Add UBI to benefits totals

[jdebacker]

This PR updates the computation of benefits totals to include UBI program benefits.

[codecov]

Codecov Report

Merging #2418 into master will increase coverage by 0.00%.
The diff coverage is 100.00%.

@@           Coverage Diff           @@
##           master    #2418   +/-   ##
=======================================
  Coverage   99.92%   99.92%           
=======================================
  Files          13       13           
  Lines        2554     2560    +6     
=======================================
+ Hits         2552     2558    +6     
  Misses          2        2           

Impacted Files	Coverage Δ
taxcalc/calculator.py	100.00% <100.00%> (ø)
taxcalc/utils.py	100.00% <100.00%> (ø)

---

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 cbaee03...4ce36e4. Read the comment docs.

[jdebacker]

This is ready for review @Peter-Metz @MattHJensen.

Note, I'm assuming we don't need a consumption benefits scale parameter for UBI since it's a cash transfer.

[Peter-Metz]

@jdebacker, thanks a lot for the PR. It looks like benefit_value_total and benefit_cost_total aren't getting updated with the new implementation because the BenefitPrograms() function is using an array of zeros for UBI, as opposed to the calculated amounts. The easiest solution I could come up with is to add a few lines to _calc_one_year() which is called during calc_all():

benefit_value_total_ubi = self.array('ubi') + self.array('benefit_value_total')
self.array('benefit_value_total', benefit_value_total_ubi)
benefit_cost_total_ubi = self.array('ubi') + self.array('benefit_cost_total')
self.array('benefit_cost_total', benefit_cost_total_ubi)

This solution feels a little hardcoded, but we wouldn't have to modify calcfunctions.py.

Also, to avoid double counting when calculating expanded income, we need to remove this line.

Finally, I was looking at the three Tax-Calc tables -- difference, distribution, and diagnostic -- and it looks like the first two have line items for UBI and benefits totals. We might want to consider adding those variables to the diagnostic table as well.

[jdebacker]

@Peter-Metz Thanks for those comments. Changing this to a WIP while I resolve those issues you raise.

[jdebacker]

@Peter-Metz Can you review and see that my latest commits address the issues you raise about computing UBI for benefits totals and adding to the diagnostic tables. Thanks!

[jdebacker]

Also, @Peter-Metz, can you share code you used to find that benefit_value_total and benefit_cost_total aren't getting updated? I could add that as a test.

[Peter-Metz]

Hey @jdebacker, these latest changes look good to me at first glance, but I'll take a closer look tomorrow. Here's how I tested if benefits_cost_total is getting updated:

from taxcalc import *

ubi_ref = {'UBI_21': {2020: 1000}}    

pol = Policy()                                                                             
recs = Records()                                                                           
calc_base = Calculator(pol, recs) 
calc_base.advance_to_year(2020)                                                                
calc_base.calc_all()
                                                          
pol_ubi = Policy()                                                                            
pol_ubi.implement_reform(ubi_ref)
calc_ubi = Calculator(pol_ubi, recs)                                                                                                                      
calc_ubi.advance_to_year(2020)                                                               
calc_ubi.calc_all()

(calc_ubi.weighted_total('ubi') - calc_base.weighted_total('ubi'))/1e9
### 241.02410248

(calc_ubi.weighted_total('benefit_cost_total') - calc_base.weighted_total('benefit_cost_total'))/1e9  
### 0.0     

[jdebacker]

Thanks @Peter-Metz. Just added a test that confirms UBI is included in the benefits totals coming out of Calculator.calc_all() as you laid it out.

[Peter-Metz]

@jdebacker all looks good to me, thanks again for the PR. I'll leave this open for @MattHJensen to review and merge.

[MattHJensen]

Thanks a lot for the PR, @jdebacker and review @Peter-Metz. This looks good to me, too. Merging.