Mortality models incorporating long memory for life table estimation: a comprehensive analysis

2021 ◽  
pp. 1-38
Author(s):  
Hongxuan Yan ◽  
Gareth W. Peters ◽  
Jennifer Chan
Keyword(s):  
Time Series ◽  
Life Expectancy ◽  
Long Memory ◽  
Age Groups ◽  
Deviance Information Criterion ◽  
Information Criterion ◽  
Life Tables ◽  
Mortality Data ◽  
Model Parameters ◽  
Mortality Models

Abstract Mortality projection and forecasting of life expectancy are two important aspects of the study of demography and life insurance modelling. We demonstrate in this work the existence of long memory in mortality data. Furthermore, models incorporating long memory structure provide a new approach to enhance mortality forecasts in terms of accuracy and reliability, which can improve the understanding of mortality. Novel mortality models are developed by extending the Lee–Carter (LC) model for death counts to incorporate a long memory time series structure. To link our extensions to existing actuarial work, we detail the relationship between the classical models of death counts developed under a Generalised Linear Model (GLM) formulation and the extensions we propose that are developed under an extension to the GLM framework known in time series literature as the Generalised Linear Autoregressive Moving Average (GLARMA) regression models. Bayesian inference is applied to estimate the model parameters. The Deviance Information Criterion (DIC) is evaluated to select between different LC model extensions of our proposed models in terms of both in-sample fits and out-of-sample forecasts performance. Furthermore, we compare our new models against existing models structures proposed in the literature when applied to the analysis of death count data sets from 16 countries divided according to genders and age groups. Estimates of mortality rates are applied to calculate life expectancies when constructing life tables. By comparing different life expectancy estimates, results show the LC model without the long memory component may provide underestimates of life expectancy, while the long memory model structure extensions reduce this effect. In summary, it is valuable to investigate how the long memory feature in mortality influences life expectancies in the construction of life tables.

scholarly journals The decomposition of life expectancy for age and cause of death in Tuscany, Italy

2019 ◽  
Vol 29 (Supplement_4) ◽  
Author(s):  
N Nante ◽  
L Kundisova ◽  
F Gori ◽  
A Martini ◽  
F Battisti ◽  
...  
Keyword(s):  
Life Expectancy ◽  
Causes Of Death ◽  
National Level ◽  
Age Groups ◽  
Resistant Bacteria ◽  
Mortality Data ◽  
Life Expectancy At Birth ◽  
Icd 10 ◽  
Reduction Of Mortality ◽  
The Impact

Abstract Introduction Changing of life expectancy at birth (LE) over time reflects variations of mortality rates of a certain population. Italy is amongst the countries with the highest LE, Tuscany ranks fifth at the national level. The aim of the present work was to evaluate the impact of various causes of death in different age groups on the change in LE in the Tuscany region (Italy) during period 1987-2015. Material and methods Mortality data relative to residents that died during the period between 1987/1989 and 2013/2015 were provided by the Tuscan Regional Mortality Registry. The causes of death taken into consideration were cardiovascular (CVS), respiratory (RESP) and infective (INF) diseases and cancer (TUM). The decomposition of LE gain was realized with software Epidat, using the Pollard’s method. Results The overall LE gain during the period between two three-years periods was 6.7 years for males, with a major gain between 65-89, and 4.5 years for females, mainly improved between 75-89, <1 year for both sexes. The major gain (2.6 years) was attributable to the reduction of mortality for CVS, followed by TUM (1.76 in males and 0.83 in females) and RESP (0.4 in males; 0.1 in females). The major loss of years of LE was attributable to INF (-0.15 in females; -0.07 in males) and lung cancer in females (-0.13), for which the opposite result was observed for males (gain of 0.62 years of LE). Conclusions During the study period (1987-2015) the gain in LE was major for males. To the reduction of mortality for CVS have contributed to the tempestuous treatment of acute CVS events and secondary CVS prevention. For TUM the result is attributable to the adherence of population to oncologic screening programmes. The excess of mortality for INF that lead to the loss of LE can be attributed to the passage from ICD-9 to ICD-10 in 2003 (higher sensibility of ICD-10) and to the diffusion of multi-drug resistant bacteria, which lead to elevated mortality in these years. Key messages The gain in LE during the period the 1987-2015 was higher in males. The major contribution to gain in LE was due to a reduction of mortality for CVS diseases.

Influence of some ARFIMA model parameters on the accuracy of financial time series forecasting

2021 ◽  
Vol 62 ◽  
pp. 85-100
Author(s):  
Robert Garafutdinov ◽  
Keyword(s):  
Time Series ◽  
Long Memory ◽  
Financial Time Series ◽  
Time Series Forecasting ◽  
Model Parameters ◽  
Memory Length ◽  
Financial Time ◽  
True Value ◽  
Arfima Model ◽  
Practical Recommendations

The influence of ARFIMA model parameters on the accuracy of financial time series forecasting on the example of artificially generated long memory series and daily log returns of RTS index is investigated. The investigated parameters are deviation of the integration order value from its «true» value, as well as the memory «length» considered by the model. Based on the research results, some practical recommendations for modeling using ARFIMA have been formulated.

scholarly journals Research on RMB Exchange Rate Volatility Risk Based on MSGARCH-VaR Model

2020 ◽  
Vol 2020 ◽  
pp. 1-10
Author(s):  
Xiaofei Wu ◽  
Shuzhen Zhu ◽  
Junjie Zhou
Keyword(s):  
Exchange Rate ◽  
Value At Risk ◽  
Financial Risk ◽  
Deviance Information Criterion ◽  
Exchange Rate Volatility ◽  
Information Criterion ◽  
Model Parameters ◽  
Volatility Risk ◽  
Rmb Exchange Rate ◽  
Trade Friction

This paper captures the RMB exchange rate volatility using the Markov-switching GARCH (MSGARCH) models and traditional single-regime GARCH models. Through the Markov Chain Monte Carlo (MCMC) method, the model parameters are estimated to study the volatility dynamics of the RMB exchange rate. Furthermore, we compare the MSGARCH models to the single-regime GARCH specifications in terms of Value-at-Risk (VaR) prediction accuracy. According to the Deviance information criterion method, the research shows that MSGARCH models outperform the single-regime specifications in capturing the complexity of RMB exchange rate volatility. After the RMB exchange rate reform in 2015, the volatility is more asymmetric and persistent, and the probability of being in the turbulent volatility regime is significantly increased. The continuous escalation of Sino-US trade friction has increased the VaR of RMB exchange rate log-returns. From the evaluation results of the actual over expected exceedance ratio (AE), the conditional coverage (CC) test, and the dynamic quantile (DQ) test, we find strong evidence that two-regime MSGARCH models could forecast VaR more accurately, which provides practical value for China’s foreign exchange management authorities to manage the financial risk.

scholarly journals Identifying Structural Breaks in Stochastic Mortality Models

2015 ◽  
Vol 1 (2) ◽  
Author(s):  
Colin O’Hare ◽  
Youwei Li
Keyword(s):  
Life Expectancy ◽  
Structural Breaks ◽  
Structural Break ◽  
Mortality Rates ◽  
Insurance Companies ◽  
Mortality Data ◽  
Stochastic Mortality ◽  
Stochastic Mortality Models ◽  
Mortality Models ◽  
Mortality Forecasts

In recent years, the issue of life expectancy has become of utmost importance to pension providers, insurance companies, and government bodies in the developed world. Significant and consistent improvements in mortality rates and hence life expectancy have led to unprecedented increases in the cost of providing for older ages. This has resulted in an explosion of stochastic mortality models forecasting trends in mortality data to anticipate future life expectancy and hence quantify the costs of providing for future aging populations. Many stochastic models of mortality rates identify linear trends in mortality rates by time, age, and cohort and forecast these trends into the future by using standard statistical methods. These approaches rely on the assumption that structural breaks in the trend do not exist or do not have a significant impact on the mortality forecasts. Recent literature has started to question this assumption. In this paper, we carry out a comprehensive investigation of the presence or of structural breaks in a selection of leading mortality models. We find that structural breaks are present in the majority of cases. In particular, we find that allowing for structural break, where present, improves the forecast result significantly.

Time Series Decomposition into Oscillation Components and Phase Estimation

2017 ◽  
Vol 29 (2) ◽  
pp. 332-367 ◽  
Author(s):  
Takeru Matsuda ◽  
Fumiyasu Komaki
Keyword(s):  
Time Series ◽  
State Space ◽  
Empirical Bayes ◽  
Real Data ◽  
Information Criterion ◽  
State Space Models ◽  
Model Parameters ◽  
Seasonal Adjustment ◽  
Phase Dynamics ◽  
Time Series Decomposition

Many time series are naturally considered as a superposition of several oscillation components. For example, electroencephalogram (EEG) time series include oscillation components such as alpha, beta, and gamma. We propose a method for decomposing time series into such oscillation components using state-space models. Based on the concept of random frequency modulation, gaussian linear state-space models for oscillation components are developed. In this model, the frequency of an oscillator fluctuates by noise. Time series decomposition is accomplished by this model like the Bayesian seasonal adjustment method. Since the model parameters are estimated from data by the empirical Bayes’ method, the amplitudes and the frequencies of oscillation components are determined in a data-driven manner. Also, the appropriate number of oscillation components is determined with the Akaike information criterion (AIC). In this way, the proposed method provides a natural decomposition of the given time series into oscillation components. In neuroscience, the phase of neural time series plays an important role in neural information processing. The proposed method can be used to estimate the phase of each oscillation component and has several advantages over a conventional method based on the Hilbert transform. Thus, the proposed method enables an investigation of the phase dynamics of time series. Numerical results show that the proposed method succeeds in extracting intermittent oscillations like ripples and detecting the phase reset phenomena. We apply the proposed method to real data from various fields such as astronomy, ecology, tidology, and neuroscience.

scholarly journals Cohort Working Life Tables for Older Canadians

2010 ◽  
Vol 37 (1-2) ◽  
pp. 175 ◽  
Author(s):  
Frank T. Denton ◽  
Christine H. Feaver ◽  
Byron G. Spencer
Keyword(s):  
Time Series ◽  
Life Expectancy ◽  
Labour Force ◽  
Working Life ◽  
Life Tables ◽  
Labour Force Survey ◽  
Participation Rates ◽  
Men And Women ◽  
Male Cohort ◽  
Statistics Canada

We construct cohort working life tables for Canadian men and women aged 50 and older and, for comparison, corresponding period tables. The tables are derived using annual single-age time series of participation rates for 1976-2006 from the master files of the Statistics Canada Labour Force Survey. The cohort calculations are based on stochastic projections of mortality coupled with alternative assumptions about future participation rates. Separate tables are provided for the years 1976, 1991, and 2006, thus spanning a period of substantial gains in life expectancy and strong upward trends in female participation. Life expectancies based on the cohort tables are greater than those based on the period tables, for both men and women, and that is reflected in increased retirement expectancies. For example, a male aged 50 in 1976 could have expected to live three years longer and to have almost four more years in retirement, based on the male cohort table under medium assumptions, as compared with the corresponding period table.

scholarly journals Effects of covid-19 pandemic on life expectancy and premature mortality in 2020: time series analysis in 37 countries

BMJ ◽  
2021 ◽  
pp. e066768
Author(s):  
Nazrul Islam ◽  
Dmitri A Jdanov ◽  
Vladimir M Shkolnikov ◽  
Kamlesh Khunti ◽  
Ichiro Kawachi ◽  
...  
Keyword(s):  
Time Series ◽  
New Zealand ◽  
Life Expectancy ◽  
Confidence Interval ◽  
Time Series Analysis ◽  
South Korea ◽  
Mortality Data ◽  
Years Of Life Lost ◽  
The Us ◽  
The Difference

Abstract Objective To estimate the changes in life expectancy and years of life lost in 2020 associated with the covid-19 pandemic. Design Time series analysis. Setting 37 upper-middle and high income countries or regions with reliable and complete mortality data. Participants Annual all cause mortality data from the Human Mortality Database for 2005-20, harmonised and disaggregated by age and sex. Main outcome measures Reduction in life expectancy was estimated as the difference between observed and expected life expectancy in 2020 using the Lee-Carter model. Excess years of life lost were estimated as the difference between the observed and expected years of life lost in 2020 using the World Health Organization standard life table. Results Reduction in life expectancy in men and women was observed in all the countries studied except New Zealand, Taiwan, and Norway, where there was a gain in life expectancy in 2020. No evidence was found of a change in life expectancy in Denmark, Iceland, and South Korea. The highest reduction in life expectancy was observed in Russia (men: −2.33, 95% confidence interval −2.50 to −2.17; women: −2.14, −2.25 to −2.03), the United States (men: −2.27, −2.39 to −2.15; women: −1.61, −1.70 to −1.51), Bulgaria (men: −1.96, −2.11 to −1.81; women: −1.37, −1.74 to −1.01), Lithuania (men: −1.83, −2.07 to −1.59; women: −1.21, −1.36 to −1.05), Chile (men: −1.64, −1.97 to −1.32; women: −0.88, −1.28 to −0.50), and Spain (men: −1.35, −1.53 to −1.18; women: −1.13, −1.37 to −0.90). Years of life lost in 2020 were higher than expected in all countries except Taiwan, New Zealand, Norway, Iceland, Denmark, and South Korea. In the remaining 31 countries, more than 222 million years of life were lost in 2020, which is 28.1 million (95% confidence interval 26.8m to 29.5m) years of life lost more than expected (17.3 million (16.8m to 17.8m) in men and 10.8 million (10.4m to 11.3m) in women). The highest excess years of life lost per 100 000 population were observed in Bulgaria (men: 7260, 95% confidence interval 6820 to 7710; women: 3730, 2740 to 4730), Russia (men: 7020, 6550 to 7480; women: 4760, 4530 to 4990), Lithuania (men: 5430, 4750 to 6070; women: 2640, 2310 to 2980), the US (men: 4350, 4170 to 4530; women: 2430, 2320 to 2550), Poland (men: 3830, 3540 to 4120; women: 1830, 1630 to 2040), and Hungary (men: 2770, 2490 to 3040; women: 1920, 1590 to 2240). The excess years of life lost were relatively low in people younger than 65 years, except in Russia, Bulgaria, Lithuania, and the US where the excess years of life lost was >2000 per 100 000. Conclusion More than 28 million excess years of life were lost in 2020 in 31 countries, with a higher rate in men than women. Excess years of life lost associated with the covid-19 pandemic in 2020 were more than five times higher than those associated with the seasonal influenza epidemic in 2015.

scholarly journals Gender gap in life expectancy in India and role of age groups: A comparison between before and after male – female life expectancy at birth crossover

PLoS ONE ◽  
2021 ◽  
Vol 16 (12) ◽  
pp. e0260657
Author(s):  
Girimallika Borah
Keyword(s):  
Life Expectancy ◽  
Gender Gap ◽  
Age Groups ◽  
Absolute Number ◽  
Reproductive Age ◽  
Sample Survey ◽  
Mortality Data ◽  
Female Life Expectancy ◽  
Life Expectancy At Birth ◽  
Before And After

To assess the gender gap in life expectancy at birth in India and its major states as well as the timing of male-female life expectancy at birth crossover. To analyze the age-specific contributions to the changing gender differences before and after the crossover at the national and sub-national levels. We have used sample-survey-based age-specific mortality data available for the periods 1970–2018 to construct abridged life tables. The contribution of different age groups to the gender gap is estimated by using Arriaga’s method of decomposition. During 1981–85 female life expectancy at birth caught up with male life expectancy at birth for India and by 2005 all major states completed the crossover. The male-female crossover in life expectancy at the national level in the early 80s is remarkable in the face of continued female disadvantage from birth till adolescence, even for some richer states. We provide evidence that gender difference in longevity in favour of females is largely a function of adult age groups and younger age groups contribute negatively to the gender gap in life expectancy at birth in most states. Juxtaposing the results from contribution in an absolute number of years and their relative contribution change before and after the crossover, it is established that although the adult and old age groups contribute the highest in the absolute number of years before and after the crossover, the contribution of the reproductive age groups and childhood years in the recent time is most relevant in relative terms.

scholarly journals Allowing for shocks in portfolio mortality models

2022 ◽  
Vol 27 ◽  
Author(s):  
Stephen J. Richards
Keyword(s):  
Information Criterion ◽  
Exposure Period ◽  
Mortality Data ◽  
Time Dimension ◽  
Mortality Trends ◽  
Mortality Models ◽  
Local Flexibility ◽  
The Uk ◽  
Actuarial Model

Abstract The COVID-19 pandemic creates a challenge for actuaries analysing experience data that include mortality shocks. Without sufficient local flexibility in the time dimension, any analysis based on the most recent data will be biased by the temporarily higher mortality. Also, depending on where the shocks sit in the exposure period, any attempt to identify mortality trends will be distorted. We present a methodology for analysing portfolio mortality data that offer local flexibility in the time dimension. The approach permits the identification of seasonal variation, mortality shocks and occurred-but-not reported deaths (OBNR). The methodology also allows actuaries to measure portfolio-specific mortality improvements. Finally, the method assists actuaries in determining a representative mortality level for long-term applications like reserving and pricing, even in the presence of mortality shocks. Results are given for a mature annuity portfolio in the UK, which suggest that the Bayesian information criterion is better for actuarial model selection in this application than Akaike’s information criterion.

scholarly journals Serotype patterns of pneumococcal disease in adults are correlated with carriage patterns in older children

2019 ◽  
Author(s):  
Anne L. Wyllie ◽  
Joshua L. Warren ◽  
Gili Regev-Yochay ◽  
Noga Givon-Lavi ◽  
Ron Dagan ◽  
...  
Keyword(s):  
Invasive Pneumococcal Disease ◽  
Pneumococcal Disease ◽  
Age Groups ◽  
Deviance Information Criterion ◽  
Information Criterion ◽  
Older Children ◽  
Disease Patterns ◽  
Pneumococcal Carriage ◽  
Age Related ◽  
Carriage Prevalence

ABSTRACTBackgroundThe importance of specific serotypes causing invasive pneumococcal disease (IPD) differs by age. Data on pneumococcal carriage in different age groups, along with data on serotype-specific invasiveness, could help to explain these age-related patterns and their implications for vaccination.MethodsUsing pneumococcal carriage and disease data from Israel, we evaluated the association between serotype-specific IPD in adults and serotype-specific carriage prevalence among children in different age categories, while adjusting for serotype-specific invasiveness. We used a sliding window approach to estimate carriage prevalence using different age groupings. Deviance Information Criterion was used to determine which age groupings of carriage data best fit the adult IPD data. Serotype-specific disease patterns were further evaluated by stratifying IPD data by comorbidity status.ResultsThe relative frequency of serotypes causing IPD differed between adults and children, and also differed between older and younger adults and between adults with and without comorbidities. Serotypes over-represented as causes of IPD in adults were more commonly carried in older children as compared to younger children. In line with this, the serotype-specific frequency of carriage in older children (aged 36-59 months), rather than infants, best correlated with serotype-specific IPD in adults.ConclusionsThese analyses suggest that older children, rather than infants, are the main drivers of disease patterns in adults. These insights could help in optimizing vaccination strategies to reduce disease burden across all ages.40-word summary of the article’s main pointSerotype-specific rates of invasive pneumococcal disease in adults are better correlated with serotype-specific carriage patterns in older children (36-59 months of age) than those in infants.

Sign in / Sign up

Export Citation Format

Share Document