Idea and help file tweak to get_cpiu

[shingtgen]

Is your feature request related to a problem? Please describe.

1. I found the help file confusing regarding the "base year" (specifically the 2nd paragraph in the description). I think of the base year as being the starting year, but the description seems to refer to the base year as being the year you're adjusting to. So if I were inflating from 2018 to 2020, I would call 2018 the "base year" and 2020 the "adjusted year" (though there might be other terms for that). I think it's also unclear in the description how the starting year is determined.

2. I'd like to incorporate get_cpiu into the CPS tax model (using merge), but there are two issues:

• The tax model uses CBO inflation projections, not BLS actuals.
• The tax model doesn't have a variable named "year."

Describe the solution you'd like

1. I think the Fed Fiscal RAs would be a great group to review/weigh in on the help file wording; I'd be happy to work with them on that if it would be helpful.

2. To be able to use get_cpiu in the tax models:

• If you'd like to incorporate CBO projections, we use CBO's 10-Year Economic Projections that are published each January (see https://www.cbo.gov/about/products/budget-economic-data#4). I have some instructions on this saved in the Tax Doc if helpful (having trouble pasting a succinct link but it's under Tax Models > CPS Tax Model > Usage Instructions > Updating the model).

• Ideally it would be great if the program allowed me to specify the year variable (in CPS, would be h_year).

Describe alternatives you've considered
Will continue to incorporate cpi-u data in the tax model as we have done until a new approach is available.

Additional context
n/a

Thanks for your great work on this!

[c-zippel]

This Khan Academy post gives a good overview of what "base year" means in the inflation-adjustment context. Would it be helpful to link to this in the help file?

Also, could you say more about what you mean by "I think it's also unclear in the description how the starting year is determined"?

The second part of this has been implemented in v 0.1.4. New argument yearvarname() lets you specify the variable in your dataset that contains the year.

On the first part: Great idea, and I think this enhancement would be a longer-term project because it'd be somewhat tricky to implement. The data in the CBO spreadsheet is not rectangular, so it'd take some work to get Stata to read in the right part; and get_cpiu would need to "splice" the projected CPI-U onto the CPI-U-RS like the inflation spreadsheet does.