精算论坛第138期讲座-冯润桓(5月30日)

报告人：冯润桓教授  （美国伊利诺伊大学香槟分校）

Dr. Runhuan Feng is an Associate Professor and the Director ofActuarial Science at the University of Illinois. He is a Fellow of the Societyof Actuaries and a Chartered Enterprise Risk Analyst. He is also the State FarmCompanies Foundation Scholar in Actuarial Science. Dr. Feng has published extensivelyon computational techniques of risk metrics for investment-combined insuranceproducts. Most recently, he was commissioned by the Society of Actuaries toconduct the rst North America industry-wide survey on nested stochasticmodeling and performed subsequent research study to create resources for financialreporting actuaries on computational methods to speed up nested simulations.

报告一：

报告题目：条件亚式期权  

报告时间：2018年5月30日下午2:00—2:50

摘要：

ConditionalAsian options are recent market innovations, which offer cheaper and long-datedalternatives to regular Asian options. In contrast with payoffs from regularAsian options which are based on average asset prices, the payoffs fromconditional Asian options are determined only by average prices above certainthreshold. Due to the limited inclusion of prices, conditional Asian optionsfurther reduce the volatility in the payoffs than their regular counterpartsand have been promoted in the market as viable hedging and risk managementinstruments for equity-linked life insurance products. There has been noprevious academic literature on this subject and practitioners have only beenknown to price these products by simulations. We propose the first analyticalapproach to computing prices and deltas of conditional Asian options incomparison with regular Asian options. In the numerical examples, we put to thetest some cost-benefit claims by practitioners. As a by-product, the work alsopresents some distributional properties of the occupation time and thetime-integral of geometric Brownian motion during the occupation time.

报告二：

报告题目：高效嵌套模拟的新技术—样本回收法  

报告时间：2018年5月30日下午3:00—3:50

摘要：

As moreregulatory reporting requirements in the regulatory regimes around the worldmove towards dependence on stochastic approaches, insurance companies areexperiencing increasing difficulty with detailed forecasting and more accuratevaluation and risk assessment based on Monte Carlo simulations.

Stochasticmodeling is commonly used by financial reporting actuaries whenever reporting procedures,such as reserving and capital requirement calculation, are performed undervarious economic scenarios, which are stochastically determined. Nestedstochastic modeling is required whenever modeling components under eacheconomic scenario are determined by stochastic scenarios in the further future.

Many existingtechniques to speed up runtime of nested simulations is based on the reductionof inner loop calcuations by curve fitting techniques. The essence of thesetechniques is to develope a functional relationship between risk factors(equity values, interest rates, etc) and target features (insurance liabilityor their greeks) of inner loop calculations. Such functional relationship canbe approximated by multivariate interpolation or smoothing techniques such asleast squares Monte Carlo. Nonetheless, these techniques often require a largesize of economic scenarios to develop accurate enough functional relationships,which could also be very costly to begin with. The proposed new technique is basedon an entirely different strategy, which is to avoid ``approximate"functional relationship but instead to save time by recycling a limited set ofeconomic scenarios. We shall demonstrate with a variety of examples theefficiency and wide applicability of this new method.

报告地点：中央财经大学学术会堂南楼506（精算院会议室）

欢迎各位老师和同学积极参加！