Description: Basic feature engineering uses input data to create dummy variables, log transforms, and other single-value transformations. This talk builds upon these univariate transformations and presents several strategies for creating complex multivariate features that can enhance your model's lift. Topics such as target encoding, dimension reduction/projection, row/column, and other feature engineering topics will be covered.

Description: Risk management techniques such as stress testing and setting capital targets are fairly well established for life insurers; however, the implementation of these techniques can vary considerably. A fundamental question is how to manage multiple valuation perspectives, and the list of possibilities keeps growing. Companies can potentially consider some or all of the following: U.S. reporting measures (Stat, GAAP), internal measures (EV, VNB, economic capital), international reporting measures (IFRS), rating agency models or other developing standards (ICS, NAIC group capital). This session will begin with a quick overview of the defining characteristics of different valuation frameworks. We will then provide more focused and practical considerations for managing risk and capital under different frameworks, using specific examples to highlight key trade offs.

Description: Data is an essential component of financial analysis. In this session, the presenters will discuss data governance/management framework and how it can be applied to US GAAP LDTI and VM-51 regulations. This session will also discuss the important relationship between actuaries and IT and how they can work together towards best practices for extracting, managing and using data needed for various actuarial models. Real life case studies relevant for valuation actuaries (e.g. data needs for LDTI and VM-51) will also be presented and examined, highlighting best practices as well as challenges and pitfalls for setting up and governing data processes along with areas for where good communication and strong relationships are important to success.

Description: This speech shares the practical experience of platform construction in the actuarial practice work of Ping An People Life, at the same time, shares the development mode of the actuarial system combining online and offline, and summarizes and expounds the advantages and disadvantages of this mode.

Description: For insurance companies, modeling the future values of assets and their interactions is useful for setting up asset strategies and managing reserves for liabilities. These are the presentation slides for a research paper that evaluates the ability of Wilkie’s stochastic investment models to predict Toronto Stock Exchange [TSX] price index yield. In the Wilkie Model, correlated econometric indices are modeled through a cascade structure.

Description: This presentation examines the measurement and hedging of synthetic CDO tranche spread risks based on market spread data following the sub-prime crisis. A range of methods proposed for pricing bespoke CDOs are examined to assess their ability to hedge the credit spread risk. The methods assessed are calibrated to the traded CDO index spread and then compared based on the mean absolute pricing errors over a time period including the sub-prime crisis. Standard pricing methods and variations used to price bespoke CDOs generally perform poorly in hedging credit spread risk. Past data can be used to improve the performance of the methods. The results of this analysis also raise concerns with the accuracy of 'mark-to-model' valuations of bespoke CDOs using standard market methods.

Description: In December 2007, Goldman Sachs launched a product called QxX index swap, which is designed to allow market participants to hedge or gain exposure to longevity and mortality risks. This presentation offers a quantitative analysis of this brand new financial innovation. First of all, the presentation sets up a risk-neutral framework to price QxX index swaps. This framework, which is based on the dynamics of death rates under a two-factor stochastic mortality model in a risk-adjusted probability measure, yields prices [spreads] that are fairly close to the spreads that Goldman Sachs currently offers. The presentation then explores the uncertainty involved in this model-based pricing framework. Specifically, it studies parameter risk by using Markov Chain Monte Carlo [MCMC] and model risk by examining structural changes in mortality dynamics. The results indicate that both model risk and parameter risk are significant. Actuaries should therefore be aware of these issues when placing a value on a longevity index swap.

Description: Characteristics of Asian Indexed Executive Options. Presented at August 2011 Actuarial Research Conference.

Description: This presentation illustrates a calibration model against 60 years of historical data using a pragmatic mixture of filtering, maximum likelihood and method of moments techniques, and compare the resulting model with interest rates in contemporary economic scenario generators.

Description: Predictive modeling has ranked highly in terms of technical skills desired by employers of actuaries. In recent surveys it has been identified as a key opportunity for actuarial professionals to add more value to their organizations. While there has been an increased focus on predictive modeling at industry conferences, the subject lends itself very well to a hands-on learning experience. That is one of the motivations behind organizing this boot camp.