MSF AFTS KP 2014 Lectures 2 Stud

download MSF AFTS KP 2014 Lectures 2 Stud

of 52

Transcript of MSF AFTS KP 2014 Lectures 2 Stud

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    1/52

    Krzysztof Piontek

    Department of Financial Investments and Risk Management

    1

    Analysis of Financial Time Series

    statistics - revision

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    2/52

    Krzysztof Piontek

    Department of Financial Investments and Risk Management

    2

    Course Requirements

    The basic knowledge of:

    financial markets,

    financial instruments,

    risk management,portfolio management,

    mathematics,

    statistics,

    probability calculus,

    econometrics.

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    3/52

    Krzysztof Piontek

    Department of Financial Investments and Risk Management

    3

    Two main tasks in financial data modeling:

    - Modeling dynamics (econometrics)

    - Modeling statistical distributions (statistics)

    the order of the observations over the time

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    4/52

    Krzysztof Piontek

    Department of Financial Investments and Risk Management

    4

    Analysis of financial time seriesis main part of the

    discipline called financial econometrics

    Two main objectives of financial econometrics:

    1) Verification of different models developed in theory of finance

    (financial economics) by using financial time series models

    2) Analysis of financial data to identify main characteristics

    (for example risk characteristics)

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    5/52

    Krzysztof Piontek

    Department of Financial Investments and Risk Management

    5

    Time series:

    A sequence of random variables measuring certain quantity of

    interest over time

    Financial Time Series:

    collection of a financial measurement over time.

    Example: log return r

    t

    Data: {r

    1

    , r

    2

    , ..., r

    T

    } (t data pints)

    Futures prices or returns are random variables since we do not

    know their values until we observe theirs.

    A random variable can be thought of as an unknown value that may

    change every time it is inspected.

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    6/52

    Krzysztof Piontek

    Department of Financial Investments and Risk Management

    6

    Name, Date, Open, High, Low, Close, Volume

    KGHM, 20090904, 83.45, 85.70, 83.00, 83.00, 661561

    KGHM, 20090907, 84.50, 85.95, 84.40, 85.95, 389379

    KGHM, 20090908, 87.00, 89.60, 87.00, 88.90, 1251472

    KGHM, 20090909, 87.70, 89.30, 87.70, 88.90, 701582

    KGHM, 20090910, 89.90, 90.40, 86.15, 88.00, 970480

    KGHM, 20090911, 88.00, 88.00, 85.00, 85.20, 865165KGHM, 20090914, 84.00, 84.65, 82.50, 83.60, 401915

    KGHM, 20090915, 84.60, 85.40, 82.90, 84.00, 495518

    KGHM, 20090916, 85.00, 86.95, 84.95, 86.95, 722730

    KGHM, 20090917, 88.30, 88.85, 86.10, 87.00, 454126

    KGHM, 20090918, 86.00, 86.90, 84.20, 84.60, 1359126

    KGHM, 20090921, 85.10, 85.80, 84.00, 84.00, 448032KGHM, 20090922, 85.50, 90.50, 85.50, 90.50, 2367387

    KGHM, 20090923, 90.50, 92.40, 89.40, 91.00, 1273626

    KGHM, 20090924, 89.50, 92.25, 88.00, 88.00, 950405

    KGHM, 20090925, 88.00, 88.45, 85.20, 86.50, 1388713

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    7/52

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    8/52

    Krzysztof Piontek

    Department of Financial Investments and Risk Management

    8

    Main index of Warsaw Stock Exchange - WIG

    A graph of daily levels

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    9/52

    Krzysztof Piontek

    Department of Financial Investments and Risk Management

    9

    Characteristics of financial time series

    usually we think about

    returns

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    10/52

    Krzysztof Piontek

    Department of Financial Investments and Risk Management

    10

    Rates of return

    (clusteringof variance)

    Quantile-quantile plot

    (normal quantiles)

    Autocorrelation of squared returns

    (lags)

    Autocorrelation of returns

    (lags)

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    11/52

    Krzysztof Piontek

    Department of Financial Investments and Risk Management

    11

    1 0,2264b

    Skewness: 30.10.9519.08.04

    2200 observations

    Histogram of COMPUTERLAND rates of return vs. normal density

    Compute the intervals x + s, x + 2s, x + 3s and compare the percentage of

    data in these intervals to the Empirical Rule (68%, 95%, 99.7%)

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    12/52

    Krzysztof Piontek

    Department of Financial Investments and Risk Management

    12

    Statistics

    short revision

    Descriptive Statistics

    Provides numerical and

    graphic procedures to

    summarize theinformation of the data in

    a clear and

    understandable way

    Inferential Statistics Provides

    procedures to

    draw inferences

    about a populationfrom a sample

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    13/52

    Krzysztof Piontek

    Department of Financial Investments and Risk Management

    13

    Numerical Data Properties

    Central Tendency

    (Location)

    Variation

    (Dispersion)

    Shape

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    14/52

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    15/52

    Krzysztof Piontek

    Department of Financial Investments and Risk Management

    15

    Bell-shaped

    Bimodal

    (Multimodal)

    Uniform

    (Unimodal)

    Right-skewed Left-skewed

    Truncated

    a histogramis a graphical display of

    tabulated frequencies, shown as bars.

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    16/52

    Krzysztof Piontek

    Department of Financial Investments and Risk Management

    16

    Percentiles and Quartiles

    The pth percentile is the value of a givendistribution such that p% of the

    distribution is less than or equal to that

    value.

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    17/52

    Krzysztof Piontek

    Department of Financial Investments and Risk Management

    17

    Quantile

    Some quantiles have special names:

    The 2-quantile is called the median

    The 4-quantiles are called quartilesThe 10-quantiles are called deciles

    The 100-quantiles are called percentiles

    Quantiles provide information about the shape of data as well as its

    location and spread.

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    18/52

    Krzysztof Piontek

    Department of Financial Investments and Risk Management

    18

    0.00

    0.10

    0.20

    0.30

    0.40

    0.50

    0.60

    -5 -4 -3 -2 -1 0 1 2 3 4 5

    10th

    percentile=-1.2816

    10 percent under curve

    (shaded red)

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    19/52

    Krzysztof Piontek

    Department of Financial Investments and Risk Management

    19

    a Q-Q plot( Q stands for quantile) is a probability plot, a kind

    of graphical method for comparing two probability

    distributions, by plotting their quantiles against each other.

    If the data sets agree or the observed set matches the

    theoretical, the plot will be on a straight line.

    If the plot is not on a straight line, then the model is a poor fit.

    Quantile-quantile plots are used to determine whether two samples come from the same

    distribution family. They are scatter plots of quantiles computed from each sample, with a

    line drawn between the first and third quartiles. If the data falls near the line, it is

    reasonable to assume that the two samples come from the same distribution. The method

    is robust with respect to changes in the location and scale of either distribution.

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    20/52

    Krzysztof Piontek

    Department of Financial Investments and Risk Management

    20

    Normal quantile plot

    -2 -1 0 1 2

    -2

    -1

    0

    1

    2

    Normal q-q plot

    Theoretical Quantiles

    SampleQuantiles

    q-q plot 100 sample

    observations from a normaldistribution with mean 0 andstandard deviation 1

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    21/52

    Krzysztof Piontek

    Department of Financial Investments and Risk Management

    21

    Normal quantile plot

    -2 -1 0 1 2

    40

    45

    50

    55

    Normal q-q plot of Height of our Sample Data

    Theoretical Quantiles

    SampleQuantiles

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    22/52

    Krzysztof Piontek

    Department of Financial Investments and Risk Management

    22

    Descriptive statistics

    Types of Descriptive Statistics. 1. Measures of Central Tendency

    2. Measures of Dispersion of Variability 3. Measures of distribution Shape.

    4. Quantiles

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    23/52

    Krzysztof Piontek

    Department of Financial Investments and Risk Management

    23

    Measures of Central Tendency

    Mean:

    Sum of all measurements in the data divided by

    the number of measurements.

    the AVERAGE (or the EXPECTED VALUE)

    Median (2nd quartile):

    A number such that at most half of the

    measurements are below it and at most half of themeasurements are above it.

    Mode:

    The most frequent measurement in the data.

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    24/52

    Krzysztof Piontek

    Department of Financial Investments and Risk Management

    A car went a certain distance at a speed 60 kilometers perhour) and then the same distance again at a speed 40

    kilometers per hour.

    What was the car's average speed during the whole journey?

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    25/52

    Krzysztof Piontek

    Department of Financial Investments and Risk Management

    A car went a certain distance at a speed 60 kilometers per

    hour) and then the same distance again at a speed 40

    kilometers per hour.

    What was the car's average speed during the whole journey?

    average speed

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    26/52

    Krzysztof Piontek

    Department of Financial Investments and Risk Management

    A car travels for a certain amount of timeat a speed 60

    km/hand then the same amount of time at a speed

    40 km/h

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    27/52

    Krzysztof Piontek

    Department of Financial Investments and Risk Management

    A car travels for a certain amount of timeat a speed 60

    km/hand then the same amount of time at a speed

    40 km/h

    average speed..

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    28/52

    Krzysztof Piontek

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    29/52

    Department of Financial Investments and Risk Management

    29

    Measures of dispersion or variability

    Measures how spread out around the center are

    the data.

    The simplest measure of variability is the RANGE.

    This is simply the maximum value minus the

    minimum value.

    INTER-QUARTILE RANGE (IQR) which is defined as

    the value at 75% minus the value at 25% of the

    distribution. Thus the central 50 % of the

    observations fall between these two values.

    Krzysztof Piontek

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    30/52

    Department of Financial Investments and Risk Management

    30

    Measures of dispersion or variability.

    The most common measures of variability

    are the STANDARD DEVIATIONand the

    VARIANCE. The variance is the standard

    deviation squared, or the standarddeviation is the square root of the variance.

    2 2

    1

    1

    ( ) ( )1

    N

    iiS x x xN

    2( ) ( )S x S x

    Krzysztof Piontek

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    31/52

    Department of Financial Investments and Risk Management

    31

    Measures of Shape

    The two most common measures of shape are

    SKEWNESS and KURTOSIS.

    Skewness is a measure of the lack of symmetry in a

    distribution, or whether the distribution is skewed

    to the left or the right.

    Positive skewness: Values clustered toward lower

    range with a long tail extending to upper ranges.

    Negative skewness: Values clustered toward

    upper range with long tail extending to lower

    ranges.

    Krzysztof Piontek

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    32/52

    Department of Financial Investments and Risk Management

    32

    Shape

    Right-SkewedLeft-Skewed Symmetric

    Mean = MedianMean Median Median Mean

    Krzysztof Piontek

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    33/52

    Department of Financial Investments and Risk Management

    33

    Skewness

    Measures of asymmetry of data

    Positive or right skewed: Longer right tail

    Negative or left skewed: Longer left tail

    2/3

    1

    2

    1

    3

    21

    )(

    )(

    Skewness

    Then,ns.observatiobe,...,Let

    n

    i

    i

    n

    i

    i

    n

    xx

    xxn

    nxxx

    Krzysztof Piontek

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    34/52

    Department of Financial Investments and Risk Management

    34

    1 2

    4

    1

    2

    2

    1

    Let , ,... be observations. Then,

    ( )

    Kurtosis

    (

    3

    )

    n

    n

    i

    i

    n

    i

    i

    x x x n

    n x x

    x x

    Kurtosismeasures the thickness of the tails.Higher values of kurtosis indicate more extreme

    values or heavier tails. Negative kurtosis is thinnertails. Heavier or thinner than the normal curve.

    Why subtract 3? Because a normal curve has a kurtosis of 3: if wesubtract three then it is zero and thus kurtosis is expressed subject tocomparisons with the normal curve.

    excess

    kurtosis

    Krzysztof Piontek

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    35/52

    Department of Financial Investments and Risk Management

    35

    leptokurtic

    platykurtic

    mesokurtic

    Kurtosis > 3

    excesskurtosis >0fat tails (typical feature)

    Kurtosis = 3excesskurtosis = 0

    Kurtosis < 3

    excesskurtosis < 0

    Krzysztof Piontek

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    36/52

    Department of Financial Investments and Risk Management

    36

    Expected value

    risk

    (measured as volatility)

    Krzysztof Piontek

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    37/52

    Department of Financial Investments and Risk Management

    37

    Normal Distribution

    2

    2

    1( )

    21

    ( ) ,2

    x

    f x e x

    A density curvedescribes the overall pattern of a distribution.The total area under the curve is always 1.

    The normaldistribution is symmetric, single-peaked and bell-shaped.

    Mean, Median, and mode are same for a normal distribution

    A normal distribution can be described if we know their mean andstandard deviation.(If we know and , we know every thing about thenormal distribution.)

    The probability density function of a normal variable with mean and

    standard deviation can be expressed as,

    Krzysztof Piontek

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    38/52

    Department of Financial Investments and Risk Management

    38

    Normal Distribution

    .

    x

    z

    2

    21

    ( ) ,2

    z

    f z e z

    Standardizing and z-Scores

    If x is an observation from a distribution that has mean and

    standard deviation , the standardized valueof x is

    A standardized value is often called a z-score. If x is normal distribution

    with mean and standard deviation , then z is a standard normal

    variable with mean 0 and standard deviation 1.

    Krzysztof Piontek

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    39/52

    Department of Financial Investments and Risk Management

    39

    Normal Distribution

    Approximately 68% of the

    measurements lie in the

    interval to +

    Approximately 95% of the

    measurements lie in the

    interval 2to + 2

    Approximately 99.7% of the

    measurements lie in the

    interval 3to + 3

    Krzysztof Piontek

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    40/52

    Department of Financial Investments and Risk Management

    40

    Raw (crude) moment (moment about zero)

    1

    1x

    N

    k

    k i

    i

    mN

    k

    km E X

    For a sample of N observations the sample moment is

    Moments: values (statistics, quantities) used to

    characterize the probability distributions of

    random variables.

    Krzysztof Piontek

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    41/52

    Department of Financial Investments and Risk Management

    41

    Central moment (moment about its mean)

    11

    1 x

    N

    kk i

    i

    mN

    k

    k E X E X

    Sample moment

    Krzysztof Piontek

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    42/52

    Department of Financial Investments and Risk Management

    42

    1

    2

    2

    31 3 2

    2

    42 2

    2

    m

    mean

    variance

    skewness

    kurtosis

    Raw (crude) moment (moment about zero)

    Central moment (moment about its mean)

    Krzysztof Piontek

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    43/52

    Department of Financial Investments and Risk Management

    43

    Outliers

    Outliers are cases that have data values very

    different from the data values for the majority of

    cases in the data set.

    Outliers are important because they can change the

    results of our data analysis.

    Whether we include or exclude outliers from a data

    analysis depends on the reason why the case is an

    outlier and the purpose of the analysis.

    Krzysztof Piontek

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    44/52

    Department of Financial Investments and Risk Management

    44

    The higher moment (raw or central) the more

    senitive to outliers.

    outlier

    1

    observation

    1000 observations

    skewness not zero but e.g. 0.4

    kurtosis not 3 but e.g. 4.5

    Krzysztof Piontek

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    45/52

    Department of Financial Investments and Risk Management

    45

    Expected value

    risk

    (measured as volatility)

    Krzysztof Piontek

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    46/52

    Department of Financial Investments and Risk Management

    46

    The simplest model

    The world is a big gaming machine (lottery).

    Each day the rate of return is drown from the same

    distribution which is constant over time.

    (all moments of the daily return distributions areconstant over time)

    Krzysztof Piontek

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    47/52

    Department of Financial Investments and Risk Management

    47

    Discrete time model (simple rates of return)

    tt

    t

    tt

    t z

    P

    PPr

    1

    1

    ),N(~ 2tr

    )1,0N(~t

    z

    The return is decomposed into two parts as

    rt= predictable part + not predic. part

    Krzysztof Piontek

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    48/52

    Department of Financial Investments and Risk Management

    48

    VALUE AT RISK

    Rq

    np. 5 %,

    1 %

    R

    (R)

    Krzysztof Piontek

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    49/52

    Department of Financial Investments and Risk Management

    49

    VALUE AT RISK

    an approach to VaR estimation under the assumption

    of NORMAL return distribution:

    mean of the distribution of returns,

    standard deviation of returns

    constant dependent on the tolerance level,

    e.g.:

    if q = 0.05, then c = -1.65

    if q = 0.01, then c = -2.33

    , - constant in time

    cRq

    Krzysztof Piontek

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    50/52

    Department of Financial Investments and Risk Management

    50

    0 t

    Krzysztof Piontek

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    51/52

    Department of Financial Investments and Risk Management

    51

    The Black-Scholes model of the market for an equity

    makes the following explicit assumptions:

    It is possible to borrow and lend cash at a known

    constant risk-free interest rate.

    The price follows a geometric Brownian motion with

    constant drift and volatility.

    There are no transaction costs.

    The stock does not pay a dividend (see below for

    extensions to handle dividend payments).

    All securities are perfectly divisible (i.e.it is possibleto buy any fraction of a share).

    There are no restrictions on short selling.

    Krzysztof Piontek

  • 8/10/2019 MSF AFTS KP 2014 Lectures 2 Stud

    52/52

    Department of Financial Investments and Risk Management

    1 2

    ( ) ( )r T

    c S d X e N d N2

    1

    2

    ln r

    dX

    TS

    T

    2

    2

    2

    ln S

    r TX

    d

    T

    N(x) denotes the standard normal cumulative distribution function

    the Black-Scholes formula