RETAIL TRADERS CAN OUTSMART HFT ALGOS
By Rich Kambak
“It’s invisible to the naked eye, but it’s there; a string of binary code submerging us in a sea of information that isn’t necessarily pertinent to what we need to know.” Dmitry Sklyarov
Financial research firm, Dalbar, released a new study that its empirical data validated that buy-and-hold stock fund portfolio strategies earn on average a 3.7% annual return versus the S&P 500 Index at 11.1%.
What keeps the annual returns so low? Other than the market-takers model, highly acclaimed author, Michael Lewis, has published a top shelf seminal trilogy for generations to learn from; hopefully to take heed; augmented by his last installment Flashboys.” Lewis penned, “Liars Poker” and “The Big Short”.
In what we assume is a reformed and repentant financial industry that nearly bankrupt the world, this “you can’t make it up” story exposes another deliberate lack of fiduciary transparency toward public investors; focusing on the avarice of High Frequency Trading.
HFT – THE NEW FRONT RUNNING
“Flashboys” revealing story of HFT corruption is convincingly enough to have sparked an investigation by the Securities Exchange Commission and the Financial Regulatory Authority (FINRA) into the discount brokerage’s routing trade executions practices; similar to “front running” – being “price driven” exploiting profitability by having a leg up on price formation before retail traders can see it. This equates into being non-compliant to the SEC “best practices” regulations.
HFT was introduced in 2006. In 2008, HFT reportedly made US$21 billion dollars in profits, considered “surprisingly modest” when compared to the overall market. The Commodity Futures Trading Commission (CFTC) formed a special group comprised of industry and academic experts to advise CFTC on how to best define HFT. They have good reason to be concerned, HFT controls more than 84% of all US market exchanges trades according to the Financial Times.
HFT uses anonymous trading algorithms (“Ambush”, “Raider” and “Guerrilla” to name a few), while spawning opaque “dark pools” niches ( an ambiguous amount of non-proprietary orders as their strategy for risk aversion) for banks and brokerage firms in routing order prices ahead of retail traders. Beating out the price posted latency is the leverage. One could equate it to comparing the efficiency in “speed”; a Cisco circuit or one made by Juniper. Trade execution speed in say, 465 microseconds – which is rather slow- could be dramatically affected by either a Cisco circuitry versus a Juniper circuitry.
The protagonist Brad Katsuyama in “Flashboys” – an employee at the trader’s desk for the Royal Bank of Canada – turns whistleblower on HFT when he executes a “buy” trade on 10M available shares in Apple at a specific price, that price shifts upwards in less time than it takes to blink our eyelid.
MARKET MAKERS FUNCTION
Market makers, the counter-party to a retail trader’s trade, typically absorbs the fluctuations-dissipation of excess demands on share contracts during intraday trading, lowering the price when they to buy and raising the price when they have to sell to maintain a “risk neutral” equilibrium. Given the influx of individual orders, limit orders, bank and institutional orders and accumulated inventory, a market maker has to be artfully adroit in managing mega amounts of contracts at any given time when clearing orders; all in competition to drive up or drive down an asset’s price.
Price formation, when exercised in the cause-and-effect of a “best price” fill price transaction, the new price is based on the net orders, not the accumulated “unit in shares” inventory. This inventory is left to the discretion of the market maker to distribute by the time the closing bell is rung on Friday afternoon.
THE FAT FINGER TRADER AND HFT FAT CITY SCENARIO
Bring in HFT algorithms and suddenly you have a random stochastic differential deviation ; the log-linear equation is skewed. The Sharpe Ratio price action becomes either risk averse, driving investors away, or misleading up-tick on the return on capital (ROC) that herds stock pickers into a gullible “traponomics” game theory model. The market maker is faced with a net order price imbalance. The solution is a default choice; access a pre-determined routed market index exchange pathway that, incidentally, is plugged into HFT algorithms that most likely executes a flash trade.”
The common belief on Wall Street is that HFT improves market liquidity by narrowing the bid-ask offer, minimizes volatility which in turn makes investing less expensive for market participants. Yet the liquidity is manipulated and leveraged against the naïve investor whose strategy is typically the buy-and-hold model. Reducing the cost of commissions entices the public to become self-directed investors, not understanding what is actually in play; HFT is “painting the tape”.
When the retail trader puts their fat finger on the computer keyboard’s Enter button – they most likely will experience the microcosm of a flash crash on their investment. In return, the HFT profit-squatter is in Fat City.
Nanex, LLC did a 4-year study comparing quote traffic to the value of trade transactions. The empirical data showed that HFT causes a 10-fold decrease in market efficiency.
PURE PLAY OFFENSE FOR RETAIL TRADERS: SHORT PUTS
The trader puts the burden on the counter-party to carry the trade. If one trades small, risk-aversion is minimized. This blossoming strategy is proving to have an impact on Wall Street. Proprietary Traders (PPT) who are employed by banks as internal traders, have in fact complained about having to tweak their models due to increasing losses. The blame falls squarely on retail traders who predominately lean into credit option spread trades. HFT algorithms are not programmed for this trading anomaly.
CRACKING THE HFT CODE
Similar to Katsuyama’s revelation in “Flashboys” with nano second price changes; we have seen this scenario more often then not in the past year. Comparable to the thinkorswim (TOS) Market Maker Move (MMM) indicator, we built our own signal processing function based on statistical arbitrage. Unlike the TOS MMM that is “reversed engineered” to estimate potential price movement based on the underlying instrument, ours has parameters for normal distributions of price action and volatility which are equated into a “lambda” formula. The results so far have been robust.
QCOM AND AFTER MARKET EARNINGS REPORT
On April 23rd our model signaled a BUY PUT on Qualcomm, Inc (QCOM), a technology communications equipment manufacturer on the day it was to release its earnings after the market closed.
Here’s the chart:
Our model’s signal was to execute a short the option spread – at the strike price approximately 15-minutes prior to the closing bell. We posted a text note on the chart and executed a paper trade.
The earnings report outcome: Est. $1.08, Actual $1.31. A positive earnings report typically will cause the price to increase and/or make a downward spike; returning back at or near the closing price, prior to next day’s opening. That’s not the case here. The price spike drop – over $6.
We ran QCOM through Macroaxis’ engine just to get another opinion based on Modern Portfolio Theory indicators. Back testing QCOM for the past 30 days shows a +1.83% in investment gains. QCOM has excellent predictability; meaning that the serial correlation of 0.9 indicates that approximately 90.0% of current QCOM price fluctuation can be explained by its past prices.
Looking forward QCOM has a 58.0% propensity to move below its current price over the next 30 day time horizon. One can see this evidenced by the graph posted below.
The Macroaxis bell curve shows that QCOM’s price has peaked and is now on the downward slope of its price density calibration. The Alpha is 0.02; Beta is 0.76; Volatility is 1.14 and the Information Ratio is (0.0035). QCOM Value at Risk (VaR) is (3.52) and the recommendation is to “Hold”.
AND BY THE WAY…
E-Trade (ETFC), TD Ameritrade (AMTD), Charles Schwab (SCHW) and Interactive Brokers (IBKR) make up the SEC and FINRA’s A-list for subpoenas.