How to make good government policy

It can be argued that the primary business of democratic government is the development and implementation of good policy and the optimal allocation of resources to support those policy decisions.

"Good" policy has the desired effects and, where possible, does not have unanticipated secondary effects. A key factor in making good policy is the use of high- quality evidence about the issues in question so as to define or resolve the range of options available to the policymaker.

Bad policy can emerge when the evidence is ignored or compromised - indeed it is clear that such a disconnect lay at the core of the social, economic and political consequences of the mad-cow disease epidemic in Britain. In every jurisdiction there are social intervention problems that are continued because there is no substantive evidence as to either their effect or lack of effect.

One example was high school driver education in the United States, which was eventually found to increase road accidents.

The parliamentary commissioner for the environment in New Zealand concluded that the toxic problems at Mapua were aggravated by lack of technological input.

The challenges are multiple: to identify what research and knowledge is needed; to identify appropriate sources of that knowledge and ascertain the validity, quality, and relevance of the knowledge obtained; and to understand how that knowledge informs a range of potential policy options.

Yet we do not live in a technocratic state in which science defines policy goals and regulatory decisions - rather, scientifically derived knowledge provides base information on which other critical dimensions, including societal values, public opinion, affordability and diplomatic considerations must be added while accommodating the political process. ON THE other hand, policy made in the absence of information and science- based evidence can only be made on the basis of dogma and is less likely to serve the country well.

But beyond these immediate considerations, governments are increasingly required to make decisions that rely on an understanding of biological, environmental and social complexity. Decisions about climate change are an example. Our understanding of this complexity increasingly relies on scientific advances and knowledge derived through the scientific method.

Indeed, science has an impact on every aspect of our lives and offers the basis of solutions to many of the problems the world confronts - from water and food security to the issues of urbanisation and an increasingly complex social world.

Here the science can never be complete, yet action is required, and the discussion must be about risks and probabilities. Here the key challenge for the science adviser is to be free of biases and to ensure that the analysis takes into account the implications of inappropriately underestimating or over-estimating the consequences of conclusions reached. It is a waste of scarce resources to invest in policies and interventions that are not particularly effective, yet social and political pressures may mean that programmes are initiated even though their effectiveness remains uncertain.

Once in place, public opinion may make it difficult to stop ineffective yet popular programmes. For example, many social interventions work well in the pilot phase, perhaps because of the enthusiasm of their advocates, but do not work well when implemented at full scale.

Thus, ongoing evaluation to rigorous standards should be a requirement for all new and existing programmes. Only then is it possible to have the population engage in understanding which programmes are effective, what is their impact, and thus which should be core to the country's development.

Importantly there is a need to keep scientific advice separate from other considerations, and to provide protocols by which government can obtain that advice and ensure its quality.

If this is not done, trust in science is undermined. Conversely, science should not be co-opted as a proxy for debates which should belong in other domains - a classic example is the debate about abortion is a debate about values not science.

A PARTICULAR issue is how the government contracts research on its own account and ensures its quality, and in particular the need to make better use of social science disciplines in a more integrated fashion.

For example, we currently have no master register of research done to support government or to provide information to the private and voluntary sectors. There is no standard protocol for ensuring the quality of research commissioned within or by departments. The role and status of science advisers across government agencies is inconsistent.

A science adviser even within one ministry cannot be expert in all domains, but acts primarily as a trusted translator from the science community to the public and policymaker. Judgment is needed as to when there is sufficient evidence for a conclusion to be considered reliable. But science advice must also be free of bias and other agendas - science advice is essentially about summing the knowledge and providing the range of options on which other dimensions of policy formation can be conducted.

To reiterate, science does not make policy; rather it informs the policymaker, but policy made without proper assessment of the evidence can only be made on the basis of dogma and that does not serve a democracy well. These are important matters and in the past 12 months, my office has conducted an extensive review of how we can better use evidence in policy formation.

There are many practical ways we can do better and I have just released a discussion paper outlining possible steps forward.

Towards Better Use of Evidence in Policy Formation; a discussion paper, can be found at