- Email: [email protected]
072019 (M10, E10) One agency's use of risk assessment and risk communication
268
Abstracts and Reviews
COX D.R., Wermuth N., Nuffield College, Oxford, United Kingdom, University of Mainz, Germany, Journal of the Royal Statistical Society: Applied Statistics, Vol. 43, nr. 2, 1994, pp. 347-355. After some discussion of the purposes of testing multivariate normality, the paper concentrates on two different approaches to testing linearity: on repeated regression tests of non-linearity and on exploiting properties of a dichotomized normal distribution. Regression tests of linearity are used to examine the adequacy of linear scoring systems for explanatory variables, initially recorded on an ordinal scale. (Authors) Keywords: Constancy of Variance, Multivariate Normality, Ordinal Scales. 072015 (MlO) Procedures for the detection of multiple changes in series of independent observations. Antoch J., Huskova M., Charles University, Prague, Contributions to Statistics: Asymptotic Statistics, Proceedings of the Fifth Prague Symposium, 1993, pp. 3-20. The paper concerns the problem of detection and identification of one or more change points in a series of independent observations. Procedures based on Mestimators and M-residuals are introduced, their limit properties studied and typical simulation results (Authors) presented. Keywords: Change Point Problem, M-estimators, Mresiduals, Abrupt Changes, Continuous Change. 072016 (MlO) Regression rank scores scale statistics and studentization in linear models. Jureckova J., Sen P.K., Charles University, Praha, University of North Carolina, USA, Contributions to Statistics: Asymptotic Statistics, Proceedings of the Fgth Prague Symposium, 1993, pp. 111-121. In linear models, regression-invariant and scaleequivariant regression rank scores scale statistics are proposed. Among their potential applications in statistical inference in linear models, studentization and testing for homoscedasticity are considered. (Authors) Keywords: GM-Estimators, M-Estimators, RegressionRegression-Equivariance, Regression Invariance, Qua&es, Scale-Equivariance. 072017 (MlO) Significance of differences of estimates.
Bocek P., Visek J.A., Academy of Sciences of the Czech Republic, Contributions to Statistics: Asymptotic Statistics, Proceedings of the Ffth Prague Symposium, 1993, pp. 195-202. A test for evaluating the significance of difference between the estimates of regression models is proposed. The idea is applicable on any estimator having Bahadur’s representation, nevertheless for the simplicity it is explained on a concrete example of one-step Mestimates. Performance of the test is illustrated by a numerical example. (Authors) Keywords: One-Step M-Estimates, Differences between Estimates. 072018 (MlO, ESO) Use of modified profile likelihood for improved tests of constancy of variance in regression. Simonoff J.S., Tsai C.-L., New York University, USA, University of California, Davis, USA, Journal of the Royal Statistical Society: Applied Statistics, Vol. 43, nr. 2, I994, pp. 357-370. Non-constantvariance(heteroscedasticity) iscommon in regression data, and many tests have been proposed for detecting it. This papers shows that the properties of likelihood-based tests can be improved by using the modified profile likelihood of Cox and Reid. A modified likelihood ratio test and modified score tests and intuitive are derived, and both theoretical justifications are given for the improved properties of the tests. The results of a Monte Carlo study are mentioned, which show that, whereas the ordinary likelihood ratio test can be very anticonservative, the modified test holds its null size well and is more powerful than the other tests. For non-normal error distributions, Studentized tests hold their size well (without being even for long-tailed error overconservative), distributions. Under short-tailed error distributions, likelihood ratio or Studentized score tests are most powerful, depending on the degree of heteroscedasticity. The modified versions of the score tests consistently outperform the unmodified versions. The use of these tests is demonstrated through analysis of data on the (Authors) volatility of stock prices. Keywords: Heteroscedasticity, Projle Likelihood Ratio Test, Risk of a Security, Score Test, Studentized Score Test. 072019 (MlO, ElO) One agency’s use of risk communication.
assessment
and
risk
Abstracts and Reviews
Fisher A., Chitose A., Gipson P.S., Risk Analysis, Vol. 14, nr. 2, 1994, pp, 207-212.
Few organizations have the courage to evaluate their own use of risk assessment (identifying hazards and estimating their probability and magnitude) and risk communication (interacting with internal and external stakeholder groups about risks). The USDA Animal and Plant Health Inspection Service (APHIS) wants to enhance its overall risk analysis process for managing a wide range of risks to animals, plants, and human health. The authors gathered survey data for a baseline of APHIS professionals’ understanding and use of risk communication. APHIS and risk assessment professionals spend a surprisingly large share of their time communicating about risks. They perceive that risk estimates influence decisions, but that risk estimates should have more influence. Respondents reported little opposition to APHIS risk management decisions, and little use of channels such as USDA Extension Service for disseminating risk messages. Substantial variance across responses is explained mostly by differences in the roles of the 11 work units (now 10) within the agency. Location also contributes to the variance. Demographic variables seem less important. (Authors) Keywords:
Risk
Communication,
Risk Assessment,
Baseline Survey, Needs Assessment,
269
financing of the pension benefits. Finally the author examines how the real value of the pension cover level depends on both real interest rate and the basic interest rate. (Author) Keywords:
Markov
Chains,
Thiele ‘s
Equation, DefinedContributions, of
Pensions,
Bonus
and
Dif-ferential
Continuous Adjustment Salary
Increases,
The
Equivalence Principle, Real Value of Pensions. 072021 (Ml 1)
Stochastic analysis of a portfolio of endowment insurance policies. Parker G., Scandinavian Actuarial Journal, nr. 2, 1994, pp. 119-130.
Parker (1994) presents a model for the present value of insurance benefits where the interest rates and future lifetimes are random. This paper presents a generalization of this model which can be used for portfolios of identical endowment insurance contracts. Illustrations of these moments when the force of interest is modeled by an Omstein-Uhlenbeck process are presented. When of interest, some comparisons with the corresponding moments for a portfolio of temporary insurance contracts are made. (Author) Keywords:
Force
of
Interest,
Process, Endowment Insurance,
Ornstein-Uhlenbeck Portfolio of Policies,
Moments of Insurance Functions. MI:
STOCHASTIC MODELS FOR CLAIM FREQUENCY, CLAIM SIZE, AND AGGREGATE CLAIMS
072022 (Mll,
072020 (Mll) Bonus, salary increases and real value of pensions. Linnemann P., Scandinavian Actuarial Journal, nr. 2,
Journal, nr. 2, 1994, pp. 139-150.
1994, pp. 99-118.
The author considers a pension scheme with a given payment combination based on the money purchase system. The contributions to the pension scheme are typically fixed as a certain percentage of the actual salary and the contributions are used for the purchase of benefits. Bonus is used for the current purchase of supplementary insurance with the same payment combination as for the policy. The author describes how the nominal value of the pension cover level increases because of continuous bonus and salary increases. The author finds that the equivalence principle holds for the actual premium and pension payments under natural assumptions. New versions of Thiele’s differential equation are derived and new insight is gained into the
M13) Double boundary crossing result for the Brownian motion. Teunen M., Goovaerts M., Scandinavian Actuarial
In a recent paper a general bounded crossing result for the Brownian motion is obtained for a linear upper boundary based on the method of Kac (195 1). Based on his main results the authors are able to develop in the present paper some simple expressions for crossing probabilities in case of a lower and an upper linear boundary. The authors will consider a Brownian motion process for the surplus of an insurance portfolio. This surplus must stay between two given bounds. If the surplus will cross the upper boundary, we can pay a dividend. The lower boundary can reflect the influence of control authorities and regulation measurements. (Authors) Keywords: Brownian Motion, Probability, Functional Integral.
Boundary
Crossing
Abstracts and Reviews
COX D.R., Wermuth N., Nuffield College, Oxford, United Kingdom, University of Mainz, Germany, Journal of the Royal Statistical Society: Applied Statistics, Vol. 43, nr. 2, 1994, pp. 347-355. After some discussion of the purposes of testing multivariate normality, the paper concentrates on two different approaches to testing linearity: on repeated regression tests of non-linearity and on exploiting properties of a dichotomized normal distribution. Regression tests of linearity are used to examine the adequacy of linear scoring systems for explanatory variables, initially recorded on an ordinal scale. (Authors) Keywords: Constancy of Variance, Multivariate Normality, Ordinal Scales. 072015 (MlO) Procedures for the detection of multiple changes in series of independent observations. Antoch J., Huskova M., Charles University, Prague, Contributions to Statistics: Asymptotic Statistics, Proceedings of the Fifth Prague Symposium, 1993, pp. 3-20. The paper concerns the problem of detection and identification of one or more change points in a series of independent observations. Procedures based on Mestimators and M-residuals are introduced, their limit properties studied and typical simulation results (Authors) presented. Keywords: Change Point Problem, M-estimators, Mresiduals, Abrupt Changes, Continuous Change. 072016 (MlO) Regression rank scores scale statistics and studentization in linear models. Jureckova J., Sen P.K., Charles University, Praha, University of North Carolina, USA, Contributions to Statistics: Asymptotic Statistics, Proceedings of the Fgth Prague Symposium, 1993, pp. 111-121. In linear models, regression-invariant and scaleequivariant regression rank scores scale statistics are proposed. Among their potential applications in statistical inference in linear models, studentization and testing for homoscedasticity are considered. (Authors) Keywords: GM-Estimators, M-Estimators, RegressionRegression-Equivariance, Regression Invariance, Qua&es, Scale-Equivariance. 072017 (MlO) Significance of differences of estimates.
Bocek P., Visek J.A., Academy of Sciences of the Czech Republic, Contributions to Statistics: Asymptotic Statistics, Proceedings of the Ffth Prague Symposium, 1993, pp. 195-202. A test for evaluating the significance of difference between the estimates of regression models is proposed. The idea is applicable on any estimator having Bahadur’s representation, nevertheless for the simplicity it is explained on a concrete example of one-step Mestimates. Performance of the test is illustrated by a numerical example. (Authors) Keywords: One-Step M-Estimates, Differences between Estimates. 072018 (MlO, ESO) Use of modified profile likelihood for improved tests of constancy of variance in regression. Simonoff J.S., Tsai C.-L., New York University, USA, University of California, Davis, USA, Journal of the Royal Statistical Society: Applied Statistics, Vol. 43, nr. 2, I994, pp. 357-370. Non-constantvariance(heteroscedasticity) iscommon in regression data, and many tests have been proposed for detecting it. This papers shows that the properties of likelihood-based tests can be improved by using the modified profile likelihood of Cox and Reid. A modified likelihood ratio test and modified score tests and intuitive are derived, and both theoretical justifications are given for the improved properties of the tests. The results of a Monte Carlo study are mentioned, which show that, whereas the ordinary likelihood ratio test can be very anticonservative, the modified test holds its null size well and is more powerful than the other tests. For non-normal error distributions, Studentized tests hold their size well (without being even for long-tailed error overconservative), distributions. Under short-tailed error distributions, likelihood ratio or Studentized score tests are most powerful, depending on the degree of heteroscedasticity. The modified versions of the score tests consistently outperform the unmodified versions. The use of these tests is demonstrated through analysis of data on the (Authors) volatility of stock prices. Keywords: Heteroscedasticity, Projle Likelihood Ratio Test, Risk of a Security, Score Test, Studentized Score Test. 072019 (MlO, ElO) One agency’s use of risk communication.
assessment
and
risk
Abstracts and Reviews
Fisher A., Chitose A., Gipson P.S., Risk Analysis, Vol. 14, nr. 2, 1994, pp, 207-212.
Few organizations have the courage to evaluate their own use of risk assessment (identifying hazards and estimating their probability and magnitude) and risk communication (interacting with internal and external stakeholder groups about risks). The USDA Animal and Plant Health Inspection Service (APHIS) wants to enhance its overall risk analysis process for managing a wide range of risks to animals, plants, and human health. The authors gathered survey data for a baseline of APHIS professionals’ understanding and use of risk communication. APHIS and risk assessment professionals spend a surprisingly large share of their time communicating about risks. They perceive that risk estimates influence decisions, but that risk estimates should have more influence. Respondents reported little opposition to APHIS risk management decisions, and little use of channels such as USDA Extension Service for disseminating risk messages. Substantial variance across responses is explained mostly by differences in the roles of the 11 work units (now 10) within the agency. Location also contributes to the variance. Demographic variables seem less important. (Authors) Keywords:
Risk
Communication,
Risk Assessment,
Baseline Survey, Needs Assessment,
269
financing of the pension benefits. Finally the author examines how the real value of the pension cover level depends on both real interest rate and the basic interest rate. (Author) Keywords:
Markov
Chains,
Thiele ‘s
Equation, DefinedContributions, of
Pensions,
Bonus
and
Dif-ferential
Continuous Adjustment Salary
Increases,
The
Equivalence Principle, Real Value of Pensions. 072021 (Ml 1)
Stochastic analysis of a portfolio of endowment insurance policies. Parker G., Scandinavian Actuarial Journal, nr. 2, 1994, pp. 119-130.
Parker (1994) presents a model for the present value of insurance benefits where the interest rates and future lifetimes are random. This paper presents a generalization of this model which can be used for portfolios of identical endowment insurance contracts. Illustrations of these moments when the force of interest is modeled by an Omstein-Uhlenbeck process are presented. When of interest, some comparisons with the corresponding moments for a portfolio of temporary insurance contracts are made. (Author) Keywords:
Force
of
Interest,
Process, Endowment Insurance,
Ornstein-Uhlenbeck Portfolio of Policies,
Moments of Insurance Functions. MI:
STOCHASTIC MODELS FOR CLAIM FREQUENCY, CLAIM SIZE, AND AGGREGATE CLAIMS
072022 (Mll,
072020 (Mll) Bonus, salary increases and real value of pensions. Linnemann P., Scandinavian Actuarial Journal, nr. 2,
Journal, nr. 2, 1994, pp. 139-150.
1994, pp. 99-118.
The author considers a pension scheme with a given payment combination based on the money purchase system. The contributions to the pension scheme are typically fixed as a certain percentage of the actual salary and the contributions are used for the purchase of benefits. Bonus is used for the current purchase of supplementary insurance with the same payment combination as for the policy. The author describes how the nominal value of the pension cover level increases because of continuous bonus and salary increases. The author finds that the equivalence principle holds for the actual premium and pension payments under natural assumptions. New versions of Thiele’s differential equation are derived and new insight is gained into the
M13) Double boundary crossing result for the Brownian motion. Teunen M., Goovaerts M., Scandinavian Actuarial
In a recent paper a general bounded crossing result for the Brownian motion is obtained for a linear upper boundary based on the method of Kac (195 1). Based on his main results the authors are able to develop in the present paper some simple expressions for crossing probabilities in case of a lower and an upper linear boundary. The authors will consider a Brownian motion process for the surplus of an insurance portfolio. This surplus must stay between two given bounds. If the surplus will cross the upper boundary, we can pay a dividend. The lower boundary can reflect the influence of control authorities and regulation measurements. (Authors) Keywords: Brownian Motion, Probability, Functional Integral.
Boundary
Crossing