A Critique of the Productivity Commission's Cost-Benefit Analysis in the 'Disability Care and Support' Report

Article excerpt


In its 2011 NDIS report, the Productivity Commission rationalises its policy recommendation by means of a cost-benefit analysis, claiming that 'the benefits of the [National Disability Insurance] scheme would significantly outweigh the costs'. But methodology the PC adopts departs from conventional cost-benefit analysis in ways that understates costs, presumes the benefits, muddies policy comparisons, and jumbles equity and efficiency issues. These problems are traceable to the Commission's use of a 'distributional weights approach' to equity benefits. The 'basic needs approach' is an alternative way of dealing with equity considerations that better captures the underlying preferences of citizens and the rationale for disability care and support policies.


In its 2011 'Disability Care and Support' report, the Productivity Commission (PC) recommended a major overhaul of the disability support system. The PC found the current State-based disability support system is 'underfunded, unfair, fragmented and inefficient' (PC 2011a: 2) and recommended replacing the State- based schemes with a new National Disability Insurance Scheme (NDIS). In March 2013, both major political parties supported legislation to establish the scheme and a 0.5 percentage point increase in the Medicare levy to help fund it. The scheme would provide support for Australians who are born with, or acquire, a severe or profound disability. The PC estimates 410 000 people would be on the scheme (ibid.).

The PC bolstered its policy recommendation with what it describes as a cost- benefit analysis. The PC acknowledges its cost-benefit analysis is not conventional, for its analysis does not attempt the difficult task of a welfare efficiency analysis of the benefits from expenditure on the disabled, but focuses instead on equity or distributional effects. Benefits are measured using a distributional weights approach - dollars to people with disabilities have more social value than dollars to taxpayers. The methodology the PC adopts understates costs, presumes the benefits, muddies policy comparisons and jumbles equity and efficiency issues.

In cost-benefit analysis we should always ask: what is the essential benefit the project confers? Clearly any assessment of disability care and support policy options must evaluate the equity benefits. But in conventional cost-benefit analysis, equity and efficiency are strictly separated. Efficiency is all about the willingness to pay, and can, in principle, be estimated from observed behaviour using well-established economic techniques. In contrast, there is a lack of agreement on how to judge equity effects, which involve ethical or value judgments. In particular, there is no general consensus about the weights that should be attached to the welfare of different groups under the distributional weights approach. There is not even agreement on whether the distributional weights approach the PC adopts is the appropriate way to assess equity issues. I argue that although it is an approach favoured by economists, it is peculiar and does not represent the values of the general public. The basic-needs approach to evaluating equity effects is superior, especially for disability care and support policy.

The cost-benefit analysis

Chapter 20 of Disability Care and Support presents a cost-benefit analysis of the NDIS (PC 2011b). Unfortunately, it appeared for the first time in the final report and did not go through the scrutiny of the draft report hearings and consultation process. It constitutes an ex post rationalisation of the PC's recommendations, rather than being used to shape and inform policy formulation and compare options.

The PC presents the result of its cost-benefit analysis as: 'The benefits of the scheme would significantly outweigh the costs. ... The NDIS would only have to produce an annual gain of $3800 per participant to meet a cost benefit test. Given the scope of the benefits, that test would be passed easily. …