adds pareto tails for lognormal and kde

[john-p-ryan]

This PR adds several methods for calculating the income distribution with a Pareto tail. These methods include a simultaneous estimate of the cutoff and pareto parameter via maximum likelihood, as well as 2 options for sequentially estimating the upper tail, with the body being estimated by lognormal or kernel density estimation.

[codecov-commenter]

Codecov Report

Attention: Patch coverage is 12.50000% with 35 lines in your changes missing coverage. Please review.

Project coverage is 56.42%. Comparing base (10985c1) to head (634d1af).
Report is 10 commits behind head on main.

Files	Patch %	Lines
iot/inverse_optimal_tax.py	12.50%	35 Missing ⚠️

Additional details and impacted files

@@             Coverage Diff             @@
##             main      #26       +/-   ##
===========================================
- Coverage   72.81%   56.42%   -16.39%     
===========================================
  Files           3        3               
  Lines         103      140       +37     
===========================================
+ Hits           75       79        +4     
- Misses         28       61       +33     

Flag	Coverage Δ
unittests	56.42% <12.50%> (-16.39%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[john-p-ryan]

Added the notebook used to plot the income dists. The issue with very noisy g_z is resolved when the notebook is run in order. However, the discontinuity at the pareto cutoff is more pronounced. Work may be needed to smooth the derivative estimate at the cutoff.

[jdebacker]

@john-p-ryan Please sync this branch to the master, which will help pass the CI tests on GH Actions. Thanks!