Professional indemnity insurance
|
Service contracts anatomy™
JC’s practical guide to commercial contracting
|
Professional indemnity insurance
/prəˈfɛʃᵊnᵊl/ /ɪnˈdɛmnəti/ /ɪnˈʃʊərᵊns/ (n.)
business insurance that protects service providers against claims of negligence, errors, or inadequate work by their customers. This covers financial losses resulting from professional mistakes or negligence, legal costs of defending those claims, and compensation if the provider is held liable.
To what extent should a services provider arrange, or tie its liability to, professional indemnity insurance?
Remember the salient points of any insurance contract: it protects a person of limited resources from a remote but outsized risk that is beyond the person’s practical control and ability to withstand.
It does this by mutualising it: across of large group N of similar eneities, that remote risk having value R will only happen to a small number x of them.
If each member of N contributes a portion equal to say (R * 1.5x)/N, the risk of loss is spread across a small part of a large class of people each of whom paying an affordable running premium, protecting each from colossal costs that may never happen. Thus, you mutualise your risk of loss.
The risk of loss in a service contract, scaled, in the financial services industry can be yuge. That is what people are paying for— there is an argument that by offering your services you are underwriting that risk for your customer.
So take an example:
Lexrifyly provides valuation services to clients to help value their assets under management. It earns a handful of basis points running on the net asset value of customer assets under management. If its services fail or yield incorrect asset valuations that customers trade on, the customers stand to lose a great deal of money — a lot more than a couple of basis points.
These risks, especially early in the life of a service provider, can be existential. There are two pragmatic ways of managing this: the first is limiting your liability for losses as a result of failures in your product; the second is to buy insurance. A third — a hybrid — which is to limit your liability to the value of your PI insurance.
Now pragmatically, if your service is so bad that you cause outsize losses, this is existential in any case — clients are likely to walk away if your product is not fit for purpose — but as we know the kinds of unexpected contingencies that can hie a networked technology business are unlimited, they are evolving, and they cannot all be predicted or solved for. This problem becomes all the more important the more complicated technology becomes, and the more sophisticated are the things we rely upon it to do.
As the pace of technological change accelerates, the “Lindy horizon” contracts. We have had printing presses for nearly six hundred years. We have figured out their ups and downs. We have had large language models for fewer than ten.
Limiting liability to a level rate lower than the realistic expectation of potential loss — 100% of fee income is common — is great, but it leaves the customer with material functional risk to its services. What exactly is it paying for if it is not a right to sue you for actual losses if your product does not work? This is a cold rational analysis of service contracts: they are, in themselves, a form of insurance.
Limiting liability to a rate broadly equivalent to the realistic worst-case expectation of loss makes more sense, though you might ask what one gets out of this, since you are agreeing to be liable for everything for which you would ordinarily be liable. Two things: firstly, it enables you to hedge that risk entirely, by acquiring PI insurance to exactly that stipulated amount. Secondly, and even if you don’t, it knocks out “tail events” which were genuinely unexpected, and blew your liability out father than that reasonable worst case scenario.
|
Premium content
Here the free bit runs out. Subscribers click 👉 here. New readers sign up 👉 here and, for ½ a weekly 🍺 go full ninja about all these juicy topics👇
|