Search

Modeling and forecasting of mortality rates are closely related to a wide range of actuarial practices, such as the designing of pension schemes. To improve the forecasting accuracy, age coherence is incorporated in many recent mortality models, which suggests that the long-term forecasts will not diverge infinitely among age groups. Despite their usefulness, misspecification is likely to occur for individual mortality models when applied in empirical studies. The reliableness and accuracy of forecast rates are therefore negatively affected. In this study, an ensemble averaging or model averaging (MA) approach is proposed, which adopts age-specific weights and asymptotically achieves age coherence in mortality forecasting. The ensemble space contains both newly developed age-coherent and classic age-incoherent models to achieve the diversity. To realize the asymptotic age coherence, consider parameter errors, and avoid overfitting, the proposed method minimizes the variance of out-of-sample forecasting errors, with a uniquely designed coherent penalty and smoothness penalty. Our empirical data set include ten European countries with mortality rates of 0–100 age groups and spanning 1950–2016. The outstanding performance of MA is presented using the empirical sample for mortality forecasting. This finding robustly holds in a range of sensitivity analyses. A case study based on the Italian population is finally conducted to demonstrate the improved forecasting efficiency of MA and the validity of the proposed estimation of weights, as well as its usefulness in actuarial applications such as the annuity pricing.

Influential existing research has suggested that rather than being static, mortality declines decelerate at young ages and accelerate at old ages. Without accounting for this feature, the forecast mortality rates of the popular Lee–Carter (LC) model are less reliable in the long run. To provide more accurate mortality forecasting, we introduce a time-varying coefficients extension of the LC model by adopting the effective kernel methods. With two frequently used kernel functions, Epanechnikov (LC-E) and Gaussian (LC-G), we demonstrate that the proposed extension is easy to implement, incorporates the rotating patterns of mortality decline and is straightforwardly extensible to multi-population cases. Using a large sample of 15 countries over 1950–2019, we show that LC-E and LC-G, as well as their multi-population counterparts, can consistently improve the forecasting accuracy of the competing LC and Li–Lee models in both single- and multi-population scenarios.

Novel navigation applications provide a driving behavior score for each finished trip to promote safe driving, which is mainly based on experts’ domain knowledge. In this paper, with automobile insurance claims data and associated telematics car driving data, we propose a supervised driving risk scoring neural network model. This one-dimensional convolutional neural network takes time series of individual car driving trips as input and returns a risk score in the unit range of (0,1). By incorporating credibility average risk score of each driver, the classical Poisson generalized linear model for automobile insurance claims frequency prediction can be improved significantly. Hence, compared with non-telematics-based insurers, telematics-based insurers can discover more heterogeneity in their portfolio and attract safer drivers with premiums discounts.

Over-dispersed Poisson chain-ladder models are widely used in general insurance claims reserving. Although such models can accommodate the over-dispersion frequently observed in run-off triangles, they also impose an additional constraint of fixed variance to mean ratio across cells. In this paper, we relax this constraint and develop a flexible dispersion structure in a double Poisson chain-ladder model. The proposed model nests the classic over-dispersed Poisson model as a special case. A generalized likelihood ratio test is further proposed to compare different dispersion structures. In contrast to the existing claims reserving methods, our proposed method is more flexible in terms of the dispersion modelling. Simulation and empirical studies are conducted to demonstrate the importance of flexible dispersion modelling.

This paper proposes a multi-population Global Vector Autoregression (GVAR) mortality model, in which the joint mortality dynamics of all populations are formulated in a large vector autoregression model with a set of contemporaneous global mortality factors. By decomposing the large VAR regression into population-wise local mortality systems containing a population-specific autoregressive component and a cross-sectional component of global mortality factors, the proposed GVAR model could substantially reduce the extra estimation cost of including additional populations. Consequentially, the new GVAR model provides an efficient tool to analyze the joint mortality dynamics of a large group of populations, which has been a rather challenging task for existing VAR-type mortality models. Moreover, under fairly general assumptions, the proposed GVAR model could produce coherent projected mortality rates for any two ages and populations. Using single-age mortality data of 15 low-mortality countries as considered in Li and Lee (2005), we find that the global factors have substantial impacts on mortality improvements of individual populations. Furthermore, we show that the proposed GVAR model could produce satisfying mortality forecasts under various settings.

Despite the claims that China has been implementing mercantilism policies to support its exports, recent studies argue that the dominant factor affecting China’s foreign exchange reserves is the precautionary motives. Apart from the economic uncertainty, political instability may be another risk source to consider by those motives. This study intends to empirically investigate how a particular factor originated in the US such as partisan conflict or economic uncertainty explains the foreign exchange reserves hoarding in China. Using SVAR models we find strong evidence suggesting that partisan conflict is positively associated with China’s reserve level, and has a primarily dominant role in predicting the surge in China’s foreign exchange reserves. In contrast, the impact of the US economic uncertainty is found negative and less persistent. The Forecast-error-variance decomposition (FEVD) analysis confirms partisan conflict’s dominant role in explaining the variations in China’s reserve accumulation, in comparison with other variables such as US exports to China and US economic policy uncertainty. The results are robust and consistent under different scenarios and robustness checks. The findings have important policy implications and contribute to the existing literature on the global impacts of US political instability.

The vector autoregressive (VAR) model has been popularly employed in operational practice to study multivariate time series. Despite its usefulness in providing associated metrics such as the impulse response function (IRF) and forecast error variance decomposition (FEVD), the traditional VAR model estimated via the usual ordinary least squares is vulnerable to outliers. To handle potential outliers in multivariate time series, this paper investigates two robust estimation methods of the VAR model, the reweighted multivariate least trimmed squares and the multivariate MM-estimation. The robust information criteria are also proposed to select the appropriate number of temporal lags. Via extensive simulation studies, we show that the robust VAR models lead to much more accurate estimates than the original VAR in the presence of outliers. Our empirical results include logged daily realized volatilities of six common safe haven assets: futures of gold, silver, Brent oil and West Texas Intermediate (WTI) oil and currencies of Swiss Francs and Japanese Yen. Our sample covers July 2017–June 2020, which includes the history-writing price drop of WTI on April 20, 2020. Our baseline results suggest that the traditional VAR model may significantly overestimate some parameters, as well as IRF and FEVD metrics. In contrast, robust VAR models provide more reliable results, the validity of which is verified via various approaches. Empirical implications based on robust estimates are further illustrated.

U(VI) was found to form intrinsic colloids in the natural groundwater. Fluorescence spectra, TEM, SEM techniques and thermodynamic calculation were used to identify the species of U(VI) and characterize the intrinsic uranium colloidal particles in the groundwater. The effect of pH on the formation of U(VI) colloidal particles was studied. It was found that under near-neutral and strong alkaline conditions, several elements, mainly Ca, Na and Si were responsible for the formation of intrinsic U(VI) colloid, and under weak alkaline conditions, U(VI) colloidal particles were not formed due to the formation of calcium uranyl carbonate complex.

Age coherence describes the property that forecast mortality rates across ages will not diverge in the long run. Although intuitively and biologically reasonable, this property is lost when the sem...

Customer requirements (CRs) have a significant impact on product design. The existing methods of defining CRs, such as customer surveys and expert evaluations, are time-consuming, inaccurate and subjective. This paper proposes an automatic CRs definition method based on online customer product reviews using the big data analysis. Word vectors are defined using a continuous bag of words (CBOW) model. Online customer reviews are searched by a crawling method and filtered by the parts of speech and frequency of words. Filtered words are then clustered into groups by an affinity propagation (AP) clustering method based on trained word vectors. Exemplars in each clustering group are finally used to define CRs. The proposed method is verified by case studies of defining CRs for product design. Results show that the proposed method has better performance to determine CRs compared to existing CRs definition methods.

Accurate forecasts of mortality rates are essential to various types of demographic research like population projection, and to the pricing of insurance products such as pensions and annuities. Recent studies have considered a spatial–temporal vector autoregressive (STVAR) model for the mortality surface, where mortality rates of each age depend on the historical values for that age (temporality) and the neighboring cohorts ages (spatiality). This model has sound statistical properties including co-integrated dependent variables, the existence of closed-form solutions and a simple error structure. Despite its improved forecasting performance over the famous Lee–Carter (LC) model, the constraint that only the effects of the same and neighboring cohorts are significant can be too restrictive. In this study, we adopt the concept of hyperbolic memory to the spatial dimension and propose a hyperbolic STVAR (HSTVAR) model. Retaining all desirable features of the STVAR, our model uniformly beats the LC, the weighted functional demographic model, STVAR and sparse VAR counterparties for forecasting accuracy, when French and Spanish mortality data over 1950–2016 are considered. Simulation results also lead to robust conclusions. Long-term forecasting analyses up to 2050 comparing the four models are further performed. To illustrate the extensible feature of HSTVAR to a multi-population case, a two-population illustrative example using the same sample is further presented.

This paper investigates a high-dimensional vector-autoregressive (VAR) model in mortality modeling and forecasting. We propose an extension of the sparse VAR (SVAR) model fitted on the log-mortality improvements, which we name “spatially penalized smoothed VAR” (SSVAR). By adaptively penalizing the coefficients based on the distances between ages, SSVAR not only allows a flexible data-driven sparsity structure of the coefficient matrix but simultaneously ensures interpretable coefficients including cohort effects. Moreover, by incorporating the smoothness penalties, divergence in forecast mortality rates of neighboring ages is largely reduced, compared with the existing SVAR model. A novel estimation approach that uses the accelerated proximal gradient algorithm is proposed to solve SSVAR efficiently. Similarly, we propose estimating the precision matrix of the residuals using a spatially penalized graphical Lasso to further study the dependency structure of the residuals. Using the UK and France population data, we demonstrate that the SSVAR model consistently outperforms the famous Lee–Carter, Hyndman–Ullah, and two VAR-type models in forecasting accuracy. Finally, we discuss the extension of the SSVAR model to multi-population mortality forecasting with an illustrative example that demonstrates its superiority in forecasting over existing approaches.

Accurate forecasts of mortality rates are essential to various types of demographic research such as population projection, and the pricing of insurance products such as pensions and annuities. Recent studies have considered a spatial temporal autoregressive (STAR) model for the mortality surface, where mortality rates for each age depend (temporally) on their historical values as well as (spatiality) on those of neighboring age cohorts. This model has sound statistical properties including cointegrated dependent variables and the existence of closed‐form solutions. Despite its improved forecasting performance over the famous Lee–Carter (LC) model, the constraint that only the effects of the same and neighboring cohorts are significant can be too restrictive. In this study, we adopt a data‐driven adaptive weighted structure and propose the adaptive STAR (ASTAR) model. Retaining all the desirable features of the STAR, our model uniformly outperforms the LC and STAR counterparts in terms of forecasting accuracy, when mortality data for ages 0–100 from the UK, France, Italy, Spain, and Japan over the period 1950–2016 are considered. Two sensitivity tests and additional simulation results also lead to robust conclusions. The proposed ASTAR model could therefore be a widely useful tool for modeling and forecasting mortality rates in other contexts, and may be extended to multipopulation modeling.

It is well known that the US political parties hold opposing views on many economic policies and foreign policies. Since China has become the largest trading partner of the US, the operation of the US–China bilateral relationship has developed into an essential and sensitive political issue that has been widely discussed in the recent US presidential election campaigns. Therefore, the understanding of the effects of partisan conflicts on the US–China trade is crucial to improving our knowledge of the international transmission of political uncertainty. Using a recent US partisan conflict index provided by Azzimonti (2014), we investigate those effects employing a structural VAR model and find that a one standard deviation shock to US partisan conflict is associated with a 2% increase in the US exports to China and a 2% reduction in its imports from China. Such effects are still significant after one year. The results are robust across various combinations of variables, models and sample periods. Further, our findings indicate that partisan conflicts in the US may reduce its trade deficit with China.

Using the comprehensive RavenPack Dow Jones News Analytics (DJNA) database that captures firm-specific news releases and their sentiment scores at high frequencies, we examine the contemporaneous correlation as well as the lead-lag relation between news and return volatility of major commercial banks listed on the Chinese stock market. Contrary to the Sequential Information Arrival Hypothesis (SIAH), most of the Chinese bank stocks do not exhibit significant lead-lag relations between news and volatility. However, there is substantial evidence that news is strongly correlated with return volatility in all the stocks, consistent with the Mixture of Distributions Hypothesis (MDH). Further analysis based on news sentiment scores suggests that positive news arrivals influence return volatility more strongly, compared with negative news. In addition, there is some evidence indicating that news arrivals contribute to the persistence in return volatility.

The fourth outbreak of the Coronaviruses, known as the COVID-19, has occurred in Wuhan city of Hubei province in China in December 2019. We propose a time-varying sparse vector autoregressive (VAR) model to retrospectively analyze and visualize the dynamic transmission routes of this outbreak in mainland China over January 31 - February 19, 2020. Our results demonstrate that the influential inter-location routes from Hubei have become unidentifiable since February 4, 2020, whereas the self-transmission in each provincial-level administrative region (location, hereafter) was accelerating over February 4-15, 2020. From February 16, 2020, all routes became less detectable, and no influential transmissions could be identified on February 18 and 19, 2020. Such evidence supports the effectiveness of government interventions, including the travel restrictions in Hubei. Implications of our results suggest that in addition to the origin of the outbreak, virus preventions are of crucial importance in locations with the largest migrant workers percentages (e.g., Jiangxi, Henan and Anhui) to controlling the spread of COVID-19.

Existing literature argues that the mortality rate of a specific age is affected not only by its own lags but by the lags of neighbouring ages, known as cohort effects. Although these effects are a...

Quality function deployment (QFD) is a method commonly used in manufacturing industries to identify customer requirements (CRs) and the manufacturing capacity to meet CRs. Importance rate (IR) of CRs plays an important role in the QFD method. The existing methods in determining IR of CRs cannot consider all related factors of customer satisfaction, need importance, personal information, and relationship between customer satisfaction and function implementation. This paper proposes a method to improve the IR accuracy. Based on comments of customers for a product, importance rates of CRs are defined using integrated importance-performance analysis (IPA) and Kano models. IPA and Kano models are integrated by spectral clustering where a similarity matrix W is formed to balance the influence proportion between the IPA and Kano models considering comments of different customers for the product. IR of CRs is used in the QFD method to define functions and structures of the product. The proposed method is compared with several existing methods in case studies of designing an upper limb rehabilitation device and a feed drive system of the CNC machine. Results show that the proposed method has improved accuracy of IR of CRs for the product design and manufacturing.