Today’s Content

 Discussing what we have covered and what to be covered.

 Reviewing the theorems of MLR
 Regression with the time series data. (Ref. Woolridge)
 Normality assumption (Ref. Woolridge)
 Hypothesis testing (Ref. Gujrati)
 Confidence Interval Approach and Test of significance approach

 74558716191
 4R9ca4
 Course: Econometrics I
• Content Covered:
 Introduction
 What is Econometrics?  Estimation of the OLS estimators.
 Methodology of Econometrics.  Algebraic properties of OLS estimators.
 Different types of data and pros and cons  SST, SSE and SSR. Derivation and
of using those data. interpretation of of R square.
 Regression and correlation.  Impact of change in unit of measurement
in dependent and independent variable.
 Chapter2: SLR  Linear and non-linear regression.
 Simple linear regression.  Assumptions of SLR [ emphasize the
assumption 4 and 5]
 Understanding of PRF and SRF.
 Proof of unbiasedness of OLS estimator
 OLS and relevance of using OLS. using assumption 1 to 4.
 Distinguish between Estimator and  Derivation of the variance of the OLS
estimate. estimator. Proof of unbiasedness of the
variance of the OLS estimators using
assumption 1 to 5.
Chapter 3: Multiple Linear regression:
Formation and relevance of multiple linear regression and interpretation of the
estimators.
‘Partialling out’ interpretation.
R square and its interpretation.
Assumptions of MLR [ discussion of MLR 3].
 Proof of theorems 3.1 to 3.4 (Gauss Markov) using the assumptions.

To be Covered:
Hypothesis Testing and Prediction:
Test of hypothesis: restricted vs. non-restricted regression, Neyman-Pearson
methodology; point and interval estimations; Two approaches to testing
hypothesis: confidence interval approach, test of significance approach,
prediction and forecasting: prediction interval, prediction variance.
Multicollinearity:
The nature of multicollinearity, estimation in the presence of multicollinearity,
detection of multicollinearity and its remedial measures.

Heteroscedasticity:
The nature of heteroscedasticity, consequence of using OLS in the presence of
heteroscedastic disturbances, detection of heteroscedasticity and its remedial
measures, the method of weighted least squares (WLS).

Autocorrelation:
The nature of the problem of autocorrelation, consequence of using OLS in the
presence of autocorrelated disturbances, detection of autocorrelation and its
remedial measures, the method of generalized last squares (GLS).

Model Specification and Diagnosis:

Specification errors: Consequences and detection, errors of measurement, Model
specification criteria: R square criterion, AIC and BIC, Ramsey’s RESET test.
 Regression with the time series data
Cross Section and Time Series Data:
Cross-Sectional Data Set on Wages and Other Minimum Wage, unemployment, and Related Data for
individual Characteristics Puerto Rico
obsn wag edu expe fema marri obsn year avgmin avgco prune prgnp
o e c r le ed o v mp
1 3.10 11 2 1 0 1 195 0.20 20.1 15.4 878.7
0
2 3.24 12 22 1 1 2 195 0.21 20.7 16.0 925.0
3 3.00 11 2 0 0 1
3 195 0.23 22.6 14.8 1015.9
4 6.00 8 44 0 1 2
. . . . . .
5 5.30 12 7 0 1
. . . . . . . . . . . .
. . . . . .
. . . . . .
. . . . . .
37 198 3.35 58.1 18.9 4281.6
525 11.5 16 5 0 1 6
6
526 3.50 14 5 1 0 38 198 3.35 58.2 16.8 4496.7
7
The Nature of the time series data
 Temporal ordering
 we must know that the data for 1970 (for example) immediately precede the
data for 1971.
we must recognize that the past can affect the future, but not vice versa.

 Randomness of the Time series

Today we do not know what the Dow Jones Industrial Average will be at the
close of the next trading day.
We do not know what the annual growth in output will be in Canada during the
coming year.
Since the outcomes of these variables are not foreknown, they should clearly
be viewed as random variables
 Static Model
A static model relating y to z is

Usually, a static model is postulated when a change in z at time t is believed to

have an immediate effect on y.
An example of a static model is the static Phillips curve, given by

where inft is the annual inflation rate and unemt is the unemployment rate. This
form of the Phillips curve assumes a constant natural rate of unemployment and
constant inflationary expectations, and it can be used to study the
contemporaneous tradeoff between inflation and unemployment.
A static model with multiple variables:
Finite Distributed lag models
In a finite distributed lag (FDL) model, we allow one or more variables to affect y
with a lag. For example, for annual observations, consider the model

where gfrt is the general fertility nand pet is the real dollar value of the personal
tax exemption. Equation (10.4) is an example of the model

which is an FDL of order two

CLM Assumptions
Under CLM, OLS is not only BLUE, but is the minimum variance unbiased estimator
 We can summarize the population assumptions of CLM as follows
y|x ~ Normal(b0 + b1x1 +…+ bkxk, s2)
 While for now we just assume normality, clear that sometimes not the case
 Large samples will let us drop normality
The homoskedastic normal distribution with a single explanatory variable
Normal Sampling Distributions
Under the CLM assumption s, conditiona l on
the sample values of the independen t variable s

ˆ j ~ Normal  j , Var ˆ j , so that  
 ˆ j   
 
j
~ Normal 0,1
sd ˆ j
ˆ j is distribute d normally because it
is a linear combinatio n of the errors
Economics 20 - Prof. Anderson 10
HYPOTHESIS TESTING: GENERAL COMMENTS

1. Confidence interval approach

2. Test of significance approach
Both these approaches predicate that the variable (statistic or estimator) under
consideration has some probability distribution and that hypothesis testing
involves making statements or assertions about the value(s) of the parameter(s)
of such distribution.
Hypothesis Testing: Confidence Interval Approach (Two-Sided or Two-Tail
Test)

To illustrate the confidence interval approach, once again we revert to our wages-education
example. From the regression results given in Eq. (3.6.1), we know that the slope coefficient
is 0.7240. Suppose we postulate that

Null Hypothesis: H0: β2 = 0.5 and Alternative

Hypothesis: H1: β2 = 0.5

Actually it is what is known as a two-sided hypothesis. Very often such a two-sided

alternative hypothesis reflects the fact that we do not have a strong a priori or theoretical
expectation about the direction in which the alternative hypothesis should move from the
null hypothesis.
Hypothesis Testing: Confidence Interval Approach (Two-Sided or Two-Tail Test)

Is the observed ˆ β2 compatible with H0?

To answer this question we need to construct confidence Interval for β2.
We will do this- by using t distribution. The t variable can be defined as-

This t variable follows a t distribution with n-2 df.

Two-Sided or Two-Tail Test

Now lets define the 100(1-α) % confidence interval for

β2 as follows:

Substituting the value of t it can be rewritten

as follows:

Rearranging the above equation we

can get,
Hypothesis Testing: Confidence Interval Approach (Two-Sided or Two-Tail Test)

Or more compactly the 100(1-α) %

confidence interval for β2 is

Similarly the 100(1-α) %

confidence interval for β1 is

Notice an important feature of the confidence intervals given in Equations

5.3.6 and 5.3.8: In both cases the width of the confidence interval is proportional to the
standard error of the estimator. That is, the larger the standard error, the larger is the
width of the confidence interval. Put differently, the larger the standard error of the
estimator, the greater is the uncertainty of estimating the true value of the unknown
parameter. Thus, the standard error of an estimator is often described as a measure of
the precision of the estimator (i.e., how precisely the estimator measures the true
population value).
Hypothesis Testing: Confidence Interval Approach (Two-Sided or Two-Tail Test)

Now Suppose that ,

ˆβ2 =0.7240 , se (ˆβ2)=0.0700, α= 5%s0 1-α =95% confidence
interval.
n=13, so degrees of freedom (df) = n-2=13-2=11
and α/2 =0.05/2=0.025 (since two tail test we should divide alpha by 2)

• Now at 11 df and 0.025 level of significance the critical value of t

which is defined as tα/2 can be obtained from t chart given in the
appendix.
• tα/2 = 0.201
Two-Sided or Two-Tail Test

Therefore the 95% confidence

interval for β2

0.7240- (2.20*0.070) ≤ β2 ≤ 0.7240+(2.201*0.0700)

0.5700 ) ≤ β2 ≤ 0.8780
The 95 cases out of 100 will contain the true value of β2 as described by the above
equation. If the value of β2 under null hypothesis falls within the 100(1-α) %
confidence interval we do not reject null hypothesis. If it lies outside the interval we reject
the null hypothesis.
Two-Sided or Two-Tail Test
 Hypothesis Testing: Confidence Interval
Approach (Two-Sided or Two-Tail Test)

• Following this rule our hypothetical example null hypothesis that H0: β2 =
0.5 clearly lies outside of the 95% confidence interval . So reject the
null hypothesis with 95 % confidence.
One-Sided or One-Tail Test
Sometimes we have a strong a priori or theoretical expectation (or expectations
based on some previous empirical work) that the alternative hypothesis is one-
sided or unidirectional rather than two-sided, as just discussed. Thus, for our
wages-education example, one could postulate that

• H0: β2 ≤ 0.5 and H1: β2 > 0.5

Perhaps economic theory or prior empirical work suggests that the slope is greater
than 0.5. Although the procedure to test this hypothesis can be easily derived
from Eq. (5.3.5), the actual mechanics are better explained in terms of the test-
of-significance approach discussed next
Hypothesis Testing: The Test-of- Significance Approach

Testing the Significance of Regression Coefficients: The t Test An alternative but

complementary approach to the confidence- interval method of testing statistical hypotheses is
the test-of- significance approach developed along independent lines by R.

A. Fisher and jointly by Neyman and Pearson.10 Broadly speaking, a test of significance is a
procedure bywhich sample results are used to verify the truth or falsity of a null hypothesis.
The key idea behind tests of significance is that of a test statistic (estimator) and the sampling
distribution of such a statistic under the null hypothesis. The decision to accept or reject H0 is
made on the basis of the value of the test statistic obtained from the data at hand.
Hypothesis Testing: The Test-of- Significance Approach
Recall that the t variable can be
defined as

It follows the t distribution with n − 2 df. If the value of true

β2 is specified under the null hypothesis, the t value can
readily be computed from the available sample, and therefore
it can serve as a test statistic.
Hypothesis Testing: The Test-of- Significance Approach

• We know that ˆ β2 = 0.7240, se ( ˆ β2) = 0.0700, and df = 11. If we

assume α = 5%, tα/2 = 2.201.
• If we assume H0: β2 = ˆ β2 = 0.5 and H1: β2 = 0.5
• One can compute the t value in the middle of the double inequality given by Eq.
(5.7.1) and see whether it lies between the critical t values or outside them. For our
example,
The value of t static is , t = (0.7240 − 0.5)/ 0.0700 = 3.2
And the critical value of t obtained from t table at 11 df and α/2 level of significance is
tα/2 = 2.201.
Rules: since t > critical value of t , we reject null hypothesis and our test
is statistically significant.
Hypothesis Testing: Confidence Interval VS The Test-of-Significance
Approach
• Once again let us return to our wages-education example. We know that ˆ β2 =
0.7240, se ( ˆ β2) = 0.0700, and df = 11. If we assume α = 5%, tα/2 = 2.201. If
we assume H0: β2 = β∗2 = 0.5 and H1: β2 = 0.5, Eq. (5.7.2) becomes
Pr (0.3460 ≤ ˆ β2 ≤ 0.6540)
Hypothesis Testing: Confidence Interval VS The Test-of-Significance
Approach

The value of t static is , t = (0.7240 − 0.5)/ 0.0700 = 3.2

which clearly lies in the rejection region.
 Hypothesis Testing: The Test-of- Significance Approach

• Since we use the t distribution, the preceding testing procedure is called appropriately the
t test. In the language of significance tests, a statistic is said to be statistically significant if
the value of the test statistic lies in the critical region. In this case the null hypothesis is
rejected. By the same token, a test is said to be statistically insignificant if the value of the
test statistic lies in the acceptance region. In this situation, the null hypothesis is not
rejected. In our example, the t test is significant and hence we reject the null hypothesis.
 Hypothesis Testing: The Test-of-Significance Approach (One Tail Test)

Suppose that
H0: β2 ≤ 0.5 and H1: β2 > 0.5.

Although H1 is still a composite hypothesis, it is now one-sided. To test this hypothesis, we use the
one-tail test.

The test procedure is the same as before except that the upper confidence limit or critical value now
corresponds to
tα= t.05,

that is, the 5 percent level. From t table the critical value of t at 5 % level of significance and at 11 df
is obtained as
tα= 1.796
Clearly t=3.2 is greater than critical t. so reject null hypothesis.
Hypothesis Testing: The Test-of-Significance Approach (One Tail Test)
Hypothesis Testing: The Test-of-Significance Approach
Summary