Professor Frino has designed a mathematical model (an econometric time-series model) that demonstrates the impacts on lit markets when trading moves into dark pools, showing how bid-ask spreads change relative to volume. Designed for the Australian market, it is also applicable to other small markets with fast trading technology, such as Canada and Singapore.
According to Frino’s model, if 20% of trading moves into the dark, it will increase trading costs on the lit exchange by almost 1 basis point. This is many times greater than the round trip ASX fee of 0.3 basis points.
“When trading moves off-exchange, trading costs on the lit exchange rise and it becomes more costly for buyers and sellers to find each other” said Professor Frino. “This implies that dark liquidity has a significant negative impact on liquidity, and on price discovery.”
Professor Frino pointed out that his research was particularly focused on smaller markets. “In big, liquid markets like the US, or even Japan, moving some trading off-exchange may make little difference,” he said. “There, where the top five stocks on NASDAQ trade five times the volume of the whole Australian market every day, taking a portion of that trading activity away isn’t that meaningful for trading costs and volumes. But in Australia or Singapore it’s a very different story because volumes and liquidity are low to begin with.”
The problem for smaller markets is that they’re following the same trends as the US and Europe, but aren’t suitable for them, Frino says.
“We’re seeing markets fragment in Asia as we did years ago in the US and Europe,” he said. “The issue is that some Asian markets aren’t big enough to handle it. Pulling liquidity off the lit exchanges, and the subsequent increase in trading costs, only further injures already struggling volumes.”
While claims are often made about dark pools hurting retail investors, Frino cautions against drawing simplistic conclusions.
“On the face of it, it makes sense to assume that retail investors suffer most from these kinds of impacts,” he said. “But the fact is that a huge percentage of retail money is invested via institutions, so what’s good for institutional investors is good for retail investors overall. However, the research indicates that dark trading is harmful for lit markets – whether that trade-off is worth it for investors, who believe they get extra liquidity and some price improvement inside dark venues, is a debate that needs to be had within the framework of this new empirical evidence.”
Frino also points out that there are different types of dark pools – broker crossing networks, where investment banks cross client orders without ever exposing them to outside traders; buy-side crossing networks, which are open to all institutional investors; internalisers, where client flow is crossed against the broker’s own flow; venue dark pools, such as ASX CenterPoint, and dark orders on lit order books, such as iceberg orders.
“The debate about the sins and virtues of different types of dark liquidity is yet to begin,” he said. “But it’s one that regulators – and indeed investors – should be very interested in.”