Seeking advice: Best-latest puf.csv to use for analysis?

[donboyd5]

I started down the road of running taxdata this week to create puf.csv using the main/master branch because I want to use the latest/best puf.csv in analysis. Using taxdata to create puf.csv has been successful so far and I appreciate all you have done, @andersonfrailey, to make it work.

However, I realize now that because I am not deeply familiar with git and GitHub, and don't know enough about versioning, I am not 100% sure if this is the right path. I'm seeking a little advice. I'm hoping someone might answer these questions:

• Am I correct in thinking that the last official puf.csv release and (associated growfactors.csv, puf_weights.csv, and puf_ratios.csv) was in August 2020?
• If not, is there a later one that is considered official?
• Is there a branch name or commit number that is associated with the current official puf.csv (and associated files)?
• Could you please tell me what puf.csv is being used in TaxBrain / ComputeStudio (tagging @hdoupe in case this is a question for you)?
• Are people who are analyzing current federal tax policy proposals using this latest puf.csv, or something else? Do they run the main/master branch of taxdata to create a latest-best puf.csv? Or do they use a specific branch or commit for that? It seems like there has been a lot of great development work since Aug 2020 and I want to make sure I take advantage of whatever is the latest/best that people are using.

Many thanks.

[andersonfrailey]

Am I correct in thinking that the last official puf.csv release and (associated growfactors.csv, puf_weights.csv, and puf_ratios.csv) was in August 2020?
If not, is there a later one that is considered official?

We've started a release system for taxdata so you can use the latest release for the official version. You can find that under the 0.2.0 tag, but also the current master branch works just fine because we haven't made any changes to the logic to make the file.

Could you please tell me what puf.csv is being used in TaxBrain / ComputeStudio

It's the one created from version 0.2.0

Are people who are analyzing current federal tax policy proposals using this latest puf.csv, or something else? Do they run the main/master branch of taxdata to create a latest-best puf.csv? Or do they use a specific branch or commit for that? It seems like there has been a lot of great development work since Aug 2020 and I want to make sure I take advantage of whatever is the latest/best that people are using.

That would depend on the version of Tax-Calculator they're using. To my knowledge, they're using the latest version of PUF.csv, but I can't say with certainty

[donboyd5]

@andersonfrailey, thanks!

[donboyd5]

@MattHJensen, @jdebacker, and @rickecon, would you mind telling me what version of puf.csv you and (to the extent you know) others analyzing currently relevant federal tax proposals are using in the analysis, and what version of Tax-Calculator?

cc @andersonfrailey

• Are you using the latest official puf.csv (created from version 0.20) or possibly something later, and if so what would be different about it since @andersonfrailey notes that the logic has not changed - would it have updated SOI and/or CBO estimates (via a different growfactors.csv file), for example, or updated underlying CPS data?

• Are you using the latest official release of Tax-Calculator, or something else?

I am trying to gear up to do some state-specific analysis (@MattHJensen would be great to talk sometime in next week or two if interested) and I want to use the best/latest taxdata and Tax-Calculator.

Many thanks.

Don

[MattHJensen]

@donboyd5, I've only used Tax-Cruncher recently (no TaxData involved) with the tip of Tax-Calculator's master branch as a dependency. To facilitate the tax-cruncher x tax-calculator interaction, I make a local package of Tax-Calculator from the tip of master with make package.

It seems to me that the state project should use the forthcoming TaxData release that Anderson is working on now with the tip of Tax-Calculator master branch.

I'll send you an email to set up a time to talk soon. It sounds like you've got the new machine up and running!