Expanded income and aftertax income should be options in distribution tables

[MattHJensen]

Users may want to see average expanded income by expanded income bin or decile in TaxBrain diagnostic tables. Same with aftertax income.

[martinholmer]

I've looked into Tax-Calculator issue #1374, which says:

Users may want to see average expanded income by expanded income bin or decile in TaxBrain diagnostic tables. Same with aftertax income.

The construction in dropq code of a TaxBrain "diagnostic" table (which is called in Tax-Calculator a "distribution" table) is fairly involved. The data structures that are created by the dropq code have been carefully designed to meet the needs of TaxBrain developers, which is exactly as it should be.

My conclusion is that the TaxBrain developers need to start out this enhancement process by specifying exactly how they want the thirteen objects returned from the dropq.run_nth_year_tax_calc_model function to be changed in order to implement the #1374 enhancement. Then, with this change specification in hand, Tax-Calculator developers can revise taxcalc code to supply revised versions of those thirteen objects.

Does this approach make sense?

@MattHJensen @feenberg @Amy-Xu @GoFroggyRun @PeterDSteinberg @brittainhard

[feenberg]

On Wed, 28 Jun 2017, Martin Holmer wrote: I've looked into Tax-Calculator issue #1374, which says: Users may want to see average expanded income by expanded income bin or decile in TaxBrain diagnostic tables. Same with aftertax income. The construction in dropq code of a TaxBrain "diagnostic" table (which is called in Tax-Calculator a "distribution" table) is fairly involved. The data structures that are created by the dropq code have been carefully designed to meet the needs of TaxBrain developers, which is exactly as it should be.

The important thing is that we drop 3 meaningful records from each row returned to the user. So if we tab by expanded income we drop 3 records from each expanded income class, not each AGI class. Otherwise I think the logic should be exactly the same. dan

…

My conclusion is that the TaxBrain developers need to start out this enhancement process by specifying exactly how they want the thirteen objects returned from the dropq.run_nth_year_tax_calc_model function to be changed in order to implement the #1374 enhancement. Then, with this change specification in hand, Tax-Calculator developers can revise taxcalc code to supply revised versions of those thirteen objects. Does this approach make sense? @MattHJensen @feenberg @Amy-Xu @GoFroggyRun @PeterDSteinberg @brittainhard — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or mute the thread.[AHvQVSGO4UXyZrqRvPOXC9ycSuhKSx_Hks5sIqYFgaJpZM4NhBUp.gif]

[martinholmer]

Dan @feenberg said with respect to issue #1374:

The important thing is that we drop 3 meaningful records from each row returned to the user. So if we tab[ulate] by expanded income we drop 3 records from each expanded income class, not each AGI class.

As far as I can tell, TaxBrain never shows distributional tables with AGI bins; it shows only distributional tables with expanded-income bins. And the dropq code does remove three records from each expanded-income bin. So, I think your concerns are being met.

@MattHJensen @PeterDSteinberg @brittainhard

[martinholmer]

I've moved a question about the Brown-Khanna GAIN Act that was submitted in this issue to its own issue #1581.

[martinholmer]

@MattHJensen said in issue #1374:

Users may want to see average expanded income by ... bin or decile in TaxBrain diagnostic tables. Same with aftertax income.

The current TaxBrain version (1.0.3) allows users to build a "Distribution Table" or a "Difference Table".

So, is your request to add to the Distribution Table two new columns: average expanded income and average after-tax expanded income?

Does it make sense to add these two new columns to the Difference Table as well?
I'm asking this because the 10/17 New York Times graphs are difference tables with decile bins as seen in this example:

[MattHJensen]

@martinholmer asked two questions:

So, is your request to add to the Distribution Table two new columns: average expanded income and average after-tax expanded income?

Does it make sense to add these two new columns to the Difference Table as well?

Yes to both.

[martinholmer]

Is the labeling of the income groups in this New York Times (10/17/2017) graph slightly incorrect?

Shouldn't the left-hand column read like this?

99 ("top 1%")
95-98
90-94
80-89
70-79
60-69
50-59
40-49
30-39
20-29
10-19
0-9

As it appears in the newspaper, there are 101 percentiles (not 100).

@MattHJensen @feenberg @Amy-Xu @andersonfrailey @hdoupe @GoFroggyRun

[MattHJensen]

I don't think it is incorrect. It looks like they are doing (x-y] and therefore starting with 1 instead of 0.

[martinholmer]

@MattHJensen said:

I don't think it is incorrect per se. It looks like they are doing (x-y] and therefore starting with 1 instead of 0.

I guess numbers in the left-hand column supposed to be quantiles that are the boundaries of the percentiles. Does that make sense?

[MattHJensen]

I guess numbers in the left-hand column supposed to be quantiles that are the boundaries of the percentiles. Does that make sense?

Yep, that's what I'm thinking. The left-hand column is a an exclusive boundary, and the right-hand column is an inclusive boundary.

[MattHJensen]

It lines up better with the way people talk about the top 1% and bottom 99%, I suppose.

[martinholmer]

Issue #1374 has been resolved by merge of #1602.