Posted: February 8th, 2015

Econometrics group report using the computer package eviews,also estimation of OLS,misspecifacion testing and interpretation of results and procedures.

Econometrics group report using the computer package eviews,also estimation of OLS,misspecifacion testing and interpretation of results and procedures.

Project description
(1) Economic theory suggests that consumption is positively related to income
and negatively related to unemployment, , (proxying income uncertainty) and real interest rates, , (intertemporal substitution). Each group should estimate the following four models (based upon this consumer theory) by ordinary least squares (OLS) using the sample period 1957 1994. Each individual student in the group is responsible for estimating, reporting and discussing one of these four models it is up to the group to decide which group member will be responsible for each model.

Econometrics (EC6005): assignment 2014 – 2015
This assignment provides an opportunity to develop and demonstrate skills using the
computer package E-Views, which is a useful tool for econometric analysis. The
exercise will also provide an insight into the application and interpretation of
fundamental econometric methods such as estimation by OLS, misspecification
testing and interpretation of results and procedures. It will also provide an
opportunity for students to work in small groups.
1. Instructions for forming and managing the small groups
(a) The purpose of the assessment is to give students the opportunity to work with
others in the production of a report which is a skill that is required in many work
environments. This will involve the division and coordination of tasks within the
group, time management, discipline and consideration of, and cooperation with,
other members of the group. This group assessment also seeks to entrench
subject knowledge through the discussion and exchange of ideas within the
group.
(b) Each individual member within the group will be responsible for the estimation,
interpretation and reporting of one of the four models specified in (1) below to
be estimated. A mark for each individual member of the group will be assigned
for this work and each member of the group may receive a different mark for this
component of the work. There will also be a group component to the work for
which all group members will receive the same mark and are expected to
allocate the division of work to reflect this. All members of a group must specify
their student numbers on the coversheet of the submitted coursework (student
names should not appear anywhere on the report to help ensure anonymity of
marking). Each individual group member should also write their student number
at the start of their section of work and each student’s individual component of
work should be written on a new page in the report. Only one piece of work
should be submitted for each group – each member of the group should not
submit separate assessments.
(c) Students will be given the opportunity to specify their group containing exactly
four members (there is NO negotiation on this) by the end (Friday, 5pm) of
teaching week 9 (28 November 2014). This should be done by email to Chris
Stewart ([email protected]) and must have all members of the group
copied in to the email. Any students who have not correctly provided notification
of their group by this deadline will be allocated to a group by the teaching team.
Any group specified with more or less than 4 students or without all members
copied in to the notification email will be regarded as not providing correct
notification of the group. (If total student numbers are not divisible by four the
teaching team may allocate one or two groups with 3 members and will advise
the affected group of how the assessment will be modified to accommodate
this).
1
(d) Each member of the group is required to negotiate their roles and
responsibilities in the execution of the task with other members of the group.
(e) Group members should hold regular group meetings.
(f) Any problems arising within the group that cannot be resolved by the group must
be communicated to the teaching team by the end (Friday 5pm) of teaching
week 13 (9 January 2014). The teaching team’s decision on the resolution of such
problems will be final. Problems arising within the group after the end of
teaching week 13 must be dealt with by the group. The teaching team must be
notified of any member of the group who cannot contribute to the work at all
due to illness or other genuinely serious circumstance. In this case the teaching
team may decide that that student should be excluded from the group (and so
not receive a mark for the assessment), and this student will not be permitted to
join any other group.1 The remaining members of the group will submit the
assessment and only they will receive a mark – no other student will be allowed
to join their group. The teaching team will advise the remaining group members
of how the assessment will be modified to accommodate the reduction in group
numbers. Only the teaching team can make decisions on whether a member may
drop out from a group for this reason.
(g) Groups are required to produce and tabulate econometric results (the 4
estimated models produced by each individual group member plus the model to
be estimated by the whole group). These should be discussed in a report of a
maximum of 3500 words (excluding the results table, bibliography and EViews
computer output in the appendix). Each group member can write a maximum of
700 words on the discussion of the model that they have been assigned to
estimate and the discussion in the group component can be a maximum of 700
words. A word count must be provided for each of the 4 individual components
of the report and for the group component of the report (giving 5 word counts).
Any text for which you wish to receive marks must be included in the main body
of the work and in the word counts. Including text in appendices that is not
included in the word counts will not receive any marks. What is expected from
the task is specified below.
The reports submitted by different groups must be distinct and are subject to
considerations of collusion and/or plagiarism according to standard practice (do not
leave yourself open to accusations of academic misconduct).2 This assignment
represents 40% of each student’s overall mark for the module – the individual
student component represents 20% of the overall mark and the group component
20% of the overall mark.
1
A student who has been established to have made no contribution at all to the work submitted by
the group (for whatever reason) may also be awarded a mark of zero.
2
Be careful to properly attribute any material you use from existing sources, including texts and
tables. Remember that all work is electronically submitted through Turnitin which will compare how
similar your work is with other current and previous students’ work and published books and journal
articles worldwide. Similarities are easily detected.
2
2. General Information on the assignment
The deadline for handing in this assignment is 11am Wednesday, 11 February
2015. It should be submitted electronically through Turnitin according to University
procedures. Specify the student numbers of all four members of the group on the
coversheet of the report and each student should specify their student number at
the top of the section of the report for which they are solely responsible.
Annual data from 1955 to 1994 for 10 OECD countries will be provided in an Excel
file on the Econometrics StudySpace website. Each group will be allocated a different
data set, where each data set is contained in its own folder in the Excel spreadsheet.
Each group will be required to estimate consumption functions for one country. The
variables are the natural logarithm of real per-capita consumption
, the
natural logarithm of real per-capita income
, the real interest rate
and
the rate of unemployment
– the latter two series are expressed as
proportions. The * denotes the country identifier which is different for each folder. 3
You should not transform any of this data (for example you should not take logs of
the data). You are to transfer this data to an E-Views workfile.
3. Assignment component for individual group members
(1) Economic theory suggests that consumption is positively related to income
and negatively related to unemployment,
, (proxying income uncertainty)
and real interest rates,
, (intertemporal substitution). Each group should
estimate the following four models (based upon this consumer theory) by
ordinary least squares (OLS) using the sample period 1957 – 1994. Each
individual student in the group is responsible for estimating, reporting and
discussing one of these four models – it is up to the group to decide which group
member will be responsible for each model.
(a)
(Model 1)
(b)
(Model 2)
(c)
(Model 3)
(d)
(Model 4)
3
The 10 countries for which data is provided (along with the country identifier, *, given in
parentheses) are: Australia (AUL), Belgium (BEL), Denmark (DEN), France (FRA), Germany (GER),
Greece (GRE), Japan (JAP), Netherlands (NET), Norway (NOR) and Spain (SPA).
3
where the
are population coefficients to be estimated and
is a stochastic
error term. The subscripts, and
, refer to the time period (this period and
last period, respectively) and is the difference operator such that
.
(2) After estimating their model each group member should test for the following
forms of misspecification in their model:
?
?
?
?
?
?
?
first-order autocorrelation (DW and Breusch-Godfrey tests, as appropriate),
second-order autocorrelation (Breusch-Godfrey test),
non-linear functional form (Ramsey’s RESET test with one fitted term),
non-normally distributed residuals (Jarque-Bera test),
heteroscedasticity (White’s test, without cross terms),
first-order ARCH effects,
structural instability (use Chow’s test with the breakpoint in 1976).
Each group member should summarise the results of the model that they are
responsible for in a table (where all four models should be reported in this single
table along with the model to be estimated by the whole group). The main
results to be tabulated for all models include all estimated coefficients and their
corresponding t-ratios, the coefficient of determination, the adjusted coefficient
of determination, the regression standard error, the F-test for the overall
significance of the regression (along with its probability value) and the eight
misspecification tests (listed above), where the probability values for these
misspecification tests should also be reported. This table should be placed at the
front of the report. See the example table in the Advice on Summarising Your
Results section below for guidance on how to report your estimation results.
(3) Each group member should also fully discuss and interpret all of the reported
estimated results for their model (in both economic and statistical/econometric
terms). This means:
?
?
?
?
?
testing whether each estimated coefficient is statistically significant (using a
two-tailed test),
testing whether the model has significant explanatory power (using an Ftest),
commenting on whether the model fits the data well according to the
coefficient of determination, the adjusted coefficient of determination and
the regression standard error,
discussing whether the estimated coefficient signs are consistent with
economic theoretical expectations (as outlined in (1) above) and how much
the dependent variable changes when each explanatory variable changes
(the marginal effects),
discussing whether the model exhibits evident misspecification according to
the eight misspecification tests discussed above. Based upon these tests
comment on whether their model satisfies the underlying assumptions of the
4
linear regression model and whether, and in what sense, OLS estimation of
these models is valid.
Use a 5% level for all tests (misspecification tests, t-tests and F-tests).
4. Assignment component for the whole group
(4) As a group you should do the following.
(a) Using Model 3, specified in 1(c) above as your starting point, apply the
general-to-specific methodology to obtain a parsimonious model (Model 5)
that is valid for inference. If more than one parsimonious model is valid for
inference, select only one of them as your favoured specification for drawing
inferences – discuss how you decided which one to choose. Do not exclude
the intercept from your parsimonious model (even if it is statistically
insignificant). You should:
?
?
?
?
?
add the main estimation results of Model 5 to the table where the results
of the four models specified in (1) are reported,4
apply all eight misspecification tests indicated in (1) to this parsimonious
model,
apply an F-test for the removal of variables from Model 3 specified in 1(c)
required to obtain the parsimonious specification (Model 5). This should
also be reported,
fully justify why you have chosen the parsimonious model that you
report,
fully discuss and interpret all of the reported estimated results for this
parsimonious model (in both economic and statistical/econometric
terms).
(b) Using Model 5 test whether the coefficient on
different from one.
is significantly
(c) Derive the long-run static solution of this model (the equilibrium). Show the
steps of this derivation and report the final equilibrium equation. Using these
results for the equilibrium model comment on the long run coefficients and
provide an economic interpretation of them.
(d) Provide a critical analysis of what you have done by discussing any potential
shortcomings of the modelling strategy employed and suggest any possible
improvements.
4
If there is more than one competing model for the parsimonious model or you feel it is useful in
justifying why you select the parsimonious model that you do you may report two versions of model
5. However, you must select one as the parsimonious model for deriving the equilibrium and you
must discuss why you select the model that you do.
5
5. Instructions on requirements for the report
(5) Write a report presenting the main results in a table and provide an analysis of
these results. Your report should include the following.
(a) A single table summarising the main estimation results of the five models
that you are asked to report. Place this at the front of the report.
(b) Five sections discussing the reported results – there should be 1 section for
each of the models estimated by the four individual group members and
one section for the group task specified in (4). The discussion should be
based on the issues and questions outlined above. This group section should
include a discussion of how you obtained your Model 5 (refer to the
general-to-specific method) and commentary on the derived equilibrium for
your favoured model for inference.
(c) Computer output of the results for Model 5 (only) should be placed in an
appendix. No computer output should enter the main text of your report –
you will lose marks if it does (without exception).
(d) A bibliography listing any sources cited in your report or referred to in
producing your work.
The whole group report should be no longer than 3,500 words (excluding computer
output, the table summarising statistical results and bibliographical references). Each
of the 5 sections of the report (one for each of the 4 individual group members and
one for the group as a whole) should be no longer than 700 words. These maximum
word limits will be strictly applied. Any reports exceeding these word limits will
automatically lose marks. The following rule will apply to any section exceeding its
word limit (or not reporting an accurate word limit): its mark will be reduced by up
to ten percentage points (depending upon the number of words over the limit). For
example, a section receiving a mark of 60% that exceeds the word limit can have its
mark reduced to 50%. In particular, you will lose one percentage point for up to each
50 words over the limit. For example, a section that is 850 words long will receive a 3
percentage point penalty while a report that is 851 words long will have its marked
reduced by 4 percentage points. You are REQUIRED to provide a screen shot of the
word count in your assignment (this will not include words featured in your table of
results, the bibliography or the computer output in the appendix, however, other
written work for which you wish to receive marks, even if included in an appendix,
should be included in the word count). This word count can be obtained using Word
by highlighting the relevant section of text and viewing the word count in the
bottom left-hand side of the screen. When the word count is displayed press the
Print Screen button to copy the screen and paste this onto a page at the front of the
section to which it refers. Assignments without word counts or with incorrect word
counts will not receive a mark. Do not reproduce proofs of econometric results or
overly long discussions of econometric theory in your report. Show that you
understand econometrics in your interpretation of the results.
6
6. Advice on Summarising Your Results
You are asked to summarise the results of your analysis in a table. An example of
how to tabulate the estimated results for all four models in a single table is as
follows (the reported numbers are just for illustration):
Intercept
LG*t
RI*t
U*t
LC*t–1
LG*t–1
RI*t–1
U*t–1
LC*t–1
LG*t–1
RI*t–1
U*t–1
Model 1
1.374 (1.161)
0.867 (11.578)
–0.126 (–0.183)
0.335 (0.348)
Model 2
–0.125 (–0.215)
0.285 (3.294)
0.651 (1.939)
1.247 (1.277)
–0.625 (–7.236)
0.081 (0.821)
–0.111 (–0.235)
–0.937 (–1.127)
Model 3
–0.709 (–0.984)
0.232 (3.685)
0.838 (2.004)
1.032 (1.684)
–0.589 (–9.571)
0.043 (0.201)
–0.190 (–0.395)
–1.819 (–2.882)
0.250 (1.376)
–0.156 (1.261)
–0.011 (–0.035)
–0.376 (–0.745)
Model 4
–0.107 (–0.194)
0.332 (5.365)
0.563 (1.404)
1.819 (1.604)
–0.658 (–10.051)
1.376 (1.517)
Model 5
0.789 (0.904)
0.328 (5.605)
–0.597 (–10.132)
1.268 (2.018)
–1.908 (–2.275)
Fit Measures
0.989753
0.997657
0.967543
0.985473
0.975803
0.976537
0.963647
0.980467
s
0.020450
0.020138
0.020146
0.020111
2
F(R =0)
271.776
392.225
199.050
219.570
[0.000]
[0.000]
[0.000]
[0.000]
Misspecification Tests
DW
1.653
1.890
1.930
1.746
2.032
LMA1
3.976
6.074
2.725
0.428
0.280
[0.044]
[0.018]
[0.106]
[0.513]
[0.618]
LMA2
0.007
1.239
1.005
1.502
1.025
[0.933]
[0.272]
[0.322]
[0.195]
[0.235]
LRF1
1.247
2.114
1.502
4.732
1.162
[0.212]
[0.347]
[0.220]
[0.094]
[0.281]
LMN2
5.648
1.088
2.114
0.430
4.732
[0.077]
[0.389]
[0.347]
[0.825]
[0.094]
LMH
4.233
0.413
7.678
0.638
2.339
[0.023]
[0.524]
[0.362]
[0.429]
[0.801]
LMARCH1
3.678
1.502
0.428
1.244
0.657
[0.031]
[0.195]
[0.513]
[0.308]
[0.418]
FCHOW
1.234
6.468
1.502
2.991
1.244
[0.543]
[0.011]
[0.195]
[0.084]
[0.308]
Table notes. The dependent variable is LC*t and figures in parentheses below coefficients are
estimated t-ratios.
is the coefficient of determination,
is the coefficient of determination
2
adjusted for degrees of freedom, s is the regression standard error and F(R =0) is the F-test for
deleting all slope coefficients from the model. DW is the Durbin-Watson statistic, LMA1 is the
lagrange multiplier (LM) test for first order autocorrelation and LMA2 is the LM test for second order
autocorrelation. LRF1 is the LR-test for non-linear functional form, LMN2 is the LM test for departures
from normality, LMH is the LM-test for heteroscedasticity, LMARCH1 is the LM-test for first-order
ARCH effects and FCHOW is Chow’s first (F-) test for structural stability with a breakpoint in 1976.
Probability values for selected tests are reported in squared brackets below the statistic to which they
correspond.
0.912546
0.883264
0.044918
119.540
[0.000]
7

LCAUL    LGAUL    RIAUL    UAUL
1955    -4.963391912    -4.529147609    -0.00365047    0.008
1956    -4.979143214    -4.534989112    -0.016473327    0.014
1957    -4.968408807    -4.535553087    0.02527399    0.02
1958    -4.95766514    -4.484919837    0.020099058    0.026
1959    -4.913508318    -4.454582167    0.012332731    0.025
1960    -4.916787261    -4.442255375    -0.00034099    0.019
1961    -4.91735895    -4.45629921    0.037887135    0.035
1962    -4.874894184    -4.408189306    0.029734697    0.028
1963    -4.834601754    -4.358899511    0.020612768    0.027
1964    -4.808839902    -4.309709375    0.005266639    0.017
1965    -4.799443478    -4.309099342    0.00976653    0.015
1966    -4.769729736    -4.263336731    0.01213376    0.017
1967    -4.73319618    -4.243716003    0.006752932    0.019
1968    -4.699700487    -4.177233949    0.00740025    0.018
1969    -4.662664517    -4.143238402    0.006277374    0.018
1970    -4.641505856    -4.1092026    -0.007159499    0.016
1971    -4.634188408    -4.092994624    -0.012328351    0.019
1972    -4.598598504    -4.069422508    -0.019016604    0.026
1973    -4.561439422    -4.045059112    -0.064096603    0.023
1974    -4.552716252    -4.049910064    -0.078482879    0.026
1975    -4.533308384    -4.036259052    -0.073490133    0.048
1976    -4.519417619    -4.017557766    -0.036854883    0.047
1977    -4.512149768    -4.0191861    -0.006330422    0.056
1978    -4.489481574    -3.983160784    -0.005790306    0.062
1979    -4.477070183    -3.971849048    -0.013020583    0.062
1980    -4.449347435    -3.949837904    0.008168508    0.06
1981    -4.423848906    -3.942254219    0.034398207    0.057
1982    -4.423238472    -3.9761962    0.039072541    0.071
1983    -4.41712538    -3.931871053    0.028066489    0.099
1984    -4.393121373    -3.893645356    0.045661545    0.089
1985    -4.370916949    -3.871323628    0.065494058    0.082
1986    -4.378741151    -3.859139216    0.06032164    0.08
1987    -4.353564697    -3.823630357    0.051730597    0.08
1988    -4.32643979    -3.797068095    0.048596145    0.072
1989    -4.296850373    -3.784298147    0.093225172    0.061
1990    -4.301749488    -3.806288292    0.083119594    0.069
1991    -4.289582511    -3.816031766    0.071424029    0.095
1992    -4.276121645    -3.795690296    0.044071681    0.107
1993    -4.257964651    -3.770257643    0.028194171    0.108
1994    -4.225106295    -3.729690774    0.041646798    0.096