Energy use and economic growth
To demonstrate how to complete the analysis for assignment 1, this page draws on data for a fictional set of countries.
- Apollo is a highly industrialised economy, with a large manufacturing sector enabled by good access to energy sources (principally fossil fuels). Competitive pressures from cheaper overseas producers are responsible for periodic downturns in production.
- Boreas is a similarly advanced economy, though less reliant on manufacturing. Financial and consulting services account for a larger share of the country’s overall economic activity.
- Cronos is a developing country that has experienced substantial growth since economic reforms in the 1980’s, including the opening of trade relations with the rest of the world. While GDP per capita remains low, what was once an agrarian society has become a global hub for low-cost industrial activity.
The fictional data in this analysis are based on the discussion on energy use and GDP growth from Our World in Data. Figure 1 presents an equivalent visualisation to Our World in Data for the three fictional countries.
1 Average growth rates
The line charts presented by Our World in Data show changes in GDP per capita and energy use relative to a base year, 1995. However, the tables provide the raw data which can be used to calculate annual changes in the variables. Note that you can download data from Our World in Data by clicking the Download button, and then selecting the Data tab at the top of the popup. The data will be in a CSV format.
The visualisations on Our World in Data typically store many details, including descriptions of the variables used. Take some time to explore these!
Table 1 presents the example data I use for the three fictional countries, which you can also download. I have also produced some guidance on working with CSV files, if you are not already familiar with the format.
The raw data also give a better picture of the baseline conditions for the three countries. GDP per capita is relatively high for both Apollo and Boreas, but considerably lower for Cronos. That is, Cronos is evidently a developing country, while Apollo and Boreas are high-income, advanced economies.
I have stated that Apollo and Boreas are developed economies, while Cronos is developing. But these are fictional countries: you are working with data from the real world. When using these terms, you should think about how they are defined. There is no one definition for which countries are ‘developed’ or ‘advanced’, with different statistical agencies and international institutions employing different criteria. You may wish to reflect on how useful these labels are in the context of the countries you have chosen. It may be enough to comment on relative living standards across your countries using the available data on GDP per capita.
From these raw data, it is possible to calculate the average annual growth rates in GDP per capita and the two energy use measures. There are at least two different ways to approach this:
- Measure the annual changes for each year, and then take the average of these.
- Calculate compound annual growth rates based only on the first and last years observed.
Either approach can be justified, and will (in most cases) give broadly similar — though not identical — results. The first approach (table 2) requires calculating the annual percentage change for each year:
\[ g_{x_t}= \frac{x_t - x_{t-1}}{x_{t-1}} \]
Where \(g_{x_t}\) is the growth rate of variable \(x\) (for example, GDP per capita) in year \(t\) relative to year \(t-1\). From here, it is a straightforward process of taking the average of all the measured annual growth rates.
| Country | Energy use (consumption) | Energy use (production) | GDP |
|---|---|---|---|
| Apollo | 0.99 | 0.68 | 1.71 |
| Boreas | -0.80 | -1.23 | 1.01 |
| Cronos | 1.67 | 1.54 | 4.97 |
The second approach (table 3) requires only the first and last years, where the compound annual growth rate of \(x\) (\(\bar{g}_x\)) is estimated using:
\[ \bar{g}_x = \left(\frac{x_{t=n}}{x_{t=0}}\right)^\frac{1}{n} - 1 \]
Where \(n\) represents the number of years. So, for the period 1995 (\(t=0\)) to 2020 (\(t=n\)), \(n=25\).
| Country | Energy use (consumption) | Energy use (production) | GDP |
|---|---|---|---|
| Apollo | 0.58 | 0.27 | 1.46 |
| Boreas | -1.05 | -1.46 | 0.85 |
| Cronos | 1.20 | 1.04 | 4.48 |
The values in the two tables differ, as the second considers only the trend based on the first and last observations. It does not account for any variation over the period (for example, increasing initially before decreasing later).
You are not expected to discuss the pros and cons of these two approaches in your assignment. Rather, it is sufficient to decide on and use one approach that makes sense in the context of the countries you have chosen.
Whichever approach you choose, you should think about what the average growth rate might ‘hide’. For example, in a given case, the average growth rate of a variable might be zero if a period of growth was later reversed by strong declines. Highlighting this is arguably more interesting than stating there was no growth at all on average.
2 Plotting trends
Figure 2 plots the annual changes in GDP and energy use. While this is not guaranteed to be the case for all countries, for the fictional examples here, the annual changes in the variables appear to track each other closely — at least for the earliest years in the series. For Apollo and Boreas, the relationship appears to weaken somewhat after (respectively) 2010 and 2006.
The above is just one way to illustrate the trends. It is not ‘wrong’ if you have chosen to do something else! The point of this exercise is to use the data to tell a story.
Depending on the countries you have chosen, you may observe differences between the two measures of energy use. For some countries, the consumption-based measure may be higher than the domestic use measure; for others, the reverse. In some cases, you may even observe a switch between which of the two is greater. These patterns could be relevant to comment on in your assignment.
Software
The guidance here has been produced using the statistical programming language R and published using Quarto. However, there is nothing in this task that cannot be accomplished in Microsoft Excel or similar tools.
Generative AI has been used in the production of this guidance in the following ways:
- Assistance with coding tasks for the production of figures and summary statistics.
- Creative input on the description and discussion of the fictional countries used.