Order-to-Trade Ratio (OTR)

The order-to-trade ratio (OTR) is a measure of how many order messages a trading participant sends to a venue for each trade it actually executes. A participant that enters, amends and cancels a great many orders but trades rarely has a high ratio. Exchanges and regulators use it to monitor, charge for or limit messaging activity that consumes system capacity without resulting in trades.

The measure came into use with the growth of algorithmic trading and especially high-frequency trading, whose strategies involve constant revision of quotes. It is known by several names, including order-to-execution ratio and message-to-trade ratio, and the precise definition varies from one venue to another.

In its simplest form the ratio is:

• $\text{Number of order messages}$ counts new orders, modifications and cancellations sent by the participant in an instrument over the period, usually a trading day.
• $\text{Number of trades}$ counts the executions resulting from that participant's orders in the same instrument and period.

A firm that sends 50,000 messages in a share during a day and executes 400 trades has a ratio of $50{,}000 / 400 = 125$, often written 125:1. A second version uses quantities instead of counts, dividing the total volume of shares or contracts ordered by the volume traded. The two can differ widely: a firm that repeatedly quotes large size and trades small amounts has a higher volume-based ratio than its message count suggests.

High ratios are a normal feature of electronic market making. A market maker posts a bid and an offer in each instrument and must move both whenever the market moves, to avoid trading at stale prices. Each move is a cancel order followed by a new order, or a single replace order. If the price of an underlying share changes several times a second, a firm quoting hundreds of options on that share must send thousands of updates, while only an occasional quote is hit. Options and other derivatives therefore tend to have far higher ratios than shares. Ratios of several hundred to one are not in themselves a sign of wrongdoing.

• Capacity management. Every message uses processing power and network throughput at the exchange and at every firm that receives market data. Venues set ratio thresholds, and charge fees or restrict access when they are exceeded, to keep systems stable.
• Deterring manipulation. Practices such as spoofing, layering and quote stuffing involve entering orders with no intention of trading. All produce very high ratios, so the measure is one input to market surveillance.
• Policy limits. After the flash crash of May 2010, regulators in several countries introduced formal requirements. In the European Union, MiFID II obliges trading venues to calculate ratios of unexecuted orders to transactions for each member in each instrument and to set maximum levels. Indian exchanges, under rules of the securities regulator, levy graded charges on algorithmic orders above set ratios.

Rules usually contain exemptions. Orders placed close to the prevailing market price, or by firms with formal market-making obligations, are often excluded or treated more leniently, since the purpose is to deter excessive messaging without penalising the provision of liquidity.

Supporters of limits argue that fleeting quotes add little to the market. Orders that are cancelled within milliseconds cannot be traded against by most participants, yet they crowd data feeds and can mislead others about the depth available. Charging for excess messages makes those who impose the cost bear it.

Critics reply that the ratio is a crude instrument. A low ratio can be achieved by quoting less often or less tightly, so a strict cap may widen the bid-ask spread and reduce displayed depth, the opposite of what was intended. A market maker in a volatile instrument needs to update quotes frequently simply to manage risk. Academic studies of ratio-based fees introduced in various markets have reported mixed results, with some finding little change in market quality and others finding thinner order books. The measure also cannot by itself distinguish a manipulator from an active liquidity provider, which requires examining the pattern and placement of the orders.

The ratio is one of a family of statistics used to describe message traffic. Others include the cancellation rate, which is the share of orders withdrawn before trading; the average lifetime of an order in the book; and peak message rates expressed as orders per second. Exchanges combine these with pre-trade controls on each participant, such as message throttles that slow or reject orders above a set rate, as part of the wider framework of algorithmic trading controls.