Numbers and statistics are wheeled out to support both sides of the telecommunications regulation argument in New Zealand at the moment. Are they actually helpful to either side though, or for that matter, do they clarify the argument? I'm not convinced they do.
Proponents of the new regulation point to the OECD broadband uptake figures, which have New Zealand pegged at the bottom quarter, or 22nd out of 30 countries. Opponents of the new regulation say "yes but... the OECD broadband statistics correlate with the GDP figures". New Zealand is at number 23 in the GDP rankings which explains why we place at number 22 for broadband uptake in the OECD.
While the OECD broadband figures have become a regular (and depressing) feature, I believe the first time the correlation between those and GDP was made by Telecom Director and Business Roundtable Chairman Rob McLeod in the Dompost:
Many urban myths surround the broadband debate.
The reality is that almost 100 per cent of homes and businesses in New Zealand can access broadband capability. New Zealand compares favourably with the OECD average on this indicator.
The issue might therefore be one of affordability. The current level of penetration (about 20 per cent of households) puts New Zealand 22nd in the OECD, as the prime minister noted in parliament. This is about the same as New Zealand's ranking for income per capita, and broadband uptake is positively correlated internationally with incomes.
InternetNZ president Colin Jackson didn't agree with McLeod, and attempted to rebut his arguments in this follow-up opinion piece, also published in the DomPost.
The statistics salvos didn't end there though. National Party ICT spokesman Maurice Williamson went on RadioNZ's The Panel programme, and repeated McLeod's statement that broadband uptake is low because GDP is low too. Note that the URL to the RadioNZ programme points to an .ASX file; you may have to copy the link and paste it into a media player that understands those manually.
Williamson also stated that unbundled Germany had much lower broadband uptake than "bundled" Switzerland, the last OECD country now to maintain a telecommunications monopoly on the last mile. This despite the two countries having similar GDP.
Jackson meanwhile didn't mince his word on the InternetNZ blog: the May 7 entry is called "The Lie: we have lousy broadband uptake because we are poor." He states that:
Running the numbers gives a correlation of 59%. Anyone with high school statistics will tell you this is insignificantly low, ie the numbers aren’t correlated.
A stubborn commenter on the blog refuses to agree with Jackson however, and says the correlation is actually quite strong. What's the truth here?
I asked a mathematician friend, David Broadfoot, for an opinion. In David's words:
59% indicates reasonably strong correlation. 80% is fairly strong. 20% is fairly weak. If numbers "aren't correlated", their correlation coefficient is zero. Numbers between 0 and minus one indicate negative correlation.
As for South Korea (high broadband uptake, low GDP per capita) and Luxembourg (middling broadband uptake, highest GDP per capita), David says they are "outliers" or abberations that do not invalidate the correlation.
So does that mean Telecom and the Business Roundtable are right, and we are where we should be in terms of broadband uptake? Not quite: in order to have broadband, which for most of us means Telecom-supplied DSL, you need a $42/month voice line on top of the $39.95-$59.95 ADSL connection. Add to the above the cost of one or maybe two mobile phones per household - and cellular calls are amongst the most expensive in the OECD - and you get a better idea as to why broadband uptake is low here. It's a service that the "high value demographics" can indulge in, by and large.
McLeod can afford broadband, but for Joe and Joelene Average it's a luxury item.
This is one area where new regulation should help. It features DSL-only lines should go a long way to reduce the overall telecommunications costs for NZ households and bump up the broadband uptake as well.
What about Germany versus Switzerland then? Well, they're not actually directly comparable. Williamson conveniently omitted to mention that Germany is still struggling with the reunification of its eastern parts, which are poor and with deficient infrastructure:
A major issue of concern remains the persistently high unemployment rate and weak domestic demand which slows down economic growth. However, according to Bert Rürup, head of Germany's Council of Economic Advisers, reunification is to blame for two-thirds of Germany's growth lag compared to its EU neighbours.
In particular, eastern Germany lacks a solid base of small and medium-sized companies, which provided the foundation for West Germany's economic prosperity. Domestic demand has stagnated for many years due to wage stagnation and zealous cost-cutting by the federal state.
So yes, we probably have lousy broadband uptake because we're poor, but are we poor because of our lousy broadband uptake? Very likely, if the figures from the HiGrowth project are to be believed. Get the broadband uptake up, and by 2030 we'll be $13 billion better off. Here's hoping that's not a meaningless statistic as well.
[See More Information link below for PDF file with the HiGrowth study.]
Thanks go to Russell Brown over at Public Address, who sent the link to the RadioNZ The Panel Programme. Russell's on after Maurice Williamson so make sure you listen to the whole programme.
Update: Every caveat you have ever heard about statistics and the application of them apply...
Other related posts:
The problem with VDSL2, part 2
The problem with VDSL2
The mysterious Dynamic Line Management on VDSL2
comments powered by Disqus