The heat wave model for constructing two-dimensional mortality improvement scales with measures of uncertainty

Recently, the actuarial professions in various countries have adopted an innovative two-dimensional approach to projecting future mortality. In contrast to the conventional approach, the two-dimensional approach permits mortality improvement rates to vary with not only age but also time. Despite being an important breakthrough, the currently used two-dimensional mortality improvement scales are subject to several limitations, most notably a heavy reliance on subjective judgments and a lack of measures of uncertainty. In view of these limitations, in this paper we introduce a new model known as the heat wave model, in which short- and long-term mortality improvements are treated respectively as ‘heat waves’ that taper off over time and ‘background improvements’ that always exist. Using the heat wave model, one can derive two-dimensional mortality improvement scales that entail minimal subjective judgment and include measures of the uncertainty.     

Macaulay durations for nonparallel shifts

Macaulay duration is a well-known and widely used interest rate risk measure. It is commonly believed that it only works for parallel shifts of interest rates. We show in this paper that this limitation is largely due to the traditional parametric modelling and the derivative approach, the Macaulay duration works for non-parallel shifts as well when the non-parametric modelling and the equivalent zero coupon bond approach are used. We show that the Macaulay duration provides the best one-number sensitivity information for non-parallel interest rate changes and that a Macaulay duration matched portfolio is least vulnerable to the downside risk caused by non-parallel rate changes under some verifiable conditions. AMS Classification 65K10 · 90C90     

A hierarchical model for the joint mortality analysis of pension scheme data with missing covariates

A hierarchical model is developed for the joint mortality analysis of pension scheme datasets. The proposed model allows for a rigorous statistical treatment of missing data. While our approach works for any missing data pattern, we are particularly interested in a scenario where some covariates are observed for members of one pension scheme but not the other. Therefore, our approach allows for the joint modelling of datasets which contain different information about individual lives. The proposed model generalizes the specification of parametric models when accounting for covariates. We consider parameter uncertainty using Bayesian techniques. Model parametrization is analysed in order to obtain an efficient MCMC sampler, and address model selection. The inferential framework described here accommodates any missing-data pattern, and turns out to be useful to analyse statistical relationships among covariates. Finally, we assess the financial impact of using the covariates, and of the optimal use of the whole available sample when combining data from different mortality experiences.     

Forecasting mortality rate improvements with a high-dimensional VAR

Forecasting mortality rates is a problem which involves the analysis of high-dimensional time series. Most of usual mortality models propose to decompose the mortality rates into several latent factors to reduce this complexity. These approaches, in particular those using cohort factors, have a good fit, but they are less reliable for forecasting purposes. One of the major challenges is to determine the spatial–temporal dependence structure between mortality rates given a relatively moderate sample size. This paper proposes a large vector autoregressive (VAR) model fitted on the differences in the log-mortality rates, ensuring the existence of long-run relationships between mortality rate improvements. Our contribution is threefold. First, sparsity, when fitting the model, is ensured by using high-dimensional variable selection techniques without imposing arbitrary constraints on the dependence structure. The main interest is that the structure of the model is directly driven by the data, in contrast to the main factor-based mortality forecasting models. Hence, this approach is more versatile and would provide good forecasting performance for any considered population. Additionally, our estimation allows a one-step procedure, as we do not need to estimate hyper-parameters. The variance–covariance matrix of residuals is then estimated through a parametric form. Secondly, our approach can be used to detect nonintuitive age dependence in the data, beyond the cohort and the period effects which are implicitly captured by our model. Third, our approach can be extended to model the several populations in long run perspectives, without raising issue in the estimation process. Finally, in an out-of-sample forecasting study for mortality rates, we obtain rather good performances and more relevant forecasts compared to classical mortality models using the French, US and UK data. We also show that our results enlighten the so-called cohort and period effects for these populations.     

Modeling mortality with a Bayesian vector autoregression

Parametric mortality models capture the cross section of mortality rates. These models fit the older ages better, because of the more complex cross section of mortality at younger and middle ages. Dynamic parametric mortality models fit a time series to the parameters, such as a Vector-auto-regression (VAR), in order to capture trends and uncertainty in mortality improvements. We consider the full age range using the Heligman and Pollard (1980) model, a cross-sectional mortality model with parameters that capture specific features of different age ranges. We make the Heligman–Pollard model dynamic using a Bayesian Vector Autoregressive (BVAR) model for the parameters and compare with more commonly used VAR models. We fit the models using Australian data, a country with similar mortality experience to many developed countries. We show how the Bayesian Vector Autoregressive (BVAR) models improve forecast accuracy compared to VAR models and quantify parameter risk which is shown to be significant.     

Forecasting and simulating mortality tables

In this paper we suggest solutions to the actuaries, facing the problem of estimating future mortality tables, especially in cases where there is a lack of relevant data and where the tendencies are not easy to estimate directly. We propose the utilization of external sources of information in the form of other, published mortality tables and use formal statistical tests to decide among these possible candidates. The procedure can also be applied for checking e.g. the goodness of mortality selection factors. We suggest the use of parametric families in modelling; for example the simple 2-parameter Azbel model. We conclude the paper by a simulation study which allows for the quantification of the possible risks related to unforeseen changes in the mortality tables in the future. To calibrate the variances of these models, initial estimates are needed, which we get by the Lee–Carter method.     

Measurement of Longevity Risk Using Bootstrapping for Lee–Carter and Generalised Linear Poisson Models of Mortality

This paper provides a comparative investigation of simulation strategies for measuring the longevity risk associated with predictions of mortality rates and derived estimates of life expectancy. The study considers the Lee–Carter framework and a generalised linear Poisson model for representing the dynamics of mortality, as well as enhancements that allow for joint modelling of the dispersion and the effect of using a negative binomial rather than a Poisson assumption.      

Automated generation of LFT-based parametric uncertainty descriptions from generic aircraft models

A computer assisted modelling methodology is developed for the generation of linearized models with parametric uncertainties described by Linear Fractional Transformations (LFTs). The starting point of the uncertainty modelling is a class of generic nonlinear aircraft models with explicit parametric dependence used for simulation purposes. The proposed methodology integrates specialized software tools for object-oriented modelling, for simulation, and for numerical as well as symbolic computations. The methodology has many generic features being applicable to similar nonlinear model classes.     

Constructions of economic functions and calculations of marginal rates in DEA using parametric optimization methods

The aim of the paper is to present and substantiate a technique to visualize DEA modelling results without any loss of mathematical rigour. The proposed family of parametric optimization methods allows one to construct an intersection of the multidimensional frontier with a two-dimensional plane determined by any pair of given directions. This approach reduces the efficiency analysis of production units to the investigation of well-known functions in economics. We also propose constructive methods to calculate marginal rates of substitution, marginal rates of transformation and so on.     

Parallel Dantzig–Wolfe decomposition of petroleum production allocation problems

This article discusses the optimization of a petroleum production allocation problem through a parallel Dantzig–Wolfe algorithm. Petroleum production allocation problems are problems in which the determination of optimal production rates, lift gas rates and well connections are the central decisions. The motivation for modelling and solving such optimization problems stems from the value that lies in an increased production rate and the current lack of integrated software that considers petroleum production systems as a whole. Through our computational study, which is based on realistic production data from the Troll West field, we show the increase in computational efficiency that a parallel Dantzig–Wolfe algorithm offers. In addition, we show that previously implemented standard parallel algorithms lead to an inefficient use of parallel resources. A more advanced parallel algorithm is therefore developed to improve efficiency, making it possible to scale the algorithm by adding more CPUs and thus approach a reasonable solution time for realistic-sized problems.     

Mortality density forecasts: An analysis of six stochastic mortality models

This paper develops a framework for developing forecasts of future mortality rates. We discuss the suitability of six stochastic mortality models for forecasting future mortality and estimating the density of mortality rates at different ages. In particular, the models are assessed individually with reference to the following qualitative criteria that focus on the plausibility of their forecasts: biological reasonableness; the plausibility of predicted levels of uncertainty in forecasts at different ages; and the robustness of the forecasts relative to the sample period used to fit the model. An important, though unsurprising, conclusion is that a good fit to historical data does not guarantee sensible forecasts. We also discuss the issue of model risk, common to many modelling situations in demography and elsewhere. We find that even for those models satisfying our qualitative criteria, there are significant differences among central forecasts of mortality rates at different ages and among the distributions surrounding those central forecasts.     

Discussion of “Copulas: Tales and facts”, by Thomas Mikosch

As Prof. Mikosch correctly points out, there exists very little sound statistical theory on modelling dependence using copulas. In this contribution, an open problem is presented concerning the efficient estimation of the parameter of a copula when no parametric assumptions are made regarding the marginal distributions.     

Valuation of contingent claims with mortality and interest rate risks

We consider the pricing of life insurance contracts under stochastic mortality and interest rates assumed not independent of each other. Employing the method of change of measure together with the Bayes’ rule for conditional expectations, solution expressions for the value of common contracts are obtained. A demonstration of how to apply our proposed stochastic modelling approach to value survival and death benefits is provided. Using the Human Mortality Database and UK interest rates, we illustrate that the dependence between interest rate and mortality dynamics has considerable impact in the value of even a simple survival benefit.     

A three-step local smoothing approach for estimating the mean and covariance functions of spatio-temporal Data

Spatio-temporal data are common in practice. Existing methods for analyzing such data often employ parametric modelling with different sets of model assumptions. However, spatio-temporal data in practice often have complicated structures, including complex spatial and temporal data variation, latent spatio-temporal data correlation, and unknown data distribution. Because such data structures reflect the complicated impact of confounding variables, such as weather, demographic variables, life styles, and other cultural and environmental factors, they are usually too complicated to describe by parametric models. In this paper, we suggest a general modelling framework for estimating the mean and covariance functions of spatio-temporal data using a three-step local smoothing procedure. The suggested method can well accommodate the complicated structure of real spatio-temporal data. Under some regularity conditions, the consistency of the proposed estimators is established. Both simulation studies and a real-data application show that our proposed method could work well in practice.

     

Spatial patterns of mortality in the United States: A spatial filtering approach

Within the field of spatial analysis, filtering has emerged as a powerful modelling technique to retrieve systematic patterns of geographical variation. In this work, spatial filtering is introduced in the context of modelling mortality for the first time in actuarial literature. Specifically, the objective of this research is to identify patterns in the spatial distribution of mortality rates for the sixty-five and older age group at the county level for the contiguous United States. The analysis carried out pertains to the spatial autocorrelation of these rates over time. The spatial filtering methodology is applied to uncover latent spatial patterns. Furthermore, the analysis reveals the extent to which these spatial patterns remained consistent across the years in the study. Results show the existence of spatial dependencies leading to an accompanying spatial filter. Thus, incorporating spatial filters as an exploratory tool in spatial analysis proves of considerable use when it comes to enriching traditional mortality models.     

Computing parallel curves on parametric surfaces

Generating parallel curves on parametric surfaces is an important issue in many industrial settings. Given an initial curve (called the base curve or generator) on a parametric surface, the goal is to obtain curves on the surface that are parallel to the generator. By parallel curves we mean curves that are at a given distance from the generator, where distance is measured point-wise along certain characteristic curves (on the surface) orthogonal to the generator. Except for a few particular cases, computing these parallel curves is a very difficult and challenging problem. In fact, only partial, incomplete solutions have been reported so far in the literature. In this paper we introduce a simple yet efficient method to fill this gap. In clear contrast with other existing techniques, the most important feature of our method is its generality: it can be successfully applied to any differentiable parametric surface and to any kind of characteristic curves on surfaces. To evaluate our proposal, some illustrative examples (not addressed with previous methods) for the cases of section, vector-field, and geodesic parallels are discussed. Our experimental results show the excellent performance of the method even for the complex case of NURBS surfaces.     

Survival Analysis of Latvian Population Using Tchebyshev's Polynomials

Several laws have been used in history to characterize mortality. Some of them are rather simple, like Gompertz's law, or quite complicated, like Heligman and Pollard's law. But no law fits very well for all countries and in all time intervals. A similar situation exists in Latvia too. Therefore the main idea of this paper was not to find concrete mortality law but to find its approximation by Fourier and Tchebyshev's series to use an appropriate number of terms. Experimentally, was found that good approximation can be reached using nine terms. Further behavior of constants with the aim of projecting them, was investigated in relation to the approximation of mortality law by the Fourier–Tchebyshev series for a large range of data (1973–1997).     

Forecasting mortality in subpopulations using Lee–Carter type models: A comparison

The relative performance of multipopulation stochastic mortality models is investigated. When targeting mortality rates, we consider five extensions of the well known Lee–Carter single population extrapolative approach. As an alternative, we consider similar structures when mortality improvement rates are targeted. We use a dataset of deaths and exposures of Italian regions for the years 1974–2008 to conduct a comparison of the models, running a battery of tests to assess the relative goodness of fit and forecasting capability of different approaches. Results show that the preferable models are those striking a balance between complexity and flexibility.     

A health insurance pricing model based on prevalence rates: Application to critical illness insurance

The Italian health insurance market is currently undersized. The paucity of assured data and the discontinuous statistical surveys carried out by the National Institute of Statistics (ISTAT) represent one of the main obstacles to the insurance market development. The paper sets forth a parametric model to estimate technical basis for health insurance policies when data are limited and only aggregated information on mortality and morbidity is available. The probabilistic framework is based on a multiple state continuous and time inhomogeneous Markov model. We provide an estimate of transition intensities from the healthy state to the sickness state when only prevalence rates of sickness are available, according to an extension and modification of the methodology proposed in Olivieri (1996) for Long Term Care insurance. We assume that mortality intensity of both healthy and sick lives is modelled by two independent Gompertz–Makeham models.