Do dark pools help or hurt price discovery and ordinary investors?

Nobody has proved a simple yes or no, and the research points in more than one direction. A 2015 study found that a little dark trading is harmless or even helpful, a lot of it is harmful, and big block trades in the dark do not seem to hurt. A 2014 model went further and argued that dark pools can help price discovery in the right circumstances. Regulators, for their part, have set limits rather than banning dark pools: the EU capped dark trading so that it would not harm price formation.

First, what is at stake: price discovery, the process of working out an asset's price through the interactions of buyers and sellers. If every trade is public, the public price has the best chance of being right. If many trades are hidden, the public price is built on less information. As long as non-public trades are only a small fraction of volume, the public price might still be considered fair; the greater the non-public proportion, the less confident we can be of it. Dark pools publish trades afterwards, so they still contribute, with a little delay.

The 2015 study, by Carole Comerton-Forde and Talis Putnins, found that dark trades are less informed than lit trades, and that high levels of dark trading raise adverse selection risk on the lit exchange by concentrating informed traders there. Their conclusion: low levels of non-block dark trading are benign or even beneficial, high levels are harmful, and block trades in the dark showed no evidence of impeding price discovery. The economist Haoxiang Zhu saw it the other way round: in his model, the dark pool acts like a screening device that siphons off uninformed traders, leaving the informed ones on the exchange, where they make the public price more accurate.

For ordinary investors the argument cuts both ways. A retail shareholder could be disadvantaged if a trader sells a large number of shares in a dark pool, because they cannot see the orders before execution. On the other side, dark pools may protect traders from market participants who use high-frequency trading in a predatory manner. Both claims are made by serious people.