It’s been two weeks since a major provision of the sprawling new EU regulation called Mifid II came into effect and its immediate impact… hasn’t exactly been as intended. In fact, it might well have helped achieve exactly the opposite of what it was meant to do.

The provision in question is a limit on the amount of equity trading that can happen in so-called “dark pools” – trading venues that preserve the anonymity of the investor. As part of the post-crisis regulatory frenzy, EU and UK regulators decided to try to push more trading of shares on to public exchanges and increase disclosure requirements.

Their thinking was that all markets should be transparent so that no investor gets short-changed. There was, in fact, no evidence of any malign effects from dark pools in the first place, but now that the crackdown has proceeded regardless, the initial signs are not promising.

Far from encouraging the use of public exchanges, all it seems to have done so far is push trading into non-public auctions.

In fact, according to Thomson Reuters data, the share of European equities being traded publicly has gone down, from 76.1% two weeks ago, before the regulation took effect, to 72.8% now. The volume concentrated in dark pools has fallen dramatically, from 6.2% to 3.4%. Trading from dark pools – and indeed from public exchanges – has migrated into auctions, which have gone from a 16.7% market share to 22.5%.

The reason for this is quite simple. Dark pools were always more expensive than regular trading venues, but investors found them to be worth it and used them for a reason. Their benefit is that they allow investors to avoid revealing the full extent of their trading intentions before they trade. In the regular “lit” book system used by exchanges, orders appear, with buyers’ and sellers’ details, before they are executed.

This means that traders, particularly high-frequency traders, can quickly work out if an investor has a very large block of shares to buy or sell and can take the other side of the trade accordingly, forcing prices in a disadvantageous direction.

Many traders and algorithms are so good at this, they can do it even if investors try dividing their orders into little blocks to disguise their intentions. Dark pools negate this problem, more than compensating for the extra expense investors pay on their slightly wider trading spread.

Even so, using dark pools still wasn’t a widespread phenomenon, as the Reuters data showed. Only 6.2% of equities by market value were traded on them before the reforms.

“We find no evidence of a negative effect of dark trading on market liquidity”The Financial Conduct Authority, before Mifid II took effect

A study by the UK’s own Financial Conduct Authority, which was a staunch supporter of Mifid II, states that this is well below the volume at which there is a risk of draining away liquidity to the extent that it disrupts price discovery. Published in August 2017, before Mifid II had taken effect, the study says: “We find no evidence of a negative effect of dark trading on market liquidity.”

It estimates that only at volumes closer to 11pc, rising to 17pc accounting for the liquidity provided by high-frequency traders, would normal market behaviour start to be affected.

Yet despite current trading volumes hovering at just over a third of the FCA’s own assessed “safe” level, they went ahead with their crackdown anyway.

Now, thanks to Mifid II, there are strict limits on the proportion of a company’s market value that can be traded in a dark pool in a given period.

What the Reuters data shows is that investors think the benefits of pre-trade privacy are so strong that they would rather wait longer to execute their trades in auctions, which gather up orders before executing and publishing information about them after the fact, than trade immediately on a public exchange. This has coincided with a three percentage point reduction in the market share of exchanges in just two weeks.

Why Mifid II should prompt investors who were actually happy using exchanges to move away we don’t yet know, but this 7,000-page regulation is so full of complicated provisions, like restrictions on the tick size of price movements, that it could be pushing trading offshore.

These sorts of unintended consequences “were entirely straightforward to see”, according to Octavio Marenzi, chief executive of Opimas Consulting. “There was no question people wouldn’t do these sorts of things,” he says.

From the very start, Mifid II has proven to be a massive headache with few obvious benefits (all of which could have been achieved in much less burdensome ways). Its start-date had to be delayed twice to give firms time to adapt; its introduction has been plagued by problems, including a data issue that forced EU regulators to delay enforcing the dark pool trading cap for two months; and it is now spraying out excessive floods of data like a firehose, which regulators have little capacity to analyse.

Some of its provisions, like a requirement to trade stocks inside the EU wherever possible, are so costly or unworkable that regulators have simply told firms to ignore them for now. So far, firms have spent some £2.5bn complying, according to Opimas.

The only good news is that, with Brexit in the works and a financial services deal looking unlikely, Mifid II could provide the UK with a huge opportunity to attract business from firms trying to avoid it, if we ever had the guts to replace it with something much leaner and more efficient.

That is unlikely for now, given the atmosphere around financial services, but over time the logic of doing so will become increasingly compelling. If there’s one saving grace from the EU’s decision to tie itself in regulatory knots, it’s that Britain has much to gain by untying itself.

It might have been our own regulators that helped create this mess, but it is now Europe’s problem. It doesn’t have to be ours.