Read part two: Peer-to-peer in the risk markets
Read part three: Blockchain in the risk markets
In the opening segment of this series on complexity, I discussed the three network graphs which have emerged in the risk markets, and which business models embody them.
In the second segment, I discussed the emergence of peer-to-peer insurance that will accomplish the three core functions of the risk markets that currently exist in a “black market” unformalised state by using distributed managerial methods.
- Risk transfer.
- Escrow of funds for a defined purpose.
- Management of reallocation of escrowed funds.
In the third segment on distributed ledger technology, we looked at how distributed ledger technology can be configured as a cohesive platform that would embody all three network graph types. I discussed how the roles of individual peers, along with carriers, and agents can work together to formalise the P2P methods in the risk markets.
In this final segment, I will contemplate the current balance coming out of the industrial age of the market share of each graph type in the risk markets, how the market’s balance may change, and what the new forthcoming equilibrium state might look like in the risk markets.
Before doing this, I’d like to discuss an important idea that emanates from the blockchain and cryptocurrency communities – the idea that there could be “one ledger to rule them all”, or, said another way, “could a single distributed risk ledger be the all-encompassing ledger, accounting for all value?”. The simple answer here is no. No single ledger, technology or network will ever be all-encompassing. That would be silly, as it would reintroduce the systemic weakness inherent of centralised system structures, namely the risk that by taking out a single central node (or ledger, in this case), the whole system could collapse.
Just as was realised in the blockchain and cryptocurrency communities, the idea of a “risk ledger to rule them all” isn’t a desired structure. In the risk markets, a single distributed risk ledger to account for all funds escrowed against all risk types is also not a desired structure. Due to the nature of risk and the diverse set of risk exposures in the world, there will need to be a diverse set of risk ledgers. We may see something materialise that looks like this example of four distributed risk ledgers, each for a specific category of risk exposure:
Hold on to that thought for now …
II would like to again reference some of the work done by the Ripple team, and their thought leadership towards a solution to address the concern of “one risk ledger to rule them all”. The Ripple team has introduced a protocol that enables value to move in a cryptographically secure way between two or more distributed ledgers. It’s called the Interledger Protocol, and more information can be found on their site.
Using the Interledger Protocol, the Ripple team has articulated and demonstrated how various types of distributed ledgers, each engineered for a specific strength or market, can be networked together to create a term they have coined the “internet of value”. Without a single shred of doubt, it’s a true statement that “finance is getting its internet”, and it’s already here, albeit in a state of maturity similar to the internet circa late 1990s.
Unlike the slow pace of the internet’s growth, however, finance’s internet will not take as long to mature mainly because it is advantaged by the preexistence of the information internet itself, and all that has been learned in its construction. Insurance and the risk markets, of all the various financial services, are the lowest hanging fruit.
This may seem like a stretch in today’s environment, but it’s not hard to imagine that by connecting many risk ledgers, each escrowing funds against a specific risk type, and using the methods outlined with the Interledger protocol, that we will see the emergence of an ‘internet of risk’. Just like with the internet of value we see emerging today, it will be made of many different distributed risk ledgers networked together. Expect to see many different distributed risk ledgers emerge in the risk markets, as well as the development of an internet of risk networking all of them together.
I would define an internet of risk as a network of distributed risk ledger networks. The technical name for a “network of networks” in complexity science is called a multiplex. Indeed, the risk markets have been operating with an informal and non-digital multiplex structure for some time. Since each insurance company manages a risk ledger, and reinsurance companies function to connect insurance companies’ risk ledgers together, the reinsurance industry effectively embodies a decentralised network of insurance companies, both graphs combine to embody a multiplex of risk ledgers.
In all likelihood, we will observe over the coming years the digitisation of the existing multiplex of risk ledgers that is the risk market of today, into a network of digitally interconnected, centralised and distributed risk ledgers, with each individual risk ledger serving the specific needs of a specific risk exposure.
KarmaCoverage is intended to be this “multiplex of risk”, to organise the interconnections between the risk ledgers of all types of P2P risk sharing, towards the goal of ensuring that as the P2P segment of the risk markets grow, that it maintains a high degree of resilience, enabling society to transfer risk efficiently among individual peers, successfully addressing the various risk exposures of those peers. You would expect to ultimately see this play out, creating an internet of P2P risk ledgers, and looks something like this:
To be fair, it isn’t possible to know the ultimate structure (or graph) of this ‘multiplex (or internet) of risk’. It will emerge by a process of self-assembly. It must employ distributed managerial methods to avoid reintroducing the fragility inherent in its overly centralised structure.
That said, many portions of an internet risk can and should be centralised for efficiency and standardisation purposes. Distributed systems have weaknesses as well, one of which is the introduction of some degree of inefficiency into the system. We wouldn’t want to act out the behaviour pattern of, “if all you have is a hammer, everything looks like a nail” and pursue a “distributed everything” type of internet of risk. The functions that should be centralised combine to make the business case for KarmaCoverage.
Now let’s take a look at how this may impact the existing balance of market share that each graph serves as a percentage of total risk. Using data on the currently formalised methods of total risk, and assigning a percentage to each graph in the risk markets, you find that the industrial age state of equilibrium between the graphs settled at roughly these percentages:
- Reinsurance: 40%
- Insurance: 60%
- P2P coverage: 0%
- This doesn’t account for all the risk transfer activity that occurs informally in the existing P2P black market.
There are two factors to consider when thinking about how the equilibrium state of the risk markets will balance out in the information age. To answer this, we need to consider market growth and how the size of the risk markets will grow as a result of formalising P2P black market activity. Secondly, we need to consider the market share split between the three graphs, given that P2P will no longer continue to be 0% of the formalised market.
Let’s look at Airbnb and the short-term real estate rental market for a benchmark. Why use Airbnb out of all the various markets with new P2P entrants for a comparison? Because real estate has an inherent limit on one side of the market: supply. There is only so much land and only so many rental units for any given geographic market. This is similar to the risk markets, where there are only so many risk exposures to be covered for each client. Each person may have various risk exposures, but each exposure only needs to be covered once, limiting the demand side of the risk markets.
In a mere eight years, Airbnb has already increased the market’s supply of rental units listed worldwide. Currently, Airbnb has over three million rental listings, and is growing at 20-30% YOY, which means it could have roughly five million rental units in the next 2-3 years. Airbnb has already surpassed the number of rental units under management of the largest players such as Marriott and Hilton, who have a combined 1.8 million units under management. Over the next couple of years, Airbnb may have as many units available as the top nine hotels, which cumulatively have just under five million units under management.
These new P2P rental units becoming available to the market via Airbnb represent an overall market expansion for the short-term rental markets. This growth in what was previously viewed as nearly saturated supply, has been followed and slightly outpaced by growth in P2P demand (see figure 10 in same report above). At the end of the day, P2P methods being applied to real estate will have nearly doubled the market supply of the short-term rental market, as well as grown demand. A full market expansion.
I have a client who’s a global executive at one of these hotel companies, who was in my office not long ago. Before he left, I asked him: “Hey, what do you think of Airbnb?” He stopped and turned to reply: “It is the greatest threat we have seen in the hotel business in a hundred years.” I smiled and offered to share some of the complexity science and P2P network research with him, so he could see the cross-industry pattern of P2P more clearly. He also shared with me some very well thought out reasons for why there is a maximum amount of threat that Airbnb and P2P methods can pose to the hotels.
According to the STR research, there are differences in the nature of the demand that Airbnb serves and the demand hotels serve. Airbnb tends to serve more leisure travellers who stay for longer durations, while hotels are maintaining business travellers and higher revenue per night.
I believe we will see a similar situation unfold in the risk markets, as with Airbnb and the real estate market. The nature of the demand for coverage P2P methods will serve, like Airbnb, will be different than the nature of the demand traditional indemnity insurance serves. When speaking with a couple of other P2P insurance startup folks who come from an actuarial background, they referred to a problem with their financial model. I finally gathered the issue was that the system could fail in the face of a large loss event. So the easiest solution on the table is to simply not apply the financial model to large loss events.
A P2P platform will be used to provide coverage “starting from the first dollar of loss and up” on large losses, as well as fully cover small loss events, while indemnity insurance products will continue to best serve the large loss events. A P2P platform in the risk markets will expand the limited side of the risk markets, the demand side, just as Airbnb in real estate expanded the limited side of that market, the supply side. I did a similar analysis using Uber and the taxi market, which yielded similar market expansion dynamics.
Using these benchmarks, the fact that the frequency of small loss events is much higher than large and catastrophic loss events – along with my crystal ball – I’m going to guess that the formalising of P2P methods in the risk markets will result in doubling the size of the formalised risk market at some ambiguous point in the future. I will also assume that the ratio between insurance and reinsurance shown above doesn’t change. This would end up with risk markets growing to nearly $10tn, with market share split between three segments, like this:
- Reinsurance: 20%
- Insurance: 30%
- P2P coverage: 50%
Surely these assumptions and predictions are wrong, but this is more of an exercise in trend observation, not an attempt to actually predict the state of the risk markets at some specific future point in time.
There will be other drivers that will impact the shifting balance. One easy to understand, yet powerful and potentially market-driven force, would be consumers voluntarily choosing significantly higher deductibles. This trend is already in motion. One indication of this trend on home insurance policies is that, in California, on policies covering over a million dollars, the lowest a deductible can be and remain compliant with regulatory rules is a $10,000 deductible. While that example is regulation imposed upon the industry, here in Florida we saw the industry self-impose an increase in deductibles from hurricane losses after the 2004-05 seasons, while at the same time many large carriers simply pulled out of the state, leaving a vacuum to be filled by new, smaller Florida domestic carriers.
Using formalised P2P ‘networked self-insurance’ methods, it’s possible for consumers to achieve an average of $10,000 in coverage on an annual basis for under $100 a month, plugging the ‘deductible gap’ all the way down to the first dollar of loss, fully addressing total risk exposure, thereby meeting the consumer’s expectations for the risk markets. That could easily lead to (and enable) consumers to request $10,000 deductibles on all their insurance policies, which would have a material impact on gross premiums.
On home and auto insurance losses, over 90% of claims are under $10,000. If the consumer behaviour of requesting ever higher deductibles on their traditional insurance policies does occur, it becomes easy to consider that premiums on traditional insurance may currently be at (or near) their historical high.
Other key drivers
There are a few other key drivers for the industry in the forthcoming years. Automation, as seen with Lemonade’s AI quoting and adjusting bot. Aggregators, which is a strategy that Google has already tried with Compare, and should manifest some consumer experience mirroring how airline carriers sell plane tickets, or a “Zillow of Insurance”. Real-time data feeds fuelling continuous actuarial pricing of risks, which will lead to hyper-dynamic pricing (with ramifications). That could play out in a number of ways depending if you’re looking at existing policies or new policies. Expansion into unaddressed risk markets, be it new types of risk exposures, or simply new geographic markets. Those are just a few, in addition to the potentially transformational consumer behaviour shift towards the sharing economy.
Obviously, this process of formalising the P2P segment of the risk markets will face headwinds, but since I entered the industry with a firm eye on the intersection of the risk markets and crowdfunding methods back in 2013, we’ve seen the number of P2P insurance companies grow from one to dozens, all over the world, pursuing a handful of business models. So it seems the moment for the formalisation of P2P methods in the risk markets is occurring, and we will see the “Napster of risk” (or Airbnb, or Facebook, or Google) emerge. Due to the convergence of factors discussed in this series, and a few others, I believe the ‘P2P moment’ for the risk markets is now.LinkedIn Group
Image by Lightspring, Shutterstock.com