You are expected to write the report (aim for 1200 words +-10%) to read, look and feel like a FT article (20% of the marks). The do-file of STATA has to be in the appendix of the report.
The report has to answer the following questions and data provided in Blackboard:
The Solow model predicts that growth rates depend on the starting level of GDP, the population growth rate and the rate of investment such that:
π‘”π‘Ÿπ‘œπ‘€π‘‘β„Žπ‘– = 𝛼+ π›½π‘–π‘›π‘£π‘’π‘ π‘‘π‘šπ‘’π‘›π‘‘π‘– +πœ€π‘–
1) Obtain estimates of the OLS estimators Ξ± and Ξ² for the regression only for the developed (developing) countries if your ID number is odd (even) for the year 2010.
The growth rate is calculated as the average growth rate of real GDP ratio between 19702010. All variables are in natural logs.
2) Add the variable population growth, which is the average population growth rate between 1970-2010 and the GDP per capita of the year 1970 in your model and estimate it again. Comment on the sign and significance of all the coefficient estimates. Outline the economic intuition underlying your results. Are they consistent with empirical literature? Explain.
3) For each of the following questions formulate a null hypothesis and test your equation:
. β€’ Investment is the only determinant of growth
β€’ The relationship between the population growth rate and economic growth is equal to 0.2
β€’ GDP in 1970 and population growth have the same effect on economic growth.
4) After investigating the related empirical literature, add an extra variable from the dataset that you have not used in your model. What are the estimates? Are they line in it?
5) Test for Heteroskedastisity. If it is present, correct it.
6) Divide the countries of your sample in regions/continents. Estimate your model adding the continents’ variable in the model. Do you find any differences? Explain.
7) Interact the continent dummies with the variable investment. Comment and Interpret the coefficients.
8) Discuss whether endogeneity issues are present. If yes, how you can solve them, discuss using the existing literature.
9) Run your model including all the years from 1970 to 2010. Illustrate and discuss the model and the suitable estimation method you used. Do you find any different estimates from the question (2)? Explain.