To main content

Consultation on the MiFID II/MiFIR review report on Algorithmic Trading

Letter to European Securities and Markets Authority, 19 March 2021.

Norges Bank Investment Management (“NBIM”) appreciates the opportunity to respond to the European Securities and Markets Authority (“ESMA”) review report on algorithmic trading. Our response is based on our experience as an active participant in European and global asset markets.

NBIM is the investment management division of the Norwegian Central Bank (“Norges Bank”) and is responsible for investing the Norwegian Government Pension Fund Global. NBIM is a globally diversified investment manager with assets valued at EUR 1,041 billion as of 31 December 2020, of which EUR 320 billion was invested in European equities and bonds of European issuers.

As a long-term investor NBIM has a vested interest in well-functioning financial markets that facilitate the efficient allocation of capital and promote long-term economic growth and thus in a regulatory environment for trading in financial instruments that facilitates such outcomes.

Institutional investors and trading algorithms

Large institutional investors typically trade sizeable orders across many instruments. In an idealised market where all participants are present simultaneously and trading is frictionless, the execution of large institutional orders should be seamless. This is rarely the case and execution will take place over time.

Depending on liquidity needs and urgency, institutional investors will often prefer to trade against other capital owners who wish to make a corresponding adjustment to their holdings and limit the trading with intermediaries that provide risk capital or arbitrates the lack of simultaneous presence of such “natural” liquidity.

To execute sizable orders in today’s fragmented and complex equity markets, institutional investors commonly rely on investment firms to trade on their behalf. The investment firms, in turn, rely on trading algorithms to execute those trades. Institutional investors may be able to specify some key parameters of these algorithms to influence the expected trade-offs between urgency and implementation shortfall. The individual trade executions undertaken to fill the investor’s order, however, are decided by the trading algorithm.

We find that market developments such as the use of trading venues operating under transparency waivers and the increased use of periodic batch auctions and closing auctions are features of the marketplace, that when well-constructed, help us access more “natural” liquidity and thus improve overall execution.

When we compare the execution of trades in European equity markets to similar trades in the North American markets, it is generally more costly or takes more time to execute sizeable orders in Europe. While differences in investor and market structure across these markets contribute to the spread in market liquidity, we believe it should also be a subject of consideration when assessing the regulatory framework for algorithmic trading. (Question 1)

Systematic Internalisers

The institutional clients of Investment firms conducting algorithmic trading must approve applicable OTC counterparties and trading venues. An extension of the definition of algorithmic trading to cover Systematic Internalisers would facilitate the due diligence and approval of such counterparties for algorithmic trading.

The increased trading volume we observe with Systematic Internalisers implies that their role is becoming increasingly important to the overall trading system and extending the regulatory definition of algorithmic trading to such Systematic Internalisers may provide adequate regulation. (Question 9)

Trading venue fee structures and best execution

MiFID II establishes the obligation of investment firms to take all sufficient steps to obtain the best possible result when executing orders for their clients. It further specifies that the investment firm’s own commissions and the cost for executing the order on each of the eligible execution venues has to be taken into account.

Investment firms commonly receive a fixed rate of commission from their institutional investor client base. While other payment models, such as a ‘cost-plus’ model that reflects the trading fees paid by the investment firm on the client’s behalf, are becoming increasingly popular, it is our sense that they are not yet dominant.

Fixed rate commissions from clients combined with differentiated and competitive fee structures from trading venues create asymmetries between market participants leading to potential incentive misalignment. Trading venue fee structures require complex comparisons both across both across venues and between trades on the same venue. This systematic discrimination can incentivise a type of rebate maximisation and venue-fee arbitrage behaviour between investment firms.

The concern is that current structures introduces an agency issue between the investment firm and the customer. The algorithmic execution which maximises rebates or minimizes the trading fees for the investment firm, may come at the expense of best possible execution for the customer.

ESMA should consider whether moves to reduce complexity discourage asymmetries between order types and encourage competition on fee levels as opposed to fee-structures could lead to a more transparent and less discriminatory and fair market structure. (Questions 30 and 31).

Incentives for Market making

Carefully executed market making in continuously trading and orderly markets earns the spread and should be expected to produce satisfactory return on capital over time for Investment firms involved in such activity. The monetary incentivises provided to these investment firm during orderly markets probably leads to narrower spreads. The rebates constitute a transfer from other market participants with no clear benefit to the market structure.

ESMA should consider amending the requirements on the agreements between trading venues and the algorithmic traders pursuing market making strategies such that the incentives come into force in the event of volatility episodes or “stressed market conditions” and should then be aimed at maintaining presence, volumes and reasonable spreads. (Questions 42-44).

Trading venue continuity

Continuity of trading is important for institutional investors who have a continuous need to adjust or rebalance positions. The instances of trading discontinuity due to the various technological and cybersecurity incidents that occurred during 2020 did accentuate the risks associated with such disruptions. The cost of securing a more resilient trading system must be weighed against the potential economic risks investors are exposed to due to disruptions.

The recent experience has shown the continued centrality of primary exchanges and their technological platforms. The consolidation among primary exchanges in Europe has led to a limited number of physically distinct matching engines. Multilateral trading facilities and other trading venues should in principle be able to facilitate continuous trading in the instance of disruption at the primary exchange. This did not happen in practice in 2020.

Apart from the continuous trading session, the closing auctions increase in importance and volume of shares. No fallback mechanism exists if there is an outage at the primary exchange for this key feature of the trading session. The current alternative approaches to determining closing prices impose a potentially high cost and risk on institutional investors. The existence of independent trading technology platforms across Europe could provide opportunity for operational risk diversification and increased overall system resilience.

We suggest that ESMA considers if primary exchanges should be required to establish ‘continuity venues’ and particularly for the closing auction. Such a solution might potentially be a cost-efficient approach to secure necessary contingencies. Most investment firms will have the necessary connections in place and would not incur significant cost. The cost to the trading venues, as more complex settlement processes might be required for each, should compare favourably to potential cost and risk to the system from potential trading disruptions. (Question 36)

Speed bumps and market symmetry

The IEX exchange first introduced speed bumps in 2012 and has since achieved a substantial market share in the US markets. Our experience has been that the innovation and introduction of symmetric speed bumps reduced the advantage for low latency high frequency traders and created a more level playing field for market participants and contributed to improved market depth.

The justification for such intervention in the matching process should be to create a level market where all participants can act on the same information, and not have a segment of the market participants being able to respond to the actions of others. Symmetric speed bumps achieve this.

The introduction of asymmetric speed bumps on the other hand institutionalises an advantage for one type of orders or segment of the market. Any reduction in quoted spreads due to the structure can be viewed as a transitory effect and need not reflect any real increase in actual market liquidity or de facto improvement of market quality.

Regulating trading venues to require tighter spreads from market makers in exchange for the imposed advantage they receive would not change this. Regulation should aim to enact a level and competitive marketplace and avoid asymmetries, not to affirm them. (Questions 47-51)


We have an interest in encouraging regulatory innovation that contribute to the attractiveness of European markets. We therefore welcome this important consultation of ESMA on algorithmic trading in European Markets.

We appreciate this opportunity to share our perspective and remain at your disposal for further to discussion of these matters.

Yours sincerely,

Emil R. Framnes, Global Head of Equity Trading and Transition, Norges Bank Investment Management

Vegard Vik /s/, Special Advisor, Norges Bank Investment Management

Download the letter (pdf)