Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: THEend8_
EC1B1 Coursework
1 Overview of the Project
• We will ask you to replicate a number of facts about exchange rate and output volatility, before
and after various countries departed from the Bretton Woods system of fixed exchange rates, as
discussed in Lecture 5.
• This replication will be based on Itskokhi & Mukhin (2021). Itskokhi & Mukhin study the volatility
of exchange rates and output for 8 countries. In Lecture 5 we studied their results for the United
States. You will show that similar results hold for other countries.
• We have assigned students into groups of 4 in order to carry out this work, and assigned a country
whose data each group will study.
• The groups allocations are in an excel file on Moodle in the Coursework module. Please use this
to identify and liaise with your fellow group members.
• The course manager will email all groups to assign them their countries.
2 Guidance
• Free riding. Though you are assigned into groups of 4, you are free to divide the work however
you think is most efficient and fair.
– However, we will not tolerate free riding. All work will be submitted with a coversheet, signed
by each of you, that needs to describe how the project tasks were divided between the group
members.
– If some of the group feel that there has not been an equal contribution from certain members,
please liaise with the course manager.
• Submission material. One member from each group needs to upload all the following documents
onto the Moodle submission portal:
– A Jupyter notebook containing all code and output
∗ This needs to show the code for your data cleaning and also include any code to generate
graphs you use in your answers
∗ Ensure your notebook is well commented and be clear what your code is doing and how
the outputs you present are generated
– A document containing the answers to all the questions presented in Section 5 - Questions
and Discussion
∗ This can be in any format, but a word or PDF file seems sensible
1
∗ You should include your graphs where relevant
∗ Answers do not need to be lengthy; markers will be rewarding precision both in your
written responses and your figures
– The completed coursework submission coversheet, signed by all group members
• Grading. When working with data, there is no “right answer”. There are often many valid ways
to proceed, we don’t expect that the final graphs and statistics will be identical to the numbers
and graphs from the paper.
– This coursework will account for 15% of your total EC1B1 grade. Your work will
be graded based on:
1. How well documented and clear the code is (25%)
2. The quality of your responses to the questions in Sections 5.1 and 5.2 (15%)
3. The quality of your figures and responses to to the questions in Section 5.3 (50%)
∗ You should ensure your graphs follow principles of good graphical design (see Schwabish
(2014)) and that your tables are well formatted
4. Whether your results seem reasonable and are in the right “ballpark” (10%)
∗ Note that we will not deduct marks if your graph or numerical values are not exactly
identical to ours
– We will run through plagiarism detection software to ensure that each group has done the
exercise independently. We will deal with suspected plagiarism in accordance with the LSE
plagiarism policy.
• Timelines. Final submission of the coursework is due at the end of W8, Sunday 13th March
by 17:00. Penalties will be applied for late submission, please liaise with the course manager if
you encounter any major issues.
– Whilst you are free to plan out your own schedule, some indicative timings you might consider
as a starting point are:
∗ W6: Assigned groups and countries
∗ W7 (first half): First meeting, division of tasks, data cleaning
∗ W7 - W8: Second meeting, data “sense check”, answering 5.1 & 5.2, producing figures
∗ W8 (second half): Third meeting, review of figures, finalising answers to 5.3, final review
3 Preparation
• Read the Introduction and Section 2 of “Mussa Puzzle Redux” by Oleg Itskokhi and LSE’s very
own Dima Mukhin!
• Read “An Economist’s Guide to Visualizing Data” by Jonathan Schwabish.
4 Data Cleaning Guidance
• Downloading the data.
– Navigate to the International Monetary Fund’s “International Financial Statistics” website.
2
– Download monthly data from January 1960 to December 1990 for the country you have been
assigned. In particular, download the series for:
(i) Industrial production
(ii) Exchange rates relative to the US dollar
(iii) Consumer prices.
– Also, download the consumer price index for the United States.
– Hints:
∗ You may find it helpful to create a free account on this website so you can save your
queries.
∗ You will save time in Python if you transpose the table before you download it (i.e. flip
the rows and columns around).
• Cleaning the data.
– Import the dataset of variables for your country into Python. Import the dataset of consumer
prices for the United States. Merge these datasets at the month level.
– Construct variables for your country for:
(i) The log exchange rate
(ii) The difference in the log exchange rate versus the previous month
(iii) The inflation rate versus the previous month
(iv) Log industrial production
(v) Growth in industrial production versus the previous month and
(vi) Growth in industrial production versus 12 months ago.
– Also, for the US, construct: (i) log consumer prices and (ii) inflation versus the previous
month.
– Identify outliers - is there a small number of implausibly extreme observations within any
of the data series? (E.g. less than 5 extreme observations in the data series.) Identify the
outliers in each data series and set them to missing.
– Interpolate missing data - for each series, replace any gaps in the data (including the
data you set to missing because they were outliers) with the mean of the value of the series
before and after the gap.
∗ For example, if your measure of inflation is missing for January 1965, replace the value
of inflation with the mean of inflation in December 1964 and February 1965.
– Construct a series for monthly real exchange exchange rate growth, where the real exchange
rate is measured for your country relative to the United States.
– Format the monthly variable into a date format.
– Hints:
∗ The International Financial Statistics will only let you export the data into Microsoft
Excel. You can use the read_excel function from the pandas module to read the excel
file into a DataFrame object.
∗ You may want to use the datetime module to format dates.
∗ The diff function can help you difference the data - but be careful! Make sure your
dataset is sorted correctly for this function to work.
3
∗ Real exchange rates are defined in Lecture 5.
∗ You may find that the downloaded Excel data has some blank columns. Use the iloc
function to get rid of these blank columns.
• End product.
– Print in your Jupyter output the full merged dataset.
– Hint: Jupyter’s default is to only print the first 5 rows, you can use pd.options.display.max_rows
to change the default.
5 Questions and Discussion
5.1 Comprehension and Warm Up
• What was the date that your country left the Bretton Woods system?
• Describe the sense in which the departure from Bretton Woods represents a “natural experiment”
about the effects of real exchange rate fluctuations on the macroeconomy.
5.2 Cleaning
• How many monthly observations are there in your dataset? Is this the number you were expecting?
• Why are we studying monthly data? Why not some lower-frequency data such as quarterly or
annual data?
• What is industrial production? Why are we studying industrial production, instead of another
series such as GDP?
• What measure did you choose for inflation and why?
• Why are you dropping outliers?
• Discuss some pros and cons of the interpolation procedure we are using
• Discuss any other issues that arose while cleaning the data and how you dealt with them.
5.3 Analysis
5.3.1 Exchange Rate Graphs
• Plot a time series graph of the monthly growth in nominal exchange rates of your country,
versus the US dollar, for every month from the start to the end of the sample. Clearly indicate
the time at which that country left the Bretton Woods system.
• Plot a time series graph of themonthly growth in real exchange rates of your country, versus
the US dollar, for every month from the start to the end of the sample. Clearly indicate the time
at which that country left the Bretton Woods system.
• Why is it useful to plot both real and nominal exchange rate growth?
4
5.3.2 Inflation and Industrial Production Graphs
• Plot a time series graph of monthly inflation, for every month from the start to the end of the
sample. Clearly indicate the time at which that country left the Bretton Woods system.
• Plot a time series graph of the monthly growth in industrial production, for every month
from the start to the end of the sample. Clearly indicate the time at which that country left the
Bretton Woods system.
• Plot a time series graph of the growth in industrial production versus 12 months ago, for
every month from the start to the end of the sample. Clearly indicate the time at which that
country left the Bretton Woods system.
• Why are your results for the monthly versus 12 monthly growth in industrial production so dif-
ferent? Which measure is more useful?
5.3.3 Comparison Statistics
• Separately both before and after your country exited Bretton Woods, calculate the standard
deviation of:
– The monthly growth of nominal exchange rates versus the US dollar
– The monthly growth of real exchange rates versus the US dollar
– The inflation rate in your country
– The difference between the inflation in your country versus the United States
– The 12-monthly industrial production growth
• When you calculate the standard deviation of variables before Bretton Woods, exclude data from
the year immediately before the country left Bretton Woods. Likewise, when you calculate the
standard deviation of variables after Bretton Woods, exclude data from the year immediately after
the country left.
• Report these numbers in a suitably formatted table, as well as the ratio of the standard deviation
of each variable before and after Bretton Woods. Why should we exclude data from the year in
which the country departed from Bretton Woods?
5.3.4 What Are The Effects of Real Exchange Rates?
• Taken together, what do your results imply about the effect of real exchange rate fluctuations? Is
there a reason why your results are particularly compelling?