The Numbers Everyone Cites — and Where They Come From
Search for construction site theft statistics and you will find the same figure repeated across hundreds of pages: theft costs the U.S. construction industry between $300 million and $1 billion a year.
That range is real, and it is worth understanding precisely — because the way it is usually quoted obscures what it does and does not measure.
The figure originates with the National Equipment Register (NER), a Verisk company, in partnership with the National Insurance Crime Bureau (NICB).
It only covers construction and agricultural equipment. It excludes several crucial components, such as stolen tools, building materials, copper, and other metals.
That $300 million to $1 billion figure also explicitly excludes every indirect cost a contractor actually feels — rental replacements, idle crews, schedule penalties, and administrative time (NER/NICB, 2016 Equipment Theft Report).
The same report is unusually candid about its own limits, stating plainly that no centralized, complete database of equipment losses exists. NER builds its national estimate by extrapolating from regions and fleets that report reliably.
This guide works through what the available data supports, what it does not, and what the pattern in the numbers suggests about stopping construction equipment theft before it happens.
Construction Theft Statistics at a Glance
$300 million to $1 billion — estimated annual U.S. losses from construction and farm equipment theft, with most estimates clustering near $400 million (NER/NICB, 2016)
11,574 — equipment thefts reported to the National Crime Information Center in 2016, the last year NER and NICB published a detailed public report (NER/NICB, 2016)
$29,258 — average value of a stolen machine recovered with NICB and NER assistance (NER/NICB, 2016)
21% — share of stolen equipment recovered in 2016 (NER/NICB, 2016)
Under 7% — recovery rate for single-item construction site thefts in FBI NIBRS data (Shrestha & Osborne, Associated Schools of Construction, 2019)
39.8% — share of construction site theft incidents involving tools, the single largest category (Shrestha & Osborne, 2019)
$5,865 — average loss per construction site theft incident across all property types (Shrestha & Osborne, 2019)
45% — share of national equipment thefts occurring in just five states (NER/NICB, 2016)
The Measurement Problem Nobody Mentions
Three structural gaps make construction theft harder to measure than almost any comparable property crime.
There is no title system for heavy equipment.
Cars have VINs and mandatory registration. Construction equipment has manufacturer PINs in inconsistent formats, no national titling requirement, and no standardized reporting path. NER and NICB identify this as a primary driver of both underreporting and low recovery.
The most-cited datasets have aged.
NER and NICB stopped publishing detailed annual public theft reports after the 2016 edition. Most 2025 and 2026 articles quoting “current” construction theft statistics are, on inspection, citing 2016 data — or citing a page that cites a page that cites it. Where this guide uses those figures, it labels the year.
Underreporting is systematic, not random
Contractors routinely absorb losses below their insurance deductible rather than file a claim and risk a premium increase. A Builder magazine survey found that while 87% of respondents filed police reports after a jobsite theft, three out of four did not file an insurance claim — meaning insurance-derived loss totals capture only a fraction of actual incidents.
The academic literature is equally frank. A study of construction theft published through the Associated Schools of Construction notes that most prior research rests on surveys of very small contractor samples — 42 respondents in one case, 102 in another — and that NER/NICB figures combine construction and farm equipment in a way that cannot be separated (Shrestha & Osborne, 2019).
None of this means the problem is small. It means the precision implied by a figure like “$1 billion annually” is not there, and anyone presenting it as a measurement rather than an estimate is overselling it.
What Actually Gets Stolen on a Construction Site (It’s Not Excavators)
The most useful corrective in the entire dataset comes from FBI National Incident-Based Reporting System data. Researchers at East Tennessee State University analyzed 15,274 theft incidents recorded at construction sites in a single year and ranked targets by frequency (Shrestha & Osborne, 2019):
| Property type | Share of incidents |
| Tools | 39.8% |
| Heavy construction equipment | 12.1% |
| Building materials | 11.3% |
| Household goods | 5.4% |
| Vehicle parts and accessories | 4.6% |
Tools appear in roughly four out of every ten construction site theft incidents — more than three times the rate of heavy equipment. Together, tools, equipment, and materials account for over 63% of incidents.
This matters because it inverts the mental model most security planning starts from. Site security is typically designed around the machine yard, because machines carry the biggest sticker price. But the highest-frequency loss is a pickup bed of cordless tools disappearing at 11 p.m., in an incident too small to claim and too routine to investigate.
Within heavy equipment specifically, the pattern is about mobility rather than value. In NER/NICB 2016 data, mowers and riding tractors accounted for 41% of thefts and loaders for 18% — and within the loader category, skid steers made up 70%. Large excavators are valuable but rarely stolen, because moving them requires equipment which thieves do not want to be seen operating. Thieves take what fits on an ordinary trailer.
What Construction Theft Costs Per Incident
Average loss figures vary by dataset because they measure different things.
NIBRS data puts the average construction site theft at $5,865 across all property types, rising to roughly $6,700 for incidents involving multiple items (Shrestha & Osborne, 2019). That average is pulled down by the volume of tool thefts.
Sorted by target, the picture changes sharply. Trucks were the costliest single category at an average $41,853 per incident, followed by other motor vehicles at $32,386 and heavy construction equipment at $15,571 in that same NIBRS analysis. NER’s separate figure for the average value of machines recovered with NICB assistance was $29,258 — higher, because it counts insured machines rather than all reported thefts.
Both figures understate total cost, and both source reports say so. NER excludes business interruption entirely. The academic literature lists the indirect consequences contractors report: job delays, operator downtime, higher premiums, possible policy cancellation, weakened bonding capacity, and liquidated damages. For a project on a tight schedule, a two-week wait for a replacement machine can cost more than the machine.
The Recovery Rate of Construction Theft Is the Real Story
If one statistic deserves more attention than the billion-dollar headline, it is this one.
NER and NICB recorded 2,442 recoveries against 11,574 reported equipment thefts in 2016 — a 21% recovery rate. FBI NIBRS data is harsher: under 7% of single-item construction site thefts resulted in recovery, and fewer than 5% resulted in an arrest (Shrestha & Osborne, 2019). Multi-item incidents recovered at roughly 5%.
For comparison, NICB has reported that more than 85% of stolen passenger vehicles are recovered.
The academic review puts the plausible range for equipment recovery at 7% to 22% depending on methodology — a wide band, but every point in it is catastrophic next to the figure for cars.
Buried in the NIBRS analysis is the detail that should shape security decisions. Among recovered items, the median recovery time for heavy equipment was one day, and over half of all single-item recoveries occurred within 24 hours. The mean was 24 days, skewed by a small number of long-tail cases.
Read those two numbers together: recovery either happens almost immediately or it effectively never happens. The odds collapse within hours of the theft. There is no meaningful scenario in which footage reviewed on Monday morning recovers equipment taken on Friday night.
Where and When Construction Site Theft Happens
Equipment theft concentrates geographically, and it tracks construction activity rather than population. In 2016, the top five states accounted for 45% of national thefts and the top ten for 63%, with Texas alone recording 2,375 — more than the next two states combined (NER/NICB, 2016).
Less intuitively, NER’s analysis of theft rates per capita found that none of the ten highest-rate statistical areas had a population above 100,000. Remote and rural sites are not safer; they are less watched.
Two further patterns from the same report:
Construction site thieves prefer new machines
56% of equipment stolen in 2016 was manufactured within the previous five years, and 75% within ten. This is the opposite of the automotive pattern, where older vehicles dominate theft counts — and the reason is that heavy equipment manufacturers have added little anti-theft technology across model generations.
A 2016 machine was roughly as easy to steal as a 1990 machine, but it fetches a much higher price.
Most construction site theft happens at work sites, not yards.
NER’s insured-loss data consistently shows the majority of losses occurring at what the insurance category calls others’ premises — i.e. active jobsites — rather than at a contractor’s own fenced compound. This makes sense because the security at a jobsite is usually much more lax compared to a fenced compound. The difference is security, not value.
That gap shows up most clearly on sites relying on mobile surveillance trailers, which are built for physical deterrence but typically still depend on someone reviewing footage after the fact.
Copper: The Fastest-Moving Number in the Data
The equipment figures are stable and dated. Metal theft is neither.
COMEX copper broke above $6.00 per pound in early 2026 after a run driven by data center construction, EV infrastructure, grid buildout, and a 50% U.S. import tariff introduced in 2025. Scrap yards pay several dollars per pound for wire, and stripped copper is effectively untraceable once melted.
The consequences for infrastructure are well documented. According to the trade group NCTA, there were more than 15,000 destructive attacks on U.S. communications networks between June 2024 and June 2025, with copper theft a major driver, affecting more than 9.5 million customers — and California and Texas together accounted for over half of incidents, as reported by CNN. Los Angeles reported that theft- and vandalism-related streetlight outages rose roughly tenfold between 2017 and 2022.
A widely circulated claim attributes $1 billion in annual construction site copper theft to the U.S. Department of Energy. That number appears throughout the security trade press, but it is consistently cited second-hand without a locatable primary document. Treat it as unverified. The directional evidence — price data, utility outage counts, and state legislative responses such as California’s AB 476 in 2025 — is strong enough without it.
The Complication: National Property Crime Is Falling
Any honest reading of construction theft data has to sit alongside an inconvenient fact.
U.S. property crime is at or near its lowest recorded level since the 1960s. FBI data showed property crime down 8.1% in 2024, with larceny-theft down 5.5%, burglary down 8.6%, and motor vehicle theft down 18.6% — the largest single-year drop ever recorded in that category.
Preliminary FBI figures released in 2026 indicated a further estimated 12.4% decline in property crime from 2024 to 2025.
So the national trend is down while the construction and metal theft trade press describes an escalating problem.
This isn’t really a contradiction. It’s a measurement issue. Both can be true, and the reconciliation matters:
National aggregates are dominated by retail and vehicle crime, categories large enough to move the total on their own.
Construction theft is poorly captured in those aggregates — it is undercounted at the source, and equipment is frequently misfiled in law enforcement systems.
Metal theft is genuinely commodity-driven. It responds to the copper price, not to the general crime rate.
The defensible position is narrow: there is no reliable national dataset showing construction site theft rising, and there is substantial commodity-price and utility-outage evidence showing metal theft rising. Anyone claiming a precise percentage increase in construction theft is working from data that does not exist.
What the Data Implies About Construction Site Security
Set the loss totals aside and look at the operational numbers — recovery rates, recovery timing, and target mix. Together they describe a specific mismatch.
Most construction sites are protected by systems designed to document theft. Cameras record. Sensors trigger. Someone reviews the footage afterward. That architecture is built on an assumption the recovery data contradicts: that evidence collected after the fact leads to recovery. At a 7% to 21% recovery rate, it usually does not.
The alarm data compounds the problem. U.S. Department of Justice research places the false rate for police alarm responses at 94% to 98%, with false alarms consuming roughly 20 minutes of police time each and accounting for 10% to 25% of all calls for service.
Departments have responded by deprioritizing unverified alarms, and a growing number have adopted verified response policies requiring confirmation before dispatch. Peer-reviewed analysis of Salt Lake City’s verified response program found an 87% annual reduction in police alarm calls alongside a 26% reduction in burglaries.
For a remote jobsite with no power, no permanent network, and a trailer full of power tools, an unverified motion alert at 2 a.m. is close to worthless. It cannot distinguish a subcontractor from a thief, it will not produce a dispatch, and by the time anyone watches it, the one-day recovery window has closed.
| Passive recording | Proactive monitoring | |
| Detection | Motion trigger | AI behavioral analysis |
| Review | After the fact, hours or days | Live operator, in seconds |
| Response | None; records only | Voice-down, lights, verified dispatch |
| Police priority | Unverified, deprioritized | Verified, in-progress incident |
| Evidence | On-site recorder, can be stolen | Cloud-stored, tamper-resistant |
| Recovery window | Missed | Intervention before loss |
Where Proactive Visual Security Fits
The statistics point toward a specific conclusion: on a construction site, the only intervention with a meaningful success rate is one that happens while the theft is in progress.
That is the design premise behind proactive visual security. CHeKT connects to the cameras already deployed on a site and adds the layers passive recording cannot provide — AI behavioral detection that distinguishes loitering and perimeter breach from ordinary site activity, live operator verification within seconds, and immediate voice-down deterrence through IP speakers mounted on site.
Where deterrence does not resolve the situation, the operator escalates a verified, in-progress incident to law enforcement rather than an unverified alarm. Verified events are retained in cloud-based storage that survives the theft of valuable on-site equipment.
For contractors evaluating hardware, CHeKT’s smart security cameras with proactive detection are built for this workflow, and the same approach is documented in CHeKT’s high-value inventory use case.
The relevant test is not whether a system captures the incident. It is whether anything happens during it.
How to Use These Construction Site Theft Statistics
If you are building a business case for site security investment, a few practices will keep it credible:
Cite the year. Most equipment theft figures in circulation are from 2016. Say so.
Separate equipment from tools and materials. The $300M–$1B range covers equipment only. Tools are the higher-frequency loss.
Use per-incident figures for project budgets. $5,865 average and $41,853 for trucks are more defensible in a proposal than a national aggregate.
Lead with the recovery rate. It is the least contested and most decision-relevant number available.
Add your own indirect costs. Published figures exclude delay and downtime. Your schedule penalties are specific to you.
Don’t claim a trend you can’t source. No public dataset currently shows a measured national increase in construction site theft.
