Upload
econometrics-freelancer
View
218
Download
0
Embed Size (px)
Citation preview
7/30/2019 Macroeconomic Models First Session Handouts
1/12
Macroeconomic Models
Introduction to Basic Theory
7/30/2019 Macroeconomic Models First Session Handouts
2/12
Model
A Model describes the theoretical relationshipbetween variables.
Macroeconomic model would thus describe
the relation between macroeconomicvariables
Models are estimated using Statistical
Methods, Econometric Theory (MathematicalModelling of Economic Theory and estimatedusing Statistical Data)
7/30/2019 Macroeconomic Models First Session Handouts
3/12
Example
We can model the relation between GDP andInvestment following one of the most simpleprocedures.
First take, GDP(y)=f(Investment(i)) as one of thebasic and assumed model
The mathematical function will then describe theexpected model where we can findtheory/literature which will help us define thenature of the model
Is it a linear, non-linear?
Hence the linear function will be y=a +b (i)
7/30/2019 Macroeconomic Models First Session Handouts
4/12
Contd.
Then the non-linear model will be any type of
Quadratic, cubic, exponential or logarithmic
type of functions
Y=a+b(i)+c(i^2) ::: where ^ is power
Y=e^I
Y=a+b(log_i) etc.
7/30/2019 Macroeconomic Models First Session Handouts
5/12
Estimation
Estimation of the specified model is then theMathematical procedures to determine the valueof a and b of the function in example for a linear
function. The a and b can be estimated using a Procedure
known as OLS, Ordinary Least Squares or ML,Maximum Likelihood. Common is OLS. So we will
use it all the way along the course. But note that to estimate the a and b, we will
need to use Matrices, and Statistical data.
7/30/2019 Macroeconomic Models First Session Handouts
6/12
Contd.
Statistical data on GDP and Investment will beused for any country, for example Pakistan formore than 30 years or for 30 or more countries
for a single year. The datasets are attached withthe email/link you will receive for thesedocuments.
OLS details are given in the attached reading
material. So how to estimate the a and b of the example
model, we will use Eviews.
7/30/2019 Macroeconomic Models First Session Handouts
7/12
Eviews
Open the given data/workfile
Note the y and I
Select the series containing data on y and I
Click Quick
Click Estimate Equation
Type the following line
y c I Note the data points equal the number of years
for which we have the data
7/30/2019 Macroeconomic Models First Session Handouts
8/12
Estimation Checking
How to see if the estimated model is the best
We use tests which are used for the checkingof OLS assumptions BLUE
Also will check some other Econometricproblems in the estimated models
These include Autocorrelation,
heteroscedasticity, Multi-collinearity andNormality of the error terms etc.
Why?
7/30/2019 Macroeconomic Models First Session Handouts
9/12
Model Selection
For example we estimated two types of models
A Linear
A Non-linear as described above
How would you select which model is the good one(technically we will explore these and related issues innext week sessions when our main point of discussionwill be Estimation and Selection of a Model among alist of Possible Models
Why? Because we need to be sure that the model usedfor forecasting the best model used on grounds ofMathematical, Statistical and Econometric reasons.
7/30/2019 Macroeconomic Models First Session Handouts
10/12
Eviews Session
Now see how we estimate each type of model
First Estimate the Linear Models
The Estimate the Quadratic Models Signs and Interpretation of Estimates
Theoretical Justifications for the Signs
7/30/2019 Macroeconomic Models First Session Handouts
11/12
Questions & Answers
Any Questions and Answers?
7/30/2019 Macroeconomic Models First Session Handouts
12/12
HomeWork
Now select a country of your choice, collect annual orquarterly data on GDP (y), Labour Force (L), AgricultureProduction Values (ag), Industrial Values (il), Land (ArableLand, Area under Cultivation) (area) and estimate thefollowing Model. You can use World Bank databank, IMF
Statistics.
Gdp=f(L, il, ag/area)?
Interpret the estimate model?
Any theoretical Reason? Any issues you find in literature (search google scholar)
related this type of models?
What Alternatives are presented?