Tim Kiladze in the Globe introduces us to the concept of ‘inverted markets’:
As high-frequency trading and electronic trading grew, stock exchanges started experimenting with a new model that paid high-frequency traders (HFTs) to post orders on their marketplaces. Known as the “maker-taker” pricing model, traders who removed liquidity, or executed an order and removed it from the marketplace, had to pay a “taker” fee; traders who posted that order got a “maker” rebate.
That’s still the dominant model for Toronto Stock Exchange-listed stocks, but there is incredible growth of a new system, known as an inverted market. It works in the exact opposite fashion. People who execute orders and remove them from the marketplace are given a rebate, while people who post orders have to pay to do so.
The rationale: sometimes there is a plethora of orders sitting around, with 20 different investors looking to unload the same stock at the same price. Of that 20, there may be one who needs to get the sale done immediately, so he or she will be willing to pay a small fee to jump to the front of the line. That’s why inverted markets are also known as “first look” markets.
The number of investors using these orders is growing fast. In January, 2015, inverted markets made up 9 per cent of trading volumes for TSX-listed securities. Now they make up 15 per cent of all trades, according to ITG Canada.
I find the explanation a bit clumsy: the third paragraph implies that a trader can jump the queue in a time-priority stack of orders by paying a fee, which is not the case. The actual reasoning is fairly involved and comes from an Aequitas diatribe titled NOT ALL SPEED BUMP MARKETS ARE CREATED EQUAL:
OPR [Order Protection Rule] has provided a perfect eco-system for those seeking to carry out predatory trading strategies.
An example of such an OPR-enabled predatory trading strategy would be as follows:
1. Place small orders across multiple marketplaces using the speed advantage to set the National Best Bid and Offer (“NBBO”), knowing everyone else will be forced (because of OPR) to trade with these orders first;
2. Use computer algorithms to identify trades from institutional investors or large retail orders, i.e. long-term investors (“LTIs”) that frequently break up orders to reduce market impact;
3. When an LTI order trades with one of these small orders, leverage the speed advantage to receive the information about this trade before the rest of the market;
4. Use this information, and the knowledge that the LTI order will next try to trade orders on other marketplaces (because of OPR), to technologically front-run that incoming LTI order, fade displayed quotes and ultimately trade with the incoming LTI order at a less favourable price to the investor.
To understand the impact of inverted markets it is important to ask the following question: who posts orders on a marketplace where they have to pay a fee?
• It is not the cost sensitive retail dealer. Retail dealers will, however, be takers of liquidity on inverted fee model marketplaces because they will have the opportunity to receive a rebate to do so, which they cannot obtain on make/take marketplaces.
• It is typically HFTs that are prepared to pay a fee to post orders for the benefit of interacting with retail flow.
There is nothing wrong with any of this as long as the retail dealer can demonstrate best execution.
However, when we take OPR into consideration, we come to the following conclusions:
• It is not retail orders that are being protected on inverted fee model marketplaces but orders from HFTs and other technologically sophisticated intermediaries, which were not the intended beneficiaries of OPR protection.
• With OPR, all trades are required to go to the marketplace with the best prices first (regardless of size), this makes inverted fee model marketplaces the perfect place for predatory traders to post small orders to get the first look at any type of flow and then deploy the type of strategy discussed above.
Incidentally, the Aequitas paper argues in favour of their product being a ‘protected’ market, i.e., subject to the OPR.:
We believe it is possible to build a marketplace where all industry stakeholders can co-exist and flourish. The NEO BookTM was specifically designed to promote and protect liquidity formation to the benefit of all liquidity seeking investors. We believe in fair access and that all liquidity providers should be able to compete on equal terms, regardless if they are HFT firms or institutional or retail investors.
This would no longer be the case if the NEO BookTM were to become an unprotected market, for two reasons:
1. Institutional investors will be at a competitive disadvantage compared to proprietary HFT firms who are in complete control of which market they access, and have the ability to quickly post and cancel their orders on an unprotected displayed marketplace. On the other hand, dealers that trade on behalf of their clients will be hesitant to post client orders on such markets where they are not price-protected and could get traded through. This is due to the lack of a well-defined best execution regime that demonstrably takes into account, monitors and enforces all elements of execution quality.
I have no sympathy whatsoever for this position.
Institutional investors can also be in complete control of which market they access, provided they perform the highly unusual step of thinking about what they are doing. Institutional investors charge their clients fat fees for their expertise, so let’s not spend too much time wailing over their lack thereof.
The second point is entirely dependent upon regulatory vagueness. The solution for regulatory vagueness is regulatory precision, not increasing the complexity of rules and exceptions that have the objective of counterbalancing this incompetence.
However, all this is leading up to a wonderful SEC memorandum on the topic of Maker-Taker Fees on Equities Exchanges:
The purpose of this memorandum is to facilitate an objective assessment of maker-taker fees in the U.S. equity markets by outlining the development of the maker-taker fee model in the U.S. and summarizing the current public debate about its impact on equity market structure. The memorandum will present both the asserted advantages and disadvantages of maker-taker fee structures. Though less frequently the focus of contemporary debate, it is important to note the asserted advantages of the maker-taker fee model. Specifically, some believe the maker-taker model is an important competitive tool for exchanges and directly or indirectly can provide better prices for retail investors. On the other hand, some believe it may exacerbate conflicts of interest between brokers and their customers, contribute to market fragmentation and market complexity through the proliferation of new exchange order types, and undermine price transparency.
I found the discussion of the NASDAQ experiment fascinating:
To test the premise that high access fees may discourage the use of markets that publicly display their posted best bid and offer (“lit markets”), NASDAQ conducted an access fee experiment in which it significantly lowered access fees and rebates in 14 stocks for transactions effected on the NASDAQ Stock Market over a four month period. The NASDAQ Pilot began on February 2, 2015, and lowered the access fee to remove liquidity from $0.003 to $0.0005 and reduced the credit to display liquidity to $0.0004 (such credits otherwise ranged from $0.0015 to $0.00305). NASDAQ’s stated intent in conducting the pilot was to test assertions that high access fees discourage the use of public markets and to generate “much-needed data about the impact of access fees on the level of off-exchange trading and, potentially, on price discovery, trading costs, displayed liquidity and execution quality as well.” NASDAQ provided data and prepared reports of the effects of the pilot that analyzed trading in the 14 stocks compared to a set of similar non-pilot control stocks. With respect to market share, NASDAQ expected offsetting effects, where the lower taker fee would be expected to increase market share and the lower rebate would reduce market share. In the first month of its pilot, NASDAQ observed a 2.9% decrease in market share in the 14 stocks compared to a 0.9% decrease in the control stocks. With respect to displayed liquidity, NASDAQ observed an expected decrease in response to the lower rebate incentive to display on NASDAQ. For example, NASDAQ’s time at the NBBO in the 14 stocks declined 4.9% compared to 0.3% for the control group. NASDAQ’s data thus showed statistically significant effects resulting from significant reductions in the access fees to take liquidity and related credits to post liquidity on NASDAQ in the 14 pilot stocks.
And the effect of maker-taker fees on retail market orders is also discussed:
Another important potential benefit of maker-taker fee structures is that they artificially narrow displayed spreads because the liquidity rebate effectively subsidizes the posting of liquidity. Broker-dealers that today execute virtually all retail marketable order flow off-exchange either match or improve upon the best price displayed on exchanges. Thus, to the extent displayed prices are artificially aggressive, this inures to the benefit of retail investors in the form of improved execution prices.
And in a discussion of the effects of maker-taker on best-execution requirements, inverted markets get a good mention:
For marketable orders, a broker may have an incentive to route to a trading venue that charges low access fees, or so-called “inverted” markets, offering rebates to take liquidity. However, venues with low taker fees (or that pay rebates to takers) generally have lower maker rebates (or impose fees on makers), and as a consequence, all else being equal, such markets would be less attractive to traditional liquidity providers compared to markets that pay a more attractive rebate to post liquidity for a given execution probability and therefore may have less posted liquidity available at the best price. These markets’ pricing structures also may attract sophisticated market participants that are willing to post liquidity on relatively unfavorable terms for the chance that such markets’ high position on taker routing tables will allow traders to interact with the first tranche of a large market order, thus allowing the traders to detect the earliest signs of a potential price move and quickly adjust their quoting or trading strategies on other markets. Accordingly, when a broker routes marketable customer order flow to a low taker fee (or inverted) venue, there is a risk that it actually may impair the execution quality of the customer’s order, particularly for larger institutional orders, if there is a potential for market-moving information leakage.
Again, I have no sympathy for such arguments whatsoever. If an institution is trading stupidly then they – and their clients – will have to pay for their stupidity.
And market complexity is discussed:
Some have suggested that to compete with non-exchange markets, as well as other exchanges, exchanges are motivated to offer the highest rebate to attract liquidity. To fund these rebates, exchanges must charge artificially high taker fees that may approach the access fee cap of $.003 per share. According to this view, within the maker-taker fee structure, where the difference between the highest rebate and highest taker fee approaches $0.006, exchange net trading fee revenues – the difference between taker fee revenues and maker rebate expenses – is generally less than one-tenth that range, between $0.0005 and $0.001 per share. Within this narrow range of net revenues, however, exchanges compete aggressively. The pressure to establish novel and competitive pricing often leads exchanges to modify their pricing frequently, typically on a calendar-month basis, which may add uncertainty and complexity to the marketplace as market participants must regularly update their routing tables to accommodate these frequent pricing changes.
Oh, routing tables must be updated? Trading strategies must be thought through to account for novel and competitive pricing? Well, Boo-Hoo-Hoo. If you’re a big enough trader for this to matter to returns, you’re big enough to think about it. This argument is merely illustrative that the controversy is artificial; it’s merely a means for the entitled private-school crowd to maintain their fat margins without having to compete against the hoi-polloi, who set up shop with not much more than a computer and a brain.