With Australia’s internet speeds a hot election topic, figures are being thrown about to demonstrate where we sit in a global field.
For example, the World Economic Forum “league table” has been cited by many commentators as showing that Australia has the 25th lowest “internet bandwidth” in the world.
It demonstrates no such thing.
The problem with statistics — particularly in the telecommunications market — is that too often, numbers are handled by people who:
- Don’t understand the limitations of the source data;
- Don’t understand the age of the source data; or
- Are so concerned with a political point that they don’t want to know whether their data is accurate or not.
The WEF data quantifies Australia’s international connectivity. It adds up the capacity of our international submarine cables, assigns a portion to internet use, divides this capacity by our population, and ranks countries according to “International capacity per 10,000 inhabitants”.
This doesn’t tell us much about Australia’s telecommunications development, because:
- International capacity doesn’t measure the broadband speeds available to customers — it measures how much traffic carriers need to send overseas. For example, if Google puts a server in Australia, searches no longer travel to America.
- Capacity on modern fibre optic cables can be added as required, but capacity measurements generally look at the total amount of lit capacity (what’s needed now), as opposed to the potential capacity of a cable system.
- International capacity may impact the “wait time” to download a file from America or Japan. However, this is in the control of the ISP, which decides how much capacity it can afford to buy.
- The WEF report uses data published by the ITU in 2002 — the equivalent of ancient history in Internet time.
- The data used in the WEF report falls short of reality (current lit capacity) by a factor of 16. Its assertion that Australia has around 5 Mbps per 10,000 citizens is simply inaccurate. The available international capacity on Southern Cross, Jarasaurus, SeaMeWe-3, Australia-Japan Cable and the venerable Tasman-2 system is more than 240 Gbps — with more than 80 Mbps of international Internet capacity per 10,000 people.
In other words, the metric used to justify calling Australia’s broadband download speeds a “disgrace” is inaccurate and inapplicable. It also ignores the fact that market forces are addressing international capacity. Three groups (Southern Cross, Telstra, and the Pipe Networks – VSNL consortium) are planning many terabits of new capacity.
What of the famed OECD league table, favourite of politicians and journalists?
While describing the table as “a disgrace” may be too strong, it should be handled with care. In particular:
- There seems to be no single common denominator for the word “broadband”. Some countries count 64 Kbps services; the European Commission’s definition (not used by all EC countries) starts at 144 Kbps. Australia’s definition (256 Kbps) is higher than 17 out of 30 OECD nations, while another three countries don’t start counting broadband until 512 Kbps.
- The OECD uses ACCC broadband figures, yet Australian Bureau of Statistics broadband numbers are significantly higher (3.6m versus 3.9m in Sept 2006).
- There is no way for the everyday reader to confirm the accuracy of the data, since some individual countries do not release their own public broadband statistics.
Broadband statistics are good enough for an argument around the barbecue — but any serious attempt to map policy (and billions of dollars) to broadband league tables needs to start with accurate data.