Why It Matters
CBSA is the 5-digit numeric identifier behind every federal dataset aligned to a "metro." When you pull FHFA HPI for Columbus, you're pulling CBSA 18140. When HUD publishes Fair Market Rent for Dallas, it's CBSA 19100. There are 938 CBSAs in the United States — 391 MSAs with urban cores over 50,000 population and 547 smaller micropolitans (μSAs). The distinction between CBSA and county matters: counties are political boundaries; CBSAs are commuting boundaries — the set of counties that share a labor market. That's the right grain for underwriting a market.
At a Glance
- What it is: The OMB-defined geographic unit that federal agencies use as "a metro area" — covers both Metropolitan (≥50K core) and Micropolitan (10K-49K core) statistical areas.
- Why it matters: Every federal dataset aligned to a metro — FHFA HPI, HUD FMR, BEA regional data, BLS employment — is keyed to a CBSA code, not a city name.
- How to use it: Always disambiguate same-named metros by CBSA code (Columbus OH = 18140, Columbus IN = 18020). When pulling federal data, filter by CBSA, not by city name.
- Total CBSAs (2023 OMB vintage): 938 — 391 MSAs + 547 μSAs. Codes are 5-digit numeric.
- Update cadence: OMB revises delineations every ~5 years (after each decennial Census) plus ad-hoc amendments.
How It Works
What CBSA means. CBSA is an umbrella term OMB introduced in 2000 to cover two older categories. Before 2000, the federal statistical system used "MSAs" for larger metros and had a separate less-formal classification for smaller urban centers. The 2000 standards unified both under CBSA and added a formal μSA definition for urban cores between 10,000 and 49,999. Today a CBSA is either an MSA (urban core of 50,000+) or a Micropolitan Statistical Area. The boundary isn't set by politicians — it's set by commuting patterns. OMB looks at Census commuting data, identifies a central county with a qualifying urban population, then adds adjacent counties where 25%+ of workers commute to the central county or where 25%+ of the central county's workers commute in. The result: a CBSA captures the counties that share a labor market. Full history at Wikipedia.
CBSA codes and naming. Every CBSA has a 5-digit numeric code assigned by OMB. The code is the disambiguator when names collide — there's a Columbus in Ohio (CBSA 18140), a Columbus in Indiana (18020), and a Columbus in Georgia-Alabama (17980). They're entirely different labor markets with different economies, different rent levels, and different appreciation patterns. Federal datasets key everything by CBSA code, not by the text name. When you download FRED regional series and see a numeric column called "cbsa_code" or "msa_code," that's what it refers to. OMB also publishes reference maps through the Census Bureau's geography portal so you can see exactly which counties belong to which CBSA.
Why federal agencies all use CBSA. Standardization. FHFA publishes the House Price Index at the CBSA grain. HUD publishes annual Fair Market Rent at the CBSA grain. BEA publishes Regional GDP and Personal Income at the CBSA grain. The Census Bureau publishes building permits at the CBSA grain. BLS LAUS publishes unemployment at the CBSA grain. If OMB changed the boundary in 2023 — and they did, moving 103 counties between CBSAs in Bulletin 23-01 — every downstream agency refiles its historical series to match. The single shared definition is what lets an investor compare Fannie Mae conforming-loan data to HUD voucher data without having to reconcile two different "Columbus" footprints.
Vintage matters when you're comparing historical data. OMB publishes a new comprehensive delineation after each decennial Census — 2003 (from 2000 Census), 2013 (from 2010 Census), and most recently 2023 (from 2020 Census via OMB Bulletin 23-01). Between comprehensive updates, OMB issues smaller annual amendments. The practical impact: if you're comparing 2019 FHFA HPI for "Austin-Round Rock" to 2024 FHFA HPI for "Austin-Round Rock-San Marcos," the footprints aren't identical. The 2023 OMB update renamed the metro and moved 1 county. Always check the CBSA vintage when comparing multi-year series. Most federal agencies refile history to the current vintage within ~6 months of an OMB update, but some lag longer.
Real-World Example
Carmen Vargas compares two "Columbus" metros and learns why CBSA codes exist.
Carmen is evaluating markets for her next long-distance rental purchase. A friend in Ohio has been pitching "Columbus" as a growth play. Before committing capital, Carmen pulls the numbers.
She Googles "Columbus Ohio median home price" and gets a mix of results — some reference CBSA 18140 (Columbus, OH metro — population 2.14M, centered on Franklin County), and some reference the city proper (population ~910K). Those are different numbers.
Then she realizes there are other Columbuses:
- Columbus, OH MSA: CBSA 18140 — 8 counties, 2.14M population, Intel fab under construction, strong job growth
- Columbus, IN MSA: CBSA 18020 — 1 county (Bartholomew), 84K population, Cummins Engine headquarters, tiny by comparison
- Columbus, GA-AL MSA: CBSA 17980 — 5 counties, 330K population, Fort Moore Army base, different labor market entirely
She pulls FHFA HPI by CBSA code for all three:
- 18140 (Columbus, OH): +38% over 5 years — the growth story her friend mentioned
- 18020 (Columbus, IN): +24% over 5 years — solid but not the same thesis
- 17980 (Columbus, GA-AL): +31% over 5 years — reasonable, driven by military employment stability
If Carmen had just searched "Columbus metro" and gone with whatever came up, she could have underwritten a deal against the wrong set of economic data. The CBSA code is what makes the data federation work — her friend meant Ohio, but federal data doesn't know that unless she's specific.
Pros & Cons
- Single shared federal standard — every statistical agency keys its regional data to CBSA
- Captures the true economic footprint (commuting-based) rather than political boundaries
- 5-digit numeric codes disambiguate same-named metros unambiguously
- OMB reference maps and Census portals make the boundaries publicly verifiable
- Combined Statistical Areas (CSAs) exist for the handful of cases where adjacent metros share significant commuting
- Vintage changes break historical continuity — 2013 Austin isn't the same footprint as 2023 Austin-Round Rock-San Marcos
- CBSA boundaries don't match MLS coverage areas, so "MLS data for Columbus" and "FHFA HPI for CBSA 18140" cover slightly different geographies
- Large metros get split into Metropolitan Divisions that some federal datasets report separately — NYC has 5 MDs, LA has 2
- Micropolitan (μSA) data is thinner than MSA data — many federal series cover only MSAs
- The commuting-pattern threshold (25%) is somewhat arbitrary and can move a county between CBSAs across vintages
Watch Out
- City name ≠ CBSA name: "Columbus" in a news headline could be any of three metros. Always verify the 5-digit code.
- 2023 vintage changes: OMB Bulletin 23-01 moved 103 counties between CBSAs. If you're running historical analysis back beyond 2023, confirm whether your data uses the 2013 or 2023 vintage.
- Metropolitan Divisions complicate large-metro analysis: FHFA doesn't publish parent-MSA HPI for the 11 largest MSAs — you have to use Freddie FMHPI as the fallback or aggregate MD-level data yourself.
- CSA vs CBSA: Combined Statistical Areas are larger groupings of adjacent CBSAs. Some datasets report at CSA level instead of CBSA — always confirm which one your source uses.
- μSA data is thinner: Many federal series (FHFA HPI, BLS CES) cover only MSAs, not μSAs. If your target market is a micropolitan area, some data sources simply don't have it.
Ask an Investor
The Takeaway
CBSA is the canonical federal definition of a metro. Every data pull you do from FHFA, HUD, BEA, BLS, or Census keys back to a CBSA code. When you underwrite a market, use the CBSA — not the city name, not the ZIP code, not the county alone. The code is what keeps Columbus-OH separate from Columbus-IN, and what lets you compare five years of historical data across sources without reconciling three different "metro" definitions.