier 2 suppliers.