Defining proprietary trading firms (or “props”) is challenging. Their business models are – as the name may suggest – proprietary and closely held. While some props are quite small, others have vast balance sheets. Props may execute their business manually or by using technologically sophisticated trading tools. They utilise a wide spectrum of strategies, with varying levels of complexity and holding periods.
In this our latest LME Insight piece, David Lutz and Joe Vu look at the origin of props, how they have evolved over time, some general examples of the strategies they run and their role on the London Metal Exchange today.
Many props can trace their origins back to open outcry trading floors, where they were often referred to as “locals”. As a result, most prop trading communities have emerged in areas with the most active trading venues historically: Amsterdam, Chicago, New York, and London. Many of the founders of proprietary trading firms started on the floor and moved “upstairs” (ie from the trading floor to their company’s office) as more trading activity became electronic.
There are two common traits that many prop business models share and both are linked to their trading-floor history. First, props were and are owned by their operators, with limited or no outside capital and therefore small balance sheets. This led to a second trait: an inability to hold large positions. A trading firm with a small balance sheet simply cannot run the large positions that a bank or asset manager is able to manage. Without the ability to run a large position, props tended to focus on the business of market making.
Market makers focus on the provision of liquidity. They aim to profit by capturing the difference between the price they buy an asset at (the “bid” price) and the price they sell an asset at (the “offer” or “ask” price). This bid-ask spread reflects the value of the service these market makers provide to the market – namely, liquidity. When open outcry trading floors were more active, they were filled with liquidity providers who serviced (or “worked”) the order flow that the brokers brought to the trading floor.
At this point, it is appropriate to point out that the evolution of props did not play out on the LME in the same way it did in other markets. Category 1 Members trading on the LME’s open outcry floor, the Ring, are both market makers, the role fulfilled by props in other markets, and clearing members. To support their clearing activities Category 1 Members must have robust balance sheets, large operations and back- or middle-office teams. This is in contrast to a local in a trading pit in Chicago, who would use the services of their clearing member and often lacked the support functions and infrastructure that a clearing membership requires.
Over the last two decades, markets have experienced a major shift towards electronic trading that has impacted almost every asset class.
Props were at the forefront of this transition. As a local operating on the floor of an exchange, trading every contract available in your market was a challenge. You would have needed to support the personnel to simultaneously cover many products and exchanges. For a prop with a small balance sheet and limited headcount, it was very difficult to cover many different markets when they had to be physically present for each contract they traded. Therefore, props were some of the first market participants to take advantage of advances in electronic trading technologies allowing them to participate in a wide range of markets efficiently and with a smaller headcount. This is how props became the pioneers of electronic trading.
As markets continued their relentless evolution towards digitisation, it was the props pursuing a more technical approach to electronic trading who were often responsible for driving innovation in trading technology. Some examples include the use of microwave towers to transmit market data at the speed of light, field programmable gate array (FPGA) technology, and other advancements in microprocessor technology.
As their business and markets evolved, props continued to act as both market makers and takers, with high-volume, low-margin, technology-driven business models.
It is a near-impossible task to comprehensively catalogue prop strategies, given that they can take many forms. At a high level, props often incorporate trading signals to drive their trading strategies.
“Day trading” is very common as props generally operate with smaller balance sheets and typically do not hold significant overnight risk. Holding periods are often measured in minutes or seconds. Nevertheless, they often provide vital liquidity to markets, and support efficient pricing that reflects all available information.
Some of the most common strategies applied by props include: statistical arbitrage, momentum trading, scalping and bridging liquidity.
Many of the quantitative trading indicators, or signals, used by proprietary trading firms feed statistical arbitrage strategies. The models employed by firms are closely held. In general, props will model price movements with a distribution curve. They will not always use a normal distribution curve, but it helps to use this standard tool for illustrative purposes.
Figure 1: A single statistical model is applied to an asset
A distribution curve of prices will be centred on a theoretical price. As the market price deviates, up or down from the theoretical price, props will buy (highlighted above in blue) and sell (highlighted above in red) the contracts based on the statistical deviations they have identified. These strategies create a result where the participant is systematically buying low and selling high (as long as prices continue following a similar distribution curve) with activity on both sides of the market, reinforcing the role of props as liquidity providers.
As more participants running similar strategies enter the market, they often tend to employ different time-horizons and nuances in the calculation of their statistical model. The net impact of these aggregated strategies tends to be tighter markets, benefiting all market participants.
Figure 2: Many statistical models being applied to an asset
Momentum trading aims to profit from transactions based on observable trends that participants believe will continue in the direction of their current movement, or momentum. Momentum strategy takes many forms, and this section will focus on a high-frequency implementation used by props.
As an example, let us assume there is 50 lots on a specific offer price and a participant enters the market to buy 45 lots. 90% of the offer has traded out. The question is, will the price increase above an offer price where the majority of the liquidity has been taken already?
The answer is not deterministic, but some may think that it could be more likely that the price will increase. A prop will analyse historical data and determine how much a price moved when similar trade characteristics existed in the past. If this analysis reveals that there may be a signal for an upward move, and there are still some contracts available on the offer to buy, the prop may attempt to enter the market and buy these lots.
By any means, this is not a riskless trade. First, the price may not continue to go up. You may buy the last five lots and then the market price may subsequently fall and selling activity may ensue. Further, the order in the market may be an iceberg order. Iceberg orders are single execution instructions that divide one large order into smaller limit orders, to obscure the true order quantity. In our example, a further 50 lots (or another amount) may be placed in the order book as soon as the visible component of the iceberg has been filled.
Additionally, there may be other firms operating a similar strategy in the market at the same time, and they will be competing to buy the remaining order on the offer. These firms will be racing to trade this last fraction of the original order. Given the importance of being the first to send an order in a market where orders are matched by their time priority, this may become a latency sensitive trade. The success or failure of the trade may depend on the speed of a participant’s trading setup versus that of their competitors.
There may be extreme situations in which one trade triggers an avalanche of incoming orders that seek to ride the momentum caused by the initial trade. The result may be a self-fulfilling momentum and a market that moves in a more volatile manner than it would have done otherwise.
The presence of these strategies helps the market to fully incorporate information more quickly. If the market is unable or slow to incorporate the information from the initial trade, the price becomes less efficient, competitive and relevant.
At any given time there may be firms active on the LME, and many other derivatives markets, who are running latency sensitive strategies. There is no structural advantage received by firms who engage in these strategies on the LME market. They receive market data and submit orders in a way that any participant on the LME is able to. Further, the LME has implemented a fixed minimum delay mechanism (commonly referred to as a “speed bump”) in its precious metals market, demonstrating its openness to many voices in the discussion on the appropriate balance of speed and efficiency in market structure. This was a unique solution to support the growth of a nascent contract where latency arbitrage may have discouraged onscreen liquidity provision.
Scalping is a simplistic technique that a subset of props utilise, in markets which typically trade in a tight price range. Those running these strategies work orders on both sides of the bid-ask spread and exit their positions in the shortest time frame that yields a profitable two-way trade akin to traditional market making. The key risk these participants face is the occasional movement of the price outside of the previously tight price range. This trading technique is often referred to as scalping, where the goal is to make small, incremental profits while the contract is trading in a tight price range, with the aim of accumulating enough profit to cushion against an adverse trend outside of the tight price range.
A subset of the prop community is formed by companies that have set up a collection of trading desks, known as “trading arcades”. These trading arcades are active in scalping strategies and also serve as a de facto training group to foster the next generation of prop traders for the firm. Those who are able to profitably trade a market may receive larger budgets and “graduate” to bigger markets. Trading arcade businesses tend not to be as technologically sophisticated as other types of prop firms but rely on a larger number of traders to generate returns through the aggregation of their trading profits.
Some props act as a bridge of liquidity between two segments of a market. For example, let’s focus on a commodity exchange traded fund, or ETF. A commodity ETF, trading on a stock exchange, will track the prices of its components, likely including energy, agricultural, precious metal and industrial metal products. In order to effectively provide liquidity in the ETF, a participant must price the ETF by tracking all the underlying prices that determine the ETF’s price.
Firms that act as commodity ETF liquidity providers do exactly this. They use data from all relevant commodities markets (often including the LME) to price the ETF contract trading on the stock exchange. When they trade the ETF, they are now exposed to commodity price risk, and use the LME (and other exchanges) to offset this risk. In this sense, ETF market makers, many of whom are props, are using their presence in both equity and the commodity derivative markets to supply liquidity in one market based on the liquidity available in the other.
A consequence of this is that the impact of orders in one market may have the potential to impact prices in a related market, where bridging liquidity has been supported by a prop. By taking on large block sizes and trading out of these positions on screen, props provide a service in facilitating execution quality.
It is significant to note that robust markets often rely on bridge liquidity that comes from separate (but related) markets - a role often fulfilled by props.
LMEselect, the LME venue for electronic trading, launched in 2001, creating an electronic venue where props have become active in the metal derivatives market. Today props are active in almost all LME contracts, and their activity is focused on the 3-month prompt date, where the majority of electronic trading activity occurs.
The LME desires the increased participation of props where their activity contributes to a healthy order book for all market participants. As such, the LME has established a number of liquidity provider programmes (and one specifically for props) to encourage positive trading behaviour and liquidity additive activity.
Some of the strategies that have been pioneered by props have also been adopted by other participants, including asset managers, banks, and LME members. The use of technology and market data as drivers for trading strategies is now standard in most financial markets, including the LME, and, as we have seen, many LME participants are using electronic trading methods that originated in and evolved from decades of open outcry trading on trading floors and data centres around the world. The LME values engagement with all user types to ensure that the market continues to develop, garners wider participation and, simultaneously, maintains high quality prices and orderly markets.
If you have any questions, please contract email@example.com.