August 2026 Serious Mortgage Delinquencies

August 2026 serious mortgage delinquencies reached 574,000 loans, up 92,000 from a year earlier. Early-stage delinquencies also rose during the month but remained below year-earlier levels, according to ICE’s August First Look. For servicing teams, these figures describe the size and severity of delinquency queues; they cannot show how individual borrowers moved through them.

Serious delinquency and early-stage stocks

ICE published the August estimates on September 28; its table specifies August 30, 2026, as the reference date. The delinquency categories exclude loans in foreclosure.

Source: ICE’s August estimates. Early-stage counts and changes are calculated by subtracting serious delinquency from total delinquency, using rounded inputs.

A team contacting recently delinquent borrowers needs a different view from one reviewing unresolved, longer-running cases. The national figures provide a benchmark for each group’s size, but neither team can use those stocks alone to estimate how many new assignments will arrive.

Read the monthly increase with ICE’s calendar caveat

The overall delinquency rate rose 14 basis points to 3.53%. ICE describes the rate as essentially unchanged after allowing for July’s calendar-driven decline. ICE’s July release, published August 25, had reported a 16-basis-point decline.

That qualification concerns the comparison base. Reading the releases together helps avoid treating every movement away from July’s lower level as new deterioration. ICE does not supply a numerical calendar adjustment for every severity category, so the same adjustment should not be assumed for serious delinquency or foreclosure activity.

Foreclosure inventory versus starts and sales

August had 37,000 foreclosure starts, down about 6% monthly, and 7,800 foreclosure sales, down about 2%. The foreclosure inventory rate remained 0.54%, according to ICE’s release.

ICE’s September methodology, page 22, distinguishes three measures:

• Inventory covers loans referred to foreclosure counsel and still in the process before sale.
• Starts count active loans entering foreclosure inventory during the reporting month that were outside foreclosure in the prior month.
• Sales identify loans in foreclosure the preceding month that move to post-sale status or are flagged as foreclosure liquidations.

Inventory describes the outstanding caseload; starts describe incoming foreclosure cases; sales describe one type of exit. Fewer starts can coexist with a growing inventory. Subtracting sales from starts does not necessarily reconcile the inventory: other exits and changes in the measured population would also need to be identified.

Our mortgage servicing coverage follows the operational questions behind these measures.

What the pre-pandemic comparison adds

The serious-delinquency rate was 1.04% of active loans, compared with an August average of 1.03% for 2017–2019. Foreclosure sales ran at 57% of August 2019’s pace, according to ICE’s historical comparisons.

These comparisons put August’s levels in context without forecasting what happens next. A rate measures prevalence within the active mortgage population; a count describes the volume of loans or events. Neither determines workload for a particular servicer. Before using a national benchmark for staffing, review your portfolio’s product mix, geography and the age of unresolved cases.

Coverage and measurement limits

ICE scales statistics from its McDash database to estimate the wider mortgage market. Its active-loan denominator includes delinquent loans and loans in foreclosure, but excludes post-sale loans and real estate owned. These definitions appear in the September Mortgage Monitor, released September 10; its mortgage-performance section covers July.

ICE’s First Look notes specify rounding to the nearest thousand, except starts and sales, which use the nearest hundred. Derived counts should therefore be treated as approximate. Keep the reporting period, definitions and source vintage together when comparing this release with other housing data and reports.

Why these totals cannot establish cure or roll rates

A cure rate follows delinquent loans that return to current status. A roll rate follows loans moving between payment-status categories, such as from 60 to 90 days delinquent. Both require a defined starting group and observations of what happened to those same loans.

Consider a hypothetical queue that grows by ten loans. Twenty entries and ten exits produce that result; so do 100 entries and 90 exits. The ending stock cannot distinguish those different workloads, and the number of exits alone does not identify how many borrowers cured.

ICE’s July First Look separately reported 64,100 cures from serious delinquency and 102,000 borrowers becoming 90-plus days delinquent. Those flow figures apply to July and cannot be carried forward as August outcomes. Forecasting foreclosure completions would require further assumptions about outcomes and timing; these aggregate figures cannot establish an approaching foreclosure wave.

Before changing staffing, request opening cases, new entries, returns to current status, other exits and closing cases for your own portfolio. Match loans across reporting dates before calculating transition rates. That is the evidence needed to distinguish a larger queue from faster movement through it.