HMDA Purchase Share
When rates are low, the refi line takes over and purchase share craters; when rates are high, refi dies and purchase is most of what's left. The swing between the two is the cleanest read on where the mortgage market is in the cycle. Versus FL + US: Hillsborough running higher purchase share than the nation means net in-migration is still pulling transaction volume toward real sales rather than existing-owner refinance activity.
- source
- HMDA
- frequency
- annual
- units
- percent
- geo
- Hillsborough County, Florida
Note: loan_purposes codes — 1: home purchase, 31: refinance, 32: cash-out refinance. Home improvement (2) and other/NA (4,5) excluded from the denominator to isolate the purchase-vs-refinance cycle mix. Multi-value filter returns grouped aggregations; fetcher extracts purchase count and sums denominator from the same response.
Purchase share sits at 70.18%, down 4.05 percentage points from a year ago as refinance activity has picked back up off last year's lows. Hillsborough is still running ahead of the national average (which sits at 64.23%) — a signal that in-migration and real sales volume are still outpacing the existing-owner refi activity that dominates higher-rate cycles. The three-year retreat from the 2023 peak of 78.41% tracks the Fed's rate moves in reverse: as mortgage rates fell through 2024, refis came back online and ate into the purchase share. Watch this line against MORTGAGE30US — when rates stabilize, this should stabilize too; if refis keep climbing while rates hold, you're seeing existing owners pull equity faster than new buyers are stepping in, which compresses inventory and tightens cap rates.
"""
HMDA Purchase Share — % of originations that are home-purchase loans
(vs. refi or cash-out refi). Purchase + refi + cash-out refi = denominator.
"""
import requests
YEAR = 2024
url = (
"https://ffiec.cfpb.gov/v2/data-browser-api/view/aggregations"
f"?years={YEAR}&counties=12057&actions_taken=1&loan_purposes=1,31,32"
)
response = requests.get(url, timeout=30)
response.raise_for_status()
buckets = {row["loan_purposes"]: row["count"] for row in response.json()["aggregations"]}
purchase = buckets.get("1", 0)
total = sum(buckets.values())
share = 100 * purchase / total if total else 0.0
print(f"{YEAR} Hillsborough purchase share: {share:.1f}%")
You are analyzing HMDA dataset hmda_purchase_share — HMDA Purchase Loans as Share of Originations (annual %). Context: When rates are low, the refi line takes over and purchase share craters; when rates are high, refi dies and purchase is most of what's left. The swing between the two is the cleanest read on where the mortgage market is in the cycle. Versus FL + US: Hillsborough running higher purchase share than the nation means net in-migration is still pulling transaction volume toward real sales rather than existing-owner refinance activity. Attributes - Units: percent - Frequency: annual - Source: HMDA · Hillsborough County, Florida Latest observation: 70.18% as of 2025-12-31 Year-over-year change: -5.46% Analyze this metric for a tech-forward real estate investor with 1–3 properties. What is it signaling right now? What should I watch next?