Unemployment and Life Expectancy

Modelling the relationship between unemployment and life expectancy at MSOA level.

Data

• Life expectancy – All England 2011 MSOAs for years 2016 to 2020 broken down by male and female. Source: ONS
• Claimant count – Averaged from 2016 to 2020 for all England 2011 MSOAs, broken down by male and female. Source: NOMIS
• Mid-year population estimate - Averaged from 2016 to 2020 for all England 2011 MSOAs, broken down by male and female.
• 2019 IMD domains: Income, Education, Crime, Housing, Living Environment. Averaged for each MSOA.

Model

We will fit a linear regression model with the following formula:

LE = c₀ + c₁Claimant_% + c₂(Is Male) + c₃IMD_Income + c₄IMD_Education + c₅IMD_Crime + c₆IMD_Housing + c₇IMD_LE

Where:

• LE – Life expectancy (as above)
• Claimant_% - Percentage of working age adult population claiming JSA or UC equivalent. This is equal to the average claimant count divided by the working age adult mid-year estimate for each 2011 MSOA.
• (Is Male) – Binary variable equal to 1 for males and 0 for females. (i.e. female is the reference category)
• IMD_domain – IMD score for each of the 2019 IMD domains given above (from 0 to 1)
• c_i – Coefficients for each variable.

The modelling will estimate the coefficients c_i from the data. These can be used to interpret the association between each variable and life expectancy. For example, once we have calculated c₁ we can say that “For each 1-point increase in claimant count percentage there is an associated -c₁ year decrease in life expectancy”.

This modelling will be performed using MSOA 2011 data above for all of England with a sub-analysis restricted to Birmingham MSOAs only.

License

This repository is dual licensed under the Open Government v3 & MIT. All code can outputs are subject to Crown Copyright.