OECD’s shoddy broadband stats exposed

Stuart Corner
Tuesday, 15 May 2007 12:49

IT Policy - Regulation

Every six months the OECD produces a report comparing member economies' on a range of broadband parameters. Australia's ranking is much used by all parties in the debate, but the whole exercise has been exposed by Australian research company Market Clarity as unsound because of inaccurate and non-comparable source data.

Market Clarity says it devoted close to 400 hours of analyst time scrutinising documents from more than 60 sources from all 30 OECD nations in order to try and verify the OECD data. It says it found that the OECD mis-reported broadband subscriber numbers for 28 out of its 30 member nations in its June 2006 analysis (published in September, 2006).

"We are forced to conclude that the OECD's broadband rankings, while providing an interesting snapshot of broadband adoption are not sufficiently rigorous or accurate to inform the basis of national policy making, nor should perceptions of national prestige be founded on the fine detail of a country's standing in the OECD's broadband rankings," Market Clarity concluded.

According to Market Clarity CEO Shara Evans, the OECD also fails to allow for inevitable error margins which will occur in any statistical analysis, or to provide sufficient source references to enable discrepancies between its published data and national source data to be accounted for.

 "In the OECD's much-vaunted "league table" there are distinct clusters of countries whose adoption rates are so close to each other the gap between them is smaller than a margin of error of five percent," Evans said. "Yet without a rigorous discussion of possible errors both in population statistics and broadband subscriber data, readers of the OECD tables have no way to test the fairness of their country's score."

Market Clarity also found that although the OECD sets 256kbps as its minimum benchmark for inclusion as a broadband customer, this information is not collected by 14 out of 30 OECD nations. Some nations also lag others in their broadband data-collection, with some national statisticians and / or competition regulators having no new data published since as long ago as December 2005, with a major knock-on impact on the reliability of the data.

"America and New Zealand certainly have cause to complain," Evans said. "We can see no reason why eight million American broadband subscribers recorded by the FCC and published in a downloadable spreadsheet appear to have been overlooked in the OECD's table. Likewise, 65,000 broadband customer connections in New Zealand seem to be missing from the OECD's calculations."

Evans said she had not sought comment from the OECD on the findings. In recent days the OECD's broadband figures and methodology have come under criticism from high level figures in the US Government.

Because of the importance of the broadband debate, Evans has decided to make the research available free of charge on its website, http://www.marketclarity.com.au/, it will be available from 17.00 hrs AEST on Tuesday, 15 May.