I draw on an analysis of the rise of sustainability reporting in the Netherlands, one of the early adopters and frontrunners in reporting (KPMG, 2011), from the early 1990’s until the mid-00’s of the 21st century. In the period analysed reporting has evolved from peripheral and inconsequential to both accepted and generally expected, in particular among larger enterprises. To examine the narrative of reporting I rely on a qualitative field analytic approach (Lounsbury & Crumley, 2007; Lounsbury, Ventresca, & Hirsch, 2003; Scott, Ruef, Mendel, & Caronna, 2000) that helps to track changes over time in the evolution of the practice of sustainability reporting and the co-evolutionary dynamics of actors, problems, decisions and their operating context. In the period analysed reporting has evolved from peripheral and inconsequential to both accepted and generally expected, in particular among larger enterprises, and has acquired increasing legitimacy.
Sustainability related practices are, despite their global aspects, typically local in character (e.g. Adams & Kuasirikun, 2000; Jennings & Zandbergen, 1995; Kolk, 2005) and focusing on a specific country is therefore appropriate. The Dutch sustainability reporting case can be characterised as a “complex social setting in which causal dynamics were not immediately apparent and the motivations of actors were obscure” (Greenwood & Suddaby, 2006: 31). As suggested in the theoretical section above, a focus on relatively lowly institutionalised, complex and uncertain field (Dorado, 2005) with multiple actors involved (Delbridge & Edwards, 2008; Garud & Karnøe, 2003) is most fitting for my study. A focus on sustainability reporting’s gradual evolution until the mid-00’s forms a good example of this. The sustainability reporting field (Etzion & Ferraro, 2010; Larrinaga-González, 2007) sees actors as diverse as firms, investors, accountants, government agencies, standard setters environmental and human rights activists entering and leaving its boundaries. Sustainability reporting is thus appropriate to study institutionally and forms a good exemplary case in which various subject positions, events, activities and decision moments are “transparently observable” allowing for theoretical development (Eisenhardt, 1989: 537) and a better understanding of the historically contingent elements of collective entrepreneurship in complex and uncertain environments.
I combine interviews and a range of secondary sources such as newspaper articles and reports. This triangulation enhances trustworthiness (Lincoln & Guba, 1985), makes it easier to see the data in a historical context (Yin, 2009) while at the same time tells us about the experiences of the multiple actors involved and the context in which instances of decision-making took place (Bryman & Bell, 2007).
The first step of data collection was to have conversations with four key informants as a first entry into the sustainability reporting field and its development. These four people all had a minimum of ten years experience in various facets of the Dutch sustainability reporting field, as well as extensive knowledge of its international context. These conversations, in combination with a prior reading of the phenomenon under study, provided me with a good initial overview of the field’s ‘who, what, when, how’. Based on these initial conversations, readings and conference attendance lists I drew up a first list of prospective interviewees. After contacting the first wave of interviewees and subsequently conducting the interviews, a snowball sampling technique was applied based on further recommendations of interviewees (Bryman & Bell, 2007). This approach made it possible to efficiently concentrate on contacting field actors with desired characteristics fitting the framing of the study, more akin to theoretical sampling (Strauss & Corbin, 1998).
INSERT TABLES 1 AND 2 ABOUT HERE
From February 2011 until February 2012 94 semi-structured anonymous interviews were conducted (see table 1). These interviews lasted typically between 1 and 1,5 hours, and the total hours of interview material was just over 100. All but seven interviews were recorded and transcribed. For those interviews that were not recorded extensive notes were taken. Conducting these interviews allowed for an in-depth understanding of both past and present moments, and the circumstances of these moments, that were of importance for the development of sustainability reporting. In order to avoid the risk of retrospective bias (Eisenhardt & Graebner, 2007; Morris, 1981), particularly an issue for interviews that discussed the more distant history of reporting, prior research (e.g. internet search, documents, information from previous interviewees) was undertaken to be able to ask specific questions, assist interviewees in structuring their thoughts and memories, and come up with counterfactuals to test their statements. The interviews progressed until empirical and theoretical saturation was reached. The interviews focused on the following main question areas: (a) understanding of sustainability reporting; (b) stakeholders involved in sustainability reporting; (c) important or critical moments in the development of sustainability reporting; (d) existence of conflicts or problems and ways to overcome them; (e) context in which aspects discussed under (c) and (d) took place.
Second, I used a diverse collection of secondary data sources (see table 2). Articles in main Dutch business and general newspapers, government reports, publications of professional service firms, reports and statements of NGO’s and investors, academic publications and integrated reports of companies were consulted. The main purpose of this data was first for contextual reading and field familiarization. Subsequently, it was also used as preparatory reading for interviews and, together with the interview data, for setting up an event history database of important moments in the evolution of sustainability reporting. The secondary data enhanced credibility and further validated my interpretations of the primary interview data as it was used as a form of “checks and balances” against to test insights derived from interview material (Yin, 2009). Finally, of relevant documents summary forms were prepared and subsequently analysed together with the interview data. These various data sources accumulated into a data set from which robust conclusions could be drawn.
The qualitative data analysis process for this study consisted of several stages and was assisted by the use of NVivo software. Not uncommon for more interpretative research based on rich qualitative data, I iteratively moved between the empirical material, generated concepts and prior theoretical constructs throughout the analysis (Locke, 2001). In the first stage of my data analysis I tried to depict the evolution of sustainability reporting in the Netherlands. Rather than having a very specific research question in mind here, the aim of this phase was to get a better understanding of the ‘who, what, when, where and how’ of my particular case and an overarching concern for its institutionalisation. I analysed the empirical material and arranged it to construct a chronological database of salient events (see e.g. Ansari & Phillips, 2011; Maguire & Hardy, 2013) which was used as an initial broad basis of the description of the analysis below, as well as to gain insight into the complexity of the case and of important events and changes of importance later in the analysis (see appendix A).
In the second stage, upon multiple further systematic readings of the empirical material and database it became apparent that some of the assumptions of most studies on institutionalisation and change did not match my observations. Most fundamentally, rather than a single identifiable institutional entrepreneur my case suggested the presence of multiple ‘entrepreneurs’, or perhaps rather ‘institutional ants’ (Czarniawska, 2009). In addition, analysing the rise of reporting did not seem to resemble a fairly neat linear and mechanistic process either, but rather a more muddled and random affair. This made me assess the possibility that some institutional assumptions may rather be ill-fitted for everyday reality. It was upon this realisation of the problematisation of commonly held assumptions (Alvesson & Sandberg, 2011, 2013) that I consulted the literature again in search for an alternative explanation for the development of Dutch sustainability reporting. Here I considered the potential of the garbage can model. Thus, rather than having the garbage can model as a core concept from the outset it emerged from the empirical material during the analysis of the data.
In the third stage the initial inductive approach was replaced by what can best be described as a more abductive (Peirce, 1978) approach focused on answering my research question. That is, rather than swinging the pendulum radically toward a more hypothetical-deductive strategy I relied on a more middle-ground strategy of iteratively comparing the raw empirical material with the various dimensions identified by the garbage can model, as previously discussed in the theoretical section of this paper. The analysis of the data thus resulted in a tabular format (table 3) which will be narratively discussed in the findings section. In this stage I coded the empirical material by systematically looking for instances of choice opportunities, problems, solutions and participants. First, participants were examined by analysing transcripts and reports, indicating actors entering and leaving the field or their importance thriving and waning over time. In line with earlier work using this approach (e.g. Rommetveit, 1976: 143–48) I subsequently studied transcripts and documents to find out “what did the main groups of participants in each period perceive as problematic and what solutions were put forward in answer to these problems?” I then structured my analysis around carefully scrutinising important events and developments (and confirming these with interviewees) with the event history as the starting point, as inferred from the empirical material by cross-referencing insights derived from the interviews, documents and the developed database. What became choice opportunities were moments where routine reactions or reactions as a consequence of an external event were at play. Combining these elements formed the backbone of sustainability reporting’s institutional story over time.
The resulting story thus resembles a narrative strategy, which is useful for “descriptively representing process data in a systematic organized form” (Langley, 1999: 707). As the aim of my research “is to achieve understanding of organizational phenomena not through formal oppositions but by providing "vicarious experience" of a real setting in all its richness and complexity (Lincoln & Guba, 1985: 359)” (Langley, 1999: 695) I decided to employ this strategy, following for instance Czarniawska (2009) and Delbridge and Edwards (2008). A narrative strategy generally provides a very accurate account as it is deeply rooted in the data, yet possible at the expense of offering a simple and general theory (Langley, 1999).