Money:Tech 2008 David Leinweber, dleinweber@post.harvard.edu Why am I here? If you h a d e v e rythin g c o m put a tion a lly... … w h e r e w ould you put it, fin a n c i a lly ? / ITG QuantEx • Quantex Algos • Internet Information • Jefferies Acquisition, Service, Founder David Leinweber ITG Spinoff • Integrating textual Haas Fellow in Finance, information in trading Haas School of Business strategies U.C. Berkeley • Caltech, post bubble djl@haas.berkeley.edu dleinweber@post.harvard.edu • MD for Equities • Institutional Buy side • $6 Billion • Center for Innovative • 6 countries, 27 quant Financial Technology strategies, Long & MN c. 2002-2008, D. Leinweber Talk at O’Reilly Money:Tech, Feb. 6, 2008 If you had everything, computationally, where would you put it, financially? Summary of JPM Article • Looking Back: Greatest Financial Technology Hits – Electronic Market Data A Short History of – Computerized Data Storage and Analysis Market Information – Electronic Execution Technology • Looking Forward: Past as Prelude – Algos to the Nth Power – Intelligence Amplification and Visualization – Finding Alpha in Textual and Internet Information Journal of Portfolio Management Fall 2005, pp 61-75 c. 2002-2008, D. Leinweber 1
Money:Tech 2008 David Leinweber, dleinweber@post.harvard.edu NYSE moves indoors. Tontine Coffee House. 1794 NYSE in 1792. The Buttonwood Tree Traders strap telegraph keys to their arms. 19 th century Blackberry. c. 2002-2008, D. Leinweber 2
Money:Tech 2008 David Leinweber, dleinweber@post.harvard.edu Telegraph wires at the NYSE, 1888. NYSE Before Telegraphy, c. 1865 Edison’s stock ticker 19 th Century Information Overload eliminates the need to decode dots and dashes. 1870. Market Data Archive c. 1950s NYSE Quote Board. 1930-40s c. 2002-2008, D. Leinweber 3
Money:Tech 2008 David Leinweber, dleinweber@post.harvard.edu NYSE Floor - 1963 Telephone Order Overload, Dealing Room, c.1950, Reuters NYSE Photo Beyond Ticker Tape Disintermediation of Execution Exchange President Keith Funston with first NYSE computer. 1966 An Internet-centric trade journal buried the traditional exchange in a 1999 issue. Electronic Disintermediation The exchanges are Boot your broker of Execution still here. Changing, but here. The Industry Standard isn’t. (Aug. 16, 1999) E*trade, 1999 c. 2002-2008, D. Leinweber 4
Money:Tech 2008 David Leinweber, dleinweber@post.harvard.edu London Stock Exchange. Big Bang minus 1. Oct 26, 1986 London Stock Exchange. Big Bang. Oct 27, 1986 Early adopters of quantitative trading systems Institutional Investor Alpha, February 2007 Renaissance Technologies Fischer Black: Options Maven and Founded 1982 and still not talking Pioneer Algo Trader "Markets look much more efficient from the banks of the Charles than from the banks of the Hudson." c. 2002-2008, D. Leinweber 5
Money:Tech 2008 David Leinweber, dleinweber@post.harvard.edu Information Advantage: 1980s-90s Information Advantage: 1990s + Algorithmic Market Making & Trading Algorithmic Market Making Fortune, “Playing NASDAQ like a piano” – Dave Whitcomb “UNIX and Market Data Feeds” ~ 1986 Feb. 5, 1996 What’s Next? Where to look • Exploitation of electronically delivered for new quantitative information was, and remains, a great success. information • Prices and market data convey a great deal advantages? of information. But they do not convey all information. Where does alpha come from? Information & Innovation “…profits may be viewed as the economic rents which accrue to [the] competitive advantage of… superior information, superior technology, financial innovation…” Andrew W. Lo, MIT Sloan School Editor. “Market Efficiency: Stock Market Behavior in Theory and Practice”, Elgar, 1997, c. 2002-2008, D. Leinweber 6
Money:Tech 2008 David Leinweber, dleinweber@post.harvard.edu Firehose of Information With Potential Alpha • Electronic versions of mainstream news – Minutely instead of Monthly – “Print”, Web, Broadcast captions, feeds. – Global news on global companies. Time zone advantage. • Press releases, disintermediated access – PRNewswire, BusinessWire – Specialized sector news • Electronic access to “official sources” – SEC, Courts, NIH, federal & state agencies – Management conference calls • New Media – Creatures of the net – Websites, mail, messages, chat, blogs, RSS… – “Social Media” “Model of the Internet” – Notre Dame Computer Science Dept. Investing, Trading & the Internet David Leinweber Time isn’t what it used to be. c. 2007 David Leinweber Earnings Surprises Earnings Surprises Before WWW (1983-1989) ...After WWW. (1995-1998) 1.12 1.12 1.08 1.08 Double Plus 1.04 1.04 1.00 1.00 Universe 0.96 0.96 Reaction Time: Reaction Time: Weeks Minutes to Hours 0.92 0.92 Double Minus 0.88 0.88 -80 -60 -40 -20 0 20 40 60 80 -80 -60 -40 -20 0 20 40 60 80 Number of Days (Earnings Report = 0) Number of Days (Earnings Report = 0) Source: R. Butman, DAIS Group Source: R. Butman, DAIS Group (8308-8912) and (9509-9802) (8308-8912) c. 2002-2008, D. Leinweber 7
Money:Tech 2008 David Leinweber, dleinweber@post.harvard.edu Democratization The Text Frontier: and Finding Alpha on the Web Disintermediation of Information Does news Ace Reporters? move Who needs ‘em! markets? Do it Yourself Disintermediated News Off the shelf algo news products • 2007 Product Announcements – Dow Jones – Reuters • Acquisitions & Strategic Partnerships – Clearforest/Reuters – Corpora – Ravenpack • Not cheap at all – ~ $100K per month • Can one size fit all – Does it get arbed away rapidly? c. 2002-2008, D. Leinweber 8
Money:Tech 2008 David Leinweber, dleinweber@post.harvard.edu eAnalyst Case Study News , Language Models, One day horizon UMass Working Paper, 2001 Source: Victor Lavrenko Source: Victor Lavrenko Ænalyst: Overview Language Models on Stock News Extracting Trends • Replace series by a sequence of regression lines • top-down procedure (slide a window, split, recurse) • automatic stopping criterion based on T-test surge • Estimate a language model for each trend flat • use trends instead of raw data plunge • align trends with concurrent news stories • Given a new document, predict which trend will most likely occur next Source: Victor Lavrenko eAnalyst Evaluation: Language Modeling Trading Simulation Software giant Microsoft saw its • Cumulative earnings: $21,000 Microsoft (MSFT) stock Software giant Microsoft saw its shares dip a few percentage points shares dip a few percentage points this morning after U.S. District this morning after U.S. District • 40 day simulation, out of sample Judge Thomas Penfield Jackson Judge Thomas Penfield Jackson issued his "findings of fact" in the issued his "findings of fact" in the • 100 stocks, Jan 1, 2000 government's ongoing antitrust case government's ongoing antitrust case against the Seattle wealth-creation against the Seattle wealth-creation • purchased or shorted $10,000 with each trade machine... machine... • Result significant at 1% level News: • determined through a randomization test P ( shares ) = 0.074 P ( shares | MSFT ! ) = 0.071 P ( shares | MSFT ! ) = 0.071 P ( shares ) = 0.074 Words like Jackson P ( antitrust ) = 0.009 P ( antitrust | MSFT ! ) = 0.044 P ( antitrust ) = 0.009 P ( antitrust | MSFT ! ) = 0.044 and antitrust are more P ( judge ) = 0.006 P ( judge | MSFT ! ) = 0.039 • Biggest Gainers P ( judge ) = 0.006 P ( judge | MSFT ! ) = 0.039 • Biggest Losers P ( trading | MSFT ! ) = 0.029 likely in the stories P ( trading ) = 0.032 P ( trading | MSFT ! ) = 0.029 P ( trading ) = 0.032 P ( against | MSFT ! ) = 0.027 P ( against ) = 0.025 P ( against | MSFT ! ) = 0.027 • IBM: $47,000 preceding the plunge. P ( against ) = 0.025 • Disney: -$53,000 P ( Jackson ) = 0.001 P ( Jackson | MSFT ! ) = 0.025 P ( Jackson ) = 0.001 P ( Jackson | MSFT ! ) = 0.025 • Lucent: $20,000 • AOL: -$18,000 P ( MSFT ! | Jackson ) = P ( Jackson | MSFT ! ) P ( MSFT ! ) / P ( Jackson ) P ( MSFT ! | Jackson ) = P ( Jackson | MSFT ! ) P ( MSFT ! ) / P ( Jackson ) Source: Victor Lavrenko c. 2002-2008, D. Leinweber 9
Money:Tech 2008 David Leinweber, dleinweber@post.harvard.edu News Sentiment (Tetlock) More Than Words: Quantifying Language (in News) to Measure • WSJ & DJNS Stories, S&P500, 1984-2004 • General Inquirer for sentiment Firms’ Fundamentals – Academic system, psychology/linguistics Tetlock, et al, UT Austin, Sep. 2006 – Developed over 20+ years • US NSF and Australian Research Council funds • Originally, PL/I on IBM mainframes News , Sentiment, • Try it at www.wjh.harvard.edu/~inquirer/ One day horizon • PSTV and NGTV word scores • One day LS trading simulation Top of General Inquirer NGTV Top of General Inquirer PSTV Distribution of Annual Returns to Tetlock LS Trading Simulation (1984-2004) Tetlock News Event Study (1984-2004 ) Note Huge Pre-event Information Leakage! (%) Note: Before trading costs c. 2002-2008, D. Leinweber 10
Recommend
More recommend
