DS Project

Chris Brierley

Olympinomics - The Economic Impacts of the Olympic Games

Introduction

This project aims to investigate the macro and microeconomic impacts of the Olympic games on host nations and firms, ultimately deciding whether hosting or sponsoring the Olympics is an economically viable strategy on improving various economic indicators.

Files used:

chart17_stocks.json

Stock_Regression.ipynb

project_sponsorship.csv

Files used:

chart16_regression.json

Stock_Regression.ipynb

sponsor_stocks.csv

The data I used and how I accessed it

I opted against using many APIs in my project as most of my project investigates past Olympic games, rather than future trends, and thus most future data is not relevant. Furthermore, in cases such as Japan, using the E-stat API would fetch results in Japanese which Vega-lite would not understand.

For data on GDP per capita and exports I used World Bank data to download relevant CSV files. I used FRED for all exchange rate data - again not using their API as Vega-lite would not load the data. To access Olympic medal tables and the number of athletes brought to each Olympics, I scraped Wikipedia tables using data from Olympics.com. I also used UNWTO seasonal data for my tourism chart. Finally, for stock prices and exchange rates, I used the Yahoo finance package in Python and the Alphavantage API respectively.

Files used:

chart18_TokyoMedals.json

London.ipynb

Rio.ipynb

TokyoMedals.ipynb

Tokyo2020Data.csv

Conclusions (Part 1)

For firms in the Olympic Partner Programme, I found that there is no real effect of the Olympics on the stock price of these firms. Each firm's stock price behaves very differently during the Olympic months, which implies that these fluctuations are explained by external factors. The firms whose stock price behaved most similarly during 2021 were Coca-Cola and P&G (the two biggest sponsors of the Olympics), as shown in the second graph whereby 63% of variation in P&G can be explained by variation in Coca-Cola. However, this is not statistically significant enough to conclude the Olympics have caused these two stock prices to act similarly.

Interestingly, GDP per capita does not significantly affect the number of medals won in the previous 3 Olympic games. Almost all countries at any income level won fewer than 20, while the U.S., Great Britain, China and Russia won considerably more medals. This appears to be due to the number of athletes each country brought rather than GDP per capita.

Files used:

chart19_OlympicIndex.json

Index_Final.csv

Files used:

chart20_Exports.json

France_Export.csv

Brazil_Export.csv

Japan_Export.csv

Spain_Export.csv

Turkey_Export.csv

UK_Export.csv

Files used:

chart21_Tourism.json

Tourism.csv

Conclusions (Part 2)

My Olympic index shows that a host's currency could be used as a winning foreign exchange strategy. For over half of host countries since 2000, their currency appreciated against the dollar in the 4-year build-up to their games. However, Brazil's Real crashes the Olympic index by 70 points, which shows that the build-up to the Olympics doesn't always appreciate the currency.

I also investigated how a host's exports compared to the country eliminated in the final voting round for the bid and found that winning countries have a larger export index across the period from winning the bid itself to two years after the games are held. This is excluding the UK, whose exports in the wake of the 2007-08 financial crisis fared much worse than for France, again highlighting the significance of external economic events compared to the Olympic games themselves.

Finally, it's too early to tell whether the handover of the Olympics from Tokyo to Paris resulted in an appreciation of the Euro against the Yen. While I expect over the next few years for the Euro to appreciate, a crash like the Brazilian Real could still occur for the Euro. As this chart uses an API, we will see how this relationship changes over coming months.

Overall, the economic viability of the Olympics is inconclusive. I believe that there is selection bias that the IOC uses to pick future hosts of the Olympics, whereby only countries that display encouraging economic forecasts are considered, again nullifying the significance of the Olympics itself. While the Olympics itself is a small economic event, using it as a signal for future prosperity appears a lot more appropriate than simply assuming its direct impacts on economies and firms alike.

Files used:

chart22_Future.json

Challenges in data cleaning/analysis and how I overcame them

The Exchange Rate Index chart posed the greatest challenge as it required taking daily exchange rate data of USD to 1 unit of local currency from FRED for the 5 previous currencies of Olympic hosts between 2000 and 2021 and converting them to an index. I then created a separate Olympic index inspired by Goldman Sachs (2012) which follows the currency of an Olympic host from 4 years prior until the day of the closing ceremony before changing to the next currency. Furthermore, using daily data would make the chart almost illegible, so I had to use Excel filters to only include the closest date to the first of the month.

When scraping Wikipedia in Python for Olympic medal tables and number of athletes, the rows in each table would differ considerably so time was therefore spent telling Python which rows should be deleted from each unique table. Instead of using a data join in Python, I instead opted to use Excel to make my final, cleaned spreadsheet.

List of sources:

Alphavantage

FRED (1) (2) (3) (4) (5)

Online Tutorials - Side navigation bar tutorial

Red Stapler - RGB card design inspiration

Traversy Media - CSS loading animation inspiration

UNWTO

Wikipedia (sourced from Olympics.com) (1)(2)(3)(4)(5)(6)

World Bank (1)(2)

Yahoo Finance (1)(2)(3)(4)(5)(6)(7)(8)