Going local – but does decentralisation actually make for more innovative policy?

The mayoral elections in May are part of a plan put in place by George Osborne to reinvigorate local government in the UK. “The old model of trying to run everything in our country from the centre of London is broken,” he said in 2015. But does decentralisation make government more innovative – and if so, why? Brian Adams looks at the case for (and against) decentralising policy decisions.

sheffield town hall

Detail of stonework on Sheffield Town Hall. The city will directly elect a mayor in 2018. Photo: Anthony Stewart Vardy via a CC BY 2.0 licence

Finding solutions to pressing social problems is a primary function of government. How can we organise government so it has the capacity to generate innovative policies that address previously intractable problems?

One proposal is to decentralise policy authority to local governments. Its supporters argue that local governments can experiment with new policies and learn from the successes and failures of others, with good policies diffusing across jurisdictions. Critics, on the other hand, counter that central governments are just as willing to innovate as local ones and do, in fact, engage in policy experiments all the time.

The argument for decentralisation, however, does not rest on the argument that local officials are inherently more willing to experiment than their central counterparts. Rather, the claim is that moving policy authority downward will yield more experimentation and innovation, regardless of the preferences officials hold and the incentives they face.

There are two causal mechanisms at play here. First, the number of governments that could potentially enact policy influences the likelihood of innovation – irrespective of whether they mean to. In a decentralised system, the number of jurisdictions that can innovate increases substantially—instead of one central government, you could have thousands of local governments. Even if only a small percentage, say 1%, of local governments are innovators, this will probably represent a significant increase in experimentation over what the central government does. Central governments can experiment, but their capacity to do so is limited, as they can only have so many pilot programmes at once and only so many efforts at policy change on the agenda at a given time. Even if local officials have less of a preference to experiment than their central counterparts, we may still see more experiments on a local level due to their sheer numbers. In other words, the larger number of decision-making bodies created by decentralisation increases aggregate capacity to experiment.

Second, there are political boundaries on central-level experimentation, imposed by the ideological and political preferences of government officials, that limit the range of policy experiments. A government will enact policy experiments that are consistent with its ideology and policy platform, and this significantly narrows the types of experiments they are willing to conduct; you certainly will not see conservative governments trying leftist policy experiments, and vice versa. But because local governments vary in their ideologies they will collectively enact a wider range of policy experiments than a single central government.

Decentralisation critics are correct that central governments can, and do, innovate, but political forces limit the number and range of experiments they will conduct at any given time. As Wallace Oates wrote nearly 20 years ago, “one might suspect that relatively independent efforts in a large number of state [governments] will generate a wider variety of approaches to public policy than a set of centrally designed experiments.”

Some public choice scholars argue that decentralisation can undermine innovation by creating incentives to free-ride off others’ experiments; in other words, rather than incur the time and expense of conducting their own policy experiments, local officials will sit back and wait for the results in other jurisdictions. This, however, is based on the questionable assumption that experimentation is primarily driven by a desire for better policy. Even though this is one motivation, many officials want to innovate so they can claim credit for being an innovator or developing a new idea. Furthermore, the symbolic act of trying a new approach to a social problem may be more valuable politically than the actual substantive outcomes of the policy. Politically, enacting an innovative policy will yield more benefits to elected officials than adopting a successful policy proven in another jurisdiction. This makes it less likely that local officials will prefer to free-ride off of other jurisdictions’ experiments even though that would be their most efficient option.

There is, however, a more likely barrier: decentralisation could constrain innovation by limiting the technical expertise available to policymakers. Economies of scale may mean technical expertise is aggregated centrally and lacking on a local level. Small, rural governments, for example, may not have the technical competence to implement agricultural reforms, while a central government would. On the other hand, because local governments are smaller the problems they face are on a more manageable scale. For example, a town of 100,000 residents may have an easier time experimenting with a new type of healthcare system than a larger jurisdiction of 50 million. The smaller jurisdiction would probably have less technical competence on the issue, but the smaller scale makes the logistics easier and the design choices less complicated. The loss of technical capacity may be outweighed by the reduction in the complexity and tractability of the problem the jurisdiction faces.

Local officials may have additional knowledge that could also affect the frequency and breadth of experimentation. Both central and local officials can acquire insights into policy through public opinion surveys, government reports, think tank studies, and other sources. However, officials in smaller jurisdictions have two potential advantages in acquiring knowledge specific to their locality. First, they may have more direct contact with citizens than their central counterparts, providing an opportunity for citizens to impart knowledge about their preferences and local conditions. Second, when jurisdictions get smaller, the likelihood of an official having personal experience with “conditions on the ground” are greater due to geographic proximity. This additional knowledge may prompt local officials to enact different policies than their central-level counterparts, although whether they do so will depend on whether it’s useful in understanding the issue at hand.

In sum, the potential of decentralisation to prompt innovation and experimentation varies across policy areas. There hasn’t been much research into how policy characteristics (such as complexity and tractability) affect the ability and willingness to innovate, and so for now these contentions will have to remain hypotheses. But there’s no reason to think that the effects of decentralisation will be consistent across policy areas. For certain types of policies, we should expect to see the innovative potential of local governments be realised; for others, we may see local governments fail to even attempt to address the problems they face. Once we have a better understanding of the types of issues that will prompt experimentation on the local level, we can delegate policy authority so as to maximise the likelihood that officials trying to innovate will get it right.

This post represents the views of the author and not those of Democratic Audit.

Brian Adams is Professor of Political Science at San Diego State University.