What’s wrong with a virtual broker-underwriter marketplace in large commercial insurance?

It makes complete logical sense. In this day and age, when most industries are opening up to virtual as an alternative/secondary medium for trade, then why not a virtual marketplace in large commercial (re)insurance?

To be clear, by virtual marketplace I mean a internet platform where 2 or more types of participants come on-board to trade. Generally the 2 participants represent a buying and selling side (or a demand side and supply side). Such platforms capitalise on network effects also referred to as network externalities in which positive trade benefits are realised as membership to the platform increases. A famous example is eBay which is an autioning marketplace. eBay succeeds in creating opportunity: the opportunity for sellers to get visibility with a large buyer base and the opportunity for buyers to get a bargain from a large number of sale options.

I see three major challenges for a broker-underwriter virtual marketplace in large commercial insurance. First of all, there is very little residual externality in broker-underwriter relationships to bring any benefits for trading on a common platform. Take the London Insurance market – it’s all based in the very small area of London EC3 where there are limited number of connections between broker and carrier firms. There are efficiency gains to be achieved by trading electronically but a virtual marketplace is not really the best tool to represent these limited number of connections. To quote eBay’s example again: eBay succeeds because it allows users to manage an unknown, an unmeasureable and physically unmanageable number of trade opportunities. However the magnitude is not quite the same  for broker-underwriter relationships in large commercial insurance. Moreover with consolidation affecting this industry, I suspect the number of broker-underwriter connections are fairly well defined at this point.

A major concern of trading electronically in the large commercial insurance world is achieving straight through processing (STP). This is the second area where the virtual marketplace falls short. The marketplace’s independent existence means that it relies on other pieces of the electronic jigsaw (typically gateways, adaptors and electronic repositories) to be in place if users are to make the most efficient use of their data. In such a scenario the marketplace is not the closest to its users (this is generally the space of adaptors and gateways) and therefore electronic trading is more transient. There is then very little pull for users to come on-board an independent virtual marketplace where their data maintenance spreads rather than consolidates.

The third consideration in any marketplace is price level and price structure. Price level is the total cost to the users in a trade and price structure is the actual share of the total price that each side pays. Pricing structure is a strategic issue which is always difficult to get right. After considering the reduced network effects and the fact that the marketplace may not be the closest to its users, it becomes apparent that the price level cannot be high. Most likely the price will be a small transaction charge:  there is very little case for the marketplace to charge a service/membership fee. Counter balance that with the costs involved: the cost to reach critical mass for a marketplace are quite high. Not to forget that standardisation of trading technology means that it becomes very costly to achieve user stickiness and the marketplace will eventually suffer from multi-homing (when users use multiple and competing marketplaces for their trading activities).

So is there any case for a virtual marketplace in large commercial insurance? Actually I believe there is under the following:

1. The marketplace needs to serve multiple types of participants, not just brokers and underwriters. There are greater network effects in some of the other relationships in the insurance value chain. And more likely these relationships have largely escaped electronic efficiencies which means that there are huge trading benefits to be realised as a result.
2. The marketplace needs to consider a path towards openness and continuous innovation to keep fresh. The provider of the marketplace could assume the role of a facilitator and allow natural discovery of new value propositions by the participants.
3. The marketplace needs to be realistic about costs involved. And more realistic about its price structure and price level. Profits per trade will be low, but numbers involved could be large once critical mass is achieved.
4. Also, and this goes without saying, the marketplace needs to be fair and open in its operation. This will add to its likeability, a key feature whose positive effects are too great to measure.

Any thoughts welcome.

Platform Competition, Compatibility and Social Efficiency

I came across this interesting research paper by HBS on platform competition and compatibility.

In summary: Many markets are characterised by 2 sided virtual and/or physical platforms where agents exchange and are charged a fee by the platform provider. The research by HBS scholars shows that profits are higher for providers if their platforms were compatible (such as when they are based on a common standard). Platform compatability generally leads to equilibrium in an oligopolistic market, and obviously benefits society best. Platform incompatibility on the other hand leads to lower profit margins, but allows one market player to gain dominance over others and earn more. The research thus shows that competition is a major factor preventing platform providers in agreeing to a common standard.

Looking at it in terms of pure supply and demand this makes complete sense. Airlines want to be overbooked. Supermarkets may squeeze their suppliers, but cannot afford to lose their customers. I can think of no industry where providers wouldn’t want a higher share of the demand.

In platforms, there is an interesting shift taking place though now with cloudware or Platform-as-a-service (PaaS). Without being an expert on this topic, I believe that the differentiation between users and developers is getting increasingly blurred. What dynamics shape pricing on such platforms? I think these platforms are still based on propreitary technology thus creating lock-ins. Is that still true for Google Apps Engine though? I don’t know, but I would love to hear from others who do.

On a personal note, I am currently involved with a trading platform in reinsurance. And while there is a common standard that all platforms adhere to, I still think that existing platforms are more aligned to one side. There is a definite empathy for one side of the platform user base.  Interestingly, the reasons are many. Not just competition! Trading platforms in this sector are still new. They are not the norm and not (yet) highly competitive. So it will be interesting to see how this develops. Any thoughts welcome.

You can read the research paper over here.