Dr Khalid Koser is a Lowy Institute Non-Resident Fellow and Deputy Director of the Geneva Centre for Security Policy.

Recent Government and Opposition policy announcements on asylum seekers defy conventional analysis because it is impossible to predict whether they will work.

First, each is liable to legitimate legal challenge and may never actually be implemented, at least not in their entirety. A Federal Court challenge has already been filed in Australia against the Government's plan to resettle asylum seekers to Papua New Guinea, while the Opposition in Papua New Guinea has also launched a legal challenge to the Manus Island processing centre. It seems unlikely that elements of the Coalition's policy will not also be challenged, especially the proposal to grant Temporary Protection Visas to recognised refugees, with the consequence that they will be denied the right to family reunion.

Second, both policies contain proposals that are basically unprecedented. We have no basis for assessing whether resettling asylum seekers to another country will deter them. The fact that unauthorised boats continue to arrive in Australia despite the Government's new policy suggest not; and also that my hunch in an earlier comment for the Interpreter that it would work was wrong.

Who knows whether the Opposition's proposal to enlist Indonesian villagers as informants against people-smugglers or buy back their boats will work, although at the risk of being proven wrong again I can't see how it will.

Third, both the Labor Party and Coalition have changed course so regularly on asylum that there is little confidence that these are actually their final policy positions. This matters because unless potential asylum seekers and people-smugglers are receiving a consistent message they are unlikely to take much notice.

Fourth, even if one of these policies were to be implemented as currently formulated, it is not clear how to evaluate them. The most obvious metric would be the number of unauthorised boat arrivals, although there will be disagreement about the time-lag that can reasonably be built in between the implementation of a policy and its impact. But each policy also contains other elements that should be subject to evaluation, ranging from the number of deaths at sea through to the disruption of people-smuggling, reducing costs, and building regional capacity and cooperation.

How do we evaluate a policy that reduces the number of unauthorised boat arrivals in Australia but incurs a higher ratio of deaths at sea? Or a policy that reduces boat arrivals by 10% but at the cost of a 50% increase in budget?

If there is no way of knowing whether either policy will work, how should the voting public assess them? Allow me to suggest three questions that voters may wish to ask themselves as they think about these policies. First, is either realistic, either in their objectives or in their proposed methods? Second, is either based on any evidence? Third, could either be monitored and evaluated?

For both sets of policy, my answer to all three of these questions is 'no'. And in this way both the Government and the Opposition have made the fundamental mistake of underestimating the voting public.