Cost benefit analysis can help or hinder good policy

The NBN was not subjected to a cost benefit analysis - but this may not have circumvented controversy around the merits of the project. AAP/Stefan Postles

Big infrastructure projects, such as the National Broadband Network, cost big money. So, how do we use public money wisely? For communications minister Malcolm Turnbull and the Coalition government, a cost benefit analysis provides the answer.

Indeed, Turnbull supports a proposal that all infrastructure decisions over $1 billion should be accompanied by a cost benefit analysis. This will inform us whether the infrastructure upgrade is warranted and which the most cost effective model.

Is this faith justified? Cost benefit analysis, as its name suggests, requires calculation of the costs of a particular project or policy and its benefits. For proponents, the optimal policy is the one suggested by the analysis to provide the highest benefit for the least cost most, often - but not always - expressed in dollar terms.

This simple premise is met by a difficult and complex calculation. Diligence is needed to ensure all costs are identified, and skill required to ensure cost estimates are realistic. The calculation of benefits are equally if not more challenging. It is generally more difficult to quantify benefits, while economic benefits are more accessible and quantifiable than social benefits. Skill, diligence and integrity also requires good data. A rigorous cost benefit analysis, then, depends on a robust research base.

Critical decisions lie at the heart of a cost benefit analysis. In the calculation of costs, for example, who bears the cost, and based on that decision, whether it should be included as a cost for the purposes of the final calculation is one contentious point.

In terms of benefits, the problem of “we don’t know what we don’t know” is highlighted. For the NBN, we do not know what technological innovation may arise during its lifetime. Whatever model ends up being implemented, we don’t have the luxury of trying different models to see which yields a better outcome before we make a decision.

Economists sometimes deal with this uncertainty by including in their analysis a calculation of “Net Present Value” accompanied by a discount applied to the economic estimate of benefits. The size of the discount rate also alters the value placed on the assumed benefits of a policy to future generations. Techniques such as these are based on assumptions that help tame the inevitable uncertainty that accompanies cost benefit analysis calculations.

To have policy impact, a cost benefit analysis requires political support. Even the most well-credentialed calculation can be discredited and dismissed. To see this at work one need only to look at the fate of the carbon tax. The tax was heavily supported by a similar form of economic reasoning that underpins cost benefit analysis, one that weighed economic cost and benefit.

This reasoning was rejected by the current government since the analysis did not fit with their stated policy. In contrast, calculations based on thin or selective evidence can be cited to justify policy. With cost benefit analysis, this can be easier, rather than more difficult, to achieve. A small change in assumption can make a big difference to the outcome. A strategic use of a cost benefit analysis can contribute to the problem of policy driving evidence or “policy-based evidence” as opposed to its more respected cousin, evidence-based policy.

Can a cost benefit analysis assist, then, in ensuring good policy? Interesting evidence comes from a 2009 debate about the merits or otherwise to policymaking of utilising a costing technique for human life incorporating the risk trade-offs people are prepared to make to achieve improvements, known as the value of a statistical life (VSL). VSL is commonly used in cost benefit analysis calculations, particularly in the United States.

Proponents argue that such quantification is critical to assessing the merits of proposed policies, not only does it provide a rational base for decisions but the very process of calculation allows for necessary reflection. It takes a heated political issue into a reflective realm and allows us to test our assumptions and beliefs. But, they also point to a pre-requisite for this to occur, namely a strong boundary separating the science (in this case those undertaking the technical analysis) and politics.

Critics of cost benefit analysis are more sceptical of its benefits. They argue its calculations are inevitably premised on particular values (for instance, the benefit of economic competition to human wellbeing) since neither science nor economics are value free. Further, cost benefit analysis reassures us that our policies are rational only under certain cultural conditions, namely where quantification itself confers legitimacy and signals rationality.

Yet, we are adept at building our rationality around our values, selecting numbers consistent with those values. Under polarised political conditions, expecting a cost benefit analysis to generate a rational basis to bridge disparate values and so enhance our collective future may be a tall order indeed.

Ulitmately, the role that can be assigned to a cost benefit analysis is limited. Done well, it can enhance public debate as well as inform political decisions. Done poorly, it merely masks a pre-determined political position.