The current US Trade-Through Rule principally applies to market centres – and essentially stops any market from trading outside of the NBBO. So if a marketable order cannot be filled at the NBBO in a particular market, the market must either reject the order or onward-route it to a market that can satisfy it at the NBBO. The whole ‘Flash Order’ debate in the US was about market operators trying to avoid returning or routing orders to their competitors (as the rule requires) by instead introducing a brief delay during which it would seek to match these orders against selected liquidity partners.
Some interpreted the SEC’s ‘Trade At’ proposal to imply a tightening of the Trade Through rule, effectively enforcing time-priority across the competing lit venues. This could have the same effect as mandating a virtual Central Limit Order Book (CLOB), possibly undermining competition.
But the real thrust of the debate appears to be whether the ‘Trade At’ rule should apply to non-displayed order matching, including broker internalisation. What exactly does this mean?
The proposed ‘Trade At’ rule is that, when non-displayed orders are matched, brokers and market operators should be required to price-improve on the BBO by a minimum amount - either a full price-tick or a minimum proportion of the spread. In other words, they cannot ‘Trade At’ the BBO price without trading with the best publicly displayed Bid or Offer – effectively giving priority to the participant prepared to display their limit order.
Where displayed markets include non-displayed orders they effectively have a Trade At rule, in that displayed orders typically take priority over non displayed orders to give an execution priority of price, display type. For a non-displayed order to execute, it must be a minimum price increment better than the best displayed price. The same is also true of midpoint (dark) books operated by MTFs under the reference price waiver – they never give a non-displayed order priority over a displayed order at the same price.
Brokers internalise client flow whenever they can – whether they use an (American) ATS or a (European) BCN/BCS or SI as the platform. This makes economic sense both for the broker and for the clients – the broker avoids the exchange and clearing fees associated with trading in a Public Limit Order Book (PLOB), and the clients interact with natural liquidity without signalling their intentions publicly. Often, though not always, the trade will happen inside the spread, offering both clients price improvement versus what they might have achieved in the public market. The more a broker internalises, the more competitive its commission rates for clients can be.
Those market operators accustomed to concentration rules might argue that internalisation shouldn’t be permitted. But it’s also hard to make a principled argument as to why a broker should be forced to buy services from an exchange/MTF and CCP if those services are not actually needed (which is the case if the broker has two clients willing to trade with one another). Especially where the level of post-trade transparency is adequate, it’s not clear who benefits (other than the exchange and CCP) from forcing brokers to use services they don’t need. And if the trade is being executed inside the spread, then there are not, by definition, any other market participants who are advertising their willingness to interact with either the buyer or the seller at that price.
On the other hand, many internalised transactions (especially of retail orders in US markets) happen at (or very close to) the public BBO. This raises two interesting questions;
- If brokers commit capital to internalise flow at the BBO whenever it is attractive to them to do so, what can be said about the flow that they don’t internalise?
- Some argue that such non-internalised exhaust flow is ‘more informed’, and hence makes the PLOBs less attractive for posting limit orders than they would be if the flow reaching them was ‘more balanced’.
- Is there a point at which, when internalisation reaches a certain threshold, the mix of marketable flow reaching PLOBs becomes less attractive, leading to wider spreads?
- When internalisation happens at the BBO, is this fair to the participant bidding/offering at the BBO in a lit PLOB?
- Firms post their bids & offers publicly, releasing information which may impact the price, in return for a greater certainty of trading. If the stock can trade repeatedly at your advertised price, but you don’t get filled, does this undermine the incentive to post displayed limit orders in the first place?
Of course, certainty of trade has already been undermined by having multiple competing markets (each with its own queue) – but at least in this case competition is between participants prepared to display their quotes publicly.
So there is a rational economic argument that internalisation of retail flow at the BBO will ultimately lead to wider spreads, (though proving that it’s actually happening would be rather tricky).
Proponents of a Trade At rule argue that this would continue to allow unimpeded matching of client flow within the spread, whilst driving more market-making activity (where the broker buys at the Bid or sells at the Offer) into the public markets. Greater marketable flow reaching the public markets would strengthen incentives for others to post displayed limit orders – driving tighter public spreads.
Firms that prefer to see more liquidity transact in public limit order books (exchanges, MTFs, prop-traders, brokers without the scale to internalise) therefore should be expected to support a Trade At rule. Unsurprisingly, this is indeed the case, although the politics of advocating something that’s potentially unpopular with your largest customers means that the contribution to the debate from exchanges and MTFs is somewhat subdued.
Still, collectivism isn’t a popular concept in capital markets – so arguing that brokers should consume and pay for services that they don’t need because “it’s in the public good” is contentious. And rightly so – legislating a revenue stream is hardly a recipe for competitive behaviour. Indeed, some might characterise it as a form of ‘concentration rule’ – albeit one that doesn’t mandate a single CLOB.
So how do you balance the legitimate (at least as far as I’m concerned) right of brokers to internalise client flow with the public good ensuring PLOB’s remain attractive places to display limit orders?
- You could quite reasonably take the view retail brokers are perfectly equipped to decide what is best for them and their clients – whether that is routing to an exchange or trading OTC against a broker that (through internalisation) offers lower (or zero) commissions.
- You could also argue sensibly that, absent any evidence that spreads are actually widening, there are insufficient grounds to consider such a significant change in regulation.
- If, however, regulators were serious about pursuing a Trade At rule, then one would want either
- A manageable way to exempt brokers from the requirement to interact with the public markets if they’re already contributing publicly to the BBO price at which they want to trade (because in such a scenario, it is harder to argue that the internalisation has undermined the incentive to post a limit order publicly), OR
- Alternatively, if an exemption-based approach was impossible to monitor effectively, another solution could be for regulators to mandate a Trade At rule without exemptions, but leave it to market operators to offer reduced cost (and non-cleared) ‘own firm preferencing’ within their order books such that brokers could ‘outsource’ their internalisation.
As I said, contentious. What do institutional investors think?
