IRS SOI ZIP Code Data 2017 – Individual Income Tax Statistics (no AGI breakdown)
- Rows
- 27,760
- Columns
- 153
- Data range
- 0 to 99999
Tax Year 2017 ZIP code-level aggregated income tax data, all states, no AGI class detail. Covers all major income and tax fields per ZIP code.
AI analysis
IRS SOI ZIP Code Data 2017 — Analytical Overview
1. Dataset Representation and Research Value
This dataset originates from the IRS Statistics of Income (SOI) division, publicly hosted at [https://www.irs.gov/pub/irs-soi/17zpallnoagi.csv](https://www.irs.gov/pub/irs-soi/17zpallnoagi.csv), confirming a U.S. federal government publisher with high institutional credibility. The "2017" in the filename anchors the data to Tax Year 2017 filings, meaning it reflects income and tax behavior reported roughly in calendar year 2018 — making it pre-pandemic and a useful baseline for economic comparison studies. With 27,760 rows spanning all 51 state FIPS codes and up to 99,999 ZIP codes, and 153 columns covering virtually every major line item on Form 1040 (wages, dividends, capital gains, deductions, credits, tax liability, and filing characteristics), this dataset is exceptionally rich for geographic economic analysis. The no-AGI-stub structure means each row represents a full ZIP code aggregate without income-class segmentation, simplifying joins but limiting within-ZIP income stratification. Its breadth makes it ideal for correlation studies linking tax behavior to housing markets, demographic patterns, health outcomes, or business density at the ZIP code level.
2. Data Quality Observations
The dataset presents an unusually clean quality profile: zero null cells across all 153 columns and 27,760 rows is a strong signal that the IRS pre-processed and suppressed rather than left missing values blank — small ZIP codes with fewer than 20 returns typically have data withheld entirely, so the absence of nulls likely reflects row-level suppression rather than true completeness. The AGISTUB column is a degenerate constant (min=0, max=0, σ=0, skew=NaN), contributing zero analytical value and should be dropped before modeling — it exists solely as a structural artifact of the no-AGI-breakdown version of this file. The ZIPCODE field shows 27,660 distinct values against 27,760 rows, implying approximately 100 duplicate ZIP code entries that warrant investigation — these may represent ZIP codes spanning multiple states or special aggregation rows (e.g., ZIP code "0" for suppressed or unclassified returns, given min=0). The overall type consistency appears sound with no flagged type mismatches across the numeric columns.
3. Key Column Distributions and Notable Patterns
The primary size variable N1 (number of returns) exhibits extreme right skew (skew=57.2) with a median of only 1,950 but a mean of 10,784 and a maximum of ~17.8 million, with 1,858 flagged outliers — this is the classic signature of a small number of massive urban ZIP codes (e.g., Manhattan) dominating the tail while the majority of ZIP codes are small rural areas. The total income field A00100 (AGI) mirrors this: median=$107,934 vs. mean=$797,441, σ=$14.96M, with 2,680 outliers, confirming that a handful of affluent urban ZIPs wildly inflate the mean. Capital gains (A01000) are even more extreme — median=$2,223, mean=$52,199, skew=72.7, 4,082 outliers — suggesting highly concentrated investment income geography. Wages (A00200) show median=$75,000 vs. mean=$538,228, consistent with the return-count skew. On the deductions side, A04470 (itemized deductions) has a skew of 80 and 3,311 outliers, while charitable contributions (A19700) and mortgage interest (A19500, A19530) show extreme skewness (skew=96–133), pointing to wealthy coastal ZIP codes as the primary drivers. Filing behavior variables like VITA and TCE (free tax preparation usage) have medians of 0 with long right tails, indicating these services are concentrated in specific lower-income geographies — a useful socioeconomic signal. All dollar amount columns (A-prefix) should be treated as $1,000 units per IRS SOI convention.
4. Recommended Join Key Columns
ZIPCODE is the primary join key and is well-populated (0 nulls, 27,660 distinct values), though analysts must handle the ~100 duplicate entries and the special ZIP=0 record before joining. STATEFIPS (51 distinct values, 0 nulls) enables state-level rollup joins against Census, BLS, or HUD state-level datasets. STATE (string, 51 distinct) provides a human-readable state join key compatible with most public datasets. For per-capita normalization prior to joining, N1 (return count) and N2 (exemption count, a proxy for population) are the appropriate denominators — dividing A00100 by N1 produces a per-return AGI estimate comparable to Census median household income. Analysts should be cautious that ZIP codes are not stable geographic units across years, and a crosswalk to Census tract or ZCTA (ZIP Code Tabulation Area) identifiers is recommended when joining to Census Bureau data.
5. Recommended Pairing Datasets for Correlation Discovery
Several datasets would pair powerfully with this IRS SOI file. U.S. Census Bureau American Community Survey (ACS) 5-Year Estimates at the ZIP/ZCTA level (available via Census API) would enable direct correlation of tax-reported income, deductions, and dependency counts against demographic variables like age, education, race, and housing tenure — particularly compelling for studying how itemized deductions (A04470, A19300) correlate with homeownership rates. HUD ZIP-level Fair Market Rents or USPS vacancy data could reveal relationships between high mortgage interest deductions and housing cost burdens. CDC PLACES or County Health Rankings data (crosswalked via FIPS) could probe whether ZIP codes with high VITA/TCE usage (free tax prep, a low-income proxy) show elevated rates of chronic disease or uninsurance. Small Business Administration (SBA) loan data or Census ZIP Business Patterns (ZBP) would allow testing whether self-employment income (A00900) and Schedule F/partnership income (A26270) correlate with local business density and industry mix. Finally, prior-year IRS SOI files (2015–2016, available at the same IRS pub directory) would enable temporal correlation analysis to detect income mobility, credit utilization shifts post-Tax Cuts and Jobs Act (2017 being the last pre-TCJA year), and changes in itemization behavior — a particularly timely research angle given the TCJA's near-doubling of the standard deduction that took effect in Tax Year 2018.
Columns
- ZIPCODE (integer)
- STATEFIPS (integer)
- STATE (string)
- AGI_STUB (integer)
- N1 (decimal)
- MARS1 (decimal)
- MARS2 (decimal)
- MARS4 (decimal)
- ELF (decimal)
- CPREP (decimal)
- PREP (decimal)
- DIR_DEP (decimal)
- N2 (decimal)
- NUMDEP (decimal)
- TOTAL_VITA (decimal)
- VITA (decimal)
- TCE (decimal)
- VITA_EIC (decimal)
- RAC (decimal)
- ELDERLY (decimal)
- A00100 (decimal)
- N02650 (decimal)
- A02650 (decimal)
- N00200 (decimal)
- A00200 (decimal)
- N00300 (decimal)
- A00300 (decimal)
- N00600 (decimal)
- A00600 (decimal)
- N00650 (decimal)
- A00650 (decimal)
- N00700 (decimal)
- A00700 (decimal)
- N00900 (decimal)
- A00900 (decimal)
- N01000 (decimal)
- A01000 (decimal)
- N01400 (decimal)
- A01400 (decimal)
- N01700 (decimal)
- A01700 (decimal)
- SCHF (decimal)
- N02300 (decimal)
- A02300 (decimal)
- N02500 (decimal)
- A02500 (decimal)
- N26270 (decimal)
- A26270 (decimal)
- N02900 (decimal)
- A02900 (decimal)
- N03220 (decimal)
- A03220 (decimal)
- N03300 (decimal)
- A03300 (decimal)
- N03270 (decimal)
- A03270 (decimal)
- N03150 (decimal)
- A03150 (decimal)
- N03210 (decimal)
- A03210 (decimal)
- N03230 (decimal)
- A03230 (decimal)
- N03240 (decimal)
- A03240 (decimal)
- N04470 (decimal)
- A04470 (decimal)
- A00101 (decimal)
- N17000 (decimal)
- A17000 (decimal)
- N18425 (decimal)
- A18425 (decimal)
- N18450 (decimal)
- A18450 (decimal)
- N18500 (decimal)
- A18500 (decimal)
- N18800 (decimal)
- A18800 (decimal)
- N18300 (decimal)
- A18300 (decimal)
- N19300 (decimal)
- A19300 (decimal)
- N19500 (decimal)
- A19500 (decimal)
- N19530 (decimal)
- A19530 (decimal)
- N19550 (decimal)
- A19550 (decimal)
- N19570 (decimal)
- A19570 (decimal)
- N19700 (decimal)
- A19700 (decimal)
- N20800 (decimal)
- A20800 (decimal)
- N20950 (decimal)
- A20950 (decimal)
- N04800 (decimal)
- A04800 (decimal)
- N05800 (decimal)
- A05800 (decimal)
- N09600 (decimal)
- A09600 (decimal)
- N05780 (decimal)
- A05780 (decimal)
- N07100 (decimal)
- A07100 (decimal)
- N07300 (decimal)
- A07300 (decimal)
- N07180 (decimal)
- A07180 (decimal)
- N07230 (decimal)
- A07230 (decimal)
- N07240 (decimal)
- A07240 (decimal)
- N07220 (decimal)
- A07220 (decimal)
- N07260 (decimal)
- A07260 (decimal)
- N09400 (decimal)
- A09400 (decimal)
- N85770 (decimal)
- A85770 (decimal)
- N85775 (decimal)
- A85775 (decimal)
- N09750 (decimal)
- A09750 (decimal)
- N10600 (decimal)
- A10600 (decimal)
- N59660 (decimal)
- A59660 (decimal)
- N59720 (decimal)
- A59720 (decimal)
- N11070 (decimal)
- A11070 (decimal)
- N10960 (decimal)
- A10960 (decimal)
- N11560 (decimal)
- A11560 (decimal)
- N06500 (decimal)
- A06500 (decimal)
- N10300 (decimal)
- A10300 (decimal)
- N85530 (decimal)
- A85530 (decimal)
- N85300 (decimal)
- A85300 (decimal)
- N11901 (decimal)
- A11901 (decimal)
- N11900 (decimal)
- A11900 (decimal)
- N11902 (decimal)
- A11902 (decimal)
- N12000 (decimal)
- A12000 (decimal)