„Moore’s Law versus Murphy’s Law: Algorithmic Trading and Its Discontents“ Kirilenko, Andrei A. and Andrew W. Lo Flashcards
What is the main idea?
Review of the emergence of Algorithmic trading and the associated risks
What do Moore’s Law and Murphy’s Law state?
Moore’s Law – Computational progress (computers) will become faster and more efficient over time.
Murphy Law – Whatever can go wrong will go wrong; whatever can go wrong will go wrong faster and bigger when computers are involved.
What is Algorithmic Trading (AT)?
Computers trade (use of mathematical models, computers to automate the buying and selling of financial securities).
Improves efficiency by lowering costs, reducing human error, increasing productivity.
What are the benefits of AT?
◦ Lower costs / Scalability
◦ Reducing human error
◦ Increasing productivity
Which 3 developments have started the rise of AT? (over the last 2 decades)
o Financial system is becoming more complex over time
o Breakthroughs in quantitative modelling of financial markets (formulas)
o Breakthroughs in computer technology
Which 5 major developments that have fuelled the popularity of Algorithmic Trading?
- Quantitative Finance (Portfolio optimization, CAPM, Black & Scholes)
- Index Funds (Passive investing)
- Arbitrage Trading (uses algorithms to identity arb. opportunities, “Statistical arbitrage strategies” becoming more popular)
- Automated Execution and Market Making
- High-Frequency Trading (Automated trading of millions of transactions per day, very profitable, minimal risk)
What is the recipe for an index fund?
Choose securities, weigh by market cap, add or subtract securities when needed; afterwards weights are self-adjusting
How has the meaning of passive investing changed?
Before, most investors and managers equated “passive” investing with low-cost, static, value-weighted portfolios.
With the many technological innovations, it now refers to trading on a well-defined and transparent algorithm, which could require active trading → decoupling of active trading and active investing; demand for algorithms that execute passive-active strategies.
What are the benefits of arbitrage strategy to the market?
Arbitrage strategies improve:
Liquidity (arbitragers increase trading activity -> bigger liquidity)
Informational efficiency in the stock price (arbitragers adjust the mispricing, all info available determines the price)
What is Automated Execution and Market Making?
“Execution strategy” – how to break up a large trade into smaller ones over time to reduce cost (demand and price increase when you buy a big chunk at once)
Can be automated by computers
Market making – intermediary participates in buying/selling securities to smooth out temporary imbalances in supply & demand
Market makers make profit from the spread that they ask as a commission
Autoquoting – automatically giving better trades for larger trade sizes (like bulk discounts)
Name the 5 major incidents when AT went wrong.
August 2007: Arbitrage Gone Wild
May 6, 2010: The Perfect Financial Storm – Flash Crash
March and May 2012: Pricing Initial Public Offerings in the Digital Age
August 2012: Trading Errors at the Speed of Light
September 2012: High-Frequency Manipulation
What went wrong in August 2007: Arbitrage Gone Wild?
Large hedge fund companies lost a whole lotta money (Mostly statistical arbitrage funds).
Unwind Hypotheses: Forced liquidation of 1 or more large equity market-neutral portfolios (to raise cash/reduce leverage) -> Reduction in prices -> Similar portfolios experienced losses -> They also liquidated/deleveraged their positions -> …
What went wrong in May 6, 2010: The Perfect Financial Storm – Flash Crash?
33 minutes of crazy volatility and high volume.
Apple traded at 100 000$ per share.
Reason:
Automated execution algorithm on autopilot;
A game of “hot potato” among high-frequency traders;
Cross-market arbitrage trading, and a practice by market makers to keep placeholder bid-offer “stub quotes”.
What went wrong in March and May 2012: Pricing Initial Public Offerings in the Digital Age?
NASDAQ’s glitch delayed Facebook’s IPO 30 min (too much traffic). It was an infinite loop of recalculating the price.
In March 2012 BATS announced their IPO on their own exchange website, but it failed due to a software bug.
Both the Facebook glitch and the BATS fiasco can be explained as regrettable software errors that extensive testing failed to catch.
What went wrong in August 2012: Trading Errors at the Speed of Light?
Knight Capital Group started sending incorrect/random stock orders to NY Stock Exchange due to a software glitch.
Trades went through, and they were forced to sell at a half a billion loss.
Internalization - broker-dealers like Knight are permitted to post prices that are fractions of a penny better than prevailing quotes (can use 100.011 instead of 100.01 bid and 100.019 instead of 100.02 for ask). The dealer pockets the penny difference, and buyer gets price-preference in the queue.
