Format: The first page must give the project title, your candidate number, and a word count. No personal identifiers should be included anywhere in your essay. The essay should be no less than 1500 and no more than 2,000 words long. The bibliography and the title page are not part of the word count. Any words beyond 2,000 will not be read. You must reference all sources and you must include a separate bibliography.
You are being asked to create graphs, perform calculations, and analyse the graphs based on the lecture material, as well as on relevant journal articles. Make sure that you label all graphs properly. Both the X and the two Y1 and Y2 axes should be labelled. There should be a legend showing each series on the graph. In short, someone who hasn’t read the questions should be able to understand what is shown in your graphs.
Your assignment output is a paper on a specific topic of your choice. Before you write your paper, you need to do the 9 preparatory tasks below:
- Preparation
- Based on the lecture on Economic Institutions, the accompanying videos and seminar discussions, formulate a testable hypothesis on how a certain economic institution (or its reform) affects macroeconomic outcomes, e.g. economic growth or income levels, inequality, consumption or any other variable you are interested in. Why do you think this hypothesis is interesting to look at? Write down your answer.
- Pick a country or, preferably, a set of countries of your choice. It can be the country you were born in, or any other which attracts your interest, or your own country among a larger group of similar countries.
- Justify your intuition. Explain why you think your hypothesis is correct in your country/countries of choice.
To do the explanation properly, you need to consult the economic literature, and summarise it. For example:
- if your hypothesis is that improvement in property rights raises living standards, then you need to do a literature search on ‘property rights’ and ‘economic growth’ and/or ‘living standards’.
- If your statement is that corruption reduces living standards, you need to search for ‘corruption’ and ‘living standards’ in the scientific literature databases (jstor, sciencedirect, proquest, web of knowledge, and only then – google scholar, repec). Any hypothesis in 1) contains the key words used to start your literature search.
Pick not only the most cited but also the most recent papers informing your hypothesis, ideally published in good economic journals. To get a rough idea of what a good journal is in economics, you need run a search for top economics journals. Any paper you find contains its own summary, called abstract. Read those, and then move on to the introduction and conclusion of each paper you find. This will give you a clear idea about the content of the paper, and how it has reached its conclusions. A good target number of references is 15-20.
- Download data on the macroeconomic indicator you chose about your country/countries from the World Development Indicators (WDI, https://databank.worldbank.org/source/worlddevelopment–indicators). To ensure comparability of the data over time, you need to use real rather than nominal values for GDP, GDP per capita, income, consumption, and anything which has been calculated by multiplying a quantity with a price. Let’s call this your Y variable (your outcome variable).
- Download data on the institutional indicator you would like to correlate with your main macroeconomic indicator. WDI contains many of them but if not, then use other databases, e.g. Economic Freedom of the World indices, Heritage Foundation indices, Polity IV. Let’s call this your X variable (explanatory variable).
- Merge the data sets on your Y and X variables. Merging data sets means you need to have both Y and X for the same country in the same year in the following format:
| Country | Year | Y variable | X variable |
| C1 | 1960* | [value]** | [value] |
| C1 | 1961 | [value] | [value] |
| … | … | [value] | [value] |
| C1 | 2018 | [value] | [value] |
| C2 | 1960* | [value] | [value] |
| C2 | 1961 | [value] | [value] |
* Note: 1960 is just an example. It can be any year you have data coverage for.
** Note: Your actual values will not be in brackets, they will be numbers. Consider using natural logarithms of monetary values.
The data above is what we call a tidy data: columns represent variables, rows represent observations, and you typically have a match between X and Y variables on certain identifiers. In this case, the identifier is a country-year: each country-year is being observed once in X, and once in Y. In other words, we have a unique observation for each country in each year. It is OK if some of the country-years are missing. The data will just have a missing observation for that country-year, and will not include it in any results. Please DO NOT replace any missing observation by a 0.
Describe the data and your data sources. You do not need to attach the above table in your paper, just describe your X and Y, how many observations for each, any significant years in which your X has changed its value.
- Perform a graphical analysis of the correlation between Y and X. As the scale of Y and X is likely to be very different (unless you generate a new variable Y1=ln(Y)—recall why we do logarithmic transformations from Principles of Economics and from Data Skills), you will need to plot Y and X on two separate Y-axes. Plot Year on your horizontal axis. Pay particular attention to how your Y variable changes around turning points in your X, i.e. around your institutional reforms. Based on your graphical observation, is your hypothesis confirmed or rejected? Refer back to the economic literature you have read, and discuss why you think it was confirmed/rejected.
- Are there better ways to test your hypothesis? If yes, describe them. Test your hypothesis by using simple bivariate regression analysis in MS Excel or any other software. A conventional way to do regression analysis is in STATA or R. For doing a regression in MS Excel, search how to do a regression in Excel.
- Conclude in a few sentences: What was your question? What was your hypothesis? What did you do to test it? Did you find evidence in favour of your hypothesis? What is the importance of this evidence?
- Execution
Structure your output as follows:
Title
- Introduction and hypothesis (<150 words)
- Literature (500-900 words)
- Data (<250 words)
- Graphical analysis (<250 words)
- Results from hypothesis testing (<250 words)
- Summary and conclusion (<150 words)
Bibliography
Note: the word limits in parentheses above are just a suggestion, not a mandate.
