Abstract
Discussions of failure in public policy have been hampered by a lack of consensus on a definition of the term ‘failure’. It can be shown that arguments relating to policy failure tend to conflate forms of failure that are actually discrete, such as failure to meet objectives, claims of negative distributional outcomes and negative electoral outcomes attributed to specific policy decisions. This article attempts to unify and clarify the discourse on policy failure by presenting a multi-dimensional approach that can identify separate aspects of failure within a single policy or program. This multi-dimensional approach to policy failure is then be applied to climate change policy in Australia, in order to demonstrate how some aspects of a policy can be interpreted as failed while others can simultaneously be interpreted as successful, even by the same observer. As this example illustrates, global pronouncements of a public policy as a ‘success’ or ‘failure’ should be avoided in favour of more precise evaluations of what kind of failure occurred, and who was affected and in what ways.
Introduction
Discussions of failure in public policy have been hampered by a lack of consensus on a definition of the term ‘failure’. While the term has been used in various contexts for several decades, no commonly agreed-upon definition of ‘failure’ has emerged in either the academic literature or in practitioner circles. In fact, it can be shown that arguments relating to policy failure are not only ambiguous, but they also tend to conflate forms of failure that are actually discrete. This imprecision has led to confusion in theoretical debates as well as uncertainty in policy evaluation, as opposing voices tend to talk past each other rather than contest well-defined positions.
There is therefore a pressing need to unify and clarify the discourse on policy failure so that theoretical and practical advances can be made. Of course, given the vast potential for subjective understanding of policy outcomes, it may be impossible to produce truly objective criteria for determining success or failure in public policy. Nonetheless, it is certainly possible to pinpoint particular failures and non-failures that occur within a single policy or program, so that arguments related to policy failure might offer more precise and therefore useful analyses. In this article, we develop a multi-faceted framework that scholars can use to identify separate aspects of policy failure. We then apply this framework to an analysis of climate change mitigation policy in Australia under prime ministers Rudd and Gillard, in order to demonstrate how more refined assessments can be made of the kinds of failure that occur.
There have been previous attempts to identify dimensions of success and failure in public policy (e.g. Bovens, 2010; Howlett, 2012; McConnell, 2010a). Existing frameworks for discussing policy success and failure tend to distinguish ‘political’ outcomes from substantive program outcomes. But this literature does not go far enough in delineating the various forms of failure described through evaluations of public policy, especially in establishing a difference between the achievement of stated objectives and the perceived distributional results of policy outcomes. In addition, evaluation of failure is affected considerably by time, which may cause outcomes to seem disappointing in the near term but perhaps as successes (or at least, not failures) in the long term, and vice versa. These previous frameworks can therefore be improved by further clarifying and more precisely defining the kinds of failure that can occur, and also by acknowledging the dynamic effects of time on any analysis of public policy.
Furthermore, because the formulation and implementation of policy are a complex undertaking that usually involves numerous actors and multiple processes, generalised pronouncements of a policy as a ‘success’ or ‘failure’ are difficult to support. It is, however, possible to show how some aspects of a policy can be interpreted as failed while others can simultaneously be interpreted as successful, even by the same observer. It might therefore be more productive to avoid general claims of success or failure, and instead to generate refined evaluations of the separate dimensions of a policy and its outcomes. Any framework that purports to aid in the determination of policy failure should therefore not aim to help diagnose ‘failed’ policies but should instead help to identify which aspects of a policy have failed and to explain why these aspects ought to be considered to have failed. The key is to divide the analysis into separate components so that different forms of failure are not confused.
The climate change mitigation strategies proposed, modified, and finally enacted by the Australian federal government under Kevin Rudd and Julia Gillard between 2007 and 2013 provide numerous examples that can effectively illustrate the utility of a multi-faceted approach to evaluating policy failure. While calls for action on climate change from the scientific and academic communities have been strong and persistent, the political and media-driven debates have remained polarised over which policy instruments and political tactics to employ – if any – to address the issue. However, before considering this policy example, it is necessary to establish some key features of an analytical framework.
What does ‘policy failure’ mean?
Failure is a common descriptor in the policy sciences. Unfortunately, its use is not consistent. Many authors are vague in their writing about policy failure, to the extent that it is often unclear what exactly has failed and in what way. In other cases, policy failure has been defined but different authors have produced different definitions, with no consensus emerging on the meaning of the term. Policy scientists are left without a common understanding of what it means to say that a policy has ‘failed’. Here, we use the term ‘policy’ in the sense of general public sector activity, and it should be understood to include formal policy ideas, projects being shaped for implementation, and ongoing programs to deliver services to the public (Bovens and ‘t Hart, 1996: 16).
A further point of confusion is that the term ‘political’ can connote both resource-distribution competition as well as political partisanship activities. These two meanings involve similar dynamics of competition, lobbying, agenda-setting, coalition-building, reward and punishment and with alternating factions in positions of power and authority. Although the concepts are obviously related, there is an important conceptual distinction between the ‘political’ activities that pertain to legislatures, political parties, governments and organised oppositions, and the interactions of the variously identified groups that compete for resources within the general population. In the policy failure literature, these concepts are sometimes combined, but there is some utility in analysing them separately. Here, we will refer to the competition for resources among segments of the population as ‘distributional’ and the interaction between political parties and organised movements as ‘political’ or ‘electoral’. It should be noted, of course, that elections are not the sole instrument for measuring political/electoral outcomes.
The common modes of policy failure
Most commonly, policy is seen to have failed when it has not been able to achieve its official or stated objectives. These objectives can be specified very precisely, as in the Australian government’s agricultural re-establishment grant program designed to encourage small farmers to exit the farming industry (Botterill, 2001). In other cases, policy objectives may remain generalised and imprecise, such as in America’s ‘war on drugs’ (Gray, 2012). Reasons for not being able to achieve policy objectives can include flaws in formulation or conceptual approaches (e.g. Bagley, 1988), flaws in implementation design (e.g. Logsdon, 1986), or simply because the objectives cannot be achieved at all (Kerr, 1976).
Secondly, observers tend to see failure when a policy produces negative distributional outcomes, such as financial or other material losses, reduced access to opportunity, negative public health outcomes, or infringements of rights. These negative outcomes can occur within a policy’s target population, or in another group, or can be manifested as externalities to the policy in question, in which case they might apply to the general public. Van Zwanenberg and Millstone (2003), for example, rate the UK’s response to bovine spongiform encephalopathy (BSE, commonly referred to as ‘mad cow’ disease) as a policy failure because it did not minimise the risk of exposure of the disease to the broad population (that is, the British public) that it was supposed to protect. In another example, the Canadian province of Québec’s preferential language laws may have benefited the majority French-speaking population by improving their access to high-paying employment, but these laws also resulted in a reduction of employment opportunities, feelings of alienation, and massive waves of emigration among the resident minority English-speaking community (Endleman, 1995). And policy failures related to negative externalities, particularly the environmental consequences of specific policies, have been widely reported (e.g. Hall, 2011).
In some analyses, policies are deemed to have failed when they cost too much, waste money, or do not provide adequate value (Andrews and Bonta, 2010; Evans, 1999). While it may be tempting to classify excessive cost as a separate category of policy failure, excessive cost can be more inclusively interpreted as a form of negative distributional outcome. When an observer comments that a policy failed because it wasted money, the implication is that the opportunity-cost is serious, namely, the money spent needlessly on one particular policy or program could have been spent in other policy areas where it could have benefited other groups of people.
Next, policy is often seen to have failed when it contributes to negative consequences for a particular political party or government. These negative political/electoral consequences can manifest themselves through an election result or simply in the changing court of popular opinion. In Canada, for example, the provincial government of British Columbia under Gordon Campbell’s Liberal Party suffered a major loss of popularity after instituting a new sales tax that ultimately contributed to the premier’s decision to resign (Shaw, 2010). This has been interpreted by some observers as a policy failure (e.g. McArthur, 2011), especially since the tax was repealed following a referendum. In the United States, the Iran-Contra affair, which significantly reduced President Reagan’s approval rating, has been seen as a policy failure (Brody and Shapiro, 1989), but it did not prevent Reagan’s vice president, George HW Bush, from being elected president in 1988. Thus, some authors are willing to interpret political outcomes or negative public opinion as a policy failure even when there is no change of government.
And lastly, policy is often seen to have failed when obstacles arise that prevent even a well-defined policy from being effectively implemented. In Australia, for instance, despite a long-standing formal recognition in several states of the need for protection of freshwater ecosystems, no enforceable policies to this effect were implemented (Nevill, 2007: 44). In another example, the state of California was not able to implement a plan to build a low-level radioactive waste containment facility in the 1990s – even though legislation had been passed, a site had been formally chosen, and a contractor had been selected – because the state government could not counter an aggressive campaign of disinformation and creative re-branding from the anti-nuclear lobby (Albrecht and Amey, 1999). In cases like these, authors see the inability to implement policy as a failure of the overall public policy process.
Frequently, the boundaries between these phenomena are blurred in discussion, and it becomes unclear whether, for example, an author considers a policy to have failed because it did not meet its objectives or because it resulted in a negative distributional outcome. Birkland (2009), for instance, interprets the US Federal Emergency Management Agency’s response to Hurricane Katrina as a policy failure, but it is unclear whether he is referring to a failure to achieve specific material objectives in the recovery effort, or a failure in terms of the general inability to protect the hurricane-affected population from the impacts of displacement, trauma and loss of property. In these cases, the analyst implicitly assumes that a policy’s objectives are to maximise the welfare of the policy’s target population, and distributional outcomes can therefore become confused with objective attainment. However, welfare maximisation is not always the primary objective of a policy or program, so this assumption should be considered with care.
In addition, some particular aspects of a public policy can be considered a failure while others simultaneously are not. For example, automated cameras have been used in many countries to issue penalties to drivers caught speeding or disobeying traffic lights. These have proven to be highly unpopular in some jurisdictions (Wells, 2008) and have become major election issues in others (e.g. Chen and Warburton, 2006: 666), to the extent that policy failure on political/electoral grounds could be argued. However, studies have repeatedly confirmed the success of automated traffic cameras in reducing both speeding and collisions (Wilson et al., 2010), indicating that policies to install traffic cameras can attain a stated objective of reducing vehicle accidents while also achieving distributional benefits of greater safety for motorists and pedestrians. In another example, ‘tough on crime’ legislation, which includes mandatory minimum sentences and ‘three-strikes-you’re-out’ laws, has proven to be very popular among voters in many jurisdictions (Frost, 2010), despite numerous evaluation studies questioning its effectiveness in achieving crime reduction objectives (Chen, 2008; Cook and Roesch, 2012; Turner et al., 1999) and despite documented negative distributional outcomes for minority communities in some countries (Cook and Roesch, 2012; Newell, 2013). And more famously, state-endorsed abuses of civil rights, such as the various Jim Crow laws in the US states, achieved their policy goals of repression, segregation and discrimination, but were obviously detrimental to the populations they targeted. It would be repugnant to declare these kinds of discriminatory policies to be ‘successful’, even though they may have achieved their objectives as originally stated.
The purpose here is not to highlight opposing viewpoints on controversial issues. Whether decision-makers and the public believe that traffic cameras or harsh criminal penalties are desirable policies is not the focus of this discussion. Rather, the point is that even when different observers agree about facts and interpretations pertaining to a particular policy area, there can still be multiple and conflicting interpretations of policy failure. Failure in one area of a policy does not necessarily imply failure in other areas, and so objective-attainment failures can co-exist with political ‘non’-failures and vice versa. While failure to achieve policy objectives is certainly the most frequently examined mode of policy failure, and is the main focus of evaluation studies, the possible co-existence of separate aspects of failure and success within a single policy only serves to emphasise the inadequacy of using only one dimension (whether objective attainment or any other single mode) as a single category of analysis for policy failure. Declarations of total policy failure, therefore, may be difficult to maintain.
These dynamics are compounded by the unavoidable effect of time on policy analysis. Events that may appear to point to policy failure in the short term may indicate more positive returns in the long term, and vice versa (McConnell, 2011: 66). The Sydney Opera House, for instance, was considered to be a hugely expensive and wasteful undertaking around the time it was designed and built, but is now recognised as ‘one of the most famous and iconic buildings in the world’ (McConnell, 2010a: 87). The Olympic Stadium in Montreal, by contrast, was seen as an architectural and technological achievement early on, but was not actually completed until 11 years after the 1976 Olympics and was not finally paid for until 2006, by which time its two main professional sports tenants had already ceased to use the facility (Roult and Lefebvre, 2010; Whitson, 2004). Evidently, then, it is not adequate to simply declare a policy to have failed; high levels of precision are required for meaningful explanations of what kind of failure occurred, who was affected, and how the impacts of the failure were felt by the relevant groups over time.
Deriving a framework for focusing the analysis
Several authors have already taken up the task of creating distinct categories for policy evaluation. Ingram and Mann (1980: 20), for example, point out that while a policy can be advanced on the basis of solving a pressing social or economic problem, the same policy often has additional political/electoral or distributional purposes that can sometimes be at odds with the policy’s stated socio-economic objectives. In Ingram and Mann’s example, welfare policies were developed in the US ostensibly to help the poor, but their implementation had an additional political/electoral function of quelling urban riots. In other words, public policies can have multiple, parallel and sometimes divergent or even conflicting functions, and these various functions need to be addressed and evaluated separately.
Mark Bovens and his colleagues (Bovens, 2010; Bovens and ‘t Hart, 1996; Bovens et al., 2001) have suggested that the various functions of public policy that Ingram and Mann refer to are of two general types which can be assessed by viewing them through separate lenses: there is a ‘program’ perspective, where the achievement of policy objectives and the repercussions of policy outcomes are considered, and a ‘political’ perspective, where the dynamics of political processes and electoral outcomes can be examined. Program assessments concentrate on ‘effectiveness, efficiency and resilience’ and ‘social problem-solving’; political assessments address ‘the way policies and policy makers become represented and evaluated in the political arena’ (Bovens et al., 2001: 20). The benefit of this approach is that policy evaluation efforts can be categorised according to the policy functions that they aim to evaluate, and therefore multiple assessments can be performed at once. In the US welfare example cited above, a policy’s socio-economic objective of reducing poverty could be evaluated for success or failure independently of the political role that the same policy might play in reducing civil unrest. Therefore, a single policy might be a success in one respect while simultaneously a failure in another.
In addition to considering which functions of policy to evaluate, there is the related issue of where to position an evaluation along the development trajectory of a policy. In this vein, a much larger literature has pursued a debate as to which stage of the policy cycle evaluators should focus on when determining success or failure (deLeon and deLeon, 2002; Linder and Peters, 1987; Matland, 1995). Some of these authors take a firm position on whether policies should be judged primarily at the formulation stage (i.e. by evaluating the achievement of a policy’s objectives) or at the implementation stage (i.e. by evaluating the effects of a program on target populations); however, this is a false dichotomy as there is no reason why these stages of the policy cycle cannot both be evaluated. Doing so would give two independent dimensions for policy evaluation: a policy cycle or ‘locus’ dimension, ranging from formulation to implementation, and a material-political or ‘focus’ dimension (Bovens, 2010).
This is a major advance in the creation of a framework for assessing policy success or failure. Unfortunately, a two-dimensional framework is not entirely adequate. For one thing, most descriptions of the policy cycle include more than two stages (Howlett et al., 2009). Secondly, conceptualising the dimensions of analysis in this way produces overlapping and mutually exclusive categories. For instance, it is difficult to imagine a situation in which a policy fails politically at the formulation stage but achieves political success at the implementation stage. Presumably, if a policy is politically unpopular enough to merit ‘failure’ when it is being formulated, it might not progress to the implementation stage at all. Any framework for evaluating policy success or failure must admit the possibility of overlapping dimensions, since it would be impossible to completely compartmentalise the intentions and outcomes of any particular policy. However, it is important that separate categories of analysis do not have the potential to cancel or contradict each other.
David Marsh and Allan McConnell (Marsh and McConnell, 2010; McConnell, 2010a, 2010b), in their well-developed articulation of policy success, divide evaluation into three categories: process, program and politics. These three categories encapsulate both locus and focus of analysis at the same time. The ‘process’ category addresses the act of moving a policy from idea to implementation; the ‘program’ category allows for evaluation of a policy’s outputs; and the ‘politics’ category is where evaluations of the political and electoral consequences of a policy take place. Thus a policy could be a process success, if it manages to navigate ‘a complex, veto-point-filled and multi-actor approval process to creation and implementation’ (Howlett, 2012: 545–546), while simultaneously resulting in political failure for the government that enacted it. Marsh and McConnell’s (2010: 572) example of Australian prime minister John Howard’s 2005 labour reform law (‘Work Choices’) provides a case in point, as it sailed through an expedited legislative process but ultimately contributed to Howard’s defeat in the 2007 election.
A major contribution of the Marsh and McConnell framework, therefore, is that it acknowledges that the process of enacting policy is a highly contested and precarious activity that can be judged to be successful or unsuccessful independently of the various other aspects of policy. The ‘program’ category is left intact, but the ‘political’ category of previous conceptualisations is now separated into a process category and a political outcomes category, which as the Australian labour law example illustrates, should be considered separately. This obviates the need for a debate on whether formulation or implementation is a better or more legitimate choice for policy evaluation.
However, this framework can be improved further. The ‘program’ category addresses both objective-attainment considerations and distributional outcomes at the same time. As discussed above, while some observers merge these issues, they are in fact separate public policy concerns. A policy that without question achieves all of its stated objectives can still produce negative distributional outcomes, for its target population or even for the general public. Accordingly, therefore, there are four categories of evaluation, rather than only three, in which success or failure might be identified: objective attainment, distributional outcomes, political/electoral consequences and policy process issues related to effective implementation. And thus we have arrived at the framework that is reflected in the wider literature that employs the term ‘policy failure’, in which these four categories are frequently independently cited as areas in which failure can occur. The main weakness of the existing literature is that it does not formally recognise that failures and successes in the different categories can exist simultaneously.
And lastly, as previously discussed, time is an important factor in policy evaluation. Failure is not a fixed or terminal state; policies that appear to have failed now may yield successful outcomes in the future, owing to changes in circumstances, or to shifting norms, value sets, or demographic identities. Nonetheless, time cannot be considered to be a separate category for evaluation, because it affects all areas. Objectives that appear not to be obtained in the short term can be achieved in the long term; distributional outcomes may appear to have failed a population now but may be perceived to benefit that same population later; and so on. The effects of time should therefore be considered in addition to each of the four categories outlined above.
Policy failures in Australian climate change response policy
The Australian federal Labor government under Kevin Rudd and Julia Gillard (a period of nearly six years ending September 2013) adopted several strategies to address climate change, the most notable being a system of financial penalties on companies in heavy greenhouse gas emitting industries. While some have attempted to label the entire area of greenhouse response policy as an ongoing failure (Christoff, 2005; Crowley and Walker, 2012; Mercer et al., 2007), we take a more nuanced view. By dividing the analysis into objective-attainment, distributional, political and implementation categories it is possible to show that some aspects of Australian climate change policy have resulted in failure while others have not.
The following is not intended to provide a comprehensive review of all climate-related policy activity during the tenure of the Rudd and Gillard governments. Rather, it seeks to illustrate how different categories of assessment can help refine policy evaluation efforts by facilitating more precise and meaningful analyses of success and failure.
Objective attainment
After many years in which John Howard’s conservative coalition government had refused to ratify the Kyoto Protocol, the newly-elected Labor government completed ratification in December 2007. Australia’s original carbon emissions target for the 2008–2012 commitment period of the Protocol was an increase of 8% above 1990 levels, unlike the larger reduction in emissions that applied to most of Kyoto’s other participant countries. In addition, the Kyoto Protocol allowed emissions reduction from slowing the rate of deforestation to be included in Australia’s total emissions accounts, which was a specific concession that had been negotiated by the Howard government well before Kevin Rudd’s Labor Party came to power (Christoff, 2005: 33–34). Thus the Rudd government’s decision to ratify Kyoto committed Australia to a set of well-defined and documented policy objectives.
Because the Kyoto objectives were clearly stated, ratified by the Parliament of Australia, and well-publicised, and because progress against these objectives can be readily quantified, reducing carbon emissions according to the Kyoto Protocol targets is an obvious choice for consideration as Australia’s primary climate change mitigation policy objective. Other specific policy objectives certainly existed, such as appeasing industry, obtaining the recognition of the international community, and, in the case of Julia Gillard’s government, satisfying the demands of the Australian Greens who held the balance of power in the Senate. These other policy objectives may be clearly stated, or they may be hidden or inferred. Here, we consider Australia’s success in meeting its official commitments to the first period of the Kyoto Protocol; other analyses might prefer to examine progress made against other policy objectives, but in general, arguments that intend to show that policy objectives were not met should clearly identify which objectives are being assessed and should demonstrate that these objectives were in fact intended by policy makers.
Although the Australian federal government also pursued a number of climate change adaptation initiatives and a renewable energy target (Beeson and McDonald, 2013: 336), the government’s core strategy for mitigation of climate change centred on carbon pricing in one form or another. The plan for a Carbon Pollution Reduction Scheme (CPRS), as it was initially called in 2008, began as a market-based carbon emissions trading arrangement modelled after the European Union’s Emissions Trading System, but following a series of political controversies and negotiations it finally resulted in a plan to transition to a market-based trading system after three years under a fixed-price regime. Following Rudd’s election in late 2007, an extended period elapsed while this preferred policy model was developed (Garnaut, 2008). The Labor government then encountered difficulties in passing the CPRS legislation through the Senate in late 2009, and no legislative action on climate change mitigation could be taken – although a prominent public debate developed that arguably cost Rudd his job and led to Gillard forming a minority government after the 2010 election (Curran, 2011). A compromise negotiated by Gillard with the Greens in 2011 allowed the final arrangement to come into effect in July 2012.
In total, therefore, the Rudd and Gillard governments’ achievements in the area of climate change mitigation amount to several years of prolonged debate and one year of a transitional carbon pricing system during 2012–2013, after which time Labor was voted out of office. Nevertheless, Australia’s inventory of domestic carbon emissions at the first quarter of 2013 indicated that the country had clearly met its target for the Kyoto Protocol’s first commitment period, with average annual emissions around 104% of 1990 levels (Australia, 2013: 15).
In effect, this outcome indicates that although Labor’s climate change activity under Rudd and Gillard disappointed many of their own supporters (along with supporters of the Greens who propped up the minority Labor government after the 2010 election), the delayed program on carbon pricing did not undermine the achievement of Australia’s objectives under the Kyoto Protocol. Owing to timing, there is no causal connection between the decisions of the federal government on carbon pricing and Australia’s ability to fulfil its commitments to Kyoto in the period to 2012; numerous other factors were in play during that period, especially the actions of state governments in implementing their own policies and programs, such as the New South Wales Greenhouse Gas Reduction Scheme, an emissions trading scheme active solely in the electricity generation industry between 2003 and 2012. However, it is clear that even though the Rudd-Gillard Labor governments did not enact any climate change mitigation policies until the very end of the first Kyoto commitment period, the extensive discussion raised public awareness of the issues and their policy objective of meeting Australia’s Kyoto targets was achieved. Therefore, it is not accurate to label Labor’s climate change mitigation policies as a policy failure on objective-attainment grounds.
Distributional outcomes
The impact of Australia’s climate change mitigation strategies over the long term and across a wider population will be difficult to specify with any level of accuracy. Australia’s contribution to global climate change has been estimated to be in the order of only 1.5–3.7% (Christoff, 2013: 351–352), so its capacity to affect the global climate through direct domestic and trade-related emissions reductions is small. In addition, Australia’s level of influence on the architecture and activities of any international climate regime is currently minimal (Beeson and McDonald, 2013: 340–341). This does not mean that Australian climate mitigation efforts are in vain, just that their impact on the global climate in the long term may be difficult to measure, as policies intended to have an impact on domestic emissions outputs or on global climate governance regimes will likely be obscured by the actions of bigger international players.
However, in the short term, and on a domestic level, Australia’s climate change mitigation policies can have more direct winners and losers. For example, the delay in implementing a carbon pricing regime that occurred between 2007 and 2012 granted the fossil fuels industry a reprieve from having to act to reduce carbon emissions. Carbon pricing may have long term economic benefits and this mechanism has been endorsed as such by leading economists, most notably in Australia by Ross Garnaut (2008, 2011), but carbon pricing entails short term financial penalties and adjustment costs for some companies. This is especially true for those that rely on emissions-heavy technologies such as coal extraction and metals production. These industries vigorously opposed Labor’s carbon pricing proposals and, according to some analysts, were instrumental in gaining policy adjustments including more exemptions and financial compensation for affected corporations (Crowley, 2012).
Of course, the fossil fuels and metals industries are not the only actors affected by the distributional outcomes of mitigation strategies. For example, policies that favour fossil fuels extraction are likely to disadvantage the renewable energy sector, and in fact the Australian renewables industry was surprisingly neglected during the Rudd-Gillard years when the focus was on how best to arrange a carbon pricing regime that would be acceptable to fossil fuel companies (Diesendorf, 2012). Therefore, while the climate change mitigation activities of the federal government (or more precisely, the delays in negotiating its position) may have been seen as a mixed outcome by heavy carbon emitting companies, actors in the renewable energy sector were more likely to view these activities as a policy failure and a lost opportunity.
The above analysis is not comprehensive. There are many more actors and groups of actors who have benefited or have been disadvantaged by Australia’s climate change mitigation policy. The climate change response policy field is thus inherently complex, with many trade-offs and mixed outcomes (Head, 2014); hence, any analysis of benefits and dis-benefits is bound to be both difficult and contested. The point is that in any reading of Australia’s record on climate change, failure will be perceived by some while others will perceive success, and the challenge is to identify these stakeholders and to explain how the policy has or has not failed them.
Political/electoral consequences
Climate change response has figured prominently in Australian politics since at least 2007, when it became an issue that some have argued dominated the federal election in that year (Rootes, 2008). In one sense, Labor’s direction on climate change mitigation during the Rudd-Gillard years might be seen as a political/electoral failure, as some have argued that part of Rudd’s downfall and Gillard’s minority government status after 2010 are the result of Labor’s failure to enact the climate change mitigation efforts that were promised during the 2007 election campaign (Christoff, 2013: 361; Crowley, 2012: 40–41; Curran, 2011). However, these suggestions are difficult to prove convincingly, especially given the multiple other factors that were at play in those years, including the Greens Senators voting against Labor’s legislative scheme in 2009 (on the grounds that it lacked ambition), and the macro context of the global financial crisis and attendant fears of its effects on the Australian economy.
More concretely, the government’s climate change mitigation policy in the Rudd-Gillard years should not be construed as a clear political/electoral policy failure, because public opinion on carbon pricing shifted over time. While it is true that climate change was a prominent issue during the 2007 election and that support for climate action was high at that time (Pietsch and McAllister, 2010), numerous studies have shown that the Australian public’s appetite for climate change mitigation efforts – due especially to the realisation that these efforts might have financial costs – declined rapidly and significantly after 2008 (Hanson, 2010; Leviston et al., 2011; McDonald, 2012). Furthermore, there is some evidence that politicians were aware of this decline in popular support for carbon pricing: a sample of Australian politicians at all three levels of government who were surveyed in 2009 affirmed that they knew that climate change was a middle or low priority for their constituents (even Green politicians only rated their constituents’ support for climate change issues at an average of 6 out of 10 in terms of priority) and that they themselves gave more attention to the issue than they believed was demanded by the electorate (Fielding et al., 2012). It is likely, then, that as climate change mitigation activities lost public support from 2009, it became more difficult for federal politicians to act swiftly to introduce carbon pricing. Labor’s climate change mitigation policy in the 2007–2013 years was therefore broadly in line with trends in public support, and although they were removed from office in the 2013 election, it is unlikely that this outcome was the result of policy failure in the climate change portfolio.
d. Implementation
Policy failure in climate change mitigation in the Rudd-Gillard governments.
Conclusion
The climate change mitigation strategies pursued by the Australian federal government under the Labor leadership of Kevin Rudd and Julia Gillard from 2007 to 2013 represented substantial policy failure in some areas, but less so in others. Although their carbon pricing scheme was delayed and then modified substantially from its original articulation, Australia did fully meet its declared policy objectives as a party to the Kyoto Protocol, so it is not possible to declare a policy failure on specific objective-attainment grounds. Likewise, the decline in popular support for climate action from 2009 is a reasonable justification for the delay in implementing carbon pricing, suggesting that judgements of policy failure on political/electoral grounds are not firmly based. However, the disjointed policy process, in which the government was not able to implement its chosen policy ideas, nor able to make optimal use of available expertise, may be characterised as a failure in implementation. Moreover, while some actors benefited from the delay in implementing carbon pricing, others were punished, suggesting that multiple interpretations of failure are possible in distributional terms.
These differing interpretations fuel the debate in the academic literature as to whether success and failure of policies are matters of fact (Sanderson, 2002) or of understanding (Van Assche et al., 2012). Individuals with different group-based biases can view the same events but apply different value sets when assessing the ‘success’ or ‘failure’ of those events, so blanket assessments of any complex policy initiative as a ‘success’ or ‘failure’ are likely to be untenable.
Recognising distributional outcomes as a separate category for determining success or failure becomes a powerful tool in resolving the problem of multiple and conflicting interpretations. Dividing policy failure into categories, as we have done above, allows us to identify failure as failure of what; examining distributional outcomes in particular allows for a determination of failure as failure for whom. Once this is done, steps can be made toward reconciling differences between interpretivist and rationalist perspectives. Intersubjectivity of understanding can be approached, or even reached, if observers can recognise that a policy can be a failure for one group while simultaneously existing as a success for another. The key to assessing failure in public policy lies in defining exactly which aspect of the policy is under scrutiny, and what kind of failure is being examined. Any determination of failure should specify failure of which aspects, and for whom, and the consequences of the policy’s outcomes at the distributional and political/electoral levels.