Descriptive statistical analysis of all foreign direct investments

Problem 1 (20 %)
Provide a descriptive statistical analysis of all foreign direct investments as % of GDP
(FDITOT). Include a histogram in the analysis. Define the mean, mode, median and the
standard deviation. Comment on the shape and symmetry of the distribution. Provide an
examination for extreme observations.
• Set up and interpret a 95 % confidence interval for the mean. Is FDI in the
Netherlands something special?
Assume that the variable FDITOT is normal distributed with mean and standard deviation as
found.
• What is the probability that a randomly selected country has an FDI share between
10 % and 15 % of GDP?
• What is the probability that a randomly selected country has an FDI share above 35
% of GDP?
• The World Bank wants to draw special attention to the 40 % of the countries with the
lowest share of FDI. What is the maximum share of FDI to be considered for
attention? Does the result obtained provide any information regarding the distribution
of FDI by countries over the world?
Problem 2 (15 %)
Foreign direct investments may be positively related to the GDP per capita. Examine this
issue by use of simple linear regression and let FDITOT be explained by GDP. The model
has the form:
FDITOTi = β0 + β1GDPi + εi i = 1, …, 157
Explain the applied method and assumptions. Undertake a model control, and inspect the plot
of residuals, and the plot for normality. Is white noise observed? Is this a good model when
you look at the coefficient of determination? Can the hypothesized relation among the two
variables be confirmed? Use the model to predict the share of FDI for a country with a GDP
per capita equal to 14 (1,000). This level of GDP per capita is equal to the mean world GDP
per capita.
Page 4 of 5
Problem 3 (15 %)
Expand the model and apply multiple regression to examine which variables influence on
FDITOT. Examine the impact of the variables GDP, OPEN, AGICUL, INTERNET, HEALT,
PEACE, MOBILE and LIFEEXP. Set up a matrix of correlation and comment. Undertake a
selection of the variables and find the most suitable model. Provide an interpretation of the
signs and comment on relevant plots of residuals. Compare with the model estimated under
problem 2.
Problem 4 (15 %)
The establishment of a new firm for example an FDI by a foreign firm may be harder if
bureaucracy harms the process. This can be the case in low income countries where the tax
and administrative system is not well developed.
To examine this statement, set up a cross-table where the 157 countries on one hand is divided
by the variable LEVEL into five categories, and on the other hand is divided into three
categories given by the variable STARTUP.
Set up the relevant hypothesis, describe the test by formulas, and undertake the test. Assume
a 5 percent level of significance. What is the outcome of the test? Discuss the validity of the
assumptions for the test and provide an adjustment of the test if necessary.
Finally, calculate the share of Least Developed Countries with the lowest start up procedures.
Set up a 95 % confidence interval for this share. It has been claimed that it is fairly easy to
start up businesses in the LDC’s, and the level of start up procedures is low in 40 % of the
countries. Is this statement true?
Problem 5 (10 %)
Access to a reliable network for mobile telephone has turned out to be of importance for the
establishment of especially small businesses located in many remote places in Africa and
Asia. To examine the importance of the infrastructure relative to mobile telephone networks,
consider the variable MOBILE. Use the variable GEO to create two data series for the mobile
cellular subscriptions (per 100 persons); one for Africa setting GEO = 1 and one for ASIA
setting GEO = 3.
Test a hypothesis stating the mean mobile cellular subscriptions (per 100 people) is higher in
Asia than in Africa. This implies that the investment climate is better in Asia than in Africa.
Page 5 of 5
Set up the hypotheses and perform the test. What is the outcome and how is the p-value
defined? Perform a test to examine if the variances are equal among the two data series.
Problem 6 (15 %)
The emergence of the Internet has increased the accessibility of communication all over the
globe. Just as the railway and steamers connected the world 150 years ago during the age of
colonialism. However, the Internet may be more developed in some regions than in others.
To examine this issue, use the variable INTERNET divided into the four groups given by the
variable GEO. List the assumptions of the applied method, set up the hypotheses investigated,
and conduct the test. What is the outcome? If necessary provide a supplementary analysis,
and comment on the result. Is the statement confirmed?
Problem 7 (10 %)
The themes addressed in the present exercise constitutes a comprehensive insight into the
activities of Foreign Direct Investment throughout the world. The board of the World Bank
is very interested in the results that you have provided and wants to distribute the results in a
report written in popular language. However, the board members are not competent in reading
statistical analyses. Therefore, you have to provide an easy to read and eloquent statement
summing up the main results from your investigation in order to tell them about the outcome
of your findings.
In the statement you could for example consider the mean share of the FDI over the World.
How is FDI distributed among the countries, and is this distribution symmetric? How likely
is it to have an FDI share between 10 % and 15 % of GDP or an FDI share above 35 % of
GDP?
What influences the FDI share? Is it economic conditions like GDP, the openness of the
economy or the size of the agricultural sector? Or is it more infrastructural variables like the
Internet or issues related to the political stability as the expressed by the peace index? Relative
to the infrastructural dimension how is the accessibility of telephone- and internetworks
related to the geographical dimension throughout the world?
Institutional factors may also be of importance for example relative to the level of income. Is
it likely that the bureaucracy is more significant in the least developed countries relative to
the high-income countries, or is it the other way around?