Take the 2-minute tour ×
Quantitative Finance Stack Exchange is a question and answer site for finance professionals and academics. It's 100% free, no registration required.

I am currently doing my research for my master thesis, which will clearly focus on the question of risk managment in algorithmic trading systems.

I have done research about this topic and found some valuable nuggets here:

  • Extreme Value Theory and Fat Tails in Equity Markets. Blake LeBaron and Ritirupa Samanta. May, 2004.

  • Algorithmic Trading and DMA

However, as I see, algorithmic trading is an extremely hidden topic. Therefore, I really would appreciate from you as financial professionals, a hint about papers about risk managment in hft/algorithmic trading/blackbox trading!

share|improve this question

3 Answers 3

up vote 7 down vote accepted

Indeed, algorithmic trading is a very hidden subject. Some people even say that working in the algorithmic trading sector feels quite lonely because nobody is willing to share secrets, ideas or innovations.

All I can help you with are some industry-specific terms which might speed up your search for relevant papers and information:

  • Risk of ruin tables
  • (Peak-to-valley) drawdown (maximum drawdown, duration of drawdown etc.)
  • Number of consecutive losses
  • Confidence intervals
  • Empirical distributions (for risk or P/L management)
  • Value-at-risk VaR (or likewise) measures (use a valid fat-tailed distribution)
  • Influence analysis
  • Slippage problems
  • Liquidity problems (especially in thinly traded securities or derivatives)
  • Model backtesting biases (e.g. survivorship bias)
  • Latency issues and competition (and different quotation times between exchanges)
  • Transaction costs
  • Order rate flow control
  • Order Flow Toxicity
  • Pre-trade (and post-trade) checks
  • Unexpected market conditions (unknown unknowns such as the financial crisis)
  • Flash crash recognition (and hedging or exploitation by aggresive counter strategies)
  • Outlier problems and faulty quotations (with HF tick data)
  • Quote stuffing
  • And the new one: queue jumping (but that's only for HFT; corresponding picture)

Try Google Scholar for these keywords. Good luck :)

(if anyone knows any other keywords, please append them to my answer)

share|improve this answer
3  
Do you work in this field? We're pretty loathe to have people outside the industry post total speculation about an important topic such as this. –  chrisaycock May 29 '13 at 15:17

You might want to check this:

http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2285407&download=yes

share|improve this answer
2  
Can you summarize the contents of that paper? –  chrisaycock Aug 2 '13 at 13:01
1  
Basically what it says is that order feeds and executions can get so fast (in the microseconds) that the risk monitoring (in the milliseconds) is no use, creating what he calls “shadow trading”. Regulations (like EMIR and MIFID) cannot expect to risk monitor this stuff. He thinks the exchanges ought to be responsible for risks, and has a look at how NYSE or CBOT could do that. I can’t really see how exchanges can be held responsible for the orders their client place, or cancel. But they can negotiate and agree guidelines with their clients. –  rupweb Mar 7 '14 at 16:12

Algorithmic Trading in general is no different from normal trading except all of the trading is automated. So it encompasses the same risk parameters that normal traders would.

When it comes to High Frequency Trading, the risk management checks would be at Strategy Level as well as "individual trade" level.There would be checks for sizes, values etc. However any trading is essentially a "risk management" exercise and main risk in HFT is "execution" risk.

For example, Indian Regulator SEBI has provided the following guidelines -http://www.sebi.gov.in/cms/sebi_data/attachdocs/1333109064175.pdf

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.