Transition Probabilities and Duration Analysis among Disability States: Some Evidence from Spanish Data

• Authors: Guillem López-Casasnovas.
• BSE Working Paper: 110130 | September 15
• Keywords: Spain , Markov transition , disability states , cost of disability , survival analysis
• JEL codes: J11, I18, H55
• Spain
• Markov transition
• disability states
• cost of disability
• survival analysis

In this paper we study the disability transition probabilities (as well as the mortality probabilities) due to concurrent factors to age such as income, gender and education. Although it is well known that aging and socioeconomic status influence the probability of causing functional disorders, surprisingly little attention has been paid to the combined effect of those factors along the individuals’ life and how this affects the transition from one degree of disability to another. The assumption that tomorrow’s disability state is only a function of the today’s state is very strong, since disability is a complex variable that depends on several other elements than time. This paper contributes into the field in two ways: (1) by attending the distinction between the initial disability level and the process that leads to his course (2) by addressing whether and how education, age and income differentially affect the disability transitions. Using a Markov chain discrete model and a survival analysis, we estimate the probability by year and individual characteristics that changes the state of disability and the duration that it takes its progression in each case. We find that people with an initial state of disability have a higher propensity to change and take less time to transit from different stages. Men do that more frequently than women. Education and income have negative effects on transition. Moreover, we consider the disability benefits associated to those changes along different stages of disability and therefore we o er some clues on the potential savings of preventive actions that may delay or avoid those transitions. On pure cost considerations, preventive programs for improvement show higher benefits than those for preventing deterioration, and in general terms, those focusing individuals below 65 should go first. Finally the trend of disability in Spain seems not to change among years and regional differences are not found.