Looking For Value: Drone insurance in the US

A recent PWC report estimated the commercial drone total addressable market (TAM) at $127bn (global). A quick back of the envelope calculation makes it into a ~$10bn opportunity for the insurance industry.

Drones, as the radio-controlled aircraft are called, have many potentialities, civilian and military.” Popular Science magazine 1946

Drones or Unmanned Air Vehicles (UAS) have been a thing of the future for a long time. Early UAV where in many ways similar to early PCs. They were costly, unwieldy and complicated to operate. Drones were a value proposition that only appealed to military and academic communities. But like PCs, things have changed.

DJI, the industry’s biggest player was expected to reach $1bn in sales in 2015[i]. Suppliers to the leisure drone industry are shipping components that translate into run rates of $10m units/year[ii]. It is a growing industry with some already very large players. Is there any value left to capture anywhere in the value chain?

Insurance is quickly becoming one of the industry’s biggest concerns[iii]. In 2015, Lloyd’s published a report on key issues for insurance providers in the drone industry. In its study, Lloyd’s qualified the UAV insurance coverage as “small” with “rising demand” and “significant growth” left ahead[iv].

Assessing the depth of the commercial drones opportunity for the insurance industry necessitates first building a better understanding of the drone market. What are its growth prospects? Its key drivers and constraints? With that in mind, we can then turn our sight towards the insurance industry and assess what are the prerequisite conditions for its success into what could be a very lucrative market.


1. The rise of the commercial drone

Key drivers:

Commercial UAVs are an emerging market pushed by some powerful factors among which in order of importance:

(a) A compelling value proposition: nowadays, drones combine highly relevant features for several industries with consumer-friendly user experience at very low prices. These features are relevant for a wide range of applications in agriculture, logistics, the media, infrastructure etc. and open new opportunities for significant cost savings. As an example: a standard wind turbine inspection currently costs ~$1,500 per tower. Using a drone to perform a similar task would cut its cost by 50%[v]. Drones are becoming valuable assets to many industries and regular improvements in machine vision, image processing, range sensors and robotics are constantly improving that value proposition.

(b) A rising infrastructure: drones can be seen as another set of data acquisition nodes supporting the wider big data initiative. The wider economy rush to build a ubiquitous network of smart sensors and autonomous systems is lifting UAV sells. Drones are by design great assets to the Internet of Things.

(c) Abundant funding: UAV are capital intensive by their industrial nature. Yet new entrants face no shortage of cash to fund their growing operations. Between 2014 and 2015 VC investments have tripled and show no signs of slowing down.

(d) Widespread technology: components prices are falling. Evidence suggests that the UAV drones parts price are in fact subject to Wright’s law[vi]. Open source flight control software makes stability and reliability in hardware within reach of new manufacturer without significant investments.

Both abundant funding and widespread technology contribute to significant barriers to entry erosion and should translate in smaller margins for drone manufacturers. A quick look at Parrot’s ROIC/WACC[vii] over time supports this conclusion.

Table 1.1 Parrot’s capacity to create shareholder’s value over time:


Main constraints:

But just like there are many forces pushing the market upwards, there are others limiting its expansion, they can be summed-up in one word: regulations.

(a) Regulation (safety): the first issue addressed by regulation in the US is safety. In 2016, the FAA established clear rules under which a registered commercial agent would be aloud to operate. These rules are primarily designed to minimize collision between aircrafts and with on-the-ground individuals. The FAA current throughput is of 230 authorizations per months. After two years of operations, the FAA had granted 5,551 authorizations[viii] (equal to a 8.5% CMGR). This could constitute a serious bottleneck on the commercial drones growth rate. However, it seems that a large black market of unregistered commercial drone operators has flourished. According to Stephen Hartzell, a lawyer within the industry, a lot of operators use drones commercially without authorizations. As of early 2016, the FAA had not fined a single operator solely for operating commercially without an authorization.

(b) Regulation (privacy invasion): UAVs’ military heritage and their capacity for surveillance put them at the center of a lot of legislative and social scrutiny. While the FAA has address most of the safety issue conducting tests[ix], the privacy issue could prove to be a complete different beast. Courts being slower then the FAA or Congress, debate around regulating the UAV is likely to last for years.

(c) Regulation (peripheral): beyond the FAA and the debate on privacy invasion that is likely to take place in courts, more then half of the States have enacted drone laws[x]. Yet many of these local laws are unenforceable. Federal government regulates airspace, not the State.

Overall, legal uncertainty has been impeding commercial drones development. Regulation could take a long time to be put in place.


Market sizing:

With that in mind, we now try to assess how big the US commercial drone market could be.

In 2013, the U.S. Department of Transportation (DOT), made the following projections[xi]:

These commercial numbers could be highly conservatives given the diversity of applications. In the past, the FAA has already underestimated the popularity of drones. In 2010, it said there would be a total of 15,000 drones in US skies by 2020. More then that are already been sold per month[xii]. Thinking about a few example applications put things in perspective (ranked by likelihood):

(a) Precision agriculture: a 10-acre analysis takes 10–15 minutes with a UAV down from 1h in a regular context. There are 152,000 farms that have annual income of $250,000. It is reasonable to think that these farms will have UAVs to routinely inspect their crops.

(b) Media — Company promotion: YouTube is the second largest search engine. 61% of 30,000 US enterprises used YouTube to communicate with their customers[xiii] at some point. It is reasonable to think that they could turn to UAVs if it improved content.

(c) Media — Real estate: there are 2m real estate professionals listed on Zillow that could use small UAVs to provide marketing perspective for their clients’ properties[xiv] in 2016.

(d) Property insurance claims: there were 510,000 hail loss claims alone (mostly roof damages) to be investigated in 2014[xv].

And the list of applications could go on and on. These examples give us a sense of how diverse and deep the market is. We could estimate diffusion for each one of these applications separately through analogy and then use the aggregate as a decent TAM estimate. For instance, we could use GoPro’s diffusion variables to model growth and TAM in the sports segment, or use e-sport diffusion variables to model growth and TAM in the drone-racing segment. However, there would be 2 limitations to this approach:

(a) The industry is still at the beginning of its technological S curve. Design standards are not completely set and the industry’s success is still subject to a lot of uncertainty due to evolving regulation. Hence it is hard to predict which application will/could take off. In 2013, Jeff Bezos announced in “60 minutes” that soon octocopters would be handling 86% of all of Amazon deliveries. The FAA decision to limit commercial drone usage to visual line of sight has put a serious dent in his plan.

(b) The diversity of applications imposes aggregating many separate forecasts using different diffusion models and relying on a multitude of assumptions. Too many variables would have significantly weakened the quality of our end result.

Hence, we shall limit our TAM estimates to some of the applications that we’ve identified earlier (see the last 3 paragraphs of the note for some estimates).


2. Eyes in the sky insurance

A recent PWC report estimated the commercial TAM at $127bn (global). A quick back of the envelope calculation makes it into a $10bn opportunity for the insurance industry. This calculation assumes a 20% discount to TAM to make the UAV value proposition appealing compared with traditional offerings and insurance premiums equal to 10% of operating costs.

The drone industry carries many promises. Insurance companies have started positioning themselves to benefit from them albeit very cautiously. A quick assessment of the US landscape puts forth 24 decently sized brokers but only 3 large underwriters.

Regulation is not only key for commercial drone development; it is also for insurers’ confidence. Until recently, the regulatory landscape was too uncertain for insurers to invest significant effort in the UAV market. The recent progress made by the FAA could very well lead to an infliction point in the industry’s growth.

For the industry to thrive, it still needs a better understanding of the risks it’s taking as well as what factors could be source of failure: insufficient information on risk, adverse selection and insufficient market capacity to spread risks.

So far, insurance companies standard policies have covered third party liabilities and physical loss/damages. Premiums are established yearly at an average $1,000 for $1m third party liability coverage. This translates into an estimated 5–15% of total operating cost[xvi]. Premiums have been fairly stable over the last 4 years. Low volatility in premiums can be explained by an equally low claim rate. In today’s landscape, policy availability is more the issue then premiums.

Some new entrants have started offering coverage by the hour. This model could prove extremely appealing to many operators looking for flexibility. The US market could be vast. I am going to take a wild guess and estimate the precision agriculture alone at $100m (assuming a 40% penetration rate similar to that currently in Japan[xvii] and 2h/week utilization at $15/h). Of course this is just one of the many possible applications. Commercial media could very well generate $250m (assuming that 30% of the 61% US companies that have already used YouTube would generate another average 3h of footage at $15/h using drones). Real Estate could also generate another $150m (assuming that 10% of properties listed on Zillow had some sort of aerial footage included — that’s 5 footages per listed real estate professional). Adding just these three applications could yield $500 alone.

How about private drone usage (instead of commercial)? I believe that this market is only likely to become attractive for insurers if US regulators make drone insurance mandatory. It is already the case for 7 out of 15 European countries and for other motorized vehicles in the US such as car, planes and motorcycles.

There are 300h of video uploaded every our to YouTube. If people started adopting action drones like they did with the Go-Pro and had to contract insurance for every hour of uploaded footage the potential upside could be staggering.

Insurers that position themselves like platforms could very well benefit from strong network effects where more subscribers could lead to lower premiums. If such was the case then empirical study shows that the winner could acquire an extremely strong competitive advantage and take between 50 and 90%[xviii] of the $500m mentioned above.

Thanks for reading — feel free to reach out with any thoughts by replying here or on my twitter 👋


REFERENCES

[i] Ryan Mac, “Bow to your billionaire drone overlord: Frank Wang’s quest to put DJI robots into the sky”, May 2015, Forbes

[ii] “Drone Industry Report: I’s in the sky”, Feb 2016, Oppenheimer

[iii] Jeremiah Karpowick , “7 Commercial drone predictions for 2017”, 2017, Commercial UAV expo

[iv] “Drones Take Flight: Key issues for insurance”, 2015 Emerging Risk Report, Lloyd’s

[v] “Clarity from above: PwC global report on the commercial applications of drone technology”, May 2016 PwC

[vi] Michael Mauboussin “ Clean Tech Gaining From Wright’s Law”, Feb 2017, Valuewalk

[vii] Parrot, Data Update, 3Q11, Goldman Sachs

[viii] http://www.Faa.gov/uas/beyond_the_basics/section_333/

[ix] “FAA announces six UAS test sites”, Aug 2015, Federal Aviation Administration

[x] NCSL, “Current unmanned aircraft state law landscapte”, May 2017, NCSL website

[xi] “Unmanned Aircraft System (UAS)”, Service Demand 2015–2035, Sept 2013, Department of the air force

[xii] Craig Guillot “Can pilotless planes soar over safety, privacy and regulatory barriers” , Jan 2016, Sage Business Researcher

[xiii] “YouTube Outperforms LinkedIn as Social media channel of choice for over 60% of enterprises”, Jul 2016, PR Newswire

[xiv] National Association of Realtors “NAR President testifies before house judiciary subcommittee in support of responsible commercial drone use”, Sept 2015, News release

[xv] David Fennig, “Forecast report”, May 2016, NICB

[xvi] Steer Davies Gleave, “ Study on the third-party liability and insurance requirement of remotely piloted aircraft systems”, Nov 2014, European Commission

[xvii] Colin Snow, “The truth about drones in precision agriculture”, Skylogic Research

[xviii] Michael Mauboussin, “ Total Addressable Market: Methods to estimate a company’s potential sales”, Sept 2015, Credit Suisse Global Financial Strategies