dw-data / mobile-internet

In Malawi, people have been paying around 87% of gross national income per capita for 1GB of mobile data. Rwandans only pay 2%. DW examines the reasons for this huge price discrepancy.

Home Page:https://p.dw.com/p/3knuq

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Prices for mobile internet in African countries

You can read the story here in English or German

Idea, research and writing: David Ehl

Data analysis and visualization: Gianna-Carina Grün

Data sources

Mobile data prices

The source for mobile internet prices is the British telecom provider Cable UK, given that they provided the most recent data (beginning of 2020). The data was downloaded from their website.

There, they also included more detail on the scope of their research project: Data from 5,554 mobile data plans in 228 countries were gathered and analysed by Cable.co.uk between 3 February and 25 February 2020. The average cost of one gigabyte (1GB) was then calculated and compared to form a worldwide mobile data pricing league table.

Key variable description According to their methodological description of the data the core variables in the dataset are defined as follows:

  • "Plans measured": Researchers first established the mobile data providers in each country before selecting one SIM plan from each data amount they offer. A country with eight providers, for example, each offering three different SIM data sizes, will have 24 plans recorded (...) Packages were recorded up to a maximum of 60 per country – records beyond this number have negligible impact on the average (...) This number will always be between 3 and 60.

  • "Average price": Averages are calculated as the MEDIAN of all recorded package prices/data limits. (...) The average monthly cost of 1GB of data for each country as a whole is calculated as the MEDIAN of every plan recorded

  • "Sample date" / Price data : The 'Sample date' column is very important as it represents the day the data was collected for that particular country. It is a snapshot. Prices change all the time and therefore may be different by the time you go about viewing and/or utilising our data

Gaps in the data: According to their description, the following African countries missing from the dataset:

Name Continental region Missing data reason
Eritrea NORTHERN AFRICA NO PROVIDERS
Congo (Democratic Republic of) SUB-SAHARAN AFRICA PRICES ONLY IN 'UNITS'
Zimbabwe SUB-SAHARAN AFRICA UNRELIABLE EXCHANGE RATES
South Sudan SUB-SAHARAN AFRICA NO PROVIDERS
North Korea ASIA NO PROVIDERS
St. Pierre and Miquelon CARIBBEAN NO PROVIDERS
Christmas Island CARIBBEAN NO PROVIDERS
Venezuela SOUTH AMERICA NO PROVIDERS
Wallis and Futuna OCEANIA NO PROVIDERS
Marshall Islands OCEANIA NO PROVIDERS
Cook Islands OCEANIA NO PROVIDERS
Tuvalu OCEANIA NO PROVIDERS

Data caveats: It's important to consider the high volatility of pricing data. A research analyst of CableUK comments: "When it comes to the cost of 1GB pricing can change dramatically over time since changes to the data sizes available can happen overnight." Thus, the data should be seen as snapshots of a certain point in time.

Other data sources: We also considered the mobile pricing data by the ITU (International Telecommunication Union) as well as by the A4AI (Alliance for Affordable Internet). The main reason to decide against both sources was that the CableUK data is more recent (Q1 2020 vs Q2 2019). Additionally, the ITU data is more granular: offering different "baskets" with different data volumes which was not necessary to answer our initial research question.

Monthly national income

The data on national income was taken from The World Bank, who publishes the Gross national income per capita by country and by year. We used the GNI per capita, according to the Atlas method (current US$)

Because the data for mobile prices are monthly, we divided the World Bank data by 12 to arrive at the monthly GNI per capity (in current US$) we used for our analysis.

The GNI is provided for different years for different countries; we used the latest available data for each country.

Definitions

The Alliance for Affordable Internet defines: "affordable internet is where 1GB of mobile broadband data is priced at 2% or less of average monthly income." This recommendation has also been adopted by the UN Broadband Commission for sustainable development as their new target.

Methological decisions

We excluded Somalia from our analysis, because the latest available GNI data was from 1990, which we considered too outdated. We furthermore excluded Oceania, as the GNI data here was very gappy and would not have allowed for a solid regional average. Furthermore, for the analysis on African countries, we excluded the Comoros Islands, Sao Tome and Principe and Seychelles from our analysis and focused on countries with more than 850,000 inhabitants.

We chose to use the data for the median price instead of the cheapest price for 1GB of mobile internet per country, since this is more likely to reflect the given situation for end users (since it is unlikely the cheapest price per country is available to every person within that country).

We refactored the pricing data in order to reflect the affordability target (see definition). Instead of showing the percentage share each population on average spends on 1GB of mobile internet, we calculated off of the monthly national income what the two percent affordability benchmark would equal in US dollar.

We relabelled the CableUK data labelled as 2019 by the data provider as 2018 data, since only 2 samples were collected in 2019, whereas the other 228 samples were collected in Q4 2018.

Results

Comparison to other continents

  • the median price for 1GB mobile internet per month is US$ 3.30 in African continent

  • only in the Americas 1GB mobile internet is more expensive

  • all regions are below the recommended UN target for affordability -- except for Africa. Here, people need to pay more than 2% of their monthly income

  • an "affordable" median price would be US$ 2.41

Comparing African countries

  • prices within Africa vary a lot

  • in 21 African countries, internet would be considered affordable, meaning 1GB costs less than 2% of the country's monthly gross national income.

  • in 28 African countries, internet is considerably more expensive -- on average, the price in these countries is three times as high as it should be

  • furthermore, it is particularly the poorer countries where mobile internet is more expensive than it should be: except for Botswana and Equatorial-Guinea, all countries with a too high median price are "low income" (19 countries) or "low middle income" (7 countries) respectively

Price range: actual and affordable prices

  • Mobile internet is most expensive in Malawi: for 1 GB mobile data a person in Malawi has to pay about US$ 27.41 -- in relation to a monthly median income of US$ 31,67 this price equals approx. 87 percent (in stark contrast to the recommended 2 percent). An adquate price would be US$ 0.63

  • significant price ranges between actual and affordable prices can also be found in Benin, Chad, Eswatini, Madagascar and the Central African Republic

How prices changed from 2018 to 2020

  • the market does not seem to move in the direction intended by the UN

  • the median price fell in 32 countries, but rose in 12 countries

  • despite the change, the price is still above the recommended one in 26 countries

  • in these 9 countries, the price was already too expensive in 2018 and the rose even more until 2020: Malawi, Madagascar, Mauretania, CAR, Benin, Niger, Guinea, Burundi and Senegal

  • while the price was affordable in 2018 in Cameroon and Rwanda, it rose above the recommended threshold in 2020

  • the price was considered too expensive in 2018 in Congo, Lesotho and Namibia but fell until 2020, so that the median price is now considered affordable

About

In Malawi, people have been paying around 87% of gross national income per capita for 1GB of mobile data. Rwandans only pay 2%. DW examines the reasons for this huge price discrepancy.

https://p.dw.com/p/3knuq


Languages

Language:Jupyter Notebook 100.0%