The question Who will win the fight to control smart-follow capacity? was posed this week in another insurance publication. The authors predicted “more of a clash over which side of the market – carriers, brokers or service providers – comes to essentially control or own” digitally allocated capacity.

Forgive me, but that’s a daft question.

Capacity is risk capital, nothing more. So-called ‘smart follow’ is a way of allocating capital to risk that uses technology in support of risk selection. “Smart-follow capacity” in the syndicated risk market is therefore not a thing that can be controlled.

It isn’t even a thing, in the way that ‘potluck dinner food’ (for example) isn’t a thing. Any food can be or become potluck dinner food, anyone can decide to make it so (provide they have a social life), and any potluck dinner food can be taken away from the party to become, say, picnic food.

Food is food. Potluck is simply a way of allocating it to different eaters.

Capacity is capital. Smart follow is simply a way of distributing it to risk (or vice versa).

What’s really interesting is the technology that makes certain capital “smart” (although that certainly cannot be controlled). In the case of “smart follow,” it’s one of several systems which are used:

a) from the distributors’ perspective, to assemble capacity behind a specific risk or group of similar risks, or
b) from the following risk carriers’ perspective, to find and potentially bind risks the meet certain criteria.

In both cases, specific risks (or the individual risks within portfolios of risk) are underwritten, structured, and priced by a lead underwriter (sometimes – increasingly – using ‘augmented’ underwriting, but that’s another story).

Under some smart-follow models, the capital providers are able to set their own price for their slice of the risk, whether through human intervention or algorithmic allocation, leading to a kind of risk auction process. Under others, a risk carrier puts a slug of capacity behind a specific underwriter or broking firm to provide a predetermined line on any risk under the broad heading of, say, international property. That line is automatically added to the lead underwriter’s stamp or the broker’s facility. (More than once I have heard the latter called “dumb follow,” but who’s to say?).

No matter how much of the decision-making is handed over to technology, nothing magical is happening. Any decisions taken by a platform, such as to bind 5% of a risk at a rate of 0.3%, represent only the automation of processes previously undertaken by human risk distributors of risk and capital.

As was the case in the world before any part of the syndicated underwriting process was ‘smart’, the people who hold either the capital or the risk will decide whether and how the two are matched, and how much technology is used in the process.

A broker holding a big pile of nice-looking risk is likely to allocate it to capital providers who have decided it would be cheap and easy to load risk of certain kinds onto their own balance sheet by using that broker’s smart-follow facility. There’s nothing new there, except a bit of technology that makes broker facilities more efficient.

Similarly, underwriters with capital to deploy may decide it would be inexpensive and easy to distribute capital by taking a following position on a consortium, then going to lunch. Still others, whether aspiring to be exchange operators, asset managers, or something else entirely, will set up algorithmic platforms to distribute the risk and capital of others.

Since the cyclical nature of the market means the supply of one or the other almost always exceeds demand, either those with the risk or those with the capital will almost always seek to make use of such systems, whether owned or controlled by brokers, underwriters, or service providers.

Smart or dumb, syndication in practice always comes down either to distributing capital or to distributing risk. When the market was soft, underwriters were delighted to get on board with smart platforms, in order to gain as much premium as they could, but brokers stayed away. When the market turned, it was brokers who began to see the smart in algorithmic underwriting, since it helped them gain as much risk capital as possible for the risks they needed to offload.

Over the past decade we’ve seen multiple platforms developed by underwriters, brokers, or third parties. They’ve met with varying degrees of success, but none among them can be considered to have controlled a significant slice of the algo market. Now, I argue, it is impossible to deliver a knockout blow to the competition.

For those who have set their sights on winning the smart-follow battle, I am sorry, I have some bad news. If ever that was possible, even when it was a conventional insurance technology arms race (and it probably wasn’t even then), today no one can win, because tech has moved ahead at an incredible pace since the first smart trade happened, and ruled it out.

Pretty soon, if they want it, everyone will have their own algorithmic smart follow platform. The decision-making tree behind the question ‘to write or not to write’ (as a follower, at least) is simply not particularly complicated. Given a week or two, any firm with a decent, Level 3 AI coder and a knowledge of the syndicated risk market can probably knock up a tool that works.

When everyone has their own smart follow platform, and it’s connected to the data ingestion, submission analysis, portfolio modelling, and exposure management tools that soon everyone will be running, control and ownership will go back to exactly where they are now, and have always been: in the hands of whoever has the upper hand within individual relationships at any given point in the cycle. Like an English council on the south coast, expect no overall control.

Then it goes back to basics: risk appetite and portfolio management. In that environment, everyone can win.

Guillaume Bonnissent is Chief Executive of Quotech