Via Nanex [7],
Exploratory Trading in the eMini
On November 12, 2012, Adam D. Clark-Joseph published Exploratory Trading [8], which analyzes CFTC audit level trading data in the eMini S&P 500 futures market. This is a special, "regulators-only" data-set that contains all orders and trades, and each order and trade has a trader identifier. What this paper exposes is astounding. The following is our summary of Clark-Joseph's paper.
Exploratory trading
Exploratory trading is a form of manipulation designed to test the market's reaction to a trade. Probing for stop orders would be one form of exploratory trading. This paper specifically investigates exploratory trading that attempts to determine whether the bid/ask spread is about to shift up or down a level. The impact on the market would be an increase in intraday volatility. Exploratory trading distorts the market's view of supply and demand and induces trading activity from other participants. Furthermore, as participants learn of the strategy, they will employ counter-measures - which will further muddy an accurate picture of supply and demand for everyone else. This is why regulations ban manipulation.
The Top 8 HFTs Remove Liquidity 59% of the Time
Passive market making involves buying at the bid, and selling at the ask, which earns the market maker the bid/ask spread. Passive market making provides liquidity, narrows spreads, and lowers trading costs. Aggressive trading removes liquidity: buying at the ask (removes sell orders) and selling at the bid (removes buy orders).
Between September 17, 2010 and November 1, 2010 in the eMini futures contract (December 2010 contract, symbol ESZ0):
- 41,778 accounts traded this contract
- 30 of these accounts (less than 1/10th of 1%) met criteria to be classified as HFT.
These 30 HFT accounts:
- participated in 46.7% of total trading volume.
- grossed $1.51 million per trading day.
Of these 30 HFT, the top 8:
- were aggressive 59.2% by volume (the other 22 were aggressive 35.9% by volume).
- grossed $793,342 per trading day.
Nanex Discussion
The top HFTs probe the market by aggressively pinging order books and then analyzing market reaction: a practice that allows them to get a private glimpse of the "true" supply and demand at the expense of everyone else. Once the market direction is ascertained, these HFT aggressively remove liquidity, causing an immediate market move. Since the eMini is heavily arbitraged by SPY (which in turn is arbitraged by its many components and options), these sudden moves in the eMini will set off waves of overwhelming message traffic [9] as traders and algos react and reprice thousands of instruments in milliseconds.
In light of our discovery [10] that Waddell and Reed's trades in the eMini on May 6, 2010 were entirely passive (0% aggressive), we wonder if this probing by HFTs may have set in motion the downward spiral on that day, resulting in the Flash Crash. These HFTs not only manipulated markets on that day in a disastrous way, they drove liquidity providers away from the market.
A lot of media discussion about HFT focuses on 3 benefits: they provide liquidity, narrow spreads and lower trading costs. This Harvard paper exposes some disturbing truths: the top HFT engage in a predatory market manipulation strategy that removes liquidity 59.2% of the time (by volume), causes undue intraday volatility [11] (which amounts to a tax on investors), warps the true picture of supply and demand, and raises trading costs for everyone processing market data [12].
Perhaps even more disturbing was the Bloomberg article [13] where we first learned of this paper. It appears that rather than investigate HFT manipulating the markets, the regulator is investigating academic access to their audit level data-set.
Next time the media writes about the benefits of HFT - ask them if they've read Adam Clark-Joseph's paper on Exploratory Trading [8].