Derived table · PPP and EIDL prosecutions

PPP and EIDL identity-theft prosecutions by program

Name: PPP and EIDL identity-theft prosecutions by program
Creator: Pandemic Darlings
Published: 2026-10-15

Of the 1,134 prosecutions that name PPP alone, 107 involved identity theft, 9.4%. Cases naming only EIDL ran at 19.8%, and the 104 cases whose program list also names unemployment insurance at 40.4%. Seven rows, counted from the 1,977-case identity-theft classification.

Rows7

Columns6

Files1 · 361 B

FormatCSV

Source vintageProject classification of the DOJ press-release case index, through 2026-07-10

Published byPandemic Darlings

Acquired / built2026-10-15

Built fromPPP/EIDL identity-theft classification

SHA-25656f9d51e44eb59abb17e1310c23883772c244a83537bc1faddc9e89aa590c306

RightsDerived from U.S. government public records; attribution to Pandemic Darlings requested.

Dictionarydictionary.csv

Download CSV 361 B Data dictionary CSV At a glance Preview Columns How to use Before you quote a row

At a glance

Measured on the file on 2026-10-15; the checksum above is the file these numbers describe.

Rows7

Columns6

Keyrow · unique

Cells left blank0 of 42 · 0.0%

Prosecutions1,977

Identity theft246 · 12.4%

Touching unemployment insurance104 cases · 40.4%

Not touching unemployment insurance1,873 cases · 10.9%

What it is

Each of the 1,977 prosecutions in the case-level classification is marked identity theft when another person's identity or information was used, and self-inflation when the defendant applied in their own name and misstated payroll, revenue or documents. This table counts the identity-theft cases by the programs each case names.

Four rows split the cases by program and add up to the total row. Two more rows cut across them: cases whose programs field mentions unemployment insurance, and the rest. Unemployment schemes ran on other people's identities more often: 42 of 104 cases, against 204 of the 1,873 cases that stayed with PPP and EIDL.

How it was built

The programs field of each case is split on commas and slashes, trimmed and upper-cased. A case is PPP if a piece reads PPP or PAYCHECK PROTECTION PROGRAM, EIDL if a piece reads EIDL or ECONOMIC INJURY DISASTER LOAN, and goes in PPP only, PPP and EIDL, EIDL only, or program not stated accordingly. Other programs named alongside, such as RRF or ERC, do not move a case.

A case touches unemployment insurance if the field contains the word UI or the word UNEMPLOYMENT anywhere. Shares are identity-theft cases divided by cases, in percent with one decimal.

Preview

row	row_kind	cases	identity_theft_yes	identity_theft_no	identity_theft_share_pct
PPP only	group	1134	107	1027	9.4
PPP and EIDL	group	588	86	502	14.6
EIDL only	group	243	48	195	19.8
Program not stated	group	12	5	7	41.7
All cases	total	1977	246	1731	12.4
Touches unemployment insurance (UI)	cross_cut	104	42	62	40.4
Does not touch UI	cross_cut	1873	204	1669	10.9

First 7 of 7 rows · cells longer than 60 characters are cut with an ellipsis · shown as text, unformatted.

Files

File	Rows	Size	SHA-256
`ppp-eidl-identity-theft-by-program.csv` CSV, UTF-8, header row	7	361 B	56f9d51e44eb59abb17e1310c23883772c244a83537bc1faddc9e89aa590c306	Download ↓
`dictionary.csv` data dictionary, 6 rows	6	—	built with this page	Download ↓

Columns

Column	Type, measured	Distinct	Range / values	Meaning
`row`	text	7	PPP only (1), PPP and EIDL (1), EIDL only (1)	The group of cases.
`row_kind`	text	3	group (4), cross_cut (2), total (1)	`group` = one of four program groups that add up to the total; `total` = all 1,977 cases; `cross_cut` = the unemployment-insurance split, which cuts across the groups.
`cases`	integer	7	12 – 2K · median 588	Prosecutions in the row.
`identity_theft_yes`	integer	7	5 – 246 · median 86	Cases classified as identity theft.
`identity_theft_no`	integer	7	7 – 1.7K · median 502	Cases classified as self-inflation.
`identity_theft_share_pct`	number	7	9.40 – 41.70 · median 14.60	`identity_theft_yes` as a percent of `cases`, one decimal.

Type, blanks, distinct count, and range are measured on the file. The same table as CSV: dictionary.csv.

How to use it

Specific to this file, measured on it: keys, joins, encodings, units, and the values that will trip a naive count.

Do not add the cross-cut rows to the groups. The four group rows sum to the total row; the two cross_cut rows also sum to it, but each of them contains cases from every group.
The program-not-stated row is small: 12 cases, so its 41.7% rests on 5 cases.
LF line endings, UTF-8, 361 bytes.

Before you quote a row

Prosecutions, not fraud. The cases are those in the Justice Department's press releases; schemes never charged, or charged without a release, are not here.
Charges change. A plea can drop an identity-theft count, and a release does not always mention one, so a case marked self-inflation can still have involved someone else's identity.
The rule decides the counts. A different way of reading the programs field moves cases between rows; the case-level file carries the field as written, so any other rule can be applied to it.

Cite this table. Name the source vintage; the numbers depend on it.

PPP and EIDL identity-theft prosecutions by program, from the PPP/EIDL identity-theft classification of 1,977 DOJ-announced cases. Pandemic Darlings, https://pandemicdarlings.com/data/ppp-eidl-identity-theft-by-program/, SHA-256 56f9d51e44eb59abb17e1310c23883772c244a83537bc1faddc9e89aa590c306