The Rumsfeld effect: knowns and unknowns in mobile communications

Citation

Curwen, P. (2008), "The Rumsfeld effect: knowns and unknowns in mobile communications", info, Vol. 10 No. 2. https://doi.org/10.1108/info.2008.27210baa.001

The Rumsfeld effect: knowns and unknowns in mobile communications

The Rumsfeld effect: knowns and unknowns in mobile communications

Readers are doubtless familiar with the musings of Donald Rumsfeld about the nature of what is known, known to be known, unknown, known to be unknown and so forth. The more one studies mobile communications, the more one begins to wonder whether he had a point.

On the face of it, facts about operators, regulators, networks and so forth should be easy to come by – “known knowns”, as it were. It is not as though information is in short supply on the Internet; in an annual report, say, or for a fee on a reputable for-profit website. The problem is that this is not fail-safe. It does not necessarily produce untruths – although it can – but it does not necessarily produce “knowns” either.

Consider, for example, a recent paper in info written by Harald Gruber (Gruber, 2007). Central to this paper is a table (Table IV, p. 40) that contains information on 2G and 3G licenses in Europe up to 2005. Now such licenses are surely “known knowns” – since they have to be awarded there can surely be no ambiguity about how many, to whom and whether they had been launched by a specified date.

Table IV

Number of 3G licenses and n+1 rule (revised)

Regrettably, not so. An alternative version of Table IV is appended below. In column 2 there are three differences from the original – Finland, Ireland and Norway all contain one fewer licensee than previously. By implication, there are three parallel reductions in column 3. Now column 2 must in principle be a “known known” – you either issue a license at a specified time or you do not. Since I am sure that my version of column 2 is correct, the original version must be incorrect even though the data were taken from reliable sources as cited in private correspondence with the author who has agreed the corrections. So “known knowns” are more reluctant to become “known” than you might expect.

Column 4 is trickier. There are different figures compared to the original table in respect of the UK, Denmark, Switzerland and Finland. The problem here is partly that the end-date for the original table is 30 September 2005. There are several “reliable” sources of data on high-speed network launches relating to cdma2000 and W-CDMA[1]. Unfortunately, they do not agree with one another all of the time. Of course, a couple of months either way may not seem much in the greater scheme of things, but it can be awkward when there is a specific end-date and they cannot agree on which side of that end-date the launch took place.

Launches are notoriously ambiguous events. There are, for example, “pre-commercial” launches, launches using data cards and launches using handsets. Interestingly, even the above-mentioned sources are sometimes unable to pin down their own version of the date, contenting themselves with a named quarter such as 2005 Q3 or even simply a year.

Launches, therefore, are not necessarily “known knowns”, but it should still be possible to pin down, in all bar a very few cases, on which side of a given end-date they fall. In that sense, part of column 4 is indeed a “known known” and both versions of the table cannot be correct.

Does this matter? It is in the nature of academic publications that the great majority currently favor authors who engage in lots of econometric calculations. However, if the data are “dodgy”, the precise value of the ensuing correlations is necessarily brought into question. As demonstrated above, the truth may be out there in the case of mobile communications, but it requires a lot of effort to winkle it out and ambiguities will probably always be with us.

Notes

1. For example, www.3gamericas.org, www.gsacom.com, www.cdg.org and www.3gtoday.com

Peter CurwenVisiting Professor of Telecommunications in the Department of Management Science, Strathclyde University, Glasgow, UK. E-mail: pjcurwen@hotmail.com