Non-Life Pricing Seminar 2019

12.30pm – 12.45pm: Welcome & Introduction (Cecilia Cheuk)

12.45pm – 1.30pm: Winner’s Curse Application of Game Theory to Insurance Pricing (Andrew Smith)

• Andrew Smith analyses competitive auctions where several insurers place bids to insure a risk, as can occur with internet price comparison sites
• An insurer is most likely to win the proposals where the risk is under-priced
• The winner’s curse is either losing the business, or winning the auction but then losing money
• The impact of winner’s curse is related to the standard errors of parameter estimates in modelling of expected insurance claims
• Insurers can mitigate winner’s curse by raising quotes or declining to quote
• Andrew Smith compares methods for calculating the appropriate price shading
• These are illustrated with numerical examples

1.30pm – 2.15pm: Pricing in an Increasingly Connected & Automated World (Graham Wright & Richard Clarkson)

• Automated pricing methodologies, techniques and applications
• The industrialisation of data enrichment
• Automated underwriting judgement and empirical decision support.
• The connected systems environment to support this new world.

2.15pm – 3.00pm: Parallel Breakout Sessions:

1. Commercial Pricing: from Technical Models to Trading (Matthew Brophy)

• An overview of the Commercial Pricing techniques used in the Irish Market today 
• The main challenges faced by Commercial Pricing actuaries and some of the practical considerations 
• The growing importance of interacting with the business to achieve effective Trading. 

2. Unstructured Data - Getting started with Image Recognition & Text Analytics (Eoin Ó Baoighill & Edward Roche)

• Unstructured data can be a gold mine, or a bottomless pit. How can you use unstructured data efficiently?
• This talk will cover practical applications of image recognition and natural language processing.
• We will cover the principal methods used, and the pipeline needed.
• Practical advice for getting started and sample R and Python code will be provided

3.00pm – 3.30pm: Coffee Break

3.30pm – 4.15pm: Meddling with Modelling - Who Knows Best? (Ben Symington & Simon Warsop)

• With a focus on implementation of models and impact to the end customer, we will identify and discuss some of the benefits and pitfalls of each chapter in pricing sophistication.
• With increasing scrutiny of GI pricing practice from a range of observers continuing to make this a hot topic, the presenters will role play real world man vs. machine examples to bring our views and wider perspectives to life. If you are considering motor pricing factors or pricing motor factories then this session will inform and provoke.
• Key learning points will be:
  • The pros and cons from a customer perspective of each chapter of the modelling journey.
  • Key points to consider when making changes to your model.
  • How to balance continued evolution of pricing sophistication and real customer outcomes.

4.15pm – 5.00pm: A Simple Introduction to Excess of Loss Reinsurance Pricing Using Monte Carlo Markov Chains (Chris Gibney)

• This presentation will provide a simple, intuitive introduction to MCMC.
• Through a simple excess of loss pricing case study, we will identify the limitations of traditional frequency-severity modelling.
• A simple MCMC model will be presented as an alternative to overcome these limitations.
• Finally, we will walk through the code (R or Python -TBC) implementing the MCMC model.

5.00pm: Closing Remarks (Cecilia Chuek)