£12bn cost of one file sharing server. Mmmm…..

Friday’s Guardian reported that a study conducted by researchers at University College London for the ‘Strategic Board for Intellectual Property Policy‘ had found a cost to the British economy of illegal file downloading as £12bn. Going to the UCL report, the following appears to be the source of this claim:

“Just to make it clear – this is not simply an issue of music and film downloads alone. Software losses due to what is often described as “piracy” were, for example, $48 billion worldwide in 2007 (BSA, 2007); and in the UK the figure was $1,837,000 or approximately £1.25 billion. An exploratory CIBER investigation found vast quantities of films, music, software, e-books, games and television content available to download and share without cost. On one peer-to-peer network we found that at midday on a weekday5 there were 1.3 million users, sharing content. If each “peer” from this network (not the largest) downloaded one file per day the resulting number of downloads (music, film, television, e-books, software and games were all available) would be 473 million items per year. If the figure for each individual is closer to five or more items per day, the lowest estimate of downloaded material (remembering that the entire season of the Fox television series “24”, or the “complete” works of the rock group Led Zeppelin can be one file) is just under 2.4 billion files. And if the average value of each file is £5 – that is a rough low average of the price of a DVD or CD, rather than the higher prices of software or E-books – we have the online members of one file sharing network consuming approximately £12 billion in content annually – for free. These figures are staggering.” p. 6

This reasoning puzzles me. for several reasons.

Firstly, do we really think that around 2% of the entire UK population were active on this one file sharing service at one time or even on one day? This seems very high, particularly when, according to the report, the service in question is not even the largest. More likely, I suspect, this is a world-wide figure so attributions of the impact on the British economy can’t really be inferred from what we’re told here. Also, it is my (admittedly limited) understanding of how filesharing works that ‘logged on’ probably includes people who are making files available to be shared, rather than actively ‘consuming’ (which is the implication in the paragraph above).

Secondly, the authors claim that an ‘entire series of 24 or the works of Led Zeppelin’ can be downloaded as a single file. It certainly is possible that this is the case, though I suspect probably not the norm. A quick look at a DVD from the Lord of the Rings shows it contains 22 files on one DVD (there are two DVDs in the box).

Thirdly, I suspect that most filesharing is music. Here, the price of a file is likely to be rather closer to the 79p/song charged by iTunes.

Fourthly, to be seen a ‘cost’ to either the intellectual property holders in any real sense, downloading a file has actually to be ‘consumed’ (watched/listened to). Working on the assumptions in the UCL’s report, 473 million items (which they seem to equate to perhaps an hour of ‘consumption’ on a DVD or CD) this would require that filesharers watch/listen to about 80 mins of new content every day. Possible, though it does seem on the high side (given that this is just one service, which the UCL claim accounts for around 2% of the UK population).

So,  lets assume (more accurately, guess) that half of those connected are British (arbritrary, it’s true but in UCL style I think on the high side). Further, let’s assume (guess) that my arbritrary DVD has twice the number of files than is normal, that would reduce the cost/DVD to 50p (which is  the same order of magnitude as the 79p we might use as a figure for all of the music content. These two, equally plausible guesses, combined would reduce the £12bn figure to £600m.

Now, of course it may be that the UCL haven’t made available here all of the information that they have access to, though a quick skim of the ‘methods’ section of the report doesn’t suggest that any of the methods they report would have yielded this sort of information.

Two issues follow from this:

– The figures which will be landing on policymakers’ desks will be deeply unreliable, but will be used to support policies such as making internet service providers responsible for policing copyright infringements of their customers.

– If it’s an indication of what happens when academic bodies become more closely concerned to respond to the needs of industry (as we’re constantly being told we need to), then standards of academic research will fall in response to what the customer wants.

I’m no specialist in this area, but if I were writing about it, I would want at least to report (if only to refute) findings such as Oberholzer & Strump’s (2004)‘s that the impact of illegal downloading on music sales was ‘indistinguishable from zero’.