System and method for analyzing data associated with statistical arbitrage

• US 8,688,558 B2
• Filed: 04/04/2011
• Issued: 04/01/2014
• Est. Priority Date: 10/30/2006
• Status: Active Grant

• Current Assignee: Credit Suisse Securities USA LLC (Credit Suisse)
• Original Assignee: Credit Suisse Securities USA LLC (Credit Suisse)
• Inventors: Chadwick, Stephen
• Primary Examiner(s): Kalinowski, Alexander
• Assistant Examiner(s): Vyas, Abhishek
• Application Number:
  US13/079,430
• Publication Number:
  US 20110178958A1
• Time in Patent Office: 1,093 Days
• Field of Search: 705/35, 705/36.R, 705/37
• US Class Current: 705/36.R
• CPC Class Codes: G06Q 10/00 Administration; Management
• CPC Class Codes: G06Q 40/00 Finance; Insurance; Tax strategies; Processing of corporate or income taxes
• CPC Class Codes: G06Q 40/04 Trading; Exchange, e.g. stocks, commodities, derivatives or currency exchange
• CPC Class Codes: G06Q 40/06 Asset management; Financial planning or analysis

Abstract

Providing computer-based systems and methods for analyzing historical performance of financial securities and identifying trades in those securities based on the securities'"'"' current position as compared to this historical performance. These computer-based systems and computer-implemented methods include identifying stock pairs to include in a trading portfolio, based on a measure of the pair'"'"'s relative performance, such as a modified Sharpe Ratio. The value of the stocks in each stock pair in the portfolio is assessed and deviations determined. This assessment can occur daily or at a longer or shorter time step. Stocks are bought or sold based on the current price of the stock as compared to historical performance. The present invention preferably employs a large number of stock pairs in the trading portfolio. This use of a large number of pairs results in a plurality of stocks being in more than one stock pair.

Claims

1. A method for identifying trades for a statistical arbitrage program, comprising:

determining, by at least one processor, a first relative performance of a stock pair at a current time;

determining, by the at least one processor, a historical time during which the stock pair reverts to an expected relative performance according to a cyclic performance between the stock pair;

determining, by the at least one processor, a second relative performance of the stock pair based on the historical time;

calculating, by the at least one processor, a deviation in the first relative performance and the second relative performance;

repeating the determining and the calculating for each stock pair of a plurality of stock pairs;

calculating a volume of shares to trade based on the calculated deviation; and

initiating a number of trades of a stock comprising one of the stocks of the plurality of stock pairs based on the calculated volume of shares and one or more of a weighting factor and a deviation curve function.

2. The method of claim 1,

wherein calculating the volume of shares to trade further comprises calculating according to a scaling factor.

3. The method of claim 1, wherein the historical time comprises a range of time in the past.

7. The method of claim 1, further comprising

identifying the plurality of stock pairs among a set of stocks;

and

selecting the stock pair among the plurality of stock pairs according to a price correlation of the stocks in the stock pair being greater than a correlation value.

8. The method of claim 1, further comprising calculating a performance measure of a trade of the stock pair according to a ratio of an expected return and a volatility of a trade of the stock pair.

4. A method for identifying trades for a statistical arbitrage program, comprising, for each stock pair in a portfolio:

calculating, by at least one processor, a natural logarithm of a price ratio of a first stock of the stock pair to a second stock of the stock pair at a current time;

determining, by the at least one processor, an average of a natural logarithm of a price ratio of the first stock of the stock pair to the second stock of the stock pair over a time period prior to the current time;

determining, by the at least one processor, a difference between the natural logarithm at the current time and the average of the natural logarithm over the time period prior to the current time;

calculating a volume of shares to trade based on the difference;

designating, by the at least one processor, one of the stocks in the stock pair as a stock to buy and the other stock in the stock pair as a stock to sell based on the difference; and

initiating one or more trades based on the designation of the stocks to buy and sell and the volume of shares to trade.

5. The method of claim 4, wherein the designating further comprises designating the first stock of the stock pair as the stock to sell and the second stock of the stock pair as the stock to buy if the difference comprises a positive value.

6. The method of claim 4, wherein the designating further comprises designating the second stock of the stock pair as the stock to sell and the first stock of the stock pair as the stock to buy if the difference comprises a negative value.

9. A method for identifying trades for a statistical arbitrage program, comprising:

determining, by at least one processor, a first relative performance of a stock pair at a current time;

determining, by the at least one processor, a historical time during which the stock pair reverts to an expected relative performance according to a cyclic performance between the stock pair;

determining, by the at least one processor, a second relative performance of the stock pair based on the historical time;

calculating, by the at least one processor, a deviation in the first relative performance and the second relative performance;

calculating a volume of shares to trade based on the calculated deviation; and

initiating a number of trades comprising a stock of the stock pair based on the calculated volume of shares and one or more of a weighting factor and a deviation curve function.

10. The method of claim 9, further comprising

calculating the volume of shares to trade according a scaling factor based on the calculated deviation.