• Notícies

Economics Researchers propose new statistics methods to calculate more accurately the mortality rates

Lledó, Pavía, Morillas

Josep Lledó, José Manuel Pavía and Francisco G. Morillas, researchers of the Department of Applied Economics of the Universitat de València have suggested a new method to calculate the Spanish mortality rates (decease probability for each age) more accurately than with the current system of the Spanish Statistical Office (INE). The tool foresees the variables of the migrants as well as the birth and decease dates. This is important due to the effect that will have in the viability of the public pension systems, among other applications.

In the article published in the journal Scandinavian Actuarial Journal some experts have validated this proposal by comparing it with the statistical systems that are currently used. The researchers of the Faculty of Economics note that it is necessary for the demographic system to be considered open, especially in those areas with migration tendencies. This fact implies computing the time that migrants spend inside and outside the region under study, since there are births or deceases that do not always occur in the analysed territory.

Besides, Lledó, Pavía and Morillas assure that the deceases and births occur throughout the year in a non-uniform way. For this reason it is necessary to incorporate the exact moments of these events. 

‘This new conceptualization that we incorporate to the statistical estimators has not always been accessible due to the difficulty of having sufficiently precise and detailed information. Due to the explosion of what is known as massive data or BigData there is the possibility to incorporate information which is necessary to calculate the survival and mortality rates’, explains Francisco Morillas. 

For the development of these statistical measures the experts have first used theoretical conceptual methodologies, based on the Lexis scheme. This conceptual framework is a way of representing three demographic variables: the moment of birth, the moment of migration, the date of decease and therefore the lifetime of each person of a particular population. Thus, experts have represented the dynamics of the Spanish population. From this theoretical framework, they have constructed statistical estimators using algebraic and statistical properties.

Once the estimators have been constructed, Lledó, Pavia and Morillas validate them by comparing the INE and their own indicators. Then they discard some of the hypotheses of the Spanish Statistical Office. Some of these hypothesis are usually used by the institution, according to the experts. 

To analyse the incidence of mortality or survival in a region or in a period it is mostly used a tool known as the mortality table. In this way mortality rates of each moment as well as prediction rates for the future can be estimated. ‘Not considering characteristics such as migrant status or birth and decease dates implies considering that the phenomena mentioned either do not exist or occur uniformly throughout the study period’, explains Francisco Morillas.

To carry out the validation of the hypotheses, a multitude of statistical tests were performed: spatial, functional, parametric point patterns and test to test the closed demographic system hypothesis. If we consider the variants that have been applied, the number of these rises to 19.008, 14.388, 6.322 and 3.488, respectively. ‘These contrasts provide objective information that helps to determine with exactitude the validity of the hypotheses tested" emphasises the researcher.

Research Group

The researchers Josep Lledó, José Manuel Pavía y Francisco G. Morillas are members of the Risks in Insurances and Finances Research Grou of the Department of Applied Economics of the Universitat de València which develops projects in the decease scope since 2002. It has mainly focused on the construction of demographic indicators related to mortality, birth or migration. These usually use implicitly certain assumptions which may provide inaccurate estimates. The impact of these estimates may affect insurance and the system public pension. The results are published in international journals.

Article:
Josep Lledó, Jose M. Pavía i Francisco G. Morillas (2017) «Assessing implicit hypotheses in life table construction», Scandinavian Actuarial Journal, 2017:6, 495-518, DOI:10.1080/03461238.2016.1177585     
 

Last update: 11 de august de 2017 06:30.

News release
 
This website uses proprietary and third-party cookies for technical purposes, traffic analysis and to facilitate insertion of content in social networks on user request. If you continue to browse, we consider that you are accepting its use. For more information please consult ourcookies policy