POLICY AND POLITICS: How might lower ranking officials have a greater impact on policy development than previously assumed?

POLICY AND POLITICS: How might lower ranking officials have a greater impact on policy development than previously assumed?

Andrew Connell

This section of Discover Society is provided in collaboration with the journal, Policy and Politics. It is curated by Sarah Brown.

How can small-territory, subnational governments make the most of their position? Subnational governments, like the devolved governments in the UK, combine some of the opportunities and limitations of the national and the local governments between which they sit. Some of their responsibilities and resources, like legislative authority and funding powers, look very like those of a national government. On the other hand, they govern territories which are comparatively small: Scotland has just under 5.5 million people and 32 local authorities, and Wales just over 3 million and 22, against England’s nearly 56 million people and 343 authorities. This means that they might be able to cultivate ‘local government’-type relationships with a comprehensive range of local groups– and the fact that their own resources are limited by their small size and the terms of devolution settlements means that they might need to.

So is there something about small subnational governments’ positioning which can allow them to operate in a distinctive way? In our recent Policy & Politics article, we explored this question by studying the development of homelessness policy in Wales. We found that although the Welsh Government’s own resources were quite limited, it had been able to make and implement innovative new homelessness legislation through developing networks with third party organisations which it had fostered over time.

It had practiced metagovernance– the creation and management of policy networks – very successfully, and it was exactly its combination of ‘national government’ type powers, and ‘local government’ type proximity to local groups that had enabled it to do this. Importantly, because the Welsh Government worked so closely with a comprehensive range of local authorities, service providers, and academics, much of the day to day governance of networks was done by officials who were not particularly senior in the civil service hierarchy and who themselves participated very actively in the networks that they governed.

Our study looked at the development and implementation of Part 2 of the Housing (Wales) Act, which was passed by the National Assembly for Wales in 2014. Since 1978, local authorities in Wales (and England) have had a duty to provide long-term housing for some homeless people, but the 2014 Act extended this by giving them additional, and legally enforceable, duties to help people at risk of becoming homeless to keep their homes. Independent observers such as the homelessness charity Crisis believe that the new duties have been effective in preventing people becoming homeless in Wales. Indeed, unusually, the Welsh legislation was the model for a Westminster statute, the Homelessness Reduction Act 2017, which gave English local authorities a similar prevention duty. This ambitious, and so far relatively successful measure, introduced at a time of austerity, could therefore offer a good test of the ability of a small government with constrained formal powers to do things differently.

We grounded our study in an analysis of documents on homelessness policy published and commissioned by the Welsh Government since 1999, combined with interviews with officials, politicians, and others outside government who had played significant roles in developing and implementing the legislation. We also conducted a policy reunion, bringing together some of our key informants with an academic discussant to consider our emerging findings and reflect further on their work.

Our research showed that very strong policy networks, encompassing third sector, local government, academics, and Welsh Government groups, were vital in developing shared understandings of problems and solutions in respect of homelessness policy reform. These networks had been nurtured by Welsh Government officials and ministers almost from the very beginning of devolution. Although housing, including homelessness, was one of the original policy responsibilities devolved to Wales in 1999, the new Welsh Government and Assembly could do little to change significantly the statutory service framework, embedded in Westminster legislation, which they had inherited.

This was a common problem in the first decade of devolution, when the National Assembly had only secondary legislative powers. Consequently, the Welsh Government developed a policy style, in homelessness and other policy fields, which relied heavily on strategies, funding, and promoting collaboration. In 2011, the National Assembly received full primary legislative powers in devolved matters, and this allowed the Welsh Government to innovate more radically. But in the case of homelessness, the new legislative powers worked hand in hand with the networks that had been developed to take forward strategies at a time when the Welsh Government could do little else.

To understand how the Welsh Government fostered and supported these networks, we applied a typology of metagovernance developed by Eva Sørensen and Jacob Torfing (2009). This proposes four categories of metagovernance tool. The first two, designing the network and steering the network’s goals and framework, are concerned with shaping interactions rather than participating in them, and are described as ‘hands off’ tools. In contrast, the third and fourth categories, actively managing the network and participating in the network, are considered to be ‘hands on’ tools.  Sørensen and Torfing also suggest that while any network group or individual that commands the necessary resources can act as a metagovernor, state groups/individuals tend to enjoy advantages as metagovernors because they can usually deploy more and stronger resources to underpin their authority.

Our research supported this last claim: the Welsh Government’s legislative and funding powers gave it a clear leading role in the networks it fostered. In Wales, as in the UK as a whole, homelessness services are deeply embedded in statutory frameworks and only government could drive through the new legislation essential for fundamental reform. And although the Welsh Government, like its Scottish and UK counterparts, does not directly provide housing or homelessness advice, it is a hugely important funder of services provided by third sector and local government – groups who were represented in the networks. So there were already strong incentives for other groups to participate in homelessness networks sponsored by the Welsh Government. Indeed, we found that as the Welsh Government acquired more formal powers, these incentives grew stronger still. Other groups came to see it as increasingly important that they should be involved in networks to help shape the outcomes that those powers would deliver. But we also found that in the operation of the networks, the boundaries between the categories of metagovernance activities were blurred.

Of course, these categories are to some extent ideal types, and Sørensen and Torfing suggest that, in practice, a combination of tools at different stages of the network’s operation might be most effective. But in our case, we saw that tools had been deployed concurrently, and many interventions could be fitted into more than one category. This distinction between ‘hands off’ and ‘hands on’ categories became especially blurred: officials shaped, steered, managed and participated in networks at the same time, playing what we have called a ‘governor-participant’ role. And it was a very small group of officials who did this: the Welsh Government’s homelessness policy team, which carried out most of the day to day metagovernance of the networks, consisted of no more than three officials. These were highly experienced and respected in the homelessness sector, but they were not the more senior policy officials who Sørensen and Torfing suggest are generally more likely to possess metagovernance skills.

In short, then, we found that the Welsh Government’s formal resources gave it decisive advantages as a metagovernor, and it deployed all four categories of governance tool identified by Sørensen and Torfing. But it did this in ways that had not been described in the existing literature, with the governor-participant role of relatively (hierarchically) junior officials being especially notable. In Wales, there were quite short ‘chains’ linking policy development and policy delivery, and quite small numbers of groups and individuals involved in these processes. This meant that many of these groups and individuals were involved both in shaping and implementing policy, and this was true inside as well as outside government.

It would be useful to look at other cases and other subnational governments to see if these conclusions can also be drawn elsewhere. But meanwhile, our research suggests that the size and position of networks can make a difference to how a government manages them, to who does that managing- and of course to the success of the policy outcomes pursued.

Sørensen, E. and Torfing, J. (2009) ‘Making governance networks effective and democratic through metagovernance’, Public Administration, 87(2): 234–58.


Andrew Connell is a Research Associate at the Wales Centre for Public Policy, Cardiff University