From Data to Rate

All the data in the world is of little use without the tools to analyse and apply it in our daily pricing routines.

Cuthbert Heath: data analyst

Cuthbert Heath is arguably the best known Lloyd’s innovator. His advances include burglary cover, jewellers block, and even excess of loss reinsurance, all of which were enormously successful for him, and remain mainstays of the market. He’s best known, though, for his brave instruction to pay all affected policyholders after the 1908 San Francisco earthquake (and less for selling ‘aftershock insurance’ from roadside stands in the devastated city, which meant his Names eked a small profit that year).

Heath’s innovations worked because he collected data tirelessly, and analysed it rigorously to price risks accurately. The greatest surviving evidence of this analytical acumen lies in his extant Earthquake Risk Book, one of the treasures of Lloyd’s.

Heath and his assistant spent countless hours in Lloyd’s library searching for data about historical earthquakes and the damage they caused. When he had exhausted its resources, he moved to the British Library to continue his collection of frequency and severity data on historical seismic events around the world.

Analytics the hard way

Numbers, though, are meaningless without interpretation. Heath lived in the age before Extreme Value Theory and Monte Carlo Simulations, and even before pocket calculators, so he analysed his data the old fashioned way, with his impressive brain and a pencil. When he’d calculated risk-based rates for earthquake risk in almost every conceivable location, he recorded them in his Earthquake Book. Armed with this data, his syndicate quickly rose to be Lloyd’s unrivalled earthquake slip leader.

Today underwriters are drenched with an almost ceaseless cascade of data, and able only to scratch the surface of what’s available without highly sophisticated tools. But analysis is the real challenge, as anyone who’s ever tried to find a detail in a US listed company’s annual 10K filing knows. What underwriters really need is a system which gathers the data, draws out the relevant information, and does the analysis for them.

Analytics made easy

That system is Quotech. As an underwriter, I read those 10Ks. I spent hours on analysis, even with the resources of a huge major managing agency – from actuaries to PML calculation engines – at my disposal. The experience led me to develop Quotech. Its massive repository of real-time data is placed at the fingertips of underwriters and brokers in a useful, easily understood way.

Even more powerful, though, is the platform’s analytics capabilities. The system uses Natural Language Processing to classify and parse documents. That ensures the most important ones are highlighted for underwriters, and helps by decoding, for example, the financial sentiment embedded in a 100+ page 10K document.

Quotech’s built-in modelling technology relieves busy people of the drudgery of spreadsheets. Its analysis assumptions can be modified, for example to reflect an evolving risk appetite for a specific peril or territory, to calculate PMLs, or to flag-up potential aggregation risks.

All the information, all in one place

Alongside data from paid sources, Quotech gathers relevant facts from the internet in real time. It’s used to enrich policyholder-provided information, and to check its accuracy. For example, a bar that declares it serves only drinks may have a food menu on its website, or a construction company claiming it does only decorating might advertise roofing. Quotech picks up such inconsistencies, allowing underwriters to rate the real risk.

It also permits better implementation of underwriting controls. For example, Quotech can monitor the list of high-risk kidnap countries, and ensure only underwriters with sufficient authority are able to quote and bind risks in those zones.

All the data – including elements that were not used during underwriting – is recorded and saved in one place, in a constant format, to create a risk snapshot at the time of quotation. This picture is preserved for future analysis by actuaries and data scientists to refine their pricing and risk selection model against actual losses.

Tech-savvy and interfaced

We’re fluent in modern data science programming languages, and work with clients’ in-house specialists to ensure everyone gets the data they require in the format they need (say as a spreadsheet, in json format, or as an R dataframe), and precisely where they want it (be it on a standard reporting platform or one built by Quotech). It automates the analysis, and leaves the judgement to underwriters.

After analysis, Quotech’s functionality goes a remarkable step further. In my next blog, I will delve into the achievements of another great Lloyd’s innovator from our rich history, and explain how Quotech maintains their innovative tradition, to ensure you stay at the epicentre of insurance success. Click here to be sure you receive it.