The regulatory, technology and market ‘dark arts trilogy’ of high frequency trading: a research agenda

Abstract

Computerization has transformed financial markets with high frequency trading displacing human activity with proprietary algorithms to lower latency, reduce intermediary costs, enhance liquidity and increase transaction speed. Following the “Flash Crash” of 2010 which saw the Dow Jones Industrial Average plunge 1000 points within minutes, high frequency trading has come under the radar of multi-jurisdictional regulators. Combining a review of the extant literature on high frequency trading with empirical data from interviews with financial traders, computer experts and regulators, we develop concepts of regulatory adaptation, technology asymmetry and market ambiguity to illustrate the ‘dark art’ of high frequency trading. Findings show high frequency trading is a multi-faceted, complex and secretive practice. It is implicated in market events, but correlation does not imply causation, as isolating causal mechanisms from interconnected automated financial trading is highly challenging for regulators who seek to monitor algorithmic trading across multiple jurisdictions. This article provides information systems researchers with a set of conceptual tools for analysing high frequency trading.

Appendix 1: Glossary of financial terms

Algorithmic trading	The use of computer programs to generate a strategy on how an order is executed. This program determines aspects such as quantity, time and price
Arbitrage	The simultaneous buying and selling of a differently priced financial instrument on the assumption that any differences are temporary enabling profit to be earned, when bought below or sold above the mean
Basis point (bp)	Used to measure percentage changes, with 1 bp equal to 1%
Broker	Usually licensed professionals who charge a fee for executing an investors order
Circuit breakers	Also called ‘collars’, and represents a condition that will halt the trading of a security on an exchange. For example, if a price has moved by 10% in a five minute interval, all further trading will be stopped for the next five minutes
Co-location	Exchanges are today offering large data centres where an investor can place equipment as close as is physically possible to the exchange’s order matching computers. This saves milliseconds on the time it takes to send and receive order details. Not all HFT will co-locate
Dark pool	Trading systems where an order is submitted anonymously. Once executed, trade details are then made public
High-frequency trading (HFT)	Using sophisticated algorithmic technologies and extraordinary high speed, these companies run trading strategies which generate orders at a high frequency. These orders typically have a short life (sub-second) and at the end of each day no positions are held (i.e. they close flat). Liquidity is the by-product of their strategies
Layering and spoofing	Both of these techniques refer to methods which try and artificially move a securities price, enabling profit to be made. With layering, for example, large buy orders and hidden sell orders are sent to an exchange. If the buy order moves the price then this can be captured in better sell prices by the hidden orders. Visible orders are cancelled before they are matched
Latency	The time interval between sending out a query and receiving the results. Every part of the infrastructure (network, processor, storage, etc.) contributes to this value. Reducing latency has led to the technical arms race
Limit orders	Specifies the price at which an investor or trader is willing to make a trade. Such orders provide the depth of a stock’s liquidity, the more limit orders there are, the more liquid is the stock
Lit exchange	Also called light pool market, refers to a stock exchange where the order book is made available to all who have paid for the service. The majority of volume is transacted on lit markets (70% in US)
Liquidity	The ability to buy or sell an order without affecting its market price. Investors need to be confident that they can easily buy or sell at a fair price. The largest provider of today’s liquidity is the HFT. They capture profit with a much tighter spread than that needed by the traditional market makers
Low-frequency trading (LFT)	Also known as broker-dealers and undertake to trade on both sides of the market, providing liquidity. Their profit is generated by the spread between the price that they buy and then sell a security for (the bid-ask spread)
Market efficiency	The degree to which prices reflect all of the available information. The efficient market hypothesis states that it is impossible to beat the market because all information is incorporated in the price
Market maker	The activity offered by a broker-dealer who has committed to buy and sell a particular security in order to facilitate trading. By keeping the markets liquid, it will become more efficient. They will trade from their own account. On the London Stock Exchange, they are called ‘jobbers’, whilst on the New York Stock Exchange they were formally known as ‘specialists’
Market quality	The term can be used describe a process which allows orders to be executed in a fair manner for all parties. How this is achieved is a function defined by each trading-venue’s characteristics, but the priority is towards the investors
Market stability	The lack of extreme price movements over a short time interval
Matching engine	The most common matching algorithm used by an exchange is for continuous trading. It follows a price/time (first in, first out) priority meaning that if two orders have the same price, the order which was received first is executed first
National best bid best offer (NBBBO)	In the US, brokers are required to execute trades at the best (lowest) offer price when buying and best (highest) bid price when selling securities for an investor
Order book	An electronic list used to record all of the interest of buyers and sellers in each financial instrument. It records, amongst other variables, the number of shares, the price and the time the order was received at the exchange. Those at the ‘top’ are executed first. The matching engine will process these data
Order types	Instructions which are sent to dark pools and lit exchanges adding additional information on how the order is to be processed. Some exchanges have thousands in addition to buy, sell, cancel and delete
Quote stuffing	The sending of thousands of messages by a vendor to camouflage the true purpose of a strategy. Other systems are forced to process all of these data and as a result are slowed down. In 1999, the US exchanges received 1000 quotes per second. In 2014, it was reported to have leapt to two million per second
Regulation National Market System (Reg. NMS)	Described by the United States Securities and Exchange Commission (SEC) as “a series of initiatives designed to modernize and strengthen the National Market System for equity securities”
Spread	The bid-ask spread is the difference between how much a security will cost to buy and how much will be received when it is sold. These spreads have been getting tighter over the past decade putting greater pressure on the traditional market maker model
Transparency	Helps reduce volatility because all of the information about a company is available to all investors – a corner stone of market efficiency
Volatility	High volatility suggests that the price of a security has rapidly changed over a short period of time. Market makers are tasked with providing enough liquidity to reduce sudden price changes