A weakness of both approaches is their sensitivity to the choice of (necessarily somewhat) arbitrary poverty lines and of peculiar value judgements regarding the social welfare objectives of the government—for instance, that the government cares equally for all the poor, regardless of how far from the poverty line they may be. For instance, the analysis of targeting errors focuses typically on sharp 0/1 indicators, and arguably tends to differentiate too drastically between the poor and the non poor, in particular between those in similar circumstances but who just happen to lie on opposite sides of some poverty line.The working tax benefit is available to anybody aged 25 or over who works sixteen hours a week or more. There is a basic element and a range of additional elements for single parents and couples, for people who work for thirty hours a week or more, and for people with a severe disability; there is also an element to contribute towards the costs of child care. A person earning below a threshold level of income receives the full benefit. For earnings above that, benefit is withdrawn at a rate of 37 pence per pound of earnings. Benefit is normally awarded on an annual basis; thus an increase in earnings, unless large, will not lead to a reduction in benefit until a person is reassessed. Other difficulties in the assessment of program changes come from their differential effects on average deadweight losses. Such differential effects can occur when the programs are funded from different revenue sources.: differences in the cost of public funds that arise from differences in those revenue sources must then be taken into account (Slemrod and Yitzhaki, 1996)2. Differences in the effects on average deadweight losses can also arise from the differential behavioural changes that different program reforms can generate among program beneficiaries. These differential behavioural changes can in general also affect the direct disaggregated welfare impact of program reforms.
None of these categories can readily be dealt with by private insurance; and none except the first can be helped by raising national-insurance benefits or by extending their coverage. Much poverty is associated with children and/or high housing costs, neither of which is an insurable risk. Two conclusions emerge: private insurance is not possible in most of these cases; nor is extending national insurance a complete answer.
The state could, of course, do nothing, and let people face the risk of starvation, but, even ignoring equity arguments, this has a range of efficiency costs, including social unrest/ crime among those facing starvation; the death by starvation of dependants including children (the future labour force);