Open AccessDetermining the Pension Benefit Obligation of a Defined Benefit Plan: Applying a Multivariate ARIMA Stochastic Model
Author(s) -
J. Tim Query,
Evaristo Diz
Publication year - 2022
Publication title -
ira-international journal of management and social sciences
Language(s) - English
Resource type - Journals
ISSN - 2455-2267
DOI - 10.21013/jmss.v17.n4.p5
Subject(s) - autoregressive integrated moving average , multivariate statistics , econometrics , time horizon , actuarial science , autoregressive model , pension , sample (material) , statistics , time series , computer science , economics , mathematics , finance , chemistry , chromatography
In this study we examine the robustness of fit for a multivariate and an autoregressive integrated moving average model to a data sample time series type. The sample is a recurrent actuarial data set for a 10-year horizon. We utilize this methodology to contrast with stochastic models to make projections beyond the data horizon. Our key results suggest that both types of models are useful for making predictions of actuarial liability levels given by PBO Projected Benefit Obligations on and off the horizon of the sample time series. As we have seen in prior research, the use of multivariate models for control and auditing purposes is widely recommended. Fast and reliable statistical estimates are desirable in all cases, whether for audit purposes or to verify and validate miscellaneous actuarial results.