Homework 1: PHD Student: Freddy Rojas Cama Homework 1 - Eco 322 Rutgers University

PhD student: Freddy Rojas Cama Homework 1 - ECO 322 Rutgers University
Homework 1
Due date: October 15th, 2012; 11.59 PM (ET)
Econometrics
/n 1rcddn 1o,a: Ca:a
1. A consumption function is an economic relationship between consumption and income
in aggregate terms. This function is estimated by OLS and the results are showed be-
low. Specically, the dependent variable is the household consumption (C) and the only
explicative variable is the household income (1), those observations are weekly recorded
and expressed in real terms. The sample size
1
and number of parameters to estimate are
= 15 and 1 = 2 respectively.
C
t
= 28.264 + 0.493945 1
t
+ -
t
(8.1789) (0.0461)
The standard errors are provided and they are below the parameters or point estimates.
(a) What is the type of data we are dealing with? What is the frequency of those data?.
(b) Is the coecient related to income or slope (or marginal propensity to consume)
signicant in statistical terms?.
(c) Construct a 99%, 95% and 90% condence intervals for the constant and slope.
Interpret those condence intervals.
2. Exercise 4.2 in Stock and Watsons book.
3. Exercise 4.3 in Stock and Watsons book.
4. Ordinary Least Squares (OLS) delivers the Best Linear Unbiased Estimators if only if
(()) certain assumptions are fullled. Lets consider the following linear model
n
t
=
0
+
1
r
t
+ -
t
being r
t
and n
t
the independent and dependent variables respectively and -
t
denotes a
normal random variable i.i.d, (i.e. identical and independent distribuited) with expecta-
tion value or mean equal to zero. Also, r
t
and -
t
are independent each other. Consider
the following information in order to have estimates for
0
and
1
.
T=50
P
t=1
n
t
r
t
= 144.70. n = 3.26. r = 0.71.
T=50
P
t=1
r
2
t
= 74.19
(a) Calculate the BLUE estimates of
0
and
1
.
(b) Verify the orthogonality conditions i.e
P
T=50
t=1
b-
t
= 0 and
P
T=50
t=1
b-
t
r
t
= 0
1
Or number of observations.
1
5. A production function is estimated for the US economy using a dataset for year 2011.
There are 51 manufacturing rms in this dataset ( = 51). The estimation is made by
OLS
Q = exp(1.37)1
0:632
1
0:452
where b c
k
= 0.632 and b c
l
= 0.452. There are 2 independent variables in this model: the
level of capital and labor skill are denoted as 1 and 1. The dependent variable is the
level of production (Q). Additionally, we have the following information;
1
2
= 0.98 co(b c
k;
b c
l
) = 0.055 :td(b c
k
) = 0.257 :td(b c
l
) = 0.219
(a) What is the kind of data we are dealing with? What is the frequency of those data?.
(b) Construct a 99% condence interval for c
l
.
(c) Evaluate if elasticities of labor and capital are identical in statistical terms.
2
.
(d) Evaluate if there exists evidence of increasing returns to scale (i.e c
k
+ c
l
1) in
this sample.
6. We have the following data n = f25. 1. 4. 0g. Your dear professor wants you to
estimate the following model n
t
=
0
by using the OLS estimator. Specically;
(a) What is the estimate of
0
?.
(b) If we add to this model a new variable with same values for each observation, can
we get the OLS estimates? why not?
7. We have the following result (STATA estimation output)
The variables are the state name (state), violent crimes per 100,000 people (crime),
percent of population living under poverty line (poverty), and percentage of population
that are single parents (single). It has 51 observations and this dataset is recorded in a
specic period of time. The above ouput shows the results of OLS estimation for a model
where the dependent variable is the crime rate and the dependent variables are poverty
rate and single status. Answer the following with a true or false statement and support
your choice.
2
Remember that
var(aX + bY ) = a
2
var(X) + b
2
var(Y ) + 2abcov(X; Y )
2
(a) Panel data is the type of data we are dealing with.
(b) The coecient associated or related to poverty is not signicative at 90 and 95%
of condence level.
(c) The coecient associated or related to single is not signicative at 90 and 95% of
condence level.
8. Show that if we run separated regressions in a model with 2 regressors (one regression
for each regressor), we can have the same coecients or point estimates only when the
autocorrelation between those 2 regressors is zero.
9. [Playing God and Mortals]. In this game you must choose to be a God or Mortal. If you
choose God, you must generate the following DGP
n =
0
+
1
r + -
where
0
and
1
are the true parameters, r is a random variable distribuited as N

j
x
. o
2
x
and - is distribuited as N

0. o
2
. You must choose values for the true parameters, j

x
. o
2
x
and o
2
.[Suggestion: choose those values from the following sets:
0
2 (5. 10),
1
2
(0.45. 0.80) . j
x
2 (3. 10) . o
2
x
2 (3. 6) and o
2
2 (0.1. 0.45)]. After this parametrization,
you must simulate 10000 draws from this DGP. Explain carefully all the steps.
If you choose Mortal, you must look for a God and ask for series n and r. [Warning: Gods
does not share the DGP with mortals]. Once you have those series, obtain OLS estimates
for
0
,
1
and o
2
(i.e
b
0
,
b
1
and b o
2
). Report all those estimates, variances and p-values
(related to a signicant test) for sample sizes 10, 100, 1000 and 10000. Each group must
include in its report both results (Mortal and God procedures).
10. We have the following series or data
n r
2 2
3 3
4 -2
3 -3
we consider the following model n
i
=
0
+
1
r
i
+n
i
where n
i
is an identical and indepen-
dent random variable. Additionally, the mean of n
i
is equal to zero.
(a) Calculate the OLS estimates (
b
0
and
b
1
)
(b) Verify the orthogonality conditions i.e
P
N
i=1
b n
i
= 0 and
P
N
i=1
b n
i
r
i
= 0
3

Homework 1: PHD Student: Freddy Rojas Cama Homework 1 - Eco 322 Rutgers University

Uploaded by

Copyright:

Available Formats

Homework 1: PHD Student: Freddy Rojas Cama Homework 1 - Eco 322 Rutgers University

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Homework 1: PHD Student: Freddy Rojas Cama Homework 1 - Eco 322 Rutgers University

Uploaded by

Copyright:

Available Formats

PhD student: Freddy Rojas Cama Homework 1 - ECO 322 Rutgers University

. You must choose values for the true parameters, j

You might also like