State Of The Art Of Trading

Raw News Feed

• Execution Management
• Order Management
• Trading Venue

All News

IOI’s are

• “Indications Of Interest” FIX messages
• Used by the Sell-Side to electronically notify the Buy-Side of their intentions to buy or sell selected securities

IOI’s can be broad

• Symbol and Side (Large or Small Buyer of SYM)

Have some specific parameters

• Symbol, Side and Quantity (Buyer of 10,000 SYM)
• Symbol, Side and Price (Buyer of SYM at $10.25)

Or precise details

• Symbol, Side, Quantity and Price (Buyer of 10,000 SYM at $10.25)

Common usage of “Actionable” IOI’s that use precise details is between trading systems (eg- dark pools to other trading systems) some colocated, which respond electronically based on criteria set by algorithmic strategies. The SEC questions if those Actionable IOI messages are quotes subject to quote rules. This is a controversial issue for dark pools since the messages being sent by their systems could be seen as information leakage.

Conventional IOI’s are sent from a broker to a customer where the broker thinks the customer may have an interest in a position that the broker may either have in hand or could be in touch with. If the human customer is sensitive about the source of shares that a broker is using to fill an order then the customer may request that the contra-side of the order be natural (another customer) or proprietary (utilizes a sell-side firms capital and inventory- usually involves some market impact due to the brokers activity in recovering or eliminating a position). The concern of course is the same type of information leakage which may be observed from automated IOI’s. The very act of the broker shopping around for the contra-side of an order leaks information to the market which may be used in adverse ways against the customer who has given the broker an order.

If a customer is misled by a brokers indication as to the contra-side of the trade then the customer has direct and immediate recourse with the broker. When a broker does not treat a customer well then the customer can stop trading with the broker cutting off commissions that normally would be paid for trade services. Most brokers are compelled to do the right thing. A regulator such as FINRA does not need to babysit every step in the trading process.

I read with great interest of the rise of the machine news readers.

Thomson Reuters

• NewsScope Analytics and the article on it.

Deutsche Börse

• Need to Know News

Bain Capital Ventures

• Edgar Online

Whereas the nerd in me hopes the new technology will produce high-frequency trade quality results, the geek in me wonders when the regulators may decide that it is time for them to get in front of this technology as being unfair to the average machine. Will a machine with an Intel chip have lower latency than a machine with an AMD chip?

Afterall, current machines are only so fast at seeing and passing along an RSS update from a human-programmed news alert to a human that analyzes the event and then decides on an action. Machines are now being given the ability to read and understand faster than our meager brains possibly could. Once you add the color of news to the black-and-white of market data to quantitative traders ability to program machines to recognize trends and then launch barely recognizable orders to colocated servers and you wind up with the ultimate electronic market.

So when we now talk about a two-tiered market, the average human versus the quant or institution, should we be adding a third tier that accounts for a machine participant. Thankfully, the one and maybe only remaining thing that humans will have over thinking machines is the ability to produce alpha.

Though how far behind could that really be…

Miscellaneous Notes

Three types of latency

• Transmission Latency- Convert data to bits
• Propogation Latency- Movement of bits across a network [Greatest delay due to distance (ref: InfiniBand)]
• Processing Latency- The application of data through middleware (feed handlers) to trading systems (algorithmic engines) [Second greatest delay due to infrastructure (ref: Exegy) and the largest unsettled issue]

Moving an order encapsulated in a FIX message from Point A to Point B faster than your competition creates an opportunity to maximize profit or minimize loss. Colocation is the result of a quest for zero latency which has moved execution management servers from the customers physical location to the closest possible physical proximity of an execution venues servers. The shorter the wire, the shorter the amount of time it takes to move a message from Point A to Point B.

These are no ordinary orders and they require no human intervention other than the initial instruction set. These are orders generated by machines that have complex event processors analyzing massive amounts of current and historical market data. Algorithms are formulas designed to recognize correlations or disparities in prices, volume or other statistics and the financial instruments involved. PhD’s have programmed the machines to take resulting data, formulate that data into orders and send the orders in barely noticeable quantities for execution. No trail of information leakage to give away a pattern.

Most customers and their traders are not all that concerned with latency as timing is not essential. Best execution is their concern. However, there are a class of traders that use short term alpha. The kind of alpha that lasts for nano to milliseconds for which the slightest delay is often the difference between profit and loss. Direct Market Access traders in high-frequency programs has the greatest demand and an estimated 60%+ share of the overall market volume.

Definitions

• Zero latency is the speed of light = 186,000 miles per second
• Nanoseconds = 1 billionth of a second
• Microseconds = 1 millionth of a second
• Milliseconds = 1 thousandth of a second
• One foot of cable = 1 nanosecond latency

Statistics

NYSE Technologies Universal Trading Platform

• 5 milliseconds, 3 milliseconds for data, 1 millisecond TSE colocated
• 2 milliseconds for order and cancel acknowledgements
• 650 microseconds roundtrip executions for Nasdaq-listed issues
• 950 microseconds roundtrip executions for NYSE/ Arca-listed issues

Others

• OptionsIT Transatlantic: 34 milliseconds
• Chi-X: 175 milliseconds, marketable IOC 350 microseconds
• Nasdaq: 250 milliseconds
• LSE: 10 microseconds
• BATS: 400 microseconds
• DirectEdge/29West: Persistent messaging 300-500 microseconds

Colocation

Colocation is the result of a quest for zero latency which has moved execution management servers from the customers physical location to a physical position within feet of the execution venue servers. The shorter the wire the shorter the amount of time it takes to move that message from Point A to Point B. Colocation is an answer for high use exchanges versus general purpose which may be better served by the use of a cloud or high speed T1.

NYSE Euronext is building the largest co-location facility. NYSE will use 20% for their own purposes with the rest being used by co-located customers.

The impact segment for this week is the January 4 SS&C announcement of their Tradeware acquisition. The connection reliant global trading community depends on solid connectivity and operational support for which Tradeware’s FIX is widely used thanks to it’s mostly plug ‘n play ability.

Benefits from being a Tradeware customer includes a limited trading system with rapid installation to most every customer, broker and execution venue in the connected world. SS&C contributes portfolio management, back-office processing, clearing, accounting and reporting enabling an alpha to clearing process.

Tradeware’s GlobalX enabled firms to reduce the costs of trading on global markets by pushing the clearing to the local broker excluding a US counterpart.

A US based customer seeking to trade a position of a financial instrument that is listed on a foreign exchange contacts their US broker. The US broker provides FX and their overseas
relationships to transact the requested business.

Direct clearing uses a broker of the exchange in the target country. Tradeware can provide that introduction when necessary. Tradeware’s GlobalX provides the mechanism for moving an order from the US customer to the local broker. When that order is executed the local broker provides the clearing and position keeping. Tradeware provides a clean shot at Europe, Asia, South America and Mexico.

A few words about transparency and liquidity

When a customer is seeking information about market liquidity, the US has a transparent structure even though hidden markets exist in reserve quantities and dark pools. Visibility is more limited for international exchanges and venues. Local brokers have the insight with regard to the truth depth of those markets.

In addition, fragmentation plays a larger role outside of the US. Market data vendors are attempting to fill the gap by providing a consolidated tape of executions. Exchanges, brokers and banks continue to form alliances to provide access to markets for those customers interested in cross-border trading.

Trading firms have many requirements. Some issues specifically related to trading are

• Pricing
• Risk Management
• Quoting
• Execution
• Clearing

There are four high-level items to consider

• Asset Class – What are you trading?
• Trading Venue – Where is it being traded?
• Connectivity – How can you get the order there?
• Clearing & Settlement – Who will handle the transfer of ownership?

General factors which should be considered when a trading firm is evaluating investment and trading opportunities on markets outside of their domestic market are:

• Liquidity – A market is said to be liquid, and the financial instrument marketable, if it can easily be bought or sold in quantity with little impact on market prices. Low liquidity refers to a thinly traded issue, while highly liquid refers to an issue which trades like water flows.
• Market Data – Is information about pricing and fundamentals of an issue that trades on an exchange, alternative trading venue (eg ECN’s – Electronic Communications Network, MLTF’s – Multi Lateral Trading Facility) or OTC – Over The Counter. May also be called a “Ticker Plant”.
• Accessibility – The manner in which an order or other inquiry can reach the trading venue. Examples are by phone to a broker or trader, through an agent or other authorized party, and a sponsoring broker system (aka DMA- Direct Market Access).
• Communications – Can be provided by the trading venue, through a vendor (eg Radianz), or other commercial communications provider.
• Latency – Describes the time it takes for a message to move from source to destination. “Low latency” refers to a minimal impact on transmission speed by the technology used (eg routers, optical fiber, copper wire) to move a message from point a to point b. Each hop a mesage takes adds to the latency calculation. “Colo” Co-location may be a consideration to reduce latency.
• Fees – Trading venues asses charges for services provided to market participants. The most widely used charge is a per share or ticket charge for transactions. Another model is called “maker-taker”. Exchanges may charge a market participant to publish execution information to the public (eg- last sale, volume) aka “Tape Revenue”.
• Clearing and Settlement – When a transaction is completed, aka trade or order executed, funds are required to pass from the buyer to the seller and securities are required to pass from the seller to the buyer. A bank or other financial institution, eg Agent Bank, upon receipt of the trade details, begins a transfer to the beneficial party.

Any single factor with unacceptable or unavailable capabilities could make working an order difficult enough that it is not worthwhile to pursue order execution.

The next item to consider are the assets being traded. There are many global exchanges and each may handle the same, similar or different products. A trading firm needs to evaluate the asset class targeted for trading then locate a trading venue to accomodate a transaction. There are many types of assets and the following are broad categories.

• Equities
• Options
• Futures
• Options on Futures
• Commodities
• Fixed Income
• Currencies
• Derivatives

After an asset is decided on, the next item to consider is who will be granting access to the exchange listings and the type of representation required to acquire or liquidate a position. An investment company is usually not permitted to trade directly with the exchange rather some form of representation is required. The medium could be either a sponsored electronic method or through a broker.

A conversation about trading on foreign markets can not be had without mentioning the manner in which assets are domiciled by the beneficial owner.

Trading on some markets have higher charges for foreign participants to settle trades in another currency. Something called a stamp tax can increase the carrying cost of a position to the point that the charge digs into the profit potential that a cost benefit analysis may restrict trading in that instrument.

One vehicle used by investors is a swap. A swap is a derivative that places ownership of an asset in the hands of the broker that is accepted by the country of trading origin. This broker may trade without additional penalities. Technically the beneficial owner, the investing firm is the owner of the asset but for the purpose of exchange records, the representing broker is the beneficial owner. This relationship avoids being charged the stamp tax. The swap explained here may have been replaced with an asset called a ‘Total Return Swap’ which can be used for longer term (a year or so) but requires a great deal of legal oversight. There are monthly resets of interest and principal.

Noting that there are many pre-requisites to acquire an asset, there is still the issue of clearing. Once an asset is acquired, the transaction needs to be passed through a clearing system so the asset is paid for or proceeds are credited to the beneficial owner. The US has a single point of eventual clearing, a CCP – a Central Counter-Party, which is the DTCC – The Depository Trust and Clearing Corporation. European and Asian markets settle transactions between their respective countries banks and financial institutions. European firms are coming together, albeit slowly, with a CCP as the EMCF – The European Multi-Lateral Clearing Facility.

Global trading obviously requires an understanding of more than simply a buy/sell transaction.

• Decide on an asset to acquire (accepting an exit strategy, liquidating the position)
• Determine a trading vehicle
• Execute a transaction
• Do a currency exchange
• Clear the trade through a generally accepted institution
• Settle the movement of the instrument
• Acquire a generally accepted price mark to post a value of the asset
• Decide on a liquidation time period
• Evaluate average daily trading volume
• Get out of the position

Many, if not all, hedge funds and prop shops are started by former traders or other investment professionals who employ their friends and business partners. They use a combination of strategies and equipment which they feel give them an edge over the market. They want their applications to utilize their collective knowledge but vendor supplied products lack deep customization.

Firms may use a combination of products in three categories.
1) Sponsoring broker relationships where the prime broker supplies a web-based or client application that allows for done away trade entry or DMA applications,
2) Sell-side OMS/EMS applications that are adapted to buy-side use (eg- TradingTechnologies, FlexTrade, Portware, Fidessa, Lava), and
3) Buy-side oriented execution venues (eg- Liquidnet, Millenium and other so-called dark pools)

Though vendor applications provide adequate functionality, most firms would like some level of consolidation, such as, some unifying middleware. When you build an application yourself, user requested incremental releases often produce “Oh, I forgot about this…” or “No, that’s not what I said…”.

The reasons for the middleware are
1) Mission critical applications need to be linked to provide a single point of risk management, pricing and quoting.
2) Middleware must be robust enough to grow or shrink with the firms requirements so valuable services are quickly connected and costly ones are terminated.
* API’s or some screen scraping may be required to manage the myriad of information sources for pricing, risk and trading.

Well-established and higher capitalized hedge funds and proprietary trading shops use internally developed applications and databases to
1) Minimize alpha leakage,
2) Maximize opportunity recognition, and
3) Customize functionality.

Products developed and maintained internally protect alpha. A faster turn-around time can be expected from an internal developer since they have a single customer versus a vendors multiple customers. An internal programmer is the best way of guaranteeing that the time spent on the firms projects is efficient and worthwhile. Firms with many programmers may dole out pieces to multiple programmers having a manger put the program together. That process provides additional protection.

Internal projects usually start with a trader or other revenue producer saying “I need …” for which a specific development task is arranged. Internal development can often run into time and cost overruns.

One way is to limit unexpected results is to survey systems usage by the current traders for current assets traded. This survey could be a high level roadmap which should include an analysis of workflows, priorities, feasibility and management.

While trading is necessary and internal traders are the best way of insuring adherence to a portfolio managers needs part of any analysis should include external execution venues. Planning should include a cost / benefit analysis to place a dollar amount on requirements. Requirements should be cross-referenced with the business model to produce a profitability metric. Essentially a cost/benefit calculation could produce a cost per function to determine the level of trading required to make that function part of the specifications. Comparing available products against the firms needs should result in a reasonable expectation.

A firm looking to decrease overhead considers costs, fees, charges and licensing that can be reduced or eliminated. Simple answers may be apparent but a solution requires digging into an analysis of how the business gets done.

An adequate survey need not take an extended period of time. The most direct method would be to retain an outside independent consultant to do an analysis that identifies inefficiencies affecting the bottom line. A solutions provider may then be employed to develop streamlined procedures or deploy improvements. Another group may also be necessary to install and train users. It is not recommended to circumvent experts. Experience holds knowledge in refined routines and major problems have more than likely been realized and vetted. A firm may also consider hiring a former consultant .

Talking points for a survey of a business follows.

Operational Discovery

Observe users

• Identify intellectual property resources. A limited number of users are instinctively aware of all activities and obscure required procedures. Has a list of problems been previously compiled?
• Include desks which perform peripheral tasks such as middle office reporting or back office accounting
• Identify power users and influential individuals and try to note personalities. Does a leader exist that knows who to involve in particular situations?

Identify process flows

• Record all activities and keystrokes from morning arrival to evening departure. Users may not be aware of all the clicks they make so it is important to review the functional steps following observation.
• Define the rationale for each process. One component of may expose the cause of bottlenecks. For example, A security master database lacks information for a newly traded instrument that delays trade entry affecting timely risk management.

Survey applications

• For each application service describe the business unit using it. Is there a need for proprietary middle-ware aggregation of external vendor products.
• Don’t forget to include ubiquitous applications. Many times, such applications are used to manually backup information or prove-out installed systems. Examples would be Outlook, Word, Excel.
• Identify information sources for applications in the process
  Market data, news, email, internal database queries.

Following information collection a discussion may be organized to cover work, process and system flows separately to see if there are extraneous or overlapping issues that present an opportunity for optimization.

Internal optimization is often difficult to implement. A major roadblock could be individuals reluctance to reveal procedures. An ability to solve a business process issue is a valuable commodity for a firm and the employee who is the “go-to” person may withhold information due to a fear of losing status. Employees compensation may be based on merit so any reduction in the limelight would impact the reward. By observing operations on consecutive days many if not all procedures would become visible.

External optimization is more of an issue in that vendors may be reluctant to expose the precise nature of their services for fear of exposing situations that create a concern for the customer who is then compelled to look for a more efficient provider. Many vendors appear to have unified solutions which under the covers have been cobbled together. Interoperability of desktop applications is a significant factor. A firm may find that a single vendor is not the solution and proprietary middle-ware would need to be constructed to fill selected gaps.

Optimization is problematic at best. A cost/ benefit analysis may reveal that an existing operation may not be efficient but the costs to reduce overhead may be greater than an effective optimization could produce. Though very simple, eliminating waste and redundant processes may be the best possible solution. Firms need to concentrate on strategic plans so that tactical solutions can become building blocks rather than patches.

There are formalized processes that define business analysis. Software manufacturers have capitalized on formalizing a process of interpretation that leaves little room for interpretation. A unique view of a subject becomes a footnote or at worst an anecdote. That leaves very little room for hybrid thoughts such as those that come from someone that has been exposed to and worked with the business side then had a change of focus to the development of technology that accentuated the anomolies of the business that creates wealth that others simply cannot think of since they are immersed in canned and generally accepted processes.

So a business analyst is simply one that analyzes the business. They don’t provide any value-add or solutions alpha. Stakeholders that have a difficult enough time convincing management to let them try somehting new run into additional blocks when those they are explaining their ideas to are trying to fit those ideas into categories that have been pre-defined when in actuality- the ideas themselves defy categorization. The spark literally loses it’s character in translation to some defined process.

It takes someone who has not just been exposed to the situations but has been immersed in the business to the point that their very survival depended upon a deeper level of understanding than some outside unknown entity. How often has it happened that a business person would spend the time with a developer or analyst that produces a product that has almost nothing to do with the intent of the stakeholders request. How does a project reach a successful conclusion when there is such a disconnect.

Management works very hard to instill a firm directive. They nonetheless hire business people who are adept at creating something out of nothing. They pay those individuals great sums of money to produce extraordinary profits. Yet management spends scarce little time in getting to understand that when a producer requests a process flow change or an application enhancement that the request is made only to make it a stronger possibility that greater returns will result.

Allocation of resources is not a guarantee of expected results, however- without the application of such resources the possibilities become more difficult to realize. What is needed is not only a business analysis that takes account of the nuts and bolts and the application of generally accepted principles. What is in fact required is a deeper understanding of the producers train of thought and motivating factors to give managment a better idea as to what may be needed. It is true that at times, producers will state they need more because they are failing to achieve expected results. That individual needs to be cut.

A player who blames the equipment is not a strong part of the team. A player that produces in spite of inefficiences and requests modifications that would allow for the possibility of increasing returns is the individual that needs to be catered to. In which case, introduction of a like-minded analyst, the addition of someone who has a more relative understanding of the business requirement, can serve to illustrate the need of a process improvement to both the programmers and management. The expectation that a solutions provider can contribute more than a business analyst should be at the heart of any project initiation. Business analysts may be required to define the nitty-gritty detail, but the solutions analyst is the person who represents the needs and desires of the stakeholder.

As it goes, I see a project team as consisting of
Stakeholders
Solutions Analyst
Business Analyst
Project Manager
Development Manager
Programmers
Operations people
Quality Assurance
User Acceptance Agents

Good enough rationale yields nominal results. Situational awareness yields extraordinary results. I get the job done.

So what happened here. We went from institutional BD’s keeping orders on their order books, working them internally using IOI’s and sales traders, taking along the retail (albeit at a higher commissions) to that same routine being called a “dark book” or “dark pool”. That rave then came into the light after high frequency (aka- short term alpha) trading was targeted as a villain. The label “Dark Pool” was then attached to many broker internal order books that aligned with competitors looking to complete orders to make their vig. And those pools registered as ATS’s (why?).

Now the regulators realize that brokers were yet, once again, colluding and manipulating markets (reminiscent of the late 80’s… again) so now they want to make it all fair and open, again. “Print those dark trades” is the call to action this time. Brokers have always, and will continue to, find ways of keeping their most prized asset- size institutional orders – as close to their vest as possible in order to both maximize a spread and other profit motives. In addition- the better a secret is kept the greater the likelihood that the customer will return to that very same broker for further business. Until such time as there is information leakage the customer will continue to use the broker that best protects their interests.

Whenever the regulators think they have a handle on the golden rule, the more ineffective they appear to be. They need to understand that size is a secret that will never be revealed or forced to be revealed. Look at HFT. That strategy is a result of size not wanting to be recognized so the solution was, as is, to break up that size into micro-particles that are only recognizable by the very machines that are used to execute the orders. Look at, IOI’s. That strategy is to use a special directed bid/offer to players that the originator knows, not thinks, would have an interest. Look at reserve quantities. That strategy is used to say- hey, look at me. I have something to do, but I’m going to tease you since once you get it, you’ll want more, and I have more. Keep coming to me. Oh yes. Once I realize you’re coming to me, I’m going to cancel my reserve because I wouldn’t want give away ice to eskimos in the winter.

And what is the solution. The $64,000 question. I’ll tell you a secret- there is no answer because there doesn’t need to be an answer. There are clearly two markets out there. The day trader looking to capitalize on what they think no one else knows, which everyone else already does know. And the professional trader- who has size to do, in an economies of scale kind of way, that is more than willing to take along the retail to effect what they will ultimately profit from.

I’m making this point because I think the regulators are throwing out the baby with the bath water and can better manage markets when they understand and account for the motivation of the players. I can identify five or six firms/systems that have capabilities which accommodate the needs of all the players involved. Unfortunately- the simplified term “transparency” does not take into account the one who only needs glasses to read.

HFT, IOI’s, Dark Pools, generic internal crosses with prints to accommodating venues, Algorithmic Trading … these are the buzzwords currently under attack. Next on the list? Maybe social networks for traders, StockTwits, IM’s, emails … I’m sorry- there will always be a mechanism that allows those in the know, to tell the people that compensate them for putting them in the know, that resistance is futile. Capitalism accepts a nominal charge or fee for making profits above and beyond those usually acceptable to those who don’t understand the system. Don’t place a punishing tax on prints obviously, or exposed as, institutional trades, but add those few extra pennies that allow the regulators to say that they observe and are taxing those in the know rather than catching up with them every so often.