Technical Analysis: Past, Present, and Future

Transcription

1 Technical Analysis: Past, Present, and Future Jasmina Hasanhodzic Boston University AAII Washington D.C. Meeting September 15, 2012

2 Status Quo Efficient markets Lefevre (1874) Bachelier (1900) Fama (1965) Samuelson (1965) Technical analysis Pt t Large gap between academics and practitioners

3 Broad Study of Technical Analysis [H. Lo 2003-present] Past Historical study: Place in context The Evolution of Technical Analysis, Lo H Present Interviews with practitioners: Understand what it is The Heretics of Finance, Lo H Future Science: Theory, standardize, extend Quantitative Approach to Technical Analysis, Lo H. to appear

4 Outline Past: History Present: Interviews Future: Science Theory Standardize Extend

5 From Technical Analysis 7th c. BC Babylon: Evidence from price diaries Intraday prices recorded when volatility is high [Slotsky 97] 17th c. Holland: Confusion de Confusiones [de la Vega] For on this point we are all alike; when the prices rise, we think they will run away from us. 18th c. Japan: The Fountain of Gold [Homma] When all are bearish, there is a cause for prices to rise.

6 to Behavioral Finance 19th c. China: Essential Business [Wang Bingyuan] When goods become extremely expensive, then they must become inexpensive again. 20th c. USA: The Wall Street Journal [Dow] It is a bull period as long as the average of one high point exceeds that of previous high points. 1955: A Behavioral Model of Rational Choice [Herbert Simon] Rational behavior compatible with computational capacities

7 Outline Past: History Present: Interviews Future: Science Theory Standardize Extend

8 In the Words of Masters On market inefficiency Raschke: Let's take the Renaissance Medallion Fund. What more proof do you need? Weinstein: I don t know of any successful traders who don t acknowledge that charts and trends are helpful.

9 In the Words of Masters On behavioral finance Acampora: That's the problem it's not with what we do, it's how we say it. Murphy: Academics are now basically copying what we do, renaming it, and trying to take credit for it.

10 In the Words of Masters On changing markets Dudack: There is a greater amount of noise in daily market action today, primarily generated by hedge-fund managers. We need to measure the market differently. Deemer: I am convinced that the Rydex funds reflect the hedge-fund trading activity which is the driving force in the market.

11 Interviews Topics: Beginnings, style, favorite patterns Historical value Variety of methods Elliott wave Weinstein s but ultimately converge to basics: patterns

12 Outline Past: History Present: Interviews Future: Science Theory Standardize Extend

13 Definition Technical analysis: Use of historical prices to predict the next price Past: Naked eye Future: Statistics

14 Theoretical Framework Bounded rationality: limited resources [Simon 55] Hard to make rigorous, but intuitive Efficient markets: price changes are random [Fama, Samuelson 65] Rigorous model, but counterintuitive Are stock returns really a coin flip?

15 What is Randomness? Which sequence is random? S1 = H,H,H,H,H,H,H,H,H,H S2 = T,H,T,H,H,T,H,T,T,T Paradox: Prob(S1)=Prob(S2)=1/210 Solution: Ask what looks, not is random => behavioral randomness

16 From Randomness to Finance Randomness Theory Finance Theory Classical (1812): expectations Classical ( ): efficient markets Behavioral (1960-now): computationally bounded algorithms Behavioral (1955-now): bounded rationality Future: Computational market efficiency [H. Lo Viola 11]

17 A Computational View of Market Efficiency [H. Lo Viola 11, Quantitative Finance] Our inspiration: agent-based models [Farmer 02, ] Our work: simple model of market evolution Market returns follow a pattern Memory-bounded agents predict next return sign Memory-2 agent predicts based on past 2 returns

18 Strategies Move the Market Returns predicted correctly => pushed closer to 0 Returns predicted incorrectly => pushed away from 0 Best strategy gets all correct except one, creates spike (new profit opportunity)

19 Outline Past: History Present: Interviews Future: Science Theory Standardize: Are charts useful? Extend

20 Our Motivation My IRA statement Are charts useful?

21 Debate What data to give investors, how to present it? [Bazerman 01, Kozup Howlett Pagano 08] [Consumer Financial Protection Bureau 11],... [Hung Heinberg Yoong 10]: Subjects pick funds based on summary statistics with or without charts Results ambiguous: charts affect choice, not outcomes

22 Our Work Focus on usefulness of charts Can humans tell returns from random permutation? Permutation preserves data moments (mean, var, ) If yes, humans can extract information from chart beyond summary statistics New, video-game experiment Allows for efficient collection of large amount of sound data

23 Our Motivation My IRA statement

24 Our Motivation My IRA statement Suppose I got this fake chart instead Would I be able to tell?

25 Anecdotal Evidence Humans cannot tell asset returns from random walk [Roberts '59, Malkiel '73, DeBondt '93,...] [Keogh Kasetty '03]: Which sequences are S&P 500? Different methods, objectives

26 Outline Past: History Present: Interviews Future: Science Theory Standardize: Video game to test if returns look random to humans Extend

27 Video Game ARORA [H. Lo Viola 12] Dataset A Dataset B Show 2 moving price charts, one real, the other random permutation of returns of real Permutation kills temporal order, keeps mean, var,... Players need to click on real

28 Which Chart Is Real? Dataset A WRONG! -10 Dataset B GOOD! +10 Task seems easier if playing the game dynamically Can learn from feedback on validity of guesses

29 Our Experiment 8 contests using data: NASDAQ, Russell 2000, USD Index, Gold Spot (tick) DJ Corp. Bond, DJIA, CAN/USD, Corn Index (daily) Each contest is charts 78 subjects, 8000 guesses

30 Our Results H0: humans cannot tell real from random charts Under H0, each guess is independent coin toss Strongly reject H0: 7 out of 8 p-values < 0.005

31 Biased Pool of Subjects? Dow Jones Corporate Bond Price Index Demographic group # subjects academic/other 9 7 Occup finance student 22 female 8 Sex male 30 high school/undergrad 17 Educ MS/PhD 21 >=30 14 Age <30 24 USA 25 Country other Entire sample p-value (%) Finance background does not help All datasets: 73% of guesses correct for finance experts vs. 72% for others

32 Future Work Which data properties did subjects exploit? Subject: When first viewing the two data sets, it is impossible to tell which is real, but a pattern quickly emerges & the eye can easily pick out the real array Is feedback critical? Human eye vs. computer Video games as trading platforms?

33 Outline Past: History Present: Interviews Future: Science Theory Standardize Extend

34 Extensions Technical indicators should evolve with markets Recall: The Rydex funds reflect hedge-fund activity which is the driving force in the market. (Deemer) New (first) indicators for hedge funds [H. Lo 07]

35 Hedge Funds Hedge funds are the driving force of the market Price to hedge-fund access: Secrecy, high fees, routine lock-ups Can hedge funds be captured passively? We create transparent, algorithmic portfolio with hedge-fund-like risk exposures => index, no alpha

36 Our Work [H. Lo 07, Journal of Investment Management] There are multiple betas each with its own factor: stocks, bonds, currencies, commodities, credit Express hedge-fund returns in terms of those betas Use a linear regression model Other work: [Kat Palaro 05, 06a,b] Goal is to replicate distribution, not returns

37 Our Model Estimate linear regression model Construct a hedge-fund clone Implement γ via futures and via short sales

38 Our Results Equal-weighted clones as indicator for hedge funds 2,700 hedge funds, 20 yrs of monthly data 14 Fund Clone SP Feb-86 Feb-88 Feb-90 Feb-92 Feb-94 Feb-96 Feb-98 Feb-00 Feb-02 Feb-04 Feb-06

39 Other New Indicators 130/30 assets at $50 billion and growing CS 130/30 Index, ProShares 130/30 ETF [H. Lo Patel 09, Cumulative Return Cumulative return CS 130/30 ETF XYZ 130/30 fund S&P Dynamic indexes are next generation of indicators

40 Conclusion Broad study of technical analysis [H. Lo 2003-present] Past: A force through history Present: Wisdom from the masters Future: Theory, standardize, extend

41 Thank you!