Constructing a Constructivism

Stephen Bell argues that my work, together with that of Colin Hay and Mark Blyth, has developed ‘a somewhat confused understanding of constructivism, [which] excessively privilege[s] agency, and lose[s] sight of the significance of institutional and wider structural variables’.Footnote 2 As a consequence of this, the agents analysed by Hay, Schmidt and Blyth are said by Bell to ‘construct their realities and fields of action seemingly unimpeded, or less impeded, by institutional constraints’.Footnote 3 In doing so, they are, as Bell puts it, ‘veering close’ towards post-modernism.Footnote 4 While I have no particular problem with post-modern approaches, I do feel that this is an unsustainable characterization of this body of work.

Bell's Critique of Colin Hay

Bell's critique of Hay centres on Hay's ‘emphasis on “strategic actors” ’, ‘who must rely upon perceptions’ of their environment and whose ‘desires, preferences and motivations’ are ‘irredeemably ideational’.Footnote 5 Being ‘irredeemably ideational’ is of course problematic because it would imply Hay does not take account of institutions at all, which is hard to square with his published work. Indeed, surely the burden of proof lies with Bell to show what is wrong with this position, rather than simply positing it as somehow beyond the pale? Bell draws upon a mere sliver of Hay's vast empirical and theoretical contribution to justify this claim. As such, it appears as more of an assertion than an argument backed up by evidence.

Bell's actual argument against Hay is that his ‘irredeemable ideationalism’ rests upon a contradiction. Bell quotes Hay as arguing that ‘institutional change does indeed occur in a context which is structured’,Footnote 6 but then cites as contradictory Hay's statement that the outcome of political struggles ‘can in no sense be derived from the extant institutional context itself’. But why is this contradictory? A contradiction is the juxtaposition of two simultaneous truth statements where each entails the negation of the other. In contrast, there is nothing contradictory in positing that ‘politics occurs in a context that in and of itself does not determine outcomes’, as Hay has done here.Footnote 7 There is nothing very controversial here, and it does not look to be very far from Bell's own argument concerning what Anthony Giddens used to refer to as the ‘duality of structures’, which he draws upon later in his article.

Bell's Critique of Mark Blyth

I see three problematic aspects to Bell's criticisms of Blyth's work. First of all, Bell suggests that Blyth has an ‘ideationally “primitive” [Bell's emphasis] account of institutional life and change’, which leads to ‘a lack of empirically grounded theorizing about how agents and their ideas actually connect with institutions or indeed wider structures’.Footnote 8 I find this difficult to square with the published work of Blyth with which I am acquainted. Indeed, one could argue that setting out to provide an empirically grounded account of institutional change is precisely what Great Transformations, the book Bell criticizes, and several subsequent pieces by Blyth that he does not address in his article, all set out to do.Footnote 9

Second, and building on this point, Bell argues that Blyth sees moments of institutional failure as being critical for his explanatory framework, but ‘what this notion of “failure” actually means is not explained’.Footnote 10 Yet Chapter 2 of Great Transformations, for example, unpacks how ‘institutions provide stability … by managing and coordinating agents’ expectations about the future such that they converge and become self-stabilizing over time.’Footnote 11 Critical here are the ideas about how the economy works that agents share and that act as co-ordinating conventions. Drawing explicitly on Keynes, Blyth explains that institutional failure is the failure of the conventional wisdom of the day to co-ordinate economic expectations of future likely states of the world.

Third, Bell interprets Blyth's remark that ‘ideas certainly do matter in periods when existing institutional frameworks … fail’ as ‘being tantamount to institutional erasure in crisis moments.’Footnote 12 To my knowledge, Blyth has never published anything about institutional contexts ‘dissolving’ or institutional conditions being ‘erased’ as Bell contends in his article. In my reading at least, Blyth argues that actors both lose their faith in existing institutions and gain new ideas about their interests, under the conditions of heightened uncertainty generated in the moments when existing institutions no longer promote the co-ordination of expectations. In sum, Bell's argument that Blyth's ‘mechanisms of crisis and uncertainty largely serve to erase existing institutional conditions’Footnote 13 asserts more than seems warranted from Blyth's published record.

The problems with this argument come out most clearly in Bell's reinterpretation of one of Blyth's minor empirical cases regarding the institutional clout of the US Federal Reserve during the inflationary crisis of the 1970s. Bell seeks to prove that institutional erasure did not happen, which strikes me as odd since Blyth never made any argument in favour of institutional erasure. Instead, Blyth argues that when existing institutions failed to avert a crisis, the new convention that governed the financial markets at that point, monetarism, also had to become the convention through which the Fed sought to govern, since otherwise its actions would have had no meaning to the agents involved.Footnote 14

For institutionalists like me, this demonstrates something about an institution's credibility to be sure, but as a constructivist I see that property as a prior function of the ability of agents to interpret, and thus act within the same intellectual framework as the markets they seek to govern.Footnote 15 Its material existence is not, in and of itself, explanatory. Taken together, then, Bell's problem with Blyth seems to be his focus upon ideas in preference to institutions. But why should Blyth not focus on ideas? There is nothing that mandates that one must start with institutions and then treat ideas as the residual.

Bell's Critique of Vivien Schmidt

Bell's criticism of my own work has two lines of attack. First, he argues that I ‘zero in on … a more fluid and flexible environment in which to effect change, largely because this move ostensibly allows agents to “construct” their realities and fields of action seemingly unimpeded or less impeded by institutional constraints’ (my italics). Second, he argues, ‘Schmidt thus sees constructivist institutionalism as a better alternative because it “puts agency back into institutional change.” ’Footnote 16 This juxtaposition of statements in his article leaves the impression that my notion of agency is indeed radically ideational since agency is ‘seemingly unimpeded’ or at best ‘less impeded’ by institutional constraints. I seem to be saying, according to Bell, that agents can construct the world any way that they like. Furthering this impression Bell argues that I see ‘institutions merely as arenas which “frame the discourse” ’, with the institutionalism in discursive institutionalism ‘reduced to the “constructs of meaning that are internal to sentient (thinking and speaking) agents” ’, with institutions that ‘simply appear as a “meaning context”, “background information” or as “contingent” (the result of agents’ thoughts, words and actions)’.Footnote 17 My discomfort with Bell's criticism of my work is that he is cutting up a complex set of arguments in such a way that I do not recognize my own work in his rendition of it.

To clarify, I call my approach ‘discursive institutionalism’ to highlight the need to add a fourth institutionalism to the three older ‘new institutionalisms’ that focus on rationalist interests, historical regularities and cultural frames.Footnote 18 I see this as an umbrella concept for the vast number of approaches that deal with the substantive content of ideas,Footnote 19 as well as the interactive processes of discourse, whether co-ordinative ones among policy actorsFootnote 20 or communicative ones between political actors and the public,Footnote 21 all of which take place in specific and pre-defined institutional contexts. Unlike Colin Hay's work, whose purpose in calling his approach ‘constructivist institutionalism’ is to delineate an ontological position, I leave open where the wide range of discursive institutionalist scholars fit on a continuum between positivism and constructivism.Footnote 22 Far from rejecting institutions for ideas and allowing agents to build the world as they see fit, I take a quite moderate constructivist position.

Like Bell, I have long appreciated institutions as simultaneously constraining structures and enabling constructs of meaning that are both external to and yet internal to sentient agents whose ‘background ideational abilities’ (building on Searle's work) explain how they are able to create and maintain institutions, while their ‘foreground discursive abilities’ (building on Habermas's communicative action) explain how agents communicate critically about those institutions in order to change or maintain them.Footnote 23 I take this position precisely because it avoids the epistemological problems of radical ideationalism Bell identifies. More specifically, I most often quite explicitly precede my discursive institutionalist discussion of the politics of ideas and discourse with a historical institutionalist account of crisis-driven and/or incremental changes in rules and regularities, which is precisely what Bell seeks to do.Footnote 24 For public policy in particular, moreover, I have repeatedly argued that discourse is just one explanatory factor, along with policy problems, policy legacies, policy preferences and political institutional arrangements.Footnote 25

Temporality and institutional flexibility

Bell also argues that Hay, Blyth and I understate the flexibility of the institutions developed in existing institutionalist scholarship. Yet, making such a case encounters a temporal problem. Specifically, the contributions by Schmidt and Blyth that Bell specifically focuses his criticisms upon were published around Reference Blyth2002. The date of publication is important because it is odd for me to be criticized for not recognizing, in 2002, the contribution of literatures on institutional change published in Reference Schmidt2010 (Mahoney and Thelen), 2006 (Tsai), 2005 (Streek and Thelen) and 2004 (Campbell).Footnote 26 But even if those literatures had been available to me in 2002, I would still have disagreed with them. And once they were available, I did in fact put my concerns on paper.

In my own work, for example, I argue that Streeck and Thelen have made great progress in endogenizing change, by theorizing the incremental processes resulting from actors’ use of mechanisms of layering, conversion and interpretation – a literature that Bell applauds and seeks to build upon.Footnote 27 However, although this literature may help to describe change, it does not explain it, since to explain change they would need to make reference to what actors think and say that leads to change, hence my constructivism. Indeed, while the newest of these approaches certainly provides us with welcome categorizations of the kinds of agents who have successfully overthrown, undermined or converted institutions, Mahoney and Thelen cannot, in their own theoretical terms, explain how such agents bring about change.Footnote 28 This is because they do not theorize about the ideas such agents may use to interpret what has gone wrong, what to do about it, how to mobilize or how they persuade others to join them.Footnote 29

In short, I have difficulty recognizing both the body of work with which I am associated as well as my own particular contributions in what Bell writes. And yet when Bell argues that ‘agents interpret and construct the experience of their institutional situation using subjective and inter-subjective cognitive and normative frameworks and discursive processes’,Footnote 30 I could not agree more. But if he and I are making essentially the same arguments, then I must wonder, moving from his criticisms to his preferred alternative model, what is the value added of his morphogenic and flexible historical institutionalist alternative?

Morphogenics, Change and Endogeneity

Bell's preferred alternative is derived from scholars whose constructivism is sometimes described as ‘critical realist’ (e.g., Margaret Archer) or ‘semi-rationalist’ (e.g., Wendt or Culpepper). I would argue that the key problem with the approach he borrows from is that it is hard to figure out what ‘institutionalism’ means in such a framework. From the perspective of the non-initiated, it actually looks as if institutions have whatever causal properties the author says they have at any given moment. Specifically, Bell argues that institutions can be both resources and constraints at the same time. Such a position is intuitively appealing but, I argue, it fails in practice to produce empirically verifiable results.

Bell's alternative to what he calls constructivist institutionalism is based upon the juxtaposition of three claims. First, as noted above, is the constructivist claim that ‘agents interpret and construct the experience of their institutional situation using subjective and inter-subjective cognitive and normative frameworks and discursive processes.’Footnote 31 This statement is then juxtaposed with two flexible-institutionalist claims that ‘agents still have contingently variable degrees of agential space or “bounded discretion” within institutional settings’ and that ‘institutions … [have] … important empowering and enabling effects’.Footnote 32

Bell argues that the payoff to embedding the constructivist claim in the two institutionalist claims is that this will move us beyond ‘a series of dualisms’ concerning ‘change and stasis’ that bedevil current approaches to explaining institutional change endogenously. The dualisms are, first, agents are constrained and enabled; second, institutions are structures and structures are institutions; third, constraints are empowerments and limitations. Bell sees these ‘change/stasis’ dualisms as a mistake since ‘evolution and more radical change can be handled similarly’, with the only difference being that ‘the level of actor discretion may increase under crisis conditions’.Footnote 33 Getting beyond these dualisms, he insists, allows us to see that ‘the line between stability and change is quite blurred’.Footnote 34 Going further still, Bell then argues that these flexible-institutionalist accounts must also be embedded in wider structuralist accounts. But we must also remember that, like institutions, ‘structures can both help constrain and empower agents’. Indeed, he says that ‘structures have institution-like effects’ and ‘institutions will also typically mediate structural effects’.Footnote 35

To get beyond the dualisms, Bell invokes two causal mechanisms: dialectics and mutual constitution. In a rapid three-fold move, Bell defines the world as being a place where ‘institutions and structures are not reducible to the actors that inhabit them’, and where Archer's ‘morphogenetic’ account of such dialectics usefully sees agents, institutions and structures as mutually constitutive over time, producing emergent new properties’. Bell ends his theoretical contribution by noting that ‘structures only exert an effect when mediated through the activities of people’.Footnote 36 ‘So how does this resolve the stasis-change, agent-institution-structure dualisms noted above? I, at least, am still left wondering’.

Invoking morphogenesis and dialectics sounds good, but it still suffers the same weakness as it did when Anthony Giddens introduced structuration theory thirty years ago.Footnote 37 It restates the problem: it does not solve it.Footnote 38 Moreover, if institutions are simply environmental objects agents can use (resources), then their institutional effect, their ability to structure behaviour, falls away. To get over this problem, one would have to specify the scope conditions governing how and when institutions constrain, and by what degree, over time, which is very difficult, and it is noteworthy that Bell makes no attempt to do so. Alternatively, as I have repeatedly argued, one could consider how agents get beyond their institutional constraints, with ideas conveyed through discourse having a causal effect on their environment. Bell seems to want to make such a move, given his first statement that ‘agents interpret and construct the experience of their institutional situation using subjective and inter-subjective cognitive and normative frameworks and discursive processes’.Footnote 39 But the morphogenesis/dialectics position he invokes, where institutions are structures and resources at the same time, obviates this progressive extension of existing work.

For Bell, institutions constrain, until they do not, when they become resources, or structures, which is when agents use them to change things, thereby using institutions to change institutions. This deep endogeneity problem robs institutions as a concept of its original, and valid, analytic purchase. Turning to ideas and discourse as I have chosen to do is one way out of this impasse. Despite his own constructivism, Bell's morphogenetics and dialectics block off that route, and without that escape route one needs a metric for when institutions constrain and when they do not if they are to be both resources and constraints at the same time. Without such a metric of scope conditions one is left in the position where one declares a change and invokes flexible institutions/resources to explain it, with the ‘proof’ of flexibility being other periods where the change is not observed, which in turn shows that, by logical equivalence, institutions were in these periods non-flexible. In addition to being tautological, such a stance is purely descriptive. It is neither explanatory nor analytic.

In DefenCe of Dualisms

Bell's alternative – the embrace of dialectics and morphogenesis – posits that institutions are always changing and collapses any and all temporal dualisms. Doing so not only throws into question the conditions under which institutions would ever have an institutional effect, it raises the issue of how one would even be able to recognize a change as a change. Collapse the distinction between stasis and change and one has to ask whether, for example, a revolution is just a particularly concentrated ‘evolutionary’ change. If all change is simply incremental steps, then periods when seemingly robust institutions and the ideas that underpin them do in fact utterly collapse (the international financial risk management architecture in 2008, North Africa in early 2011) become rather hard to explain. Of course, incremental change occurs, I would never deny this, but does it matter in the way that revolutionary change matters? Recent quantitative work suggests that it does not.

For example, Baumgartner et al., show how budgetary changes follow a scale-free distribution regardless of the level of government or its national origin, a finding that really challenges notions of institutional distinctiveness and incremental change.Footnote 40 The distribution is leptokurtic, with the vast majority of incremental changes clustered in the middle, doing little to shape outcomes. What matters are those defining events that are singular, large and come in the far reaches of the tails (high impact low probability events). Given this, Bell's claim that institutions are ‘emergent’ phenomena adds more opacity than clarity. For if they are emergent, how can they be ontologically prior, causally prior and emergent all at once? The effect (the institution) cannot precede the cause (the institution). It cannot emerge from itself. Invoking dialectics and mutual constitution is neither edifying nor clarifying in this regard, and it does not improve upon existing institutional scholarship.

In short, one is either making an institutional argument or one is not. Juxtaposing agency and institution under the rubric of ‘mutual constitution’ restates rather then resolves the problem. One can say that agents have ‘bounded discretion’ but, unless one can say what binds the discretion apart from the fact that something changed (so there must have been more discretion than binding), the claim again becomes methodologically unsustainable. Where one sees this most clearly is in Bell's own case study of Australian central bank politics.

A Constructivist After All?

Bell's analysis of Australian central bank politics seeks to show us ‘how agents, institutions and structures dialectically interact, with agents mediating and actualizing institutions and structures and with these in turn helping to shape the ideas and the scope of bounded discretion available to actors – all as part of a process of shaping institutional change.’Footnote 41 Bell investigates how the Reserve Bank of Australia gained operational independence, what it did with it and how this relates to the morphogenetic–dialectic alternative presented earlier in the article. I suggest that rather than proving the robustness of his new framework, it shows him to be a constructivist after all.

The first part of the case study revolves around the floating of the Australian dollar in 1983 and how a large current account deficit opened up that worried policy makers. It peaked twice, in 1986 and 1989, and Bell notes that ‘both episodes constituted current account crises in the eyes of policy makers’.Footnote 42 Yet, to back up this claim, Bell cites Colin Hay's work on crisis as a political construction: the very work he was critiquing in the earlier part of the paper.Footnote 43 Bell next refers to the politicians’ fear of the deficit as a ‘fixation’ that ‘illustrates the importance of constructivist insights, as well as insights about bounded discretion’.Footnote 44 Here he puts the emphasis on ‘the mindset of the policy authorities [remaining] locked’, with cognitive locking and ‘ideational path dependence’, concepts developed by both Blyth and Hay, coming to the fore.Footnote 45 Interestingly, when Bell tries to be more institutionalist, noting that ‘institutional and policy legacies’ are determining, one finds that they are in fact determining of ‘RBA leaders … novel interpretations’,Footnote 46 which allowed them, in this moment of uncertainty, to change policy, which sounds awfully like Blyth's argument in Great Transformations.

Later, in the case study, Bell details the battle between inflation hawks and doves, to find that the doves prevailed because of their ability to appeal to the bank's Keynesian dual mandate. Bell invokes the central bank's dual mandate as a legislative fact – an institution – that tempers the power of ideas.Footnote 47 But, drawing on my own work, I would argue that specific polity types (institutions) give rise to particular discursive processes that are causally important in their own right.Footnote 48 The dual mandate is a social fact, which means that it is also a norm about what policy makers should do, and is, therefore, as ideational, constructed and as contested as it is institutional, and that contestation is at base discursive.

Following this episode Bell again stresses ‘the ideational front’ opening up when a paper by an economist in the early 1990s apparently changes everyone's mind about the sources of the deficit.Footnote 49 As Bell summarizes, ‘the ideas of policy makers were thus a crucial element in interpreting reality and in shaping policy options … [but] … this constructivist take … is not the whole story.’Footnote 50 Again, I would never have said it was, especially when the structural factors invoked by Bell at this point to constitute ‘the whole story’ are as large and vague as the rise of China and the increasing flexibility of the Australian labour market. Finally, the Asian crisis example seems to be exactly about a moment of uncertainty when Australian policy makers ignored what their institutions told them and tried a new policy idea: face down the hedge funds.Footnote ⁵¹

Given this, Bell's own analysis is really much more constructivist than it is institutionalist, flexible or otherwise. Bell's arguments concerning dualisms, morphogenesis and dialectics inform, as far as I can see, none of the case study. Indeed, at the end of his argument, Bell insists that one must see ‘embedded agents as a key component of the analysis, albeit agents who are dialectically engaged in shaping and being shaped by their relevant contexts over time’, even if such agents ‘are interpretive agents using ideas to help define their interests, motives and strategies for action’.Footnote ⁵² To me this is a quintessential constructivist insight, the type of which I have made many times already. It seems that Professor Bell actually belongs much more with his constructivist interlocutors than he realizes. I, for one, welcome him to the conversation.