By Alan B Cantor

---

Simulation Is Not on the Exams, A Tale of Disruption

It was 1975.  The Travelers had just quoted us at Marsh (UT’s insurance broker) $24 million for renewal of United Technologies (UT) Products Liability program, which was a relatively modest increase from the previous year’s $22 million quote.  Seemed reasonable…

I was in Unit X, the Marsh Headquarters analytics consulting group. We determined to create a model of the risk.  Often in numerical problem solving, you can look at problems from more that one point of view.  This would be a model looking at hundreds or thousands of possible scenarios versus at most a few by an actuary.

Since this was before AI could do it, I used my own AI – Alan Intelligence.  I gathered historical loss and exposure data, for the three main divisions of UT: Electric Boat (nuclear submarines), Pratt & Whitney and Otis Elevator.  I wrote a simulation model and found that there was a 99.5% probability that losses would not exceed $12.4 million.  That was a far cry from $24 million.

I sent my analysis, along with a full report and extensive documentation to the Travelers suggesting a lower premium.  Time passed.

About a month later I got a call from the actuary on the account who said, “We received your report and analysis.  We’ll get back to you.”

Another month later, the Chief Casualty Actuary for Travelers called me to say, “Your analysis is very interesting.  We’re studying it and we’ll get back to you.”

Yet another month later I got a call from the Chief Actuary of The Travelers inviting me to come to Hartford for a meeting.  I packed up my reports and analyses, including the full simulation model source code (along with flow-charts and other documentation), packed it all into a box and took the early morning train to Hartford, CT. 

The Chief Actuary met me at the train station, and we drove to their headquarters where he had reserved a conference room where I was able to spread everything out on a large conference table.

After 2 ½ hours, he said, “Your analysis seems sound, but Simulation isn’t on the Exams.” 

When I asked him to explain, he told me that simulation modeling is not yet a part of Actuarial Science – it’s not included in the Exams. 

I replied that simulation modeling, which I had learned at Wharton, has been successfully applied in Aviation Products manufacturing and in Petroleum exploration for over 10 years.  He said he had heard of that. He was very gracious and took me back to the Hartford Club for a wonderful lunch, before returning me back to the train for my trip back to New York.

Several weeks later, we received a revised quote from The Travelers for $14 million for UT Products Liability – a reduction of $10 million in savings for our client UT. 

That was a lot of money then, when you could buy a mansion in Beverly Hills for $75,000.  The simulation model that I built was used for some years thereafter … and the savings for UT continued. 

It showed me that the power of a disruptive idea in almost any industry is slow to be accepted; but, can yield very meaningful and efficient results.  It requires an open mind to see the opportunity and the fearlessness to accept it.

Some time later, Simulation Modeling was included in the Actuarial Exams.