Risks of HFT
When machines trade with machines at light speed, things can break in spectacular ways.
1. The Flash Crash (May 6, 2010)
In 36 minutes, the US stock market crashed 9% and recovered. Trillions of dollars vanished and reappeared.
- Cause: A large sell order (from a Mutual Fund) triggered HFT algos. The HFTs, sensing chaos, simply "turned off" (pulled liquidity).
- Result: Liquidity evaporated. Stock prices fell to pennies (e.g., Accenture traded at $0.01).
- Lesson: HFT liquidity is "phantom liquidity". It disappears exactly when you need it.
2. Spoofing / Layering (Illegal)
- Concept: A trader places a massive Buy order (Say, 10,000 lots) at a slightly lower price to create the illusion of demand.
- Trap: Other algos see this "support" and start buying.
- Execution: The spoofer cancels the Buy order instantly and sells into the rising price.
- Status: Highly illegal (Navinder Sarao was jailed for this).
3. Runaway Algos
Coding errors ("fat finger" or infinite loops) can destroy a firm in minutes.
- Knight Capital (2012): A deployment error caused an algo to buy high and sell low repeatedly.
- Loss: $440 Million in 45 minutes. The firm went bankrupt.
Note
Circuit Breakers: Exchanges now have "Kill Switches" and volatility halts (LUDP - Limit Up Limit Down) to stop trading if prices move too fast, giving humans time to think.
Loading quiz…