ECO2008 Assignment 2
UK Trade and the Gravity Model
The UK’s decision to leave the European Union (EU) sparked fresh interest in international trade. In this
project, you are to assess the importance of EU membership on the UK’s exports (in 2014) using a linear
Trade flows are typically modeled using a gravity model. Gravity models assume that there are forces
that attract nations to trade with one another. Namely distance, economic output (GDP), and population.
Cultural and historical factors, like a shared language or colonial ties, are also important as they reduce
In the context of a gravity model, you would expect the UK to export more of its goods to other members
of the EU. Thus, membership of the European Union should act as an attractive force. However, the UK’s
European trade partners are not a random subsample of nations. They are closer and wealthier than other
The UK is also a member of the World Trade Organization (WTO). This is a larger trading bloc of countries, of
which the EU is also a member. Like with the EU membership variable, you would expect WTO membership
to be positively associated with trade flows. But which trading bloc is most important for the UK’s exports
The UKtrade2014.Rdata file is a dataset concentrated on countries that traded with the United Kingdom in
2014. Each observation in these data represents a trade partner. For example, the first row of data is for AFG
which is the isocode for Afghanistan. The variables contained in the dataset are detailed below.
iso Trade partner isocode
imports Total value (in £’s) of imports from trade partner in 2014
exports Total value (in £’s) of imports from trade partner in 2014
EU_d dummy = 1 if in the European Union
WTO_d dummy = 1 if trade partner is in the World Trade Organization
GDP_d GDP (in £’s) of trade partner
POP_d Population (in thousands) of trade partner
Dist_coord Distance to trade partner. This is the distance between London and the
main city of the trading county. So for Ireland (IRL) this will be the
distance in km between Dublin and London
Comlang dummy = 1 if at least 9 percent of the country speak English
Evercol dummy = 1 if colonial link exists
CONTI_d Continent of trade partner
Load the data via the command load("UKtrade2014.Rdata"). You are required to run the following three
ln(Exports) = β0 + β1EU + β2WTO +
ln(Exports) = β0 + β1EU + β2WTO + β3 ln(Distance)+
β4 ln(Population) + β5 ln(GDP ) +
ln(Exports) = β0 + β1EU + β2WTO + β3 ln(Distance)+
β4 ln(Population) + β5 ln(GDP )+
β6CommonLanguage+ β7EverColonized+ .
The above specifications mean that you are required to log transform some of the variables.
Once you run the three regressions create a regression output table using the package stargazer. I have
created a sample regression table, after omitting all coefficient values, standard errors, and other model
information, below. As you can see I have altered a number of the default settings to make the table easier
to read and digest.
Your task is to write a 1,000-word report on the regression model results. Be careful not to mechanically
describe the coefficients. For example, you could easily fill a word document with a 1,000-word description
of the numbers that are in the regression table. However, that would not be very informative and would
undermine the task at hand here. Another common error students make is to assume that demonstrating
that you know how to carry out a hypothesis test is worthy of a high mark. The reality is that you will be
marked on your ability to interpret any such test in a broader context and it is more important to tell me in
practical and analytic terms what any econometric result means rather than to recite the exact degrees of
freedom, or why you are using t distribution in the first place.
Instead, you should think more broadly about what the coefficients mean. Let’s say the coefficient estimate
associated with the logged distance variable is βˆ1 = −0.1. In practical terms, it means that a 1 percent
increase in distance is associated with a 0.1 percent reduction in the value of exports. However, the true
economic meaning of this coefficient depends on several factors, in particular the model specification. What
other explanatory variables are in the model? The importance of model specification implies that comparisons
between model results will provide more depth to any analysis and thus improve the report’s quality. For
example, in what way does the EU coefficient estimate change between the three models AND why does this
occur? Don’t be afraid to speculate in your report and also feel free to pepper your report with statistics and
plots from the dataset, e.g.: “85 percent of the countries in the data are members of the WTO”, and so forth.
The 1,000-word limit will be strictly enforced. You will need to be judicious in your editing and decide what
you think are the most important and insightful pieces of evidence provided by your modeling. There is no
singular “correct” analysis that I expect reports to conform to; instead, it is your task to persuade me of your
interpretation of the results.
The project should be typed in 12 point font size, double spaced, and should not exceed 1,000 words in length
(excluding graphs, tables, equations, R-script commands, and bibliographical references).
The project should contain a cover page with the following information:
• Full name(s) and degree(s)
• Email address(es)
• Student ID number(s)
Each project should have a References section if works are cited. The submission types are restricted to
word files with .doc or .docx extensions.
The R script
The working R code (the script) must be appended to your work. If you do not append your code with your
project, your work will be deemed incomplete, and you will be asked to resubmit with the relevant penalty
for late submission is applied. If any code is omitted, the project will be treated as incomplete.
In practical terms, this means that you need to copy and paste the R file commands into your Word file
containing your project write-up at the end of this project. The code text does not count towards the word
This is a group assignment. You will be assigned into groups of four based on your tutorial class in Canvas.
The assignment process is random, therefore you cannot choose your group mates. When the groups are
created it will be your responsibility to contact and connect with your group members. I recommend that
the first time you meet you specify the roles and responsibilities of each group member. You will need to
complete the “ECO2008 Coursework submission form” and append this to your final report.