I-765 final-stage wait hits 45 days, a 12-week low, as center divergence reaches 283 days

The I-765 Employment Authorization Document application recorded a final-stage wait of 45 days in the week of June 8, down 13 days from 58 days the prior week. That 22.7% single-week decline is confirmed by an outlier flag at z = -2.80, computed against the prior 4-week trailing mean, on a sample of 11,127 cases — the second-largest weekly count in the 12-week window. Against the 12-week trailing average of 81 days, this week's reading sits 36 days below the recent norm. The large sample means the reading is not a small-week artifact.

This week's result is not an isolated step down. The 8-week regression slope is -6.5 days per week with an R² of 0.81, a strong and confident trend. The median fell from 99 days in the week of March 23 to 45 days in Week 24, a cumulative decline of 54 days over the period. The 25th percentile (p25), which ranged through the mid-40s for most of the spring, compressed to 27 days this week, indicating that the fastest-clearing cases are now moving through in under a month.

The upper tail also narrowed sharply. The 75th percentile (p75) ran between 182 and 206 days for most of the spring window; this week it compressed to 92 days. That broad-based compression across the distribution, not just at the fast end, is consistent with a category-mix shift or a batch of cases maturing together, rather than a selective acceleration of one sub-group. The data does not identify the mechanism directly.

The aggregate median obscures a large divide at the center level. IOE, which processed 9,421 of this week's 11,127 approvals (84.7%), recorded a 31-day median and a downward slope of -4.4 days per week (R² = 0.58). LIN handled 209 cases at a 314-day median, the slowest center this week. EAC (144 cases) posted a 198-day median but carries the steepest downward slope in the table at -19.1 days per week (R² = 0.63). WAC (382 cases) sits at 138 days. MSC (the National Benefits Center, 124 cases) is at 95 days with a near-flat slope (-0.2 days per week). The uneven-centers flag fires when the spread between the fastest and slowest center exceeds 12 days; this week the spread is 283 days (314 minus 31), well past that threshold. The centers process different category and filing-type mixes, so the spread reflects routing structure as much as relative performance.

SRC (847 cases, 53-day median) is the one center trending in the wrong direction: its slope of +3.5 days per week with an R² of 0.77 is a statistically confident upward move. If SRC's share of weekly volume grows, it could put upward pressure on the aggregate even as IOE continues to fall. IOE's trajectory will effectively set the aggregate direction for as long as it accounts for roughly 85% of weekly approvals. Center-level category mix, where IOE routes electronically filed cases that may skew toward faster-moving sub-types, likely explains much of the spread, though the data does not provide a category breakdown by center.

The aggregate improvement is real, but it is almost entirely an IOE story; applicants routed elsewhere face a very different system.