Top 10 Data Sources for Emerging Market Underwriting
Ten data sources for underwriting secondary and tertiary markets — what each is genuinely good for, its update lag, and where it will quietly mislead you.
Data quality degrades as markets get smaller. National datasets have thin samples in small counties, commercial providers have poor coverage where there are few transactions, and the local knowledge that would fill the gap is not published anywhere.
These ten are what remains reliable, with an honest note on each about where it stops being trustworthy.
1. Census Building Permits Survey
census.gov/construction/bps — free, monthly, county and MSA.
The single most valuable free dataset for real estate underwriting, and the most under-used. Units permitted, split by structure size, monthly.
Good for: the supply pipeline 18–30 months ahead of delivery. If you check one thing before entering a market, check this. Watch out: permits are not starts and starts are not deliveries; some share never gets built. Small counties are lumpy — one project distorts a month, so use trailing twelve months.
2. Census Population Estimates, components of change
census.gov/programs-surveys/popest — free, annual, county.
Good for: the components — net domestic migration, net international migration, natural increase. The split matters far more than the total. Watch out: estimates, and they get revised. Annual, so it lags. Use a multi-year trend.
3. BLS Quarterly Census of Employment and Wages
bls.gov/cew — free, quarterly, county by NAICS sector.
Good for: employment and wages by sector at county level, which is how you measure concentration risk — the variable that most distinguishes small-market risk from large-market risk. Watch out: roughly a two-quarter lag. Some county-sector cells are suppressed for confidentiality, and in small counties those suppressions can hide exactly the employer you were trying to size.
4. American Community Survey
census.gov/programs-surveys/acs — free, annual and 5-year, down to tract.
Good for: income, household size, tenure, rent burden, age distribution. The 5-year estimates reach small geographies. Watch out: the 5-year file is a five-year average, so it lags a fast-moving market badly. Margins of error at tract level are large and are published — read them.
5. FRED
fred.stlouisfed.org — free, mixed frequency.
Good for: aggregating series from many agencies in one place, with charting and export. Rates, employment, house prices, regional series. Watch out: convenience invites using a national series where a local one exists. FRED will happily let you underwrite a tertiary market on national data.
6. FHFA House Price Index
fhfa.gov/data/hpi — free, quarterly, MSA and state.
Good for: a long, consistent house-price series based on repeat sales, which controls for mix in a way median-price series do not. Watch out: conforming-mortgage transactions only, so it under-covers cash and jumbo. Residential, so it is a leading indicator for commercial rather than a measure of it.
7. Freddie Mac Primary Mortgage Market Survey
freddiemac.com/pmms — free, weekly.
Good for: a clean, consistent benchmark for residential mortgage rates. Watch out: it is an owner-occupied conforming benchmark. It is not your commercial rate, your DSCR loan rate, or your bridge coupon. Use it as a directional reference, not a quote.
8. FDIC Quarterly Banking Profile and call reports
fdic.gov — free, quarterly.
Good for: the underused one. Bank-level CRE concentration, delinquency and capital position tell you whether the local banks financing your market are expanding or retrenching — which, in a market with three lenders, is a direct input to your refinance risk. Watch out: institution-level, so you must know which banks lend in your market. Lagged.
9. Local permit and assessor records
County and municipal portals — free, varying quality, parcel level.
Good for: what is actually being built and by whom, ownership history, assessed values, and the pipeline before it reaches any aggregated dataset. Frequently the earliest signal available. Watch out: wildly inconsistent between jurisdictions. Some are searchable databases, some are PDFs, some require a visit. Assessed value is not market value and the relationship varies by state.
10. Property managers and local brokers
Not a dataset — the substitute for the dataset you cannot buy.
Good for: current effective rents, concessions, absorption pace, turn times, contractor availability and which submarkets are actually leasing. In markets where commercial data providers have thin coverage, this is the only reliable source for the numbers your model is most sensitive to. Watch out: every one of them has an incentive. A broker wants a transaction; a manager wants a management contract. Ask two or three, ask for specifics rather than opinions — "what did the last three units rent for and how long did they sit" — and weight what they say against what you can verify.
What none of these gives you
Submarket-level commercial rent and absorption. Commercial data providers sell this and their coverage in tertiary markets is genuinely poor — often modelled rather than observed. Verify anything you buy against source 10.
Forward-looking anything. Every series here is historical. The pipeline in source 1 is the closest thing to a forecast you get for free, which is part of why it matters so much.
Insurance. There is no good public dataset on commercial property insurance pricing by market, and it has been the largest NOI surprise of recent years. You have to ask a broker.
A working cadence
- Monthly: permits, employment, rates.
- Quarterly: QCEW, FHFA, FDIC, re-score markets.
- Annually: population estimates, ACS.
- Continuously: conversations with source 10.
Pull in the same order each time and timestamp the file. The value is in the change over time, and you only see that if the method holds still.
What to do next
- Structure the pulls with population, jobs and supply: data framework for market entry.
- Turn them into a comparison with how to score secondary cities for rental demand.
- Watch lender capacity with the debt availability tracker by secondary market type.
Sources
Related Resources
1031 Exchange vs Capital Recycling for Portfolio Reallocation
The 45-day identification clock is a much harder constraint in a thin market. When deferring the tax is worth the deadline risk, and the three alternatives.
Best Rebalancing Models for Multi-Market CRE Portfolios
Real estate cannot be rebalanced like a stock portfolio — you cannot sell 8% of a building. Three models that work within that constraint, and when each applies.
Exit and Rebalancing Strategy for Emerging Market Portfolios
Most portfolios have an acquisition strategy and no exit strategy. Writing the sell criteria at purchase, and the sequence that follows when they trigger.
Get Real Estate Insights
Join other investors receiving actionable strategies and market analysis