Summary

The Unified BMF is a single-row-per-EIN consolidation drawn from the most-recent vintage in which each EIN appears across both the current monthly BMF pipeline and the legacy 501CX-NONPROFIT-PX pipeline. See chapter 11 for the full design.

Metric	Value
Total rows	3,687,435
Distinct EINs	3,687,435
Rows == distinct EINs (uniqueness gate)	✅ yes
Input CSV files	118
Vintages with at least one surviving row	114

Rows by source pipeline

Source	Rows	Percent
current	2,208,124	59.88
legacy	1,479,311	40.12

First / last year coverage

Metric	Year
Earliest first_year_in_bmf	1989
Latest first_year_in_bmf	2026
Earliest last_year_in_bmf	1989
Latest last_year_in_bmf	2026

Vintages observed per EIN

How many distinct BMF vintages did each EIN appear in?

Metric	Value
Minimum	1
Average	50.1
Maximum	214
EINs that appear in only one vintage	255,547

Vintage histogram

Number of surviving Unified BMF rows contributed by each vintage. A vintage with many rows contributed many EINs whose latest sighting was in that month.

Vintage (YYYY-MM)	Source	Rows contributed
2026-06	current	1,966,267
1989-06	legacy	193,984
2010-11	legacy	135,474
2016-08	legacy	97,784
2020-04	legacy	94,113
2022-01	legacy	54,697
2024-08	current	46,346
2019-08	legacy	42,702
1996-06	legacy	42,534
2011-06	legacy	42,412
2025-08	current	37,147
1998-09	legacy	34,539
2000-05	legacy	32,319
2023-08	current	28,300
2011-12	legacy	26,577
2022-08	legacy	25,074
1995-08	legacy	23,316
1997-10	legacy	22,987
2007-09	legacy	22,463
2014-06	legacy	20,208
2007-04	legacy	20,010
2012-12	legacy	19,602
2016-04	legacy	19,375
2013-07	legacy	18,728
2004-12	legacy	18,217
2003-11	legacy	18,166
2003-01	legacy	18,141
2004-04	legacy	16,696
2013-02	legacy	16,384
2018-12	legacy	16,293
1999-12	legacy	15,077
2015-07	legacy	13,427
2006-05	legacy	12,673
2023-07	current	12,471
2026-01	current	12,095
2010-08	legacy	11,730
2008-01	legacy	10,789
2002-07	legacy	10,687
2008-10	legacy	10,670
2009-10	legacy	10,532
2001-07	legacy	9,980
2026-03	current	9,574
2024-03	current	9,042
2010-07	legacy	8,989
2006-01	legacy	8,963
2010-05	legacy	8,946
2012-08	legacy	8,893
2008-04	legacy	8,790
2007-01	legacy	8,769
2002-01	legacy	8,568
2024-04	current	8,426
2024-02	current	8,276
2015-02	legacy	8,236
2009-04	legacy	8,046
2025-03	current	7,991
2012-06	legacy	7,987
2003-07	legacy	7,964
2014-02	legacy	7,690
2014-12	legacy	7,538
2023-12	current	6,971
2010-04	legacy	6,933
2015-09	legacy	6,894
2009-01	legacy	6,613
2008-06	legacy	6,585
2011-09	legacy	6,501
2025-02	current	6,347
2012-04	legacy	6,301
2015-05	legacy	6,221
2014-04	legacy	6,137
2015-12	legacy	6,056
2006-11	legacy	6,009
2009-07	legacy	5,812
2011-08	legacy	5,752
2005-07	legacy	5,634
2024-12	current	5,540
2005-11	legacy	5,423
2013-12	legacy	5,396
2016-03	legacy	5,020
2014-09	legacy	4,652
2011-07	legacy	4,607
2023-11	current	4,561
2013-05	legacy	4,528
2008-12	legacy	4,224
2025-06	current	4,102
2012-07	legacy	4,070
2013-09	legacy	3,742
2013-10	legacy	3,740
2012-02	legacy	3,626
2025-12	current	3,624
2010-01	legacy	3,578
2015-11	legacy	3,540
2024-06	current	3,532
2025-11	current	3,472
2013-06	legacy	3,408
2012-10	legacy	3,378
2025-04	current	3,338
2015-04	legacy	3,259
2025-07	current	3,142
2013-08	legacy	3,110
2025-05	current	3,093
2013-03	legacy	3,078
2011-10	legacy	3,052
2011-11	legacy	2,966
2012-03	legacy	2,874
2012-11	legacy	2,767
2026-05	current	2,738
2023-10	current	2,674
2024-11	current	2,442
2023-09	current	2,435
2013-04	legacy	2,433
2025-09	current	2,119
2024-07	current	2,059
2014-11	legacy	2,058
2016-02	legacy	1,595

Column completeness

Top 30 most-complete columns (highest non-null fraction):

Column	Non-null %
ein_raw	100.00
ein	100.00
org_name_raw	100.00
org_name_join	100.00
org_name_display	100.00
org_addr_zip_raw	100.00
org_addr_state_invalid	100.00
ntee_code_clean	100.00
ntee_common_code	100.00
ntee_common_code_definition	100.00
ntee_code_major_group	100.00
naics_code	100.00
ntee_code_definition	100.00
nteev2_code	100.00
nteev2_subsector	100.00
nteev2_org_type	100.00
nteev2	100.00
bmf_source	100.00
bmf_vintages_observed	100.00
ein_prefixed	100.00
EIN2	100.00
first_vintage_ym	100.00
last_vintage_ym	100.00
bmf_vintage_ym	100.00
first_year_in_bmf	100.00
last_year_in_bmf	100.00
org_addr_city_raw	99.92
org_addr_city	99.92
org_addr_state_raw	99.88
org_addr_state	99.81

Bottom 30 least-complete columns:

Column	Non-null %
org_parent_name	3.07
activity_code_definitions	20.87
activity_code_categories	20.87
dba_name_raw	25.01
dba_name	25.01
org_legal_suffix	36.40
in_care_of_name_clean	38.50
in_care_of_name_raw	38.58
revenue_amount	52.41
all_classifications_string	59.86
in_care_of_name_provided	59.88
org_addr_street_raw	59.88
org_addr_has_special_chars	59.88
org_addr_is_po_box	59.88
org_addr_is_rural_route	59.88
org_addr_street	59.88
org_addr_zip4	59.88
org_addr_full	59.88
org_addr_is_missing	59.88
org_addr_missing_number	59.88
classification_code	59.88
affiliation_code	59.88
affiliation_code_definition	59.88
deductibility_code	59.88
deductibility_code_definition	59.88
organization_code	59.88
organization_code_definition	59.88
status_code	59.88
status_code_definition	59.88
activity_code	59.88

Notes

All output columns are VARCHAR. The master build forces every column to string type so DuckDB’s UNION ALL BY NAME can combine the legacy slim per-vintage schema with the current full schema without per-column type-mismatch errors. Cast in your downstream tool: as.numeric(asset_amount) in R, pd.to_numeric(df['asset_amount'], errors='coerce') in pandas, etc. Lat/lon in the geocoded master output are explicitly cast to DOUBLE.
Legacy-only EINs have sparse columns. EINs whose latest appearance is in a pre-2014 legacy vintage carry only the legacy slim schema; current-pipeline-only columns (e.g. NTEEv2 fields, modern address quality flags) are NULL. The bmf_source column flags the source of each surviving row.