Scribd
Upload
40 views1 page

Panel Data Question

The document discusses the analysis of the 'xta2021' panel dataset concerning Tanzania's pension schemes from 2004 to 2015, focusing on various economic indicators. It covers topics such as the use of panel data, regression analysis, model selection between fixed and random effects, structural breaks, and estimation methods. The document also emphasizes the importance of interpreting results and understanding the implications of different statistical models in the context of pension scheme performance.

Uploaded by

omarykatunda3rd
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
40 views1 page

Panel Data Question

The document discusses the analysis of the 'xta2021' panel dataset concerning Tanzania's pension schemes from 2004 to 2015, focusing on various economic indicators. It covers topics such as the use of panel data, regression analysis, model selection between fixed and random effects, structural breaks, and estimation methods. The document also emphasizes the importance of interpreting results and understanding the implications of different statistical models in the context of pension scheme performance.

Uploaded by

omarykatunda3rd
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd

QUESTION FOUR: Panel data analysis

“xta2021” is a dataset of six Tanzania previous existing pension schemes, from 2004 to
[Link] dataset is rich in terms of the annual information related to gross domestic product
(gdp) in billions Tshs, inflation(inflation), annual operation cost in billions Tshs incurred by
each pension scheme (cost_opex) , income obtained by each scheme from investment in billions
Tshs (y_inv) , total number of members each scheme had (t_participants), schemes number of
investors (inv_participants), annual monthly average contribution (mcontr) and monetary
benefits (benefit) in billions Tshs. Using this dataset named as “xta2021” , answer the
following:

(i) Why do we use panel data? What do we mean when we say a panel data is unbalanced
and what does it inform us? Using this dataset “xta2021” we have, Is this panel
balanced? Which method is then appropriate to estimate the suitable model?
(ii) With the help of p-values, determine the integrating order of each of these variables.
Hint: present the information in an appropriate form.
(iii) Run the appropriate regression and present your results using this panel data

model format; y α β x β x +η +ε
it = 0+ 1 1 t+ 2 2t i t
, where y stands for cost_opex,
which is the dependent variable and x’s stands for independent variables named as
income obtained by each scheme from investment (y_inv) , total number of members
each scheme had (t_participants), schemes number of investors (inv_participants),
annual monthly average contribution (mcontr) and monetary benefits obtained annually
in terms of billions (benefit). What interpretations can you make regarding over all model
matters?
(iv)Under what circumstances is the Random effect model better than Fixed effect model?
What criteria would you use in choosing between FE and RE model? Distinguish
between the two. Do the results suggest random effect or fixed effect model? Give
reasons.
(v) Distinguish between static and dynamic model write up and indicate their theoretical and
practical estimation procedures/methods.
(vi) When do we test for structural break and what commands suits this test? While using
2000 as the separation time, use the chow (1960) test statistic of testing for parameter
constancy for a single known structural break in a single time series equation model.
Test for the structural break for period one of 1980 to 2000 and period two of 2000 to
2019 and comment.

(vii) η
What is the meaning of the variables i ? What do you think can be captured by this
variable and using the results above what could you conclude on our results.
(viii) Interpret results related to sigma-u, sigma-e, rho and number of
observations values and key significant values.
(ix) Explain why the first difference (FD) estimator is preferred to the within-group (WG)
estimator when considering dynamic panel data models which controls for fixed
effects. Estimate the two estimators and comment on the results.

You might also like