Risk measurement

There’s more to risk measurement than just performing value at risk calculations, so I have a lot more to talk about here than just calculating and reviewing ICAs and SCRs.

While at Gen Re (1999-2000) I discussed with potential clients how swaptions could be used to hedge the interest rate risk arising from guaranteed annuity options (GAOs).  This didn’t just require imagination and creativity in finding ways to illustrate the risks but also a keen understanding of how the firm’s interest rate risks interacted with its equity risks.  An increase in equity prices could have increased the value of some firms’ GAOs and decreased the values of others’.

While at PwC (2000-2006):
• I provided all the actuarial input to a project to build a real world stochastic model for a firm with a portfolio of shared appreciation mortgages.
• I helped a new firm develop an approach and a set of assumptions to calculate its ICA.
• I convinced a PwC director to challenge a firm’s real world economic scenario generator by producing graphs showing the “funnels of doubt” for a number of market observables.

At Prudential (2008-10), as described here, I developed an equity hedging strategy within the with profits fund.  Within this project:
• I developed the concept of solvency frontiers as a simple way to illustrate the impact of different hedges on the fund’s equity and interest rate risks.
• I calculated the impact on the benchmark option of all possible combined equity and interest rate stresses and used this to estimate the impact on the with profits fund’s solvency for all possible combined equity and interest rate stresses.
• I performed a stochastic analysis to illustrate the potential impact of all the extra gamma risks that the firm was taking by choosing to delta hedge rather than buying options.
• I recommended how to adjust the firm’s ICA to allow for the hedge being in place (without extensive work on the ICA models themselves).

While with the FSA and PRA (2010-14):
• I performed a number of ICA reviews.
• I was co-leader of the PRA’s market risk specialism, leading the team through the development of an approach to reviewing market risk modules within Solvency II internal models.
• I used risk measures like the ICA and stress and scenario testing to review the implications for policyholder security of capital management policies and Part VII transfers.

While contracting at a bulk annuity insurer:
• I reviewed the business plan assumptions and the stress scenarios that had been applied to the base run.  I recommended the investigation of a number of extra scenarios to help the firm understand particular risks.
• I made changes to the counterparty risk module within the firm’s internal model.
• I quantified the potential impact on the firm’s solvency ratios of reinsurer defaults on a range of possible dates.  To do this I needed to estimate the impact on the ICA/SCR of the changes in risk profile from recapturing large amounts of longevity risk.

As a contractor I helped a major insurer design and build an internal model for Solvency II:
• I designed and documented models for equity, gilt yield, swap spread, credit spread, default/migration and lapse risks.
• In all cases I tested the resulting probability distributions and benchmarked them against those of other firms.
• I built spreadsheet-based t-copulas to experiment with & to help the client understand their properties.
• I showed the client how to convert variances of monthly returns to variances appropriate to use with a model of annual returns and extended this to a methodology for annualising monthly correlations.

As a contractor I helped a major insurer to understand the impact on their capital requirements of changes to the migration and default methodology within their internal model.  This involved a deep dive into the mechanics of the model.

As a contractor with another major insurer, as well as documenting the data, assumptions and internal model methodology for determining capital requirements for the credit risks associated with their mortgage portfolio, I also documented their internal mortgage credit rating methodology and internal mortgage valuation methodology.  I designed and actioned test plans for their bond spread calibration model, the mortgage spread/migration model and the model that calculated the matching adjustment following stress (and portfolio rebalancing).  I designed a new approach to validating the capping of mortgage spread stresses.

As a contractor, I calibrated and validated listed equity, private equity, property and interest rate volatility internal models for a European insurer. This included identifying suitable data sources, performing statistical tests on data, unsmoothing autocorrelated returns, trying different distributions and calibration methodologies, separating volatility movements into principal components.

The use of loss functions will be important in the SCR calculation process, and least squares Monte Carlo (LSMC) is a technique that reduces the time spent on developing them.  As a contractor, I ran through a dry run process for developing loss functions (using LSMC) for a major insurer, identifying where existing Excel and R tools and the instructions for using them needed to be improved if the process were to work successfully in a production environment.