Immigration court bond records

National records by fiscal year

FY2026 includes records dated through 2026-08-31; it is an incomplete fiscal year. Each earlier row covers 10-01 through 09-30.

These are bond-table records, not unique people or all detained people. A positive NEW_BOND amount is an amount stored in a record. It does not establish that a bond was paid or a person was released. The official field key does not define the decision-code values, so this page does not calculate a bond approval or release rate.

Zero, blank and unparseable new amounts are outside the amount median; they remain in the record count. A positive amount is not necessarily an amount paid.

Records by saved court code

The court is the code stored on the bond record. It may describe a historical location and does not verify the court handling a person’s case now. Select a fiscal year to view every recorded court-code group, including small groups.

Every recorded court-code group and positive-amount count is shown. A median is undefined when no positive numeric new amount was recorded. Small groups can vary sharply; court-code rows include an UNKNOWN category for invalid or missing codes.

Decision codes saved in the source

These are literal values in the DEC field. The EOIR field key calls it a judge’s bond decision code but does not define each value. The codes below are intentionally untranslated and should not be used to infer favorable or unfavorable outcomes.

Blank or invalid values appear as UNKNOWN. Every observed code group and positive-amount count is shown. Small groups describe few records and should be read with care.

Source and counting rules

The source is EOIR’s public CASE data release, 2026-08 snapshot, and its field key. We count one accepted row from D_TblAssociatedBond when COMP_DATE parses as a date from 2020-10-01 through 2026-08-31. A case may have multiple rows. The source table has 1,643,696 accepted rows across all dates, including rows outside this page’s fiscal-year window; 36 malformed source lines were rejected during ingestion.

Fiscal years run from 10-01 through 09-30. The nationwide tables and court groups use the saved decision date; records without a parseable date or dated after the snapshot cutoff are excluded. Every observed court and code row is shown. A positive-amount median requires at least one positive numeric NEW_BOND entry; small-group medians can vary sharply. No individual case records or identifiers are published here.

Bond proceedings have different populations and decisions from asylum merits cases. These counts should not be combined with asylum grant shares or EOIR’s pending-case stock.

Complete public bond aggregate (all published groups)