What Failed
Six approaches were tried and did not work. Three failed before the treaty existed. Three are failures inside it, and two of those are still live.
Before the Treaty
Thirteen years passed between the warning and the first binding cuts. This is what was tried in the meantime.
Asking industry to act on its own, 1974 to 1987
DuPont, the biggest maker, spent a decade arguing the science was unproven, running a newspaper campaign saying nothing was settled. It reversed only after the 1988 Ozone Trends Panel, then backed a ban and brought other makers with it. [6]
Single countries banning spray cans, 1978
The US, Canada, Sweden and Norway acted alone. Production moved to other uses and other countries, and global emissions kept rising. Unilateral bans moved the problem rather than solving it. [7]
The Vienna Convention on its own, 1985
Countries could not agree on any cuts, so the treaty required none. The ozone hole was published eight weeks later and it still took until 1987 to get binding limits. A framework with nothing in it did not survive contact with the evidence. [3] [2]
Inside the Treaty
These are failures of the agreement that worked, not of the years before it.
Leaving out chemicals used to make other chemicals
The feedstock exemption was granted because leaks were thought to be under 0.5% of production. They are not. This now delays recovery by about seven years, and the exemption has never been closed. [17]
Assuming the replacements were safe
HCFCs still harmed ozone, and HFCs turned out to be strong greenhouse gases. Each substitution needed a further treaty change to undo, 29 years apart. [9] [17]
Setting quotas without enforcing them
The EU limit on HFCs created a black market. Industry put it at up to 30% of sales in 2022, with Romania identified as a major entry point for smuggled Chinese-made gas. [18]
Delay caused by feedstock leaks [17]
7 yr
Estimated illegal share of the EU market, 2022 [18]
30%
The 30% figure comes from the European F-gas industry, not from an enforcement body or an independent study. It is the best number available and it is flagged here because of who produced it. An independent estimate would be a strong contribution. [18]
Two of these six are not closed. The feedstock exemption is still open and quota enforcement is still weak. Both appear again on the still open page, because a failure nobody has fixed is also an open item.