In the late 1990s, work on instrument selection began to assess the question of the potential to develop optimal policy mixes and move away from a focus on single instrument choices (Grabosky 1994; Gunningham and Young 1997). Studies such as Gunningham, Grabosky and Young’s work on ‘smart regulation’ led to the development of efforts to identify complementarities and conflicts within instrument mixes or tool ‘portfolios’ involved in more complex and sophisticated policy designs (Barnett et al. 2008; Shore 2009; Buckman and Diesendorf 2010). For them, the key question was no longer ‘why do policy-makers utilize a certain instrument?’ as it was for earlier generations of students of policy instrument choice, but rather ‘why is a particular combination of procedural and substantive instruments utilized in a specific sector?’ (Dunsire 1993; Howlett 2000; Salamon 2002; Cubbage et al. 2007; Gleirscher 2008; Gipperth 2008; Taylor 2008; Clark and Russell 2009; McGoldrick and Boonn 2010).

This new generation of design scholars began to develop what may be described as a ‘scalpel’ – as opposed to the more blunt ‘hammer’ – approach to instrument use; one that emphasizes the importance of designing policies that employ a mix of policy instruments carefully chosen to create positive interactions with each other (Sinclair 1997). Proponents of ‘smarter’ regulation, for example, proposed the development of sophisticated policy instrument mixes in which government’s combined a range of market solutions and public and private orderings in order to overcome societal resistance and effectively attain their policy goals is an expeditious and efficient manner (Gunningham et al. 1998; Gossum, Arts and Verheyen 2010). In their 1990 study of policy targets and their behaviour, Schneider and Ingram also began to systematically pursue Doern’s insight that the extent of a government’s willingness to alter the underlying behaviour of key policy actors was a major factor affecting its choice of policy implementation tool. They argued that policy-making ‘almost always attempts to get people to do things that they might not otherwise do’ and noted that:

If people are not taking actions needed to ameliorate social, economic or political problems, there are five reasons that can be addressed by policy: they may believe that law does not direct them or authorize them to take action; they may lack incentives or capacity to take the actions needed; they may disagree with the values implicit in the means or ends; or the situation may involve such high levels of uncertainty that the nature of the problem is not known, and it is unclear what people should do or how they might be motivated. (Schneider and Ingram 1990: 513–14)

That is, they recognized that each of Hood’s ‘statecraft’ resources required not only state capacity in that area – that is, a plentiful supply of the ‘resource’ – but also a corresponding belief or endowment on the part of target groups which would allow that capacity to be utilized effectively (Schneider and Ingram 1990a; 1990b; 1993; 1994; 1997).

Thus, the effective use of ‘nodality’ for example, requires the transmission of information to targets, ‘authority’ requires the enforcement capability to coerce or force targets to do something they might not otherwise wish to do, ‘treasure’ requires having the fiscal capacity to provide targets with incentives or disincentives to act in certain ways, and ‘organization’ requires the administrative capacity to provide them with some good or service directly. But in order to be effective not only must governments have an adequate ‘supply’ of these resources, but targets must also be susceptible to their deployment: effective information transmission requires credibility or the belief among targets that a government is telling the truth; the effective use of authority requires legitimacy or the belief among the target population that the use of authority is legally and morally appropriate; the effective use of treasure resources requires cupidity or the willingness on the part of actors to accept payments or make them; and the effective use of organization requires trust on the part of the target group that an administration is competent and capable of actually delivering the promised goods or services. That is, as Schneider and Ingram noted, each ‘resource’ is not an absolute entity but a relationship, composed of both a state capacity and a target group belief creating a ‘governance’ relationship between the two parties to a policy arrangement.

Thus legal governance relies on a combination of authority and legitimacy, market governance on treasure and cupidity, network governance on information and credibility and corporatist governance on organization and trust. Policy designs are sustained and constrained by these basic relationships, which help establish longer-term tool preferences, and hence policy alternative preferences, in the minds of policy designers (Jochim and May 2010).