c1.introduction to eco no metrics

Upload: rrr

Post on 30-May-2018

213 views

Category:

Documents


0 download

TRANSCRIPT

  • 8/14/2019 C1.Introduction to Eco No Metrics

    1/29

    1

    Introduction

    to

    Econometrics

  • 8/14/2019 C1.Introduction to Eco No Metrics

    2/29

    2

    What is econometrics? Econometrics is a set of research tools employed in various

    economic fields (accounting, finance, marketing and

    management) and also in history, political science, and

    sociology.

    Econometrics means economic measurement> from the

    Greek words: eikonomia economy andmetren measure.

    It plays a special role in the training of the economists by:

    quantifying the economic reality

    bridging the gap between economic theory and the real

    world.

  • 8/14/2019 C1.Introduction to Eco No Metrics

    3/29

    3

    Econometrics may be defined as the social science that applies

    the tools ofeconomic theory,mathematicsandstatistics for

    estimating economic relationships, testing the validity of

    economic theories and validating the government policies.

    Example.

    The Central Bank rises the discount rate by 3%economic theory suggests this will influence the interest rates of

    the commercial banks, the credits, the investments, the inflation

    but how much?Only econometrics can answer this question by combining

    economic theory, statistical data and mathematical tools!

  • 8/14/2019 C1.Introduction to Eco No Metrics

    4/29

    4

    Why study econometrics?

    Economic theory makes statements about how important

    variables are related to one another, but:

    does not provide the necessary measure of strength of the

    relationship or

    the magnitudes involved (how much a change in one

    variable affects another).

    Example

    The law of demand: a reduction in price of a commodity isexpected to increase the quantity demanded of that

    commodity (but how much?).

  • 8/14/2019 C1.Introduction to Eco No Metrics

    5/29

    5

    Economic model vs econometric model

    An economic model is a set of assumptions that aproximately

    describes the behaviour of an economy.

    An econometric model gives:

    the form of the algebraic relationships among the economicvariables involved and

    a specification of the errors (due to the factors omitted from

    the model)

  • 8/14/2019 C1.Introduction to Eco No Metrics

    6/29

    6

    1.Quantity demanded, qd, for an individual commodity (Ford cars):

    qd = f( p, pc, ps, i )

    p = price of the car; pc = price of complements (like gasoline);

    ps = price of substitutes (other similar cars); i = income

    demand

    2. Quantity supplied, qs (beef):

    qs = f( p, pc, pf, ps )

    p = price of beef; pc = price of complements ;

    ps = price of substitutes (pork); pf = price of inputs (e.g. corn)

    supply

    The economic model(describes the way in which economic variables are interrelated).

  • 8/14/2019 C1.Introduction to Eco No Metrics

    7/29

    7

    The econometric model

    qd = f( p, pc, ps, i ) + e

    The random error accounts for the many factors that affectsales and miss from this model; it is a noise component thatobscures our understanding of the relationships.

    Assuming quantity demanded is a linear function of prices, theeconometric model of the demand for Ford cars is :

    qd = 1+ 2p+ 3pc+ 4p

    s+ 5i + e,

    where i are unknown parameters to be estimated usinga sample of economic data.

    The systematic part(the average behavior)

    Therandomerror(theunpredictablepart)

  • 8/14/2019 C1.Introduction to Eco No Metrics

    8/29

    8

  • 8/14/2019 C1.Introduction to Eco No Metrics

    9/29

    9

    Limits of econometrics

    any econometric model can capture only a part of thereal economic world => a simplified image of the reality;

    qualitative variables (like motivation, satisfaction,

    management) are difficult to incorporate in aneconometric model;

    imperfect identification of the contribution of each factor

    of influence (because factors interact and influences aredifficult to separate);

    forecasting based on past trends is uncertain.

  • 8/14/2019 C1.Introduction to Eco No Metrics

    10/29

    10

  • 8/14/2019 C1.Introduction to Eco No Metrics

    11/29

    11

  • 8/14/2019 C1.Introduction to Eco No Metrics

    12/29

    12

  • 8/14/2019 C1.Introduction to Eco No Metrics

    13/29

    13

    1 2 2

    2

  • 8/14/2019 C1.Introduction to Eco No Metrics

    14/29

    14

    1 2

    dependent

    (endogeneous)

    variable explanatory (exogeneous) variable

    parameters

  • 8/14/2019 C1.Introduction to Eco No Metrics

    15/29

    15

  • 8/14/2019 C1.Introduction to Eco No Metrics

    16/29

    16

  • 8/14/2019 C1.Introduction to Eco No Metrics

    17/29

    17

  • 8/14/2019 C1.Introduction to Eco No Metrics

    18/29

    18

  • 8/14/2019 C1.Introduction to Eco No Metrics

    19/29

    19

  • 8/14/2019 C1.Introduction to Eco No Metrics

    20/29

    20

  • 8/14/2019 C1.Introduction to Eco No Metrics

    21/29

    21

  • 8/14/2019 C1.Introduction to Eco No Metrics

    22/29

    22

  • 8/14/2019 C1.Introduction to Eco No Metrics

    23/29

    23

    1 2

    1 1

    2 2 2

  • 8/14/2019 C1.Introduction to Eco No Metrics

    24/29

    24

    God created the econometricians to makeGod created the econometricians to make

    weather forecasts look good .weather forecasts look good .

    The value of the explanatory variable (Yd) is known

    Estimated values of the parameters

  • 8/14/2019 C1.Introduction to Eco No Metrics

    25/29

    25

    Economic Empirical Study

  • 8/14/2019 C1.Introduction to Eco No Metrics

    26/29

    26

    Economic Empirical Study

    (anatomy of econometric modeling)

    Economic Theory; Past Experience, studies

    Formulating a model: Cause - effect C = f(Y) ==>Ct = 1 + 2Y + et

    Gathering data: Statistics monthly, quarterly, yearly data

    Estimating the model: Simple OLS method or other advances

    Testing the hypothesis:H0:2>0,

    positive relationship or not If not true

    Interpreting the results:

    Forecasting Policy implication and decisions

  • 8/14/2019 C1.Introduction to Eco No Metrics

    27/29

    27

    Types of econometric models

    1. Considering the number of influence factors simple models are based on the assumption of a single

    decisive factor of influence X, all the others havingrandom or fixed effects captured by the error term e.

    y = f(x)+e

    multiple models: use two or more factors of influence;better but more complicated => the number of factorshave to be limited.

    y = f(x1,x

    2,...,x

    p)+e

    2. Considering the form of the relationship betweenthe dependent variable y and the factors ofinfluence linear models

    non-linear models: exponential, parabolic, etc.

  • 8/14/2019 C1.Introduction to Eco No Metrics

    28/29

    28

    Types of econometric models (contd)

    3. Considering the time factor- static models: the factors influence the dependentvariable y over the same period of time:

    y = f(x1t,...,xjt,...,xkt) + et

    - dynamic models include the time variable: the time variable is one of the factorsy = f(xt,t) + et

    autoregressive models: the lagged dependent variablebecomes a factor of influence itself

    y = f(xt,yt-k) + et lagged models: the factor x influences y over several

    periods of time:

    y = f(xt,xt-1,... xt-k) + et

  • 8/14/2019 C1.Introduction to Eco No Metrics

    29/29

    29

    Types of econometric models (contd)

    4. Considering the number of equations

    - models having a single equation

    - multiple equation models: consist of a system of

    equations

    -> structural form:

    dependent (endogenous) variabiles

    independent (exogenous) variabiles

    ===

    nmnmnnnnn

    mmnn

    mmnn

    UXcXcXcYYbYb

    UXcXcXcYbYYb

    UXcXcXcYbYbYb

    ......

    ......

    ......

    22112211

    2222212122121

    112121111212111

    niYi ,, 1

    mjXj ,, 1