I’m in Tokyo this week attending the fourth Conference of States Parties to the Arms Trade Treaty (ATT). Over the years I’ve written at length on the ATT, which was adopted as the result of a campaign by what is politely called “civil society” — in other words progressive NGOs —who were frustrated by the failure of all the other institutions focused on conventional arm to get them where they wanted to go.

The point of the ATT, in theory, is to require nations to apply humanitarian standards to arms transfers. In practice, the ATT mixes a relentless focus on the purported wrongs of the West — in particular, the U.S., Britain, and Israel — with a steadfast refusal to recognize that the other institutions failed not because they were badly drafted, but because most of the world’s nations are incompetent, malevolent, or both. The result is that the ATT is failing just as badly.

Don’t believe me? Let’s look at the evidence, taken solely from the ATT Secretariat and from the supporters of the treaty themselves — specifically, the 2018 ATT Monitor Report, released on Monday.

The Treaty’s supporters commonly argue that the ATT cannot work if nations do not fulfill the reporting requirements they have accepted under it. To quote the Monitor, page 97: “Transparency and reporting remain essential Arms Trade Treaty (ATT) objectives and are a key component of its effective implementation.” On its own, this is an entirely reasonable argument: transparency and reporting, in various forms, are common to treaties on a wide variety of subjects.

In the ATT, States Parties have pledged to report in at least two, and in practice three, ways. First, they have to submit an initial report on their efforts to implement the Treaty. But of the 92 reports due when the Monitor was published, only 67 had been submitted, and only 56 are publicly available. So only 61 percent of the States Parties have transparently reported on their implementation efforts.

Nations are then required to update their initial reports as they go about implementing the Treaty. To date, not a single nation — not one — has submitted an update. That means that the initial reports submitted in past years, which are now aging or out of date, are increasingly irrelevant or even deceptive guides to actual state practice.

Finally, nations are required to submit annual reports on arms transfers under the Treaty. Of the 89 required reports, only 36 were submitted on time, and — as of today — only 48 (covering the year 2017) have been received at all. As the number of nations party to the Treaty has grown, the number of annual reports received has actually declined: 51 reports were submitted for 2015, 50 for 2016, and only 48 for 2017. In other words, only 51 percent of States Parties have submitted a current annual report.

By the by, it’s incredible that, as the number of States Parties to the Treaty has grown since 2013, the number of annual reports submitted has fallen. I do not know of any other international institution which mandates reporting where the volume of reporting was highest in its first year. The normal pattern is that reporting rises over time, and then falls as the requirement becomes too routine to bother fulfilling. The ATT’s track record is unprecedented, and unprecedentedly poor.

Of the 97 States Parties to the Treaty, only 47 have submitted both an initial report and a current annual report. Even leaving aside the requirement to submit updates to the initial report, this means that less half the States Parties to the Treaty are actually complying with its reporting requirements in their most minimal sense. So if reporting is an “essential” objective and a “key” component of the Treaty, the only possible conclusion is that the Treaty is not working in more than half its members.

But of course, this assumes that all these reports are meaningful. One of the features of institutions like the ATT is that they often becomes elaborate exercises in box-checking, in which the goal is to pile up reports regardless of their quality or utility. Indeed, most assessments — though not all — of reporting under the ATT or other related instruments are nothing more than counts of the number of reports that have been submitted, made under the apparent assumption that one report is as good as the next.

The Monitor, to its credit, is an exception to this rule. It examined every arms transfer reported for 2016 between States Parties to the ATT, and checked to see if the reported export was mirrored by a corresponding reported import. Of the 1,923 transfers, only 172 were mirrored even in part, and only 31 were mirrored exactly — in other words, a reported export matched exactly with a reported import. So only 1.6 percent of of the reported transfers matched: the other 98.4 percent did not.

So according to the treaty’s own supporters: 1) a majority of ATT nations are not filing their reports and 2) the reports that are filed are 98.4 percent unreliable or unverifiable. If reporting is indeed “essential” and “key” to the ATT, as its supporters reasonably assert, then the ATT cannot be working.

And there is a final consideration. The point of all this transparency and reporting is, supposedly (page 99 of the Monitor), to “allow observers of the ATT to assess States Parties compliance with ATT obligations.” In other words, it is to allow observers to assess whether or not States Parties are complying with the humanitarian requirements at the heart of the ATT.

You might think, therefore, that the ATT’s supporters would eagerly comb through the annual reports looking for treaty violations. But as far as getting ahead of breaking crises goes, this would be a pointless endeavor, because by the time a report is submitted — in late May, if not later — the transactions reported (inaccurately) in it are at least 5 months old, and could be 17 months old. If you want to figure out what places in the world are likely to be buying guns, don’t bother with ATT reports: watch CNN.

So all these supposedly vital reports — these often absent and inaccurate reports— are rarely used. Instead of chasing data, the treaty’s supporters chase ambulances. They focus almost exclusively on high-profile conflicts and civil wars, on U.S. and British arms exports — with a sideways glance at the rest of Europe — and on Israel. Exports or transfers from anywhere else (like Russia) are usually ignored, or alluded to in the passive voice.

Occasionally, having found a Western sin, the treaty’s supporters will then go back to tothe ATT reports, or other related instruments, to demonstrate that the exports from Britain, France, or wherever were made, but that is where their research ends — not where it starts.

There is nothing wrong with requiring nations to report their arms exports (and imports). The U.S. itself does this better than any other nation in the world (and gets no l credit for it). It is hard to see how foreign policy can be subject to any sort of democratic oversight if the executive branch is allowed to export what it wants, when it wants.

But you will never get useful bottom-up reporting by mandating it through a top-down treaty. After all, if every nation out there wants to report their exports and imports, they are free to do so: there is no need for a treaty. The problem is that most nations don’t want to report, are incapable of reporting, or both. The ATT aspires to universality, but the closer it gets to universality, the less well it will work.

Of course, even today’s ATT requires transparent reporting to work. But that reporting is often lacking. When it is present, it is almost always inaccurate or unverifiable. The data it contains are, on average, a year old, making it useless in detecting crises before they break. In practice, the treaty’s advocates focus on sensationalizing well-known conflicts, using the ATT data only as an after-the-fact proof of the malfeasance of Western nations.

Thus, by the criteria of its own supporters, the ATT is not working, and it cannot work. That is why the U.S. should not be paying its bills, and why the U.S.’s signature should be removed from it.