Is there any research on this (apologies if I have missed it in one of the many links in this thread)? I often hear people assert that presumed liability increases cyclist safety, but the only substantive argument put forward in support seems to be to draw a correlation between the UKs lack of presumed liability and the perception that cycling is less safe in the UK.
Hembrow (I know, I know) argues that it has little effect on driver behaviour and that other factors account for increased cyclist safety in NL (he does of course assume that as a given). I can see his point in the sense that if you ask anyone in NL what the law is regarding cyclists they'll rarely be able to give you an account of presumed liability beyond a vague appreciation that cyclists are somehow privileged over cars in the eyes of the law.
Not sure I really understand your point regarding NL providing protection only to children under 14: my understanding was that presumed liability applies to everyone. I think the point with children under 14 is that they cannot be held to have contributed to an incident, so the driver's financial liability cannot be adjusted downwards.
There is a good article on the situation in the Netherlands here:
http://bicycledutch.wordpress.com/2013/02/21/strict-liability-in-the-netherlands/ The flowchart gives a summary of the position and agrees regarding the children under 14 case. What is clear from that flowchart is that, in incidents involving just a motorist and an adult vulnerable road user, if there is not sufficient evidence of fault on the part of the vulnerable road user, then the motorist is liable for all costs (that is presumed liability); if there is evidence of fault on the part of the vulnerable road user, then the motorist is still liable for 50% of the costs (so effectively 50% is strict liability), and the other 50% is determined based on the evidence. I have seen some accounts suggesting that, in practice, the bar tends to be set very high for evidence leading to some liability being placed on the vulnerable road user, but that is only anecdotal.
As far as any effect on driver behaviour is concerned, whilst it is easy to put forward a theoretical effect, I think it is virtually impossible to reach any reliable conclusion because there are so many confounding factors.
However, I don't think that is the most compelling argument for moving to presumed liability anyway; I think the best argument is simply that it is morally the right approach. In the (common) case where there is no conclusive evidence of immediate fault by either party, it fits well to the established principle of determining liability on the balance of probabilities. For example it is usually very clear that, had a collision been between two bicycles the probability is that the damages would be a tiny fraction of what they would be in the case of such a collision between a motor vehicle and a bicycle, so the very act of bringing the motor vehicle into the public space is bringing the majority of the risk to the situation. Even in the case where there is clear fault on the part of the vulnerable road user, it is still the case that the motorist brought a significant proportion of the risk to the situation, and that appears to be the rationale for the 50% strict liability element in the Dutch rules. It also seems rather unjust that a vulnerable road user who is seriously injured in an incident involving a motor vehicle should have to fight through the courts even to recover the costs of medical care, loss of earnings, physical and lifestyle adaptations to cope with resulting disabilities, etc. Finally, since such legal battles are very costly, it is possible that such an approach might actually reduce the costs of motor insurance slightly.
