A New View of Financial Risk: Watch Out for the Outliers. Raymond YUEN

Transcription

1 A New View of Financial Risk: Watch Out for the Outliers Raymond YUEN

2 Content Development of modern financial theory Assumptions of modern financial theory Anomalies Implications for leverage and risk control A new view of risk: The Fractal A new view of share price movement: The Power Law Research on the theory of predicting crashes for the Hong Kong stock market Other looming signs of crashes

3 Development of Modern Financial Theory Random walks and price movement independency 1900 Efficient market frontier and capital market line 1955 CAPM 1964 Efficient Market Hypothesis 1970 Arbitrage Pricing Theory 1976

4 Assumptions of modern financial theory Investors know what they are investing No investors can dominate the market Price moves continuously Price moves randomly The market knows all the news about investment and reflects in the price immediately

5 Investors know what they are investing Know what they are investing Are you sure the price of HSBC should be HK$80s (now) or HK$130 in 2005 or HK$38 in 2009? Challenged by behavioral finance: we do not know what should be the price tomorrow due to the uncertainty of the future Figureheads of behavioral finance research: Paper: Tversky, A. & Kahneman, D., Judgment under uncertainty: Heuristics and biases Prospect Theory: describe way behave under uncertainty

6 Investors know what they are investing: Prospect Theory Describe a human thinking about alternatives involving uncertainty Emphasis gain or loss Based on rule of thumb: not vigorous and scientific Reference point, i.e. the price that we bought the stocks and the subsequent price change affect us most We hate loss more than enjoy a gain Flip coin experience Investors could be self-contradicting: bear risk and avoid risk at the same time Drunk drive and buy car insurance at the same time

7 Investors know what they are investing We want to maximize our gain Bond market > stock market: safety of capital instead of gain Investment mandate: pension funds, endowment funds, government funds: protection of capital, then return Size of bond market: US$100 trillion Size of stock market: US$36 trillion

8 Nobody can dominate the market Perfect competition : price discovery is efficient Is it true again? Difference in wealth is disregarded, However, the actual world tell us other stories A) 大户?Familiar term. B) Gini coefficient tables of the PRC, Hong Kong, Switzerland, the USA

9 Nobody can dominate the market Gini Coefficient measures the unequal distribution of income 0 total equality, 1 total inequality are a normal figure None of the country with Gini coef = 0 Country Year Gini Cofficient USA % PRC % Hong Kong % Switzerland %

10 Price moves continuously A mis-belief is that the market moves smoothly. The fact: big movements in price are a lot. Movement of Hang Seng Index : 31 Dec 1986 to 5 Feb 2013 Mean of HSI change 3.17 A number of changes> times Big Movements Maximum Jan 2008 Minimum Jan 2008

11 Price moves randomly Random: price move independently Today s movement cannot predict tomorrow s movement Truth Price movement exhibits strong relationship: today s price movement affect tomorrow s price movement Usually it is highly positively related: i.e. rises today, highly possible rise tomorrow US Dow, Apple, Xerox, IBM Previous positive price movement predicts a following positive price movement

12 The market knows all the news about investment and reflects in the price immediately Means that historic event have no impact on price Truth IBM story on PCs Behavioral finance research also found the time lag on information on the subsequent price change; some time faster, some time later If true, what is the purpose of the inside information legislation

13 Anomalies Equity risk premium Small cap premium Value stock effect Frequency of big movement Numerous investment sages Just a general understanding, there are many in research and literature

14 Anomalies Equity risk premium If in the long run, stock is as safe as the bond per a lot of belief, why? The equity risk premium (the return to compensate the risk of stock investment) was 4-8% or even as high as 19% in certain period If people know what they are investing, this should not happen. In the long run, there should not be such high premiums for equity.

15 Anomalies Small cap premium If people know all about investing: why small cap stocks provide a much higher return per research Even if they are different, the stocks shall drive to be equal by arbitrage. Small neutral: +5.8% Small value: +8% But this is gone now after its discovery in further research by Fama in 2011

16 Anomalies that cannot be explained by the stuff taught in the textbook Value stock premium Why do not-in-favor stocks provides a higher return? Impact of value : on average: 3.2% Why there is such an anomaly? Now, it is gone again.

17 Anomalies Frequency of big movements Based on the bell shape curve and normal distribution, it is just impossible for the happening on 19 Oct 1987 (-29.2%) 1 in 10^50 Other impossible falls 31 Aug 1998 (-6.8%) : 1 in 2 * 10^7 25 Aug 1998 (-4.4%), 4 Aug 1998 (-3.5%) = three events in such as short time within 4 weeks: 1 in 5 x 10^11 In July 2002, three steep falls due to IT bubbles : 1 in 4 x 10^12 In 1997, because of Asian financial turmoil, Dow -7.7%: 1 in 5*10^10 (Mandelbrot 2008) But the US stock market just started in 1790 or just 222 year history!

18 Anomalies Frequency of big movements in Hang Seng Index Per normal distribution assumption 19 Oct 1987 (-33.33%) : 1 in 1.07 * 10^ * 10^79 = Oct 1997 (-13.7%) : 1 in 1.66 * 10^12 28 Oct 1997 (+18.8%) : 1 in 1.9 * 10^24 24 Oct 2008 (-12.7%) : 1 in 2.04 * 10^10 27 Oct 2008 (+14.3%) : 1 in 3.41 * 10^13 But the Hang Seng index tracking just started in 1969 or just 43 year history!

19 Anomalies Investment cognoscenti (Author of the book and year of the book) Mr. Warren Buffett (Buffett and Clark, 2001) Mr. Peter Lynch (Lynch and Rothchild, 2000) Sir John Mark Templeton (Templeton and Scott, 2008) Mr. Ginzo Korekawa (Ginzo, 1991) Mr. Andre Kostolany (Kostolany, 1996) Mr. Jim Slater (Slater, 2000) Mr. Jim Rogers (Rogers, 2004) Mr. George Soros (Soros and Volcker, 2003) Mr. Philip Fisher (Fisher, 1997) Why so many investors can generate the huge return compare with other investors if all investors know what they are investing?

20 Implications on Leverage and Risk Control Maximum leverage: 1/3 LTCM fiasco in Russian Bond Market Value at Risk Drawback of VAR The supplement test required: stress test

21 A new view of risk: The Fractal Fractal: broken and rough The stock market does not move in a smooth and nice way but broken and rough way

22 Daily Percentage Change of Hang Seng Index from 31 Dec 1986 to 5 Feb Daily Percentage Change of HS Index Daily Percentage Change of HS Index

23 Daily Percentage Change of Hang Seng Index from 31 Dec 1986 to 5 Feb 2013 per Modern financial Theory Daily Percentage Change Per MFT Daily Percentage Change Per MFT

24 A new view of risk: The Fractal Is there any difference between two daily percentage change?

25 31/12/ /12/ /12/ /12/ /12/ /12/ /12/ /12/ /12/ /12/ /12/ /12/ /12/ /12/ /12/ /12/ /12/ /12/ /12/ /12/ /12/ /12/ /12/ /12/ /12/ /12/ /12/2012 Axis Title A new view of share price movement: The Power Law Hang Seng Index 31 Dec 1986 to 5 Feb Adj Close

26 31/12/ /12/ /12/ /12/ /12/ /12/ /12/ /12/ /12/ /12/ /12/ /12/ /12/ /12/ /12/ /12/ /12/ /12/ /12/ /12/ /12/ /12/ /12/ /12/ /12/ /12/ /12/2012 A new view of share price movement: The Power Law 12 Log Price of Hang Seng Index LogPrice 4 2 0

27 A new view of share price movement: The Power Law Regression: normal fit of data SUMMARY OUTPUT Regression Statistics Multiple R R Square Adjusted R Square Standard Error Observatio ns 6452 ANOVA df SS MS F Significanc e F Regression E E Residual E Total E+11 Coefficients Standard Error t Stat P-value Lower 95% Upper 95% Lower 95.0% Upper 95.0% Intercept 0 #N/A #N/A #N/A #N/A #N/A #N/A #N/A

28 A new view of share price movement: The Power Law Regression: log fit of data SUMMARY OUTPUT Regression Statistics Multiple R R Square Adjusted R Standard Er Observation 6453 ANOVA df SS MS F ignificance F Regression Residual Total Coefficientstandard Erro t Stat P-value Lower 95% Upper 95%Lower 95.0%Upper 95.0% Intercept 0 #N/A #N/A #N/A #N/A #N/A #N/A #N/A X Variable E

29 Research on the theory of predicting crashes for the Hong Kong stock market A model based on the Fractal theory, power law model and the dragon king phenomenon was built to predict market crashes and subsequent trough. The next two pages are the predicted results of the model.

30 Data formatting Then, the natural log data are computed with distribution of changes of: 1 Day 10 Days 20 Days 40 Days 60 Days 120 Days 240 Days 480 Days and 960 Days

31 Distribution of Return

32 Research on the theory of predicting crashes for the Hong Kong stock market History of Bubbles and Predicted Probability by the Model Great Crash Depth of Probability of Crash Probability of the Crash Drawdown with 5% Error Possibility with 10% Error Possibility (Black Monday) Oct 1987 (51%) 100% 100% (Interest Rise of US) Feb 1994 (31%) 100% 100% (Asian Financial Turmoil) Aug 1997 (56%) 100% 100% (IT Bubble) Mar 2000 (51%) 89% 100% (Financial Tsunami and Subprime Woe) Oct 2007 (46%) 89% 89% (Europe Debt Crisis) Nov 2010 (35%) 22% 44%

33 Findings of application of the discoveries in predicting stock market crashes and subsequent trough History of Bubbles and Predicted Probability by the Model Discovery of Trough after the crash (Black Monday) Dec % 50% (Interest Rise of US) Jan % 33% (Asian Financial Turmoil) Aug % 78% (IT Bubble) Apr % 44% (Financial Tsunami and Subprime Woe) Mar % 44% (Europe Debt Crisis) Oct % 44%

34 Findings The prediction is quite good for crash but not so good for trough.

35 Consideration of Convexity Additional independent variable: log of the log of the price

36 Research results considering convexity

37 Prediction results considering convexity History of Bubbles and Predicted Probability by the Model Great Crash Depth of Probability of the Crash Probability of the Crash Drawdown with 5% Error Possibility with 10% Error Possibility (Black Monday) Oct 1987 (51%) 17% 17% (Interest Rise of US) Feb 1994 (31%) 55% 44% (Asian Financial Turmoil) Aug 1997 (56%) 11% 33% (IT Bubble) Mar 2000 (51%) 44% 66% (Financial Tsunami and Subprime Woe) Oct 2007 (46%) 89% 100% (Europe Debt Crisis) Nov 2010 (35%) 22% 44%

38 Prediction results considering convexity Discovery of Trough after the crash (Black Monday) Dec % 50% (Interest Rise of US) Jan % 33% (Asian Financial Turmoil) Aug % 78% (IT Bubble) Apr % 44% (Financial Tsunami and Subprime Woe) Mar % 44% (Europe Debt Crisis) Oct % 44%

39 Prediction results: consider convexity May be due to data snooping bias The results are not as good as the simple power law model

40 Other Looming Signs of Crashes Difficulty in finding value stocks Technical analysis sign Inflation Insider trading No effect for good news Change in leading stocks Cocktail party theory of Honorable Peter Lynch Huge number of funds launched Long time neglected stocks rally High level of market PE Wide participation of public Loose credit and imminent tightening High price This time is different Strange stock listed Right issue and new share issue

41 Difficulty in Finding Value Stocks Difficulty in finding stocks with low PE, low PB, high dividend yield

42 Technical Analysis Sign Market breath analysis Cash flow indicator Shoulder pattern

43 Inflation Indicator of economic overheat Tightening may come any time

44 Insider Trading Insider starts to sell the stocks He know more than us The 1929 sale of stock of banks but executives

45 No Effect of Good News Still remember the US debt ceiling untethering has no positive effect on HK market

46 Change in Leading Stocks Especially careful if telecom, utilities and staples become the leading stocks

47 Cocktail Party Theory of Peter Lynch No interest in the stock market Interest in the stock market View about the stock market

48 Huge Number of Fund Launched Indication of a fervent market Last force pushing the market

49 Long Time Neglected Stocks Rally There is no good and investible stocks in the market. So, the long time neglected stocks also rally.

50 High Market PE From Hang Seng Index experience: 1973 May 43x 1987 Sep 22x 1997 Sep 17x The reverse of PE is the yield of earnings 1997 Not very high market PE, why collapse Real Return inflation Rule of 20

51 Wide participation of public Wide participation of public Indicators: Market Participant A and Market Participant C in HKEX Trade Volume Rally extends GODs of Stock Market

52 Loose Credit and Imminent Tightening Asia financial turmoil HK economy: de-industrialization US$ strength after month HIBOR HK property crash in 1997.

53 High Price History Land price of Tokyo versus USA a few years after Court Accord Price of Tulip in Holland in Tulipomania Determinant of market price

54 This time is different Comment of Late Honorable Sir John Mark Templeton Interesting quip of Mark Twain about the crash 1929 Great Crashes Babson Break

55 Strange stock listed Strange stocks south sea bubble and 1973 in HK 1973 HK Stock Market Mania: - examples What is the reason for such phenomenon?

56 Right Issue and Public Offer Issue Right issue Public offer IPO Reason

57 Thank you

58 Q & A