A new report from the Justice Department’s inspector general finds that during the past 10 years, the Drug Enforcement Administration has seized more than $4 billion in cash allegedly tied to drug activity. More than 75 percent of that total never resulted in criminal charges, and 80 percent ended up in the federal forfeiture fund through administrative procedures — meaning that the owners never got a day in court.

There’s no uniform standard for what constitutes a connection to drug activity. Cash can be seized if narcotics officers think it was earned from the sale of drugs or will be used to buy drugs. At airports, bus terminals and train stations, agents have seized cash from travelers coming from or going to a city known to be a hub for drug trafficking, for buying a one-way ticket or for acting “suspiciously.” From the report:

We found that different task force officers made different decisions in similar situations when deciding whether to seize all of the cash discovered. These differences demonstrate how seizure decisions can appear arbitrary, which should be a concern for the Department, both because of potentially improper conduct and because even the appearance of arbitrary decision-making in asset seizure can fuel public perception that law enforcement is not using this authority legitimately, thereby undermining public confidence in law enforcement.

In a review of 100 seizures, the report found that 85 occurred while property or the owner of the seized property was in transit (at an airport, during a traffic stop, at a shipping center, etc.), and that 79 of those were initiated based solely on a DEA agent or cooperating drug officer’s suspicions, not on preexisting intelligence.

While civil asset forfeiture is justified on the premise that it prevents criminals from ill-gotten gains, the DEA can’t say what percentage of its seizures resulted in broader investigations, or what percentage of those investigations resulted in criminal charges — because it doesn’t keep track. Given the building momentum against these seizures, one would think that if there were a clear connection between seizures of cash and investigations that nab drug dealers or result in large seizures of illicit substances, the DEA would want to track and promote those figures to tout its success. That they don’t even bother suggests that the connection between seizures and actual criminal activity is minimal.

That lack of data is why the IG’s office selected 100 cases to review itself. Its conclusion? “This review of sampled seizures provided evidence that many of the DEA’s interdiction seizures may not advance or relate to criminal investigations.”

In other words, they’re just stealing from people, albeit under the color of law. The report also found that state and local officers on federal anti-drug task forces aren’t required to get any formal training on forfeiture laws and procedures, which again results in rather arbitrary and inconsistent standards, policies and procedures.

Defenders of forfeiture often argue that the policy is generally only used against major kingpins and drug dealers. But the figures suggest otherwise. In 2012, the average DEA forfeiture was around $47,000. Last year, it was a little more than $71,000. But at the federal level, those averages are skewed by the few cases in which they really did bring down a major dealer. For example, a 2015 report found that 40 percent of forfeitures in New York City involved amounts of less than $250. Unfortunately, this latest IG report doesn’t offer a mean forfeiture amount, or break forfeitures down in a way that would indicate how much money is involved in each one. What does seem clear is that the typical forfeiture isn’t a byproduct of any broader criminal investigation. Instead, it has become its own kind of investigation, one that usually begins and ends when the drug cops take someone’s money.