# THE SCORING-ATTRIBUTION REGISTRY — one row per winner-pattern constant.
#
# WHAT THIS FILE IS FOR
#
# tests/test_scoring_attribution.py sweeps the winner-pattern corpus --- every
# module-level dict in nmtcapp/data/historical_awards.py, plus SECTION B of
# nmtcapp/data/benchmark_thresholds.py --- and requires a row here for every
# top-level key. A key with no row FAILS. That is the whole mechanism: it makes
# SILENCE impossible, and silence is how nine of these survived nine rounds.
#
# WHY THE EXISTING GATES DID NOT COVER THIS CORPUS
#
#   test_pinned_constants  sweeps nmtcapp/data/{schema,benchmark_thresholds}.py
#                          ONLY. historical_awards.py has never been in
#                          DATA_MODULES, so its four winner dicts were in no
#                          sweep at all --- not pinned, not waived, not seen.
#   the same file's waivers  rule Section B's three dicts "reaches no rendered
#                          surface". That is a RENDERING answer to an
#                          ATTRIBUTION question. Whether a number is printed
#                          and whether anyone knows where it came from are
#                          different facts, and 1.4.0 shipped with the second
#                          one unasked because the first had been answered.
#   test_fund_attribution_source  keys on an AUTHORITY TOKEN. `"std_states":
#                          3.8` names nobody, so it is invisible there.
#
# ROW FORMAT
#
#   KIND | source | CONSTANT.key
#
#   CITED    a retrievable document, with enough detail to re-retrieve it.
#   HOUSE    this package's own construct. NOT a licence to keep a number: it
#            is a declaration that no source exists, and every HOUSE row is a
#            standing candidate for deletion. The rendered surface must say so
#            on its face --- see benchmarks._METHODOLOGY.
#
# THE STANDING RULING (1.5.0 S3): SOURCE IT OR DELETE IT. A constant that
# cannot be attributed to a retrievable document does not get to stay behind a
# disclaimer forever. Six keys with no consumer at all were DELETED this round
# rather than given a row --- std_pct_deep_or_severe, mean_pct_hmr, std_states,
# urban_pct_mean, p50_cost_per_job and min_native_area_pct.
#
# WITH SIX EXCEPTIONS, WHICH THIS HEADER USED TO DENY (1.5.0 F6). The sentence
# here read "What remains has a consumer" --- and five AWARD_SIZE_TIERS rows
# and APPLICATION_VOLUME_TRENDS.trend_note say in their own text, twenty lines
# below, that they do not. The header contradicted its own rows, which is worse
# than either answer alone: a reader who trusts the header never reads the
# rows, and the rows are the true ones.
#
#   AWARD_SIZE_TIERS (five keys) reaches only get_award_size_percentiles(),
#   which nothing in nmtcapp/, streamlit_app/ or docs/ calls --- re-verified by
#   grep over the whole tree in 1.5.0. Its only callers are two tests. A
#   constant whose entire consumer set is the tests asserting it exists is the
#   definition of a deletion candidate, and it stays this round ONLY because
#   removing it deletes public API (the function as well as the dict) from a
#   release whose scope is correcting false statements, not removing surface.
#
#   trend_note DOES have a consumer and the row saying otherwise was the
#   mistake: streamlit_app/pages/4_About_and_Methodology.py:314 renders it to a
#   CDE's screen. Its forecast was corrected in 1.5.0 F5.
#
# So: what remains has a consumer OR is named here as an exception. The rows
# are what make the candidate list explicit instead of buried, and deleting
# them is the next release, behind a written methodology.
#
# A CITED ROW IS RECORDED, NOT VERIFIED, exactly as the attribution allowlist
# concedes for its own entries. What this gate enforces is that a number cannot
# enter a score without SOMEBODY having had to write down where it came from.

CITED | CITED: CDFI Fund NMTC award announcement for that round, cdfifund.gov/programs-training/programs/new-markets-tax-credit | NMTC_AWARD_ROUNDS.CY2020
CITED | CITED: CDFI Fund NMTC award announcement for that round, cdfifund.gov/programs-training/programs/new-markets-tax-credit | NMTC_AWARD_ROUNDS.CY2021
CITED | CITED: CDFI Fund NMTC award announcement for that round, cdfifund.gov/programs-training/programs/new-markets-tax-credit | NMTC_AWARD_ROUNDS.CY2022
CITED | CITED: CDFI Fund NMTC award announcement for that round, cdfifund.gov/programs-training/programs/new-markets-tax-credit | NMTC_AWARD_ROUNDS.CY2023
CITED | CITED (1.5.0 F5 -- WAS HOUSE, and the HOUSE reason had expired eight months before 1.5.0 shipped): CY 2024-2025 NMTC Program Award Book, 8 pp., cdfifund.gov/media/8018306/download?inline, linked from the 23 Dec 2025 announcement at cdfifund.gov/news/700, retrieved first-hand 2026-08-21. Page 4 states '142 CDEs out of 216 applicants were awarded allocations by the CDFI Fund' and 'A total of $10 billion was awarded out of $19.2 billion requested'. avg_award and median_award are computed from the Award Book's own allocatee table at pp. 6-8, whose 142 amounts sum to exactly $10,000,000,000 -- the published total, so the extraction checks itself. The row previously read 'Award data pending' with $5B and three '# estimated' fields, while _round_provenance in the same commit asserted the round was awarded 23 Dec 2025 with $10B. | NMTC_AWARD_ROUNDS.CY2024-2025
HOUSE | HOUSE: the dict's own header says 'Inferred from CDFI Fund award announcements'. A CONSTRUCTED PARTITION, NOT A MEASUREMENT (1.5.0 F1): the five pct_of_awards values are 0.10 / 0.20 / 0.35 / 0.25 / 0.10 --- every one on a 0.05 grid, over a population of real awards, summing to EXACTLY 1.000. Real award data does not land on a 0.05 grid and does not sum to a round 1.000 by accident; a partition somebody wrote down does both. tests/data/test_historical_awards.py::test_award_size_tiers_pct_sum_to_one asserts the sum, so the construction is ENSHRINED --- its docstring now says that passing means the values were built to sum to one, not that any share was measured. NO CONSUMER outside its own query function get_award_size_percentiles(), which nothing in nmtcapp/, streamlit_app/ or docs/ calls --- deletion candidate, see this file's header. | AWARD_SIZE_TIERS.under_25MM
HOUSE | HOUSE: the dict's own header says 'Inferred from CDFI Fund award announcements'. A CONSTRUCTED PARTITION, NOT A MEASUREMENT (1.5.0 F1): the five pct_of_awards values are 0.10 / 0.20 / 0.35 / 0.25 / 0.10 --- every one on a 0.05 grid, over a population of real awards, summing to EXACTLY 1.000. Real award data does not land on a 0.05 grid and does not sum to a round 1.000 by accident; a partition somebody wrote down does both. tests/data/test_historical_awards.py::test_award_size_tiers_pct_sum_to_one asserts the sum, so the construction is ENSHRINED --- its docstring now says that passing means the values were built to sum to one, not that any share was measured. NO CONSUMER outside its own query function get_award_size_percentiles(), which nothing in nmtcapp/, streamlit_app/ or docs/ calls --- deletion candidate, see this file's header. | AWARD_SIZE_TIERS.25_to_35MM
HOUSE | HOUSE: the dict's own header says 'Inferred from CDFI Fund award announcements'. A CONSTRUCTED PARTITION, NOT A MEASUREMENT (1.5.0 F1): the five pct_of_awards values are 0.10 / 0.20 / 0.35 / 0.25 / 0.10 --- every one on a 0.05 grid, over a population of real awards, summing to EXACTLY 1.000. Real award data does not land on a 0.05 grid and does not sum to a round 1.000 by accident; a partition somebody wrote down does both. tests/data/test_historical_awards.py::test_award_size_tiers_pct_sum_to_one asserts the sum, so the construction is ENSHRINED --- its docstring now says that passing means the values were built to sum to one, not that any share was measured. NO CONSUMER outside its own query function get_award_size_percentiles(), which nothing in nmtcapp/, streamlit_app/ or docs/ calls --- deletion candidate, see this file's header. | AWARD_SIZE_TIERS.35_to_50MM
HOUSE | HOUSE: the dict's own header says 'Inferred from CDFI Fund award announcements'. A CONSTRUCTED PARTITION, NOT A MEASUREMENT (1.5.0 F1): the five pct_of_awards values are 0.10 / 0.20 / 0.35 / 0.25 / 0.10 --- every one on a 0.05 grid, over a population of real awards, summing to EXACTLY 1.000. Real award data does not land on a 0.05 grid and does not sum to a round 1.000 by accident; a partition somebody wrote down does both. tests/data/test_historical_awards.py::test_award_size_tiers_pct_sum_to_one asserts the sum, so the construction is ENSHRINED --- its docstring now says that passing means the values were built to sum to one, not that any share was measured. NO CONSUMER outside its own query function get_award_size_percentiles(), which nothing in nmtcapp/, streamlit_app/ or docs/ calls --- deletion candidate, see this file's header. | AWARD_SIZE_TIERS.50_to_65MM
HOUSE | HOUSE: the dict's own header says 'Inferred from CDFI Fund award announcements'. A CONSTRUCTED PARTITION, NOT A MEASUREMENT (1.5.0 F1): the five pct_of_awards values are 0.10 / 0.20 / 0.35 / 0.25 / 0.10 --- every one on a 0.05 grid, over a population of real awards, summing to EXACTLY 1.000. Real award data does not land on a 0.05 grid and does not sum to a round 1.000 by accident; a partition somebody wrote down does both. tests/data/test_historical_awards.py::test_award_size_tiers_pct_sum_to_one asserts the sum, so the construction is ENSHRINED --- its docstring now says that passing means the values were built to sum to one, not that any share was measured. NO CONSUMER outside its own query function get_award_size_percentiles(), which nothing in nmtcapp/, streamlit_app/ or docs/ calls --- deletion candidate, see this file's header. | AWARD_SIZE_TIERS.over_65MM
HOUSE | HOUSE: unsourced. historical_awards.py's own header states that the 'Source: CDFI Fund Annual Reports' comment over this dict cites a publication that DOES NOT EXIST (the real series is the NMTC Public Data Release Summary Report, cumulative FY2003-FY2023) and that 'Every value under them is unsourced'. | WINNER_DISTRESS_PATTERNS.mean_pct_deep_or_severe
HOUSE | HOUSE: unsourced. historical_awards.py's own header states that the 'Source: CDFI Fund Annual Reports' comment over this dict cites a publication that DOES NOT EXIST (the real series is the NMTC Public Data Release Summary Report, cumulative FY2003-FY2023) and that 'Every value under them is unsourced'. | WINNER_DISTRESS_PATTERNS.p25_pct_deep_or_severe
HOUSE | HOUSE: unsourced. historical_awards.py's own header states that the 'Source: CDFI Fund Annual Reports' comment over this dict cites a publication that DOES NOT EXIST (the real series is the NMTC Public Data Release Summary Report, cumulative FY2003-FY2023) and that 'Every value under them is unsourced'. | WINNER_DISTRESS_PATTERNS.p50_pct_deep_or_severe
HOUSE | HOUSE: unsourced. historical_awards.py's own header states that the 'Source: CDFI Fund Annual Reports' comment over this dict cites a publication that DOES NOT EXIST (the real series is the NMTC Public Data Release Summary Report, cumulative FY2003-FY2023) and that 'Every value under them is unsourced'. | WINNER_DISTRESS_PATTERNS.p75_pct_deep_or_severe
HOUSE | HOUSE: unsourced. historical_awards.py's own header states that the 'Source: CDFI Fund Annual Reports' comment over this dict cites a publication that DOES NOT EXIST (the real series is the NMTC Public Data Release Summary Report, cumulative FY2003-FY2023) and that 'Every value under them is unsourced'. | WINNER_DISTRESS_PATTERNS.p90_pct_deep_or_severe
HOUSE | HOUSE: unsourced. historical_awards.py's own header states that the 'Source: CDFI Fund Annual Reports' comment over this dict cites a publication that DOES NOT EXIST (the real series is the NMTC Public Data Release Summary Report, cumulative FY2003-FY2023) and that 'Every value under them is unsourced'. | WINNER_DISTRESS_PATTERNS.min_pct_deep_or_severe
HOUSE | HOUSE: unsourced. historical_awards.py's own header states that the 'Source: CDFI Fund Annual Reports' comment over this dict cites a publication that DOES NOT EXIST (the real series is the NMTC Public Data Release Summary Report, cumulative FY2003-FY2023) and that 'Every value under them is unsourced'. | WINNER_DISTRESS_PATTERNS.mean_pct_native_area
HOUSE | HOUSE: unsourced. historical_awards.py's own header states that the 'Source: CDFI Fund Annual Reports' comment over this dict cites a publication that DOES NOT EXIST (the real series is the NMTC Public Data Release Summary Report, cumulative FY2003-FY2023) and that 'Every value under them is unsourced'. | WINNER_DISTRESS_PATTERNS.mean_pct_eligible
HOUSE | HOUSE: unsourced. historical_awards.py's own header states that the 'Source: CDFI Fund Annual Reports' comment over this dict cites a publication that DOES NOT EXIST (the real series is the NMTC Public Data Release Summary Report, cumulative FY2003-FY2023) and that 'Every value under them is unsourced'. | WINNER_GEOGRAPHIC_PATTERNS.mean_states
HOUSE | HOUSE: unsourced. historical_awards.py's own header states that the 'Source: CDFI Fund Annual Reports' comment over this dict cites a publication that DOES NOT EXIST (the real series is the NMTC Public Data Release Summary Report, cumulative FY2003-FY2023) and that 'Every value under them is unsourced'. | WINNER_GEOGRAPHIC_PATTERNS.p25_states
HOUSE | HOUSE: unsourced. historical_awards.py's own header states that the 'Source: CDFI Fund Annual Reports' comment over this dict cites a publication that DOES NOT EXIST (the real series is the NMTC Public Data Release Summary Report, cumulative FY2003-FY2023) and that 'Every value under them is unsourced'. | WINNER_GEOGRAPHIC_PATTERNS.p50_states
HOUSE | HOUSE: unsourced. historical_awards.py's own header states that the 'Source: CDFI Fund Annual Reports' comment over this dict cites a publication that DOES NOT EXIST (the real series is the NMTC Public Data Release Summary Report, cumulative FY2003-FY2023) and that 'Every value under them is unsourced'. | WINNER_GEOGRAPHIC_PATTERNS.p75_states
HOUSE | HOUSE: unsourced. historical_awards.py's own header states that the 'Source: CDFI Fund Annual Reports' comment over this dict cites a publication that DOES NOT EXIST (the real series is the NMTC Public Data Release Summary Report, cumulative FY2003-FY2023) and that 'Every value under them is unsourced'. | WINNER_GEOGRAPHIC_PATTERNS.min_states
HOUSE | HOUSE: unsourced. historical_awards.py's own header states that the 'Source: CDFI Fund Annual Reports' comment over this dict cites a publication that DOES NOT EXIST (the real series is the NMTC Public Data Release Summary Report, cumulative FY2003-FY2023) and that 'Every value under them is unsourced'. | WINNER_GEOGRAPHIC_PATTERNS.mean_projects
HOUSE | HOUSE: unsourced. historical_awards.py's own header states that the 'Source: CDFI Fund Annual Reports' comment over this dict cites a publication that DOES NOT EXIST (the real series is the NMTC Public Data Release Summary Report, cumulative FY2003-FY2023) and that 'Every value under them is unsourced'. | WINNER_GEOGRAPHIC_PATTERNS.p50_projects
HOUSE | HOUSE: unsourced. historical_awards.py's own header states that the 'Source: CDFI Fund Annual Reports' comment over this dict cites a publication that DOES NOT EXIST (the real series is the NMTC Public Data Release Summary Report, cumulative FY2003-FY2023) and that 'Every value under them is unsourced'. | WINNER_GEOGRAPHIC_PATTERNS.mean_hhi
HOUSE | HOUSE: unsourced. historical_awards.py's own header states that the 'Source: CDFI Fund Annual Reports' comment over this dict cites a publication that DOES NOT EXIST (the real series is the NMTC Public Data Release Summary Report, cumulative FY2003-FY2023) and that 'Every value under them is unsourced'. | WINNER_GEOGRAPHIC_PATTERNS.rural_pct_mean
HOUSE | HOUSE: unsourced. historical_awards.py's own header states that the 'Source: CDFI Fund Annual Reports' comment over this dict cites a publication that DOES NOT EXIST (the real series is the NMTC Public Data Release Summary Report, cumulative FY2003-FY2023) and that 'Every value under them is unsourced'. A CONSTRUCTED PARTITION, NOT A MEASUREMENT (1.5.0 F1): the eight sector shares are 0.22 / 0.18 / 0.17 / 0.14 / 0.12 / 0.08 / 0.05 / 0.04 --- all on a 0.01 grid, summing to EXACTLY 1.000, for a population this package holds no sector breakdown of. test_sector_shares_sum_to_one asserts the sum and test_healthcare_is_largest_sector asserts the ordering, so both the partition and its shape are ENSHRINED; their docstrings now say that passing means the values were built to sum to one. | WINNER_SECTOR_PATTERNS.healthcare
HOUSE | HOUSE: unsourced. historical_awards.py's own header states that the 'Source: CDFI Fund Annual Reports' comment over this dict cites a publication that DOES NOT EXIST (the real series is the NMTC Public Data Release Summary Report, cumulative FY2003-FY2023) and that 'Every value under them is unsourced'. A CONSTRUCTED PARTITION, NOT A MEASUREMENT (1.5.0 F1): the eight sector shares are 0.22 / 0.18 / 0.17 / 0.14 / 0.12 / 0.08 / 0.05 / 0.04 --- all on a 0.01 grid, summing to EXACTLY 1.000, for a population this package holds no sector breakdown of. test_sector_shares_sum_to_one asserts the sum and test_healthcare_is_largest_sector asserts the ordering, so both the partition and its shape are ENSHRINED; their docstrings now say that passing means the values were built to sum to one. | WINNER_SECTOR_PATTERNS.affordable_housing
HOUSE | HOUSE: unsourced. historical_awards.py's own header states that the 'Source: CDFI Fund Annual Reports' comment over this dict cites a publication that DOES NOT EXIST (the real series is the NMTC Public Data Release Summary Report, cumulative FY2003-FY2023) and that 'Every value under them is unsourced'. A CONSTRUCTED PARTITION, NOT A MEASUREMENT (1.5.0 F1): the eight sector shares are 0.22 / 0.18 / 0.17 / 0.14 / 0.12 / 0.08 / 0.05 / 0.04 --- all on a 0.01 grid, summing to EXACTLY 1.000, for a population this package holds no sector breakdown of. test_sector_shares_sum_to_one asserts the sum and test_healthcare_is_largest_sector asserts the ordering, so both the partition and its shape are ENSHRINED; their docstrings now say that passing means the values were built to sum to one. | WINNER_SECTOR_PATTERNS.small_business
HOUSE | HOUSE: unsourced. historical_awards.py's own header states that the 'Source: CDFI Fund Annual Reports' comment over this dict cites a publication that DOES NOT EXIST (the real series is the NMTC Public Data Release Summary Report, cumulative FY2003-FY2023) and that 'Every value under them is unsourced'. A CONSTRUCTED PARTITION, NOT A MEASUREMENT (1.5.0 F1): the eight sector shares are 0.22 / 0.18 / 0.17 / 0.14 / 0.12 / 0.08 / 0.05 / 0.04 --- all on a 0.01 grid, summing to EXACTLY 1.000, for a population this package holds no sector breakdown of. test_sector_shares_sum_to_one asserts the sum and test_healthcare_is_largest_sector asserts the ordering, so both the partition and its shape are ENSHRINED; their docstrings now say that passing means the values were built to sum to one. | WINNER_SECTOR_PATTERNS.education
HOUSE | HOUSE: unsourced. historical_awards.py's own header states that the 'Source: CDFI Fund Annual Reports' comment over this dict cites a publication that DOES NOT EXIST (the real series is the NMTC Public Data Release Summary Report, cumulative FY2003-FY2023) and that 'Every value under them is unsourced'. A CONSTRUCTED PARTITION, NOT A MEASUREMENT (1.5.0 F1): the eight sector shares are 0.22 / 0.18 / 0.17 / 0.14 / 0.12 / 0.08 / 0.05 / 0.04 --- all on a 0.01 grid, summing to EXACTLY 1.000, for a population this package holds no sector breakdown of. test_sector_shares_sum_to_one asserts the sum and test_healthcare_is_largest_sector asserts the ordering, so both the partition and its shape are ENSHRINED; their docstrings now say that passing means the values were built to sum to one. | WINNER_SECTOR_PATTERNS.community_facility
HOUSE | HOUSE: unsourced. historical_awards.py's own header states that the 'Source: CDFI Fund Annual Reports' comment over this dict cites a publication that DOES NOT EXIST (the real series is the NMTC Public Data Release Summary Report, cumulative FY2003-FY2023) and that 'Every value under them is unsourced'. A CONSTRUCTED PARTITION, NOT A MEASUREMENT (1.5.0 F1): the eight sector shares are 0.22 / 0.18 / 0.17 / 0.14 / 0.12 / 0.08 / 0.05 / 0.04 --- all on a 0.01 grid, summing to EXACTLY 1.000, for a population this package holds no sector breakdown of. test_sector_shares_sum_to_one asserts the sum and test_healthcare_is_largest_sector asserts the ordering, so both the partition and its shape are ENSHRINED; their docstrings now say that passing means the values were built to sum to one. | WINNER_SECTOR_PATTERNS.mixed_use
HOUSE | HOUSE: unsourced. historical_awards.py's own header states that the 'Source: CDFI Fund Annual Reports' comment over this dict cites a publication that DOES NOT EXIST (the real series is the NMTC Public Data Release Summary Report, cumulative FY2003-FY2023) and that 'Every value under them is unsourced'. A CONSTRUCTED PARTITION, NOT A MEASUREMENT (1.5.0 F1): the eight sector shares are 0.22 / 0.18 / 0.17 / 0.14 / 0.12 / 0.08 / 0.05 / 0.04 --- all on a 0.01 grid, summing to EXACTLY 1.000, for a population this package holds no sector breakdown of. test_sector_shares_sum_to_one asserts the sum and test_healthcare_is_largest_sector asserts the ordering, so both the partition and its shape are ENSHRINED; their docstrings now say that passing means the values were built to sum to one. | WINNER_SECTOR_PATTERNS.clean_energy
HOUSE | HOUSE: unsourced. historical_awards.py's own header states that the 'Source: CDFI Fund Annual Reports' comment over this dict cites a publication that DOES NOT EXIST (the real series is the NMTC Public Data Release Summary Report, cumulative FY2003-FY2023) and that 'Every value under them is unsourced'. A CONSTRUCTED PARTITION, NOT A MEASUREMENT (1.5.0 F1): the eight sector shares are 0.22 / 0.18 / 0.17 / 0.14 / 0.12 / 0.08 / 0.05 / 0.04 --- all on a 0.01 grid, summing to EXACTLY 1.000, for a population this package holds no sector breakdown of. test_sector_shares_sum_to_one asserts the sum and test_healthcare_is_largest_sector asserts the ordering, so both the partition and its shape are ENSHRINED; their docstrings now say that passing means the values were built to sum to one. | WINNER_SECTOR_PATTERNS.other
HOUSE | HOUSE: unsourced. historical_awards.py's own header states that the 'Source: CDFI Fund Annual Reports' comment over this dict cites a publication that DOES NOT EXIST (the real series is the NMTC Public Data Release Summary Report, cumulative FY2003-FY2023) and that 'Every value under them is unsourced'. | WINNER_SECTOR_PATTERNS.mean_sectors_represented
HOUSE | HOUSE: unsourced. historical_awards.py's own header states that the 'Source: CDFI Fund Annual Reports' comment over this dict cites a publication that DOES NOT EXIST (the real series is the NMTC Public Data Release Summary Report, cumulative FY2003-FY2023) and that 'Every value under them is unsourced'. | WINNER_SECTOR_PATTERNS.max_single_sector_pct
HOUSE | HOUSE: unsourced. historical_awards.py's own header states that the 'Source: CDFI Fund Annual Reports' comment over this dict cites a publication that DOES NOT EXIST (the real series is the NMTC Public Data Release Summary Report, cumulative FY2003-FY2023) and that 'Every value under them is unsourced'. | WINNER_IMPACT_BENCHMARKS.mean_jobs_per_mm_qei
HOUSE | HOUSE: unsourced. historical_awards.py's own header states that the 'Source: CDFI Fund Annual Reports' comment over this dict cites a publication that DOES NOT EXIST (the real series is the NMTC Public Data Release Summary Report, cumulative FY2003-FY2023) and that 'Every value under them is unsourced'. | WINNER_IMPACT_BENCHMARKS.p25_jobs_per_mm_qei
HOUSE | HOUSE: unsourced. historical_awards.py's own header states that the 'Source: CDFI Fund Annual Reports' comment over this dict cites a publication that DOES NOT EXIST (the real series is the NMTC Public Data Release Summary Report, cumulative FY2003-FY2023) and that 'Every value under them is unsourced'. | WINNER_IMPACT_BENCHMARKS.p50_jobs_per_mm_qei
HOUSE | HOUSE: unsourced. historical_awards.py's own header states that the 'Source: CDFI Fund Annual Reports' comment over this dict cites a publication that DOES NOT EXIST (the real series is the NMTC Public Data Release Summary Report, cumulative FY2003-FY2023) and that 'Every value under them is unsourced'. | WINNER_IMPACT_BENCHMARKS.p75_jobs_per_mm_qei
HOUSE | HOUSE: unsourced. historical_awards.py's own header states that the 'Source: CDFI Fund Annual Reports' comment over this dict cites a publication that DOES NOT EXIST (the real series is the NMTC Public Data Release Summary Report, cumulative FY2003-FY2023) and that 'Every value under them is unsourced'. | WINNER_IMPACT_BENCHMARKS.top_decile_jobs_per_mm_qei
HOUSE | HOUSE: unsourced. historical_awards.py's own header states that the 'Source: CDFI Fund Annual Reports' comment over this dict cites a publication that DOES NOT EXIST (the real series is the NMTC Public Data Release Summary Report, cumulative FY2003-FY2023) and that 'Every value under them is unsourced'. | WINNER_IMPACT_BENCHMARKS.mean_units_per_mm_qei
HOUSE | HOUSE: unsourced. historical_awards.py's own header states that the 'Source: CDFI Fund Annual Reports' comment over this dict cites a publication that DOES NOT EXIST (the real series is the NMTC Public Data Release Summary Report, cumulative FY2003-FY2023) and that 'Every value under them is unsourced'. | WINNER_IMPACT_BENCHMARKS.mean_cost_per_job
CITED | CITED: CDFI Fund NMTC award announcement for that round, cdfifund.gov/programs-training/programs/new-markets-tax-credit --- derived from the same announcements. | APPLICATION_VOLUME_TRENDS.years
CITED | CITED: CDFI Fund NMTC award announcement for that round, cdfifund.gov/programs-training/programs/new-markets-tax-credit --- restates NMTC_AWARD_ROUNDS; the CY2024-2025 column is the double round ruled above, now CITED to the Award Book. | APPLICATION_VOLUME_TRENDS.applications
CITED | CITED: CDFI Fund NMTC award announcement for that round, cdfifund.gov/programs-training/programs/new-markets-tax-credit --- restates NMTC_AWARD_ROUNDS; the CY2024-2025 column is the double round ruled above, now CITED to the Award Book. | APPLICATION_VOLUME_TRENDS.awards
CITED | CITED: CDFI Fund NMTC award announcement for that round, cdfifund.gov/programs-training/programs/new-markets-tax-credit --- derived from the same announcements. | APPLICATION_VOLUME_TRENDS.acceptance_rates
HOUSE | HOUSE: prose, and it FORECASTS --- 'Expect 30-35% acceptance in near-term rounds'. The CDFI Fund forecasts no acceptance rate. Deletion candidate. | APPLICATION_VOLUME_TRENDS.trend_note
HOUSE | HOUSE: benchmark_thresholds.py's SECTION B header says these are 'inferred from award announcements, not published by the CDFI Fund'. Round numbers chosen by this package. Disclosed on its face by benchmarks._METHODOLOGY. | WINNER_PATTERN_THRESHOLDS.min_deep_distress_pct
HOUSE | HOUSE: benchmark_thresholds.py's SECTION B header says these are 'inferred from award announcements, not published by the CDFI Fund'. Round numbers chosen by this package. Disclosed on its face by benchmarks._METHODOLOGY. | WINNER_PATTERN_THRESHOLDS.min_geographic_states
HOUSE | HOUSE: benchmark_thresholds.py's SECTION B header says these are 'inferred from award announcements, not published by the CDFI Fund'. Round numbers chosen by this package. Disclosed on its face by benchmarks._METHODOLOGY. | WINNER_PATTERN_THRESHOLDS.min_projects
HOUSE | HOUSE: benchmark_thresholds.py's SECTION B header says these are 'inferred from award announcements, not published by the CDFI Fund'. Round numbers chosen by this package. Disclosed on its face by benchmarks._METHODOLOGY. | WINNER_PATTERN_THRESHOLDS.max_geographic_hhi
HOUSE | HOUSE: benchmark_thresholds.py's SECTION B header says these are 'inferred from award announcements, not published by the CDFI Fund'. Round numbers chosen by this package. Disclosed on its face by benchmarks._METHODOLOGY. | WINNER_PATTERN_THRESHOLDS.min_jobs_per_mm_qei
HOUSE | HOUSE: benchmark_thresholds.py's SECTION B header says these are 'inferred from award announcements, not published by the CDFI Fund'. Round numbers chosen by this package. Disclosed on its face by benchmarks._METHODOLOGY. | WINNER_PATTERN_THRESHOLDS.max_single_sector_pct
HOUSE | HOUSE: benchmark_thresholds.py's SECTION B header says these are 'inferred from award announcements, not published by the CDFI Fund'. Round numbers chosen by this package. Disclosed on its face by benchmarks._METHODOLOGY. | WINNER_PATTERN_THRESHOLDS.min_sectors_represented
HOUSE | HOUSE: benchmark_thresholds.py's SECTION B header says these are 'inferred from award announcements, not published by the CDFI Fund'. Round numbers chosen by this package. Disclosed on its face by benchmarks._METHODOLOGY. | WINNER_PATTERN_THRESHOLDS.min_eligible_pct
HOUSE | HOUSE: the points a band is worth. Carried NO comment of any kind before 1.5.0 --- one of the three Section B dicts the historical_awards header's four-dict list never covered. | BENCHMARK_SCORE_POINTS.strong
HOUSE | HOUSE: the points a band is worth. Carried NO comment of any kind before 1.5.0 --- one of the three Section B dicts the historical_awards header's four-dict list never covered. | BENCHMARK_SCORE_POINTS.competitive
HOUSE | HOUSE: the points a band is worth. Carried NO comment of any kind before 1.5.0 --- one of the three Section B dicts the historical_awards header's four-dict list never covered. | BENCHMARK_SCORE_POINTS.weak
HOUSE | HOUSE: the points a band is worth. Carried NO comment of any kind before 1.5.0 --- one of the three Section B dicts the historical_awards header's four-dict list never covered. | BENCHMARK_SCORE_POINTS.below_weak
HOUSE | HOUSE: this package's own weighting of its own bands. Carried no provenance comment before 1.5.0. | BENCHMARK_METRIC_WEIGHTS.min_deep_distress_pct
HOUSE | HOUSE: this package's own weighting of its own bands. Carried no provenance comment before 1.5.0. | BENCHMARK_METRIC_WEIGHTS.min_jobs_per_mm_qei
HOUSE | HOUSE: this package's own weighting of its own bands. Carried no provenance comment before 1.5.0. | BENCHMARK_METRIC_WEIGHTS.min_geographic_states
HOUSE | HOUSE: this package's own weighting of its own bands. Carried no provenance comment before 1.5.0. | BENCHMARK_METRIC_WEIGHTS.max_geographic_hhi
HOUSE | HOUSE: this package's own weighting of its own bands. Carried no provenance comment before 1.5.0. | BENCHMARK_METRIC_WEIGHTS.max_single_sector_pct
HOUSE | HOUSE: this package's own weighting of its own bands. Carried no provenance comment before 1.5.0. | BENCHMARK_METRIC_WEIGHTS.min_sectors_represented
HOUSE | HOUSE: this package's own weighting of its own bands. Carried no provenance comment before 1.5.0. | BENCHMARK_METRIC_WEIGHTS.min_projects
HOUSE | HOUSE: this package's own weighting of its own bands. Carried no provenance comment before 1.5.0. | BENCHMARK_METRIC_WEIGHTS.min_eligible_pct
