A new report from Norges Bank Investment Management, an asset manager responsible for managing the Government Pension Fund of Norway, offers insight into how the asset manager sources liquidity for the nearly $890 billion fund. In the paper entitled Sourcing Liquidity in Fragmented Markets, the firm uses its own experience to “evaluate whether trading venues with limited pre-trade transparency contribute to well-functioning markets.” Norges defines a well-functioning market as one “that maximises [sic] natural liquidity (long-term, natural buyers and sellers can find each other with high probability) while minimizing cost (rent extraction by intermediaries such as high-frequency traders, exchanges and broker/dealers should not be excessive).”
The paper argues that liquidity pools meet these criteria because they efficiently facilitate block trading, serve as an alternative to potentially monopolistic exchanges, and have the ability to tailor themselves to participant’s requirements and innovate quickly. However, the quality of venues can vary greatly. In the paper, Norges discusses an “execution plan” in which venues are prioritized based on the predicted quality of execution. High up on Norges plan are venues that “focus on direct block crossing between investors” whereas those that are only accessible through brokers fall lower on the list.
The report also offers statistics from seven of the algorithms most used by the firm’s trading desk since January 2014. The data indicates that the most aggressive algorithms execute almost all shares on lit exchanges (98% in the case of the most aggressive algorithm), with the balance of the orders executed on a high number of liquidity pools (18 liquidity pools in the case of the most aggressive algorithm). In contrast, the most patient algorithm was able to execute 100% of orders in liquidity pools. Norges finds in its ongoing post-trade reviews that liquidity fragmentation, or the spreading of available liquidity across multiple venues, does not appear to be much of an issue when utilizing algorithmic trading. The issue, rather, is the potential for information leakage and the “toxicity” of certain liquidity pools. As a result, Norges maintains a white list of allowed trading venues.
While liquidity pools “aid in limiting the rent extraction ability of intermediaries,” it is “important that the role of exchanges as the arbiter of the price discovery process is not impeded.” The paper suggests that the increase in use of liquidity pools means that on-exchange transactions are more informative because exchanges typically fall lower in the list of an investor’s list of preferred venues. Thus, the fact that the transaction is occurring on an exchange may mean that liquidity on dark venues is depleted. Norges suggests that further study should be done on the impact of liquidity pools on price discovery.
The report notes that fragmentation in block crossing venues can be an issue because “the probability of finding a match reduces geometrically with the number of venues.” As a result, Norges plans to “support the development of utility-like block crossing venues, which would serve to limit the possibility of rapid changes in relative market share, and to increase the fill probability.” According to Bloomberg, Norges is putting its weight behind the Plato Partnership, “a consortium of asset managers and broker dealers, which are collaborating to create a not-for-profit trading utility in Europe.”