Is Retail Theft Data Reliable Enough to Cite?

By awareretailgroup.com Editorial· Edited by Tom Walsh· Published 2026-10-25

A loss prevention director who cites the National Retail Federation’s shrink survey as “theft losses” has mislabeled it: the figure is a survey-based estimate of total shrink across every cause of loss, built from retailer self-reports. Retail theft data holds up only when matched to what it actually measures. Shrink surveys, incident-perception studies, and police crime counts each answer a different question, and a store’s own audited loss outranks all three in a finance conversation.

What does the NRF’s National Retail Security Survey actually measure?

NRF’s National Retail Security Survey is distributed by email to senior loss prevention and security executives, using the NRF and Loss Prevention Research Council’s combined distribution lists (NRF PDF). Respondents answer with percentages. NRF gives them specific shrink categories and asks them to estimate what share of their own shrink came from each one (NRF PDF). The 2023 edition, covering fiscal 2022, drew answers from 177 retail brands, representing 22% of 2022 annual retail sales and more than 97,000 US retail locations (NRF PDF). That sample produced NRF’s headline shrink rate: 1.6% of sales in FY2022, up from 1.4% in FY2021 (NRF). Each step from a respondent’s internal percentage up to NRF’s industry average is an estimate.

Why doesn’t the $112.1 billion figure mean theft?

The $112.1 billion figure is the FY2022 average shrink rate, 1.6%, applied to 2022 total retail sales, up from $93.9 billion in 2021 (NRF). Retail shrink is a bucket that groups several causes together. NRSS respondents reported that external theft, including organized retail crime, averaged 36% of total loss, and internal theft averaged 29% (NRF PDF). The dollar total is also likely underreported: 78% of respondents excluded ecommerce goods loss from their shrink calculation, and 57% excluded supply chain loss or theft (NRF PDF). A figure built from percentages of a self-defined bucket works as a planning estimate. It will not survive a request for the underlying receipts.

What happened to NRF’s annual shrink report?

Retail Dive reported in 2024 that NRF would not release an annual shrink report that year, after more than three decades of the research, and that NRF pointed to the 2023 survey, which reflects FY2022 numbers, as its latest shrink statistics (Retail Dive). NRF said a broad study about retail shrink was no longer sufficient to capture the industry’s key challenges and needs (Retail Dive). NRF now publishes the Impact of Retail Theft & Violence studies (NRF), which track reported incident changes and retailer perceptions rather than a shrink rate. The 2026 edition covers more of the market, representing 31.6% of total retail sales (NRF), but it does not replace the FY2022 shrink rate. That rate is still the latest NRF has published, so label it as dated wherever it appears in a budget request.

Why was the “nearly half of shrink is ORC” claim withdrawn?

NRF withdrew one organized retail crime claim outright. A report NRF published with private security firm K2 Integrity claimed that organized retail crime accounted for nearly half of the shrink retailers reported for 2021 (CNBC). NRF has since withdrawn the claim, saying the analysis was incorrect and based on flawed data (CNBC). NRF later said K2 Integrity “mistakenly linked the results of NRF’s 2022 National Retail Security Survey with a statement made by an expert in a 2021 Senate hearing that ORC resulted in $45 billion in annual losses” (NRF). Economist Trevor Wagener recalculated organized retail crime’s share of inventory loss using NRF’s own numbers and methodology and found it closer to 5%, versus the nearly 50% figure in the retracted report (Retail Dive). NRF itself says quantifying ORC is a known challenge, because “ORC is not a single event or act” (NRF). Treat any ORC dollar figure in a budget request as uncitable until the underlying study is named.

What can FBI and police crime data tell a store about theft?

Police data measure reported crime, a different population than a retailer’s internal count, and that count depends on which system the reporting agency uses. The Council on Criminal Justice found that the FBI’s Summary Reporting System shows 2023 reported shoplifting at the same level as 2019, while the newer NIBRS shows it 93% higher over that span (CCJ). Flash mob shoplifting is not a specific offense agencies report in NIBRS, so the FBI built its 2020-2024 flash mob report from shoplifting offenses at stores with six or more offenders (FBI). NIBRS also cannot isolate organized retail crime from ordinary shoplifting (CRS). In NRF’s 2026 survey, 63% of retailers said they report fewer than half of their store-related theft incidents to law enforcement (NRF). With most retailers reporting fewer than half of their theft incidents, cite a police count only as a trend in reported crime, and keep it out of any estimate of a store’s own loss.

How do shrink surveys and police data compare?

The Council on Criminal Justice’s mid-year 2026 update, drawn from 18 cities with shoplifting data through June, found the average reported shoplifting rate was 4% higher than the same period in 2025, and about 5% higher than in 2019 (CCJ). The same update cautions that “law enforcement data understate the total volume of crime,” because many offenses are never reported (CCJ). NRF’s 2026 study, by contrast, reported a 12.4% decrease in shoplifting incidents in 2025 compared with 2024 (NRF), a different period and a different count. A police-reported increase and a retailer-reported decrease describe two different populations of incidents, and a budget request should name which one it is citing.

How do I cite a shrink or theft number in a budget request so finance will trust it?

A round number with no source attached can lose a finance team’s trust. Run every shrink or theft figure through a short checklist before it goes into a budget line. The examples below use NRF’s headline shrink figure, because it is the one most often mislabeled.

• Source: name the organization that produced the number, not the page that repeated it. Credit the $112.1 billion figure to NRF’s own 2023 survey (NRF), rather than to whichever page found it secondhand.
• Year: state the year the data covers, separately from the year it was published. The $112.1 billion figure is sometimes repeated as 2022 theft or 2023 shrink, but the accurate label is FY2022 total shrink, published in the 2023 survey (NRF).
• Method: say whether a number is a survey average, a police count, or an audited total. The $112.1 billion is NRF’s 1.6% survey average applied to total 2022 retail sales (NRF); no one summed store losses to produce it.
• Scope: say what the figure includes. Respondents who excluded ecommerce loss (78%) or supply-chain loss (57%) from their shrink number reported a narrower figure than the headline implies (NRF PDF).
• Denominator: state what the number is a share of. Retail Dive’s figure that nearly two-thirds of lost inventory has internal causes groups employee theft with inventory errors and unexplained loss into one category (Retail Dive); it is a different number from an employee-theft rate.
• Interest disclosed: name co-authors and sponsors. The retracted organized retail crime report was co-produced with private security firm K2 Integrity (CNBC).
• Retraction checked: a figure that circulated widely can still be wrong. NRF withdrew its claim that organized retail crime caused nearly half of shrink in December 2023, after the analysis proved flawed (CNBC).
• Dated figures flagged: check which reporting system an older police shoplifting count used. A police count from before 2021 predates the NIBRS transition, and NIBRS participation has been rising since (CRS), so label its reporting system and year when you cite it.
• Store number first: open the budget request with your own audited shrink percentage from the last physical inventory, run through a shrink calculator, and cite national figures only afterward as context.

Even an outside figure that passes every check describes other retailers, so keep it in the context section of the request.