Remove use of "taxable returns" statistics from reweighting calculations

[martinholmer]

Fixes #147 and fixes #106.

These #148 changes eliminate the need for the changes made in PR #143, so those CTC-related #143 changes have been removed from the code.

In keeping with the process recommended in issue #146, here are the differences between the 2023 tax expenditure estimates on this branch (<) versus on the master branch (>):

% diff  tax_expenditures_branch  tax_expenditures_master
1,8c1,8
< YR,KIND,EST= 2023 iitax 2215.0
< YR,KIND,EST= 2023 ctc 132.2
< YR,KIND,EST= 2023 eitc 78.6
< YR,KIND,EST= 2023 social_security_partial_taxability 52.2
< YR,KIND,EST= 2023 niit -52.9
< YR,KIND,EST= 2023 cgqd_tax_preference 221.7
< YR,KIND,EST= 2023 qbid 54.1
< YR,KIND,EST= 2023 salt 20.4
---
> YR,KIND,EST= 2023 iitax 2247.0
> YR,KIND,EST= 2023 ctc 108.9
> YR,KIND,EST= 2023 eitc 82.1
> YR,KIND,EST= 2023 social_security_partial_taxability 51.9
> YR,KIND,EST= 2023 niit -53.9
> YR,KIND,EST= 2023 cgqd_tax_preference 223.4
> YR,KIND,EST= 2023 qbid 55.9
> YR,KIND,EST= 2023 salt 20.7

Now both the CTC and EITC tax expenditures are above those reported by the federal agencies. But this is to be expected because our estimates assume 100% participation in these refundable credits while the federal agencies assume IRS-provided credit-specific participation rates. Here is what a recent JCT tax expenditure report says about this:

A recent IRS table shows a national EITC participation rate of 76.3 percent for 2020 (and upwards toward 80% in years before 2020):

[donboyd5]

Thank you so much for doing this. I read through the code and see that it was a bigger job than I realized. It looks right to me. I won't be able to run anything until the morning, but trust your judgment if you are comfortable merging.