nikhilwoodruff / frs Goto Github PK

Make generate_csv.py a bit more referenced and readable.

Per https://github.com/PSLmodels/PSL-Infrastructure/blob/master/Criteria/library_criteria.md

So it can be installable. The core function could point to a local directory with the input files and either export the data or return a list of DataFrames.

Moving discussion from PolicyEngine/policyengine-uk#40 which covers implementation of the logic.

Neither FRS nor SPI microdata includes capital gains, but we could impute them from HMRC statistical tables. Here are some high level stats from 2018-19:

• 256,000 individuals paid 8.8bn tax on 59.3bn capital gains (table 1)
• More than half of gains are from people aged 55 or older (table 6)
• Table 3 has the most important information (far left column is lower limit, left of each pair of columns is individuals in 000s, right is gains in millions):
  

The Office of Tax Simplification has some additional data. For example, the below chart shows that people bunch capital gains below the 12,300 threshold; while the report includes data for some charts, it omits it from this one.