(adapted from MacRae, 2011)
A range of instruments are used by governments to advance their agenda. Instruments can be used to have direct and indirect effects, by: setting the law, or influencing directly or indirectly social norms, markets, and the architecture of an issue (Lessig, 1998). This table provides a quick summary and more details are provided under each Instrument heading in the menu. Note that although legislation and regulations affect food and agricultural law, for the most part private law is not addressed on this site. Food system actors are affected by a large body of general law (property, contracts, employment, business, tort) For more on this, see two volumes of Halsbury's Laws of Canada by Don Buckingham, Agriculture and Food.
|Policy statements||Usually high-level statements, delivered or approved by members of governing parties. They set out policy goals but can, however, be more operational, and even very target specific, with precise outputs, outcomes and timelines (Howlett, 2009a)|
|Legislation||In recent Canadian tradition, legislation is broad and enabling and is revised infrequently, yet change typically requires creation of new legislation and/or modification of existing acts|
|Regulatory changes||Senior bureaucracies create regulations that give more guidance to the legislation, and construct regulatory amendments.|
|Regulatory protocols and directives||But since regulations themselves are also relatively vague, agencies must create detailed regulatory protocols and directives to guide day to day activities and decision making. These are not always public.|
|Other instrument choices and changes||Governments often use a range of other instruments, including programs, educational mechanisms and taxes or tax incentives to support their policy statements, legislation or regulations. When bundled together, they can be called a strategy or a plan.|
|Structural changes||The loci of decision making often have a significant impact on how changes are made.|
Instruments are often interconnected, especially when governments want to generate policy coherence and substantial change. On this site, the presumption is that instrument mixes are necessary, complementary and purposeful, with multiple targets, involving multiple actors and direct and indirect effects.
The Classic dilemmas of instrument choice: what is effective (adapted from WWF, 2004)?
Instrument design and execution is often affected by trade-offs that can be overlapping. Some key ones are summarized here.
Voluntary vs. mandatory
Voluntary programs are typically more acceptable to those being targeted, so in theory buy-in is higher. But, the knock against voluntary programs is that they generally attract those who have already moved significant in the direction identified by the program. Those who are usually the biggest contributors to the problem or suffer the most from the phenomena, are less likely to participate. Uptake rates can be high, but not necessarily engage a high percentage of the priority actors. If the instrument produces benefits beyond the individual participant (referred to in economics as non-excludable), the "free rider" problem is triggered, where non-participants receive rewards without participating and often paying in some way for the benefit.
Mandatory programs, in contrast, involve all targeted actors, but there is usually significant resistance to participation and some actors actively attempt to "work around" the rules. Alternately, actors may receive a benefit that they don't really need. The wider scope of the initiative, however, often leads to more fairness and faster rates of change because it involves the most reluctant or the most in need. As a consequence, it is usually more expensive. Recently, Canadian governments have been shying away from mandatory programming, but without much evidence that this is an effective strategy.
Targeted vs. general (or selective vs. universal)
Targeted (selective) approaches concentrate instrument choice and design on those who most contribute to the problem, or those most in need. Success is very dependent on good data and the ability to reach those targeted. More general (or universal) approaches do not require the same precision and can use broader approaches to connect with potential participants. Universal approaches are often more acceptable to the public as it tends to reduce discussions about the degree to which targeted groups merit support or penalty. But, typically, they are more expensive to implement in terms of gross costs, but if the program saves public expenditure in other domains, then the net costs can be reasonable for the benefits generated.
Performance vs. practice
Performance-based supports provide rewards for documented improvements in the phenomena to be improved (e.g., pollution reduction, economic activity, social equity). Practice - based supports reward actors for adopting certain practices that are likely to generate performance improvements, though such improvements are not usually measured.Some believe that performance-based instruments are the way of the future but certain conditions have to be in place to make them viable and cost-effective. According to the Wallace Centre (2003), these include:
- a realistic understanding of the causes of the problems, including the time frame over which they evolved, appropriate potential remedies, and the time over which those remedies may produce measurable improvements.
- (for economists and policymakers) valuations of the costs of doing nothing and the improvements.
- (for farmers and firms) how various practices relate to the problems.
- (for agencies) information on actor actions and resulting positive outcomes.
There are numerous implementation challenges because of our poor record of performance data collection. The primary challenge is to identify performance indicators that are both measurable at reasonable cost and clearly linked with farm and firm management decisions (Wallace Centre, 2003). Programs that reward actors for measurable results must differentiate between improvements that result from management decisions, and those factors beyond the control of the actor (Claassen and Horan, 2000). Setting quantifiable performance standards is often very complex, due to the heterogeneity of site-specific farming and food system activity. For example, a farm with a low level of sediment loss which is located next to a water body may impose greater environmental damages than a farm with a high erosion level located further from the water body (Weersink et al, 1998).
A related part of this dilemma is point vs. non-point effects. When the problem can be isolated to a specific point, i.e., a specific polluter, or a specific economic actor, then performance-based measures are often very appropriate. But many food system problems are non-point problems. There are many actors, in a wide range of locations, that are contributing somewhat to the problem. Monitoring them is almost impossible, and certainly very expensive.
Incentives vs. penalties
There's often a debate in instrument design circles about whether those to support to make changes should be provided incentives or be penalized for non-compliance with expected befaviour. For complex areas, the discussion is often about getting right the mix of the two. Incentives can include information, subsidies, grants, tax relief or reductions, supportive programs and marketing assistance. Penalties include regulations, taxes and punitive eligibility and operational rules.
Incentives are understandably more popular, but do not always produce significant uptake and improvements. Much depends on the size of the incentive, how many actors are eligible, and the effectiveness of incentive delivery. Consequently, they can be expensive. Penalties often provoke hostility and avoidance strategies and may not always impact the biggest contributors to problems, unless combined with targeting and mandatory dimensions. The implementation costs can be relatively low, but monitoring and compliance costs can be high if there is significant opposition.
Polluter pays vs. sectoral actions
For environmental problems, there is significant debate about whether polluters should be forced to pay to prevent and clean up pollution, or if more sectorally-based support programs are warranted. This debate is somewhat connected to issues around point / non-point pollution. Point source pollutants, such as those emitted from an inputs or food manufacturing facility, are easier to monitor and the negative impacts can sometimes be (at least crudely) quantified. This lends itself to making the polluter pay for both clean up and pollution prevention. Non-point source pollutants are by definition more diffused and difficult to assign to specific actors. Soil erosion in a watershed is an example. Many farms are contributing to varying degrees based on shifting temporal and spatial conditions. There may be interactions between the farms that augment or reduce the impacts. In such cases, a sector / geographic approach can often be more effective. Sometimes the polluter pays principle can be partially brought to bear on sectoral approaches by cost sharing the clean up and pollution prevention.
Practices vs. systems
Systems are more than combinations of individual practices. As it relates to agriculture,
- Systems comprise a fundamental and permanent land use change. Ensuring long-term benefits with a practice-based approach is difficult, since changes in practices are driven by a host of environmental and financial conditions, causing a great deal of year-to-year variation.
- In a practice-based approach, growers pick and choose components, making the adoption process difficult to manage and monitor. While systems approaches also involve choosing between options, these options come in more integrated packages.
- Monitoring for compliance is very difficult in a practice-based approach. With systems, monitoring is not without challenges, but more straightforward.
- Few practices are likely to reduce producer costs, whereas systems have the potential to generate greater profitability (e.g., organic food).
- Only systems adoption will achieve significant environmental improvements (see MacRae et al., 2014a).
- Practice-based approaches often focus on a mix of (potentially incompatible) strategies. For example, efforts to improve soil moisture capacity to moderate moisture variability may result in increased N20 emissions.
Many current instruments are designed to support practice adoption which is often useful for discrete problems with clear cause and effect relationships. However, most significant food system problems have multiple causes and multiple effects which makes practice-oriented instruments less helpful, unless bundled with other measures designed to promote system change. These are sometimes referred to in the policy literature as systemic instruments (Wieczorek and Hekkert, 2012).
Participatory vs. expert driven
Policy and instrument choice has historically been seen within government as the domain of elected and unelected officials. In this view, community members and non-governmental actors are insufficiently knowledgeable to participate in instrument design. This is often the case, but community members and NGOs typically have extensive information on on-the-ground conditions that affect what instruments are most appropriate and how they should be deployed. Often effective deployment is dependent on community participation. All this speaks to the need for collaboration, but government officials are typically unwilling to fully engage.
Rowing vs. steering
Popularized by Osborne and Gaebler (1992), the analysis suggests that governments should focus primarily on steering and leave the rowing to the private sector. In a neoliberal interpretation, this effectively means reducing regulatory interventions and creating more space for private firms. An approach more consistent with regulatory pluralism, however, sees the state maintaining a regulatory presence in areas where private firms have historically failed to protect the public interest, and playing a steering role on complex issues that require pluralistic implementation. This trade-off is connected to the traditional economic narrative of "command and control" (rowing) vs. "market instruments" (steering). Economists typically favour the latter, ostensibly because they have lower administrative costs / unit of change, and they contribute to firm innovations. Market-based instruments are typically seen by economists as more efficient because they don't "overpay" for a given action to be undertaken. For example, for producer subsidies, an auction scheme is preferred to a fixed rate. Pollution trading schemes are preferred by neoliberal economists than taxation schemes.
One change area where a steering approach is more likely to be successful is food and culture (Goal 9). Although direct interventions are sometimes required to create greater equity especially in areas where firms will fail to act in societal interests, in general, the role of the state in a pluralistic capitalist democracy is to creating supportive environments for positive individual and collective choices around food and culture. By altering the numerous factors that influence personal food / culture choices so that they better favour sustainability, justice and health, positive changes can result.
Elite accommodation vs. equity
A long-standing critique of many governments and their interventions is their inclination to accommodate political, economic and cultural elites at the expense of lower and middle income groups, black, indigenous and people of colour. This has sometimes been justified as a "trickle down" strategy, the assumption that favouring certain elite groups will spread benefits to others. However, the concept is highly contested, with some arguing that it just reinforces existing inequalities (cf. Summers and Balls, 2015). See Frameworks, Food Justice.
Other interventions have equity as a specific objective. However, the design of the intervention can be a significant determinant of whether improved equity is a real outcome. McGill et al. (2015), for example, assess the equity dimensions of healthy eating interventions. Upstream interventions that focused on structural elements of the food environment, such as price, were more likely to have positive equity impacts than downstream interventions that focused on individual behaviour modification.
Regulatory vs. precautionary science (adapted from MacRae and Alden, 2002)
According to Salter et al (1988) , “the science [policy-makers] seek is one that is capable of being justified and explained to a wide variety of publics.... It must facilitate clear choices. It must represent a body of evidence on which decisions can rest and be seen to be rational.” Regulators attempt to minimize the likelihood of concluding there is an effect when one doesn’t exist, avoiding the possibility of “unnecessary” regulation (known in statistics as minimizing the possibility of a Type I error). However, this increases the likelihood of creating a different kind of error: concluding there is no effect when one actually exists (or a Type II error) (Tickner, 1997). The likelihood of a Type II error can be high; in some studies representing up to a 50 per cent possibility (Schrecker, 1984). Thus, there is a greater likelihood that policymakers will claim that a risk doesn’t exist when it does, than the other way around. Thus, policy-makers often seek a high level of certainty before acting, ie, do not practice precaution. But our understanding of potential problems is usually in evolution, especially in the early stages of its emergence and definitive conclusions often take years to emerge. Rather than deal with scientific ambiguity, regulators generally treat the absence of evidence as evidence that there is no relationship. If a phenomenon has yet to be observed, then it does not exist. There is little room for the possibility that the effect has yet to be observed because we do not know how to “see” it. Although there is now significant (though incomplete) use of precautionary approaches, especially in Europe (cf. Science for Environment Policy, 2017), little of it as applied in Canada.
Single jurisdiction implementation vs. leakage
Leakage occurs because of the multiple jurisdictions in Canada but also internationally (particularly trading partners). An instrument deployed in one jurisdiction is often not widely deployed elsewhere creating advantages and disadvantages for those actors in that jurisdiction. The most dramatic form of leakage occurs in cases where the instrument is seen to add costs for business and, consequently, a firm decides to leave one jurisdiction for another where the instrument has not been deployed and no measures are in place to create equal conditions. Consequently, additional instruments that fall under the category of border adjustments may need to be enacted to create a level playing field across borders. Many international agreements are meant to address the leakage issue but are not always successful and for many issues there are no widespread multilateral agreements. However, leakage may also be countered, at least in the consumer goods sector, by the bad press that can accompany flight from regulatory requirements.
Demand-side vs. supply side
Most Western governments preference supply side instruments and are more likely to use substantial interventions to effect supply than demand (see Get Started, History of attempts to create food policy). Daugbjerg and Sønderskov (2012) have created a typology of demand side / supply side mixes, using interventions to support organic farming as a case. Using their typology, it is clear that Canada pursues the most passive demand side interventions for organic farming development. And they conclude that countries using the most active market development interventions are more likely to have the highest organic food consumption. All this suggest that the current supply-side focus of most governments is misdirected.
Why are instruments to solve Canadian food system problems often poorly designed and executed (based on MacRae's experiences with instrument designers) ?
Canadian food system interventions are regularly criticized for poor design and ineffectiveness. And most instruments are rarely well evaluated. Instrument mixes, such as those required to advance a joined up approach to food policy, are even more rarely assessed.
"Ottawa is a town that is in love with policy ideas and bored to tears by policy implementation - by the details of how things actually get done" (Testimony by former Chinese ambassador David Mulroney before a Parliamentary Committee on Government Operations, cited in Bergen, 2020). This problem is particularly acute among elected officials who typically don't have the time, training, staff resources or inclination to understand the details. The details are then left to civil servants, and often the senior bureaucracy is equally uninterested in the details. Oversight by elected officials of the details of implementation is relatively limited in the Canadian governance system (cf. MacRae et al., 2012 for a pesticide legislation case study of how this plays out).
These are the most common contributing factors to poor design:
1. Insufficient data to inform design parameters
2. Political ambivalence about aggressive interventions
3. Insufficient budget allocations that result in partial designs, poor execution and very limited monitoring and enforcement. See for example, the discussion under Subsidies and blog posts on the national food policy funded initiatives.
4. The 80/20 factor - decision makers are only interested in the 20% of actors who generate 80% of the activity
5. Inherent biases among the experts that limits how they see the problem and the solutions, including biases against instruments and designs that have been used successfully in countries not deemed like Canada
6. Ineptitude - designers don't have the skills or knowledge to well address the complexity of the situation
7. Risk aversion - fear of taking an approach that isn't well proven, is politically difficult, or that might be contestable under free trade agreements.
8. Self-interest - designs are employed that ensure on-going work for the designers which doesn't necessarily line up with what is effective.
9. Design delusion - managers see themselves as good at design so they alter work done by their staff that is actually superior
10. Failure to consult well those most affected by the interventions which often leads to negative secondary effects that compromise the outcomes; related is a focus on the "usual suspects" for consultation, often elites