A copula-based model for multivariate ordinal panel data: application to well-being composition
arXiv:1604.05643
Abstract
A novel copula-based multivariate panel ordinal model is developed to estimate structural relations among components of well-being. Each ordinal time-series is modelled using a copula-based Markov model to relate the marginal distributions of the response at each time of observation and then, at each observation time, the conditional distributions of each ordinal time-series are joined using a multivariate t copula. Maximum simulated likelihood based on evaluating the multidimensional integrals of the likelihood with randomized quasi Monte Carlo methods is used for the estimation. Asymptotic calculations show that our method is nearly as efficient as maximum likelihood for fully specified multivariate copula models. Our findings highlight the importance of one's relative position in evaluating their well-being with no direct effects of socio-economic characteristics on well-being but strong indirect effects through their impact on components of well-being. Temporal resilience, habit formation and behavioural traits can explain the dependence in the joint tails over time and across well-being components.
References in corpus (4)
- On the estimation of normal copula discrete regression models using the continuous extension and simulated likelihood
- Efficient estimation of high-dimensional multivariate normal copula models with discrete spatial responses
- Correlation structure and variable selection in generalized estimating equations via composite likelihood information criteria
- Weighted scores method for longitudinal ordinal data