Fix spurious negative auto_loan_interest from SCF -1 sentinels#1171
Open
MaxGhenis wants to merge 1 commit into
Open
Fix spurious negative auto_loan_interest from SCF -1 sentinels#1171MaxGhenis wants to merge 1 commit into
MaxGhenis wants to merge 1 commit into
Conversation
add_auto_loan_interest cleaned the SCF 'not applicable' sentinels with auto_df[AUTO_LOAN_COLUMNS].replace(-1, 0, inplace=True), which runs on a column-slice copy and is a silent no-op. The -1 sentinels therefore survived into the per-car balance*rate product: a real loan balance times a -1 rate sentinel (after the /10,000 scaling) yields a small negative interest, so a $90k loan with an unreported rate became -$9 of interest. Extract the balance/interest computation into _summarize_auto_loans and floor the raw balance and rate columns at zero (clip(lower=0)) before combining — which also assigns back, unlike the old in-place call, and is robust to any negative sentinel code. Positive magnitudes are unchanged (rate /10,000 kept; median interest/balance stays ~0.04). Adds a unit test. Surfaced by the PolicyBench populace refresh: 844 of 75,112 populace_us_2024 households carried impossible negative auto_loan_interest traced to this. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
4 tasks
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Problem
add_auto_loan_interestcleans the SCF "not applicable" sentinels with:That runs on a column-slice copy, so it is a silent no-op — the
-1sentinels survive into the per-carbalance * rateproduct. A real loan balance times a-1rate sentinel (after the/10,000scaling) yields a small negative interest: e.g. a $90k loan with an unreported rate becomes -$9 of interest.This is invisible inside policyengine-us-data but propagates:
populace-build's SCF wealth stage imputesSCF_2022().load_dataset()onto households, so 844 of 75,112populace_us_2024households carry impossible negativeauto_loan_interest. Surfaced by the PolicyBench populace refresh, whose prompts then showed models facts likeauto loan interest: -$4.50.Fix
Extract the balance/interest computation into
_summarize_auto_loansand floor the raw balance and rate columns at zero withclip(lower=0)before combining. This assigns back (unlike the old in-place call) and is robust to any negative sentinel code, not just-1.Positive magnitudes are unchanged — the
/10,000rate scaling is preserved and the real median interest/balance stays ~0.04 (a sane ~4% APR). Adds a focused unit test covering both the sentinel-balance and unreported-rate cases.🤖 Generated with Claude Code