I thought that that issue might come up. Obviously, it provokes a great deal of discussion. I have certainly reflected at great length about the various options that exist.
We all recognise that the current system is not perfect. However, we should not change it simply for change’s sake: any change should be made for good reason, such as to make the system more effective, fairer or less biased. If we were to go into that in detail, a lot of work would need to be done on the pros and cons of the various alternatives.
The committee will be aware that the Scottish Government has been considering two options. The first is drawing names by lot; the second is having two ballot papers, one of which would list candidates from A to Z while the other would list them from Z to A. We asked the Electoral Commission to look at those options in detail, and we have not yet reached a decision on whether to adopt either of them—or, indeed, any other option. It is clear that any option would involve both pros and cons. There is an argument for testing some options in local government by-elections to see what actually works in practice in a Scottish context.
However, I agree with Bob Posner of the Electoral Commission that we should not rush into changing the system and thereby risk unintended consequences. I do not mean to sit on the fence by saying that. If we are to make changes, we need to get them right. As I said earlier, the fact that we believe that the present system is flawed does not mean that we should change it simply for the sake of doing so.
I know that others consider randomisation to be an option. By way of background, I add that the Scottish Government estimates that implementing such an approach would add £2 million to the cost of conducting an election—which is not to mention the administrative burden that it would create. I am also trying to work through in my head how such a system would not create another bias somewhere along the line, depending on factors such as whether we had 100 per cent uptake of ballot papers.
Other issues add to the confusion in that area. For example, to assist people with visual impairments, we are currently required to have large-print sample ballot papers at polling stations. We might ask how we could possibly have those samples if the ballot paper were to be randomised, as opposed to having one straightforward paper—or two papers if the A-to-Z and Z-to-A option were to be adopted. Further, some local authorities still want to count by-election ballot papers manually, and having multiple forms of ballot paper would create a great deal of difficulty for them.
In making those points, I am not suggesting that we should not change the system; I am simply laying out the pros and cons for the various options, which I am sure the committee will also have considered.