Update policy parameter with 2019 IRS final values

[Peter-Metz]

This PR updates policy_current_law.json and associated test files to reflect final IRS values for 2019. Values for indexable parameters 2020 and beyond were calculated with ppp.py. Note that Tax-Data has not yet been updated to reflect the 2020 CBO inflation assumptions, so ppp.py will have to be run again once Tax-Data (and growfactors.csv) is updated.

Relevant IRS forms:

f1040.pdf
f1040sa.pdf
i1040gi.pdf
i1040s8.pdf
i1040sca.pdf
f6251.pdf
f8863.pdf
f8995a.pdf
f461.pdf
iw2w3.pdf

cc: @MattHJensen

[codecov]

Codecov Report

Merging #2399 into master will not change coverage by %.
The diff coverage is n/a.

@@            Coverage Diff            @@
##            master     #2399   +/-   ##
=========================================
  Coverage   100.00%   100.00%           
=========================================
  Files           13        13           
  Lines         2751      2751           
=========================================
  Hits          2751      2751           

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update f6f2cd9...80d763c. Read the comment docs.

[MattHJensen]

@Peter-Metz, thanks a lot for opening this PR. I will review this week.

[Peter-Metz]

@MattHJensen and I had a question regarding changes to the reform output files (e.g. 2017_law.out.csv) so we thought we'd document the history of thought here:

We were wondering why some reform output test files include changes to most/all output variables (e.g. Larson2019.out.csv) while others only include changes to payrolltax (e.g. Renacci.out.csv)?

The reason for this is that Larson2019.json uses current law as the baseline while Renacci.json uses 2017_law.json as its baseline. Since the modifications in this PR relate to policy_current_law.json, they do not affect most of the parameters in 2017_law.json. The exception to this is SS_earnings_c, which does not appear in 2017_law.json. For this reason, reforms using 2017 law as their baseline will still see a change in payrolltax whenever SS_earnings_c is changed.

[MattHJensen]

The exception to this is SS_earnings_c, which does not appear in 2017_law.json. For this reason, reforms using 2017 law as their baseline will still see a change in payrolltax whenever SS_earnings_c is changed.

@Peter-Metz and my plan is to review whether SS_earnings_c or any other parameters ought to be included in 2017_law.json in a PR that follows this one. So the reform tests results may change again in that PR.

[MattHJensen]

@Peter-Metz, thanks again for this PR. Merging.