Learning accountability in the public sector: The experience of Kerala

Correspondence PawanAdhikari, EssexBusiness School,University ofEssex. Email: padhik@essex.ac.uk Abstract Given that learning has become a core element in addressing the ever-increasing complexities in public sector accountability, we investigate two issues in this paper: (a) how learning was involved in the process of disseminating a competency framework (CF) in four state departments in Kerala, India; and (b) an analysis of the challenges involved in this process, using the theoretical insights from learning accountability. Although the CF would revitalise governance and accountability mechanisms by establishing a conducive environment for administrators to learn in, the nuances of the existing administrative structure limited the scope of this learning exercise. The CF in Kerala generated fear, tension, and resistance amongst theadministrators, due toaperceivedgapbetween thevery objectives of the CF and the administrative settings, in which learningwas insignificant. The paper argues that imposing public administrators’ accountability reforms, such as theCF,would bemeaningless until an emphasis is placed on learning accountability. Furthermore, there is a need to “unlearn” or “discard” the prevailing traditional learning and accountability practices; unless this is achieved, a conducive environment for learning accountability cannot be institutionalised, nor can reforms such as the CF bematerialised.

finds that the underperformance of public administrators in the delivery of assigned jobs in India has been attributed to a lack of clarity of tasks to be performed by administrators. A blurring job responsibility has made it difficult to identify the required competencies for different tasks at various levels. This has created obstacles in designing appropriate training and coaching programmes for public administrators, regarding their capacity development (Budhwar & Boyne, 2004;Budhwar & Varma, 2010). Recurring criticisms state that the existing HRM in the public sector, being a remnant of the colonial legacy, has failed to promote citizen-centric governance and accountability for results. This has been even more distinct in the state of Kerala, which was once perceived to be the most progressive state in India in terms of sociopolitical and economic development. For instance, the extent to which decentralised structures in the state contribute to asset accumulation amongst disadvantaged groups is tied to the wider context of participatory institutions and opportunities that allow social actors to transform assets into meaningful livelihood outcomes (Arun, Arun, & Devi, 2011). Irrespective of the high level of social development and civil society, the high level of unionisation and politicisation in the bureaucracy has resulted in the administration being trapped within a system of patronage and red tape, in which the introduction of reforms and learning has become insignificant.
In 2014, the Institute of Management in Government (IMG) 1 initiated a project of introducing a CF for public administrators in the selected departments of the state of Kerala. The significance of this initiative was predicated on the assumption that it would help redress the perceived gap between citizens' expectations and accountability in practice (IMG, 2018), and thereby strive for good governance. However, the dissemination of the CF has not proceeded as intended in these departments, as a result of resistance and inadequate support from the key stakeholders, including the state mechanisms. Our aim in the paper is to unfold the responses of diverse stakeholders at different administrative levels in order to understand the challenges to achieving learning accountability.
The remainder of the paper is structured as follows: Section 2 discusses public sector accountability, focusing on the learning aspects. This is followed by Section 3. Section 4 highlights the context of Kerala as providing an impetus for the CF projects. Section 5 presents our findings, revealing the views of public administrators and other state actors with regard to the dissemination of the CF, and its rationale for improving accountability. Finally, Section 6 offers discussion and concluding remarks.

LEARNING ACCOUNTABILITY IN THE PUBLIC SECTOR
Generally, the notion of accountability has remained elusive, as it is continually evolving and suggests different meanings to different people (Bovens, 2007(Bovens, , 2010Koppell, 2005;Luke, 2010;Sinclair, 1995). However, despite the variations in its interpretation, accountability has proved to be a powerful tool in the public sector in terms of ensuring good governance and advancing economic growth. Bovens (2007, p. 450) defines public accountability "as a relationship between an actor and a forum, in which the actor has an obligation to explain and to justify his or her conduct, the forum can pose questions and pass judgement, and the actor may face consequences." What makes public sector accountability more complex is the presence of multiple stakeholders: not least the parliament, but also executive branches, oversight agencies, external resource providers and citizens, exercising the authority of the forum under certain circumstances asking accounts from their subordinates, and for service provision they are expected to deliver (Luke, 2010). Bovens (2010) also mentions legal, professional, and other types of public accountability, whose functioning extends beyond the necessity to have clearly defined principal-agent relationships.
A major critique that pervades the public sector relates to an excessive focus on budgetary compliance and the discharging of hierarchical accountability to parliament (the forum), while there are no direct mechanisms to promote efficiency and effectiveness in resource mobilisation and service delivery, or to address the concerns of the citizens and communities. Reforms triggered in the public sector in the last three decades have at their heart a desire to extend both the scope and quality of public sector accountability beyond budgetary compliance, in order to improve the results and outcomes of the resources consumed, and to connect the reforms with the welfare of citizens and communities (Hagbjer et  For instance, NPM emphasises the adoption of private-sector managerial techniques in the running of public entities, whereby the public administrators and citizens are envisaged more as managers and customers, respectively (Hood, 1995). Hierarchical accountability to top-level authorities has continued under NPM, but public administrators are also required to discharge their accountability to clients (citizens) through efficient service delivery, and by ensuring the results and outputs of resources entrusted to them (Hagbjer et al., 2017;Hood, 1995;Hyndman & Lapsley, 2016). NPG is envisaged as more of a movement towards governance that promotes trust, networks, and collaboration at grassroots levels. A greater involvement of citizens and interest groups in governance, service delivery, and decision-making are considered paramount at a time when many local governments are experiencing budget cuts and austerity . Different notions of accountability are therefore predicated with a view to strengthening the governance momentum; not least social accountability and output-based accountability, assuming democratic deliberation and the delivery of public value (Alawattage & Azure, 2019;Almquist et al., 2013;Bracci, Garliardo, & Bigoni, 2014;Ferry, Ahrens, & Khalifa, 2019;Hyndman & Liguori, 2016). In general, both the managerial (output-based) and social accountabilities advocated by such reforms have drawn considerable academic criticism (Cooper & Lapsley, 2019;Grubnic & Cooper, 2019;Wiesel & Modell, 2014). One of the few exceptions is perhaps the study by Ferry and Ahrens (2017), in which the authors delineated how localism combined with austerity has served as a powerful external source of change for a UK local authority (Newcastle City Council), and resulted in enabling corporate governance practice to serve the public interest. In another study conducted in the Newcastle City Council, Ferry et al. (2019) showed how austerity localism has forced the council to employ public value strategy.
For instance, while discussing the outcomes of managerialism in the United Kingdom, Lapsley (2009) has delineated how the NPM toolkit tends to prevent rather than deliver intended results and accountability. Narayan, Northcott, and Parker (2017) state that accountability demands imposed upon universities as part of their commercialisation has limited researchers' autonomy, thereby damaging the culture of innovation and creativity. Employing Simon's "lever of control" framework,  show how austerity has made redundant the very objectives of localism introduced as part of NPM in the United Kingdom, thereby restricting innovations and preserving local authorities' preference to maintain budgetary control and hierarchal accountability. Scholars such as Kim and Han (2015) and Cooper and Lapsley (2019) argue that NPM has further increased the level of bureaucracy in discharging accountability, as the administrators themselves are required to design and manage the process of changes.
In a similar vein, Grubnic and Cooper (2019) discuss the challenges that members of Health and Wellbeing Boards (HWBs) in the United Kingdom have experienced in discharging multiple accountabilities within the wider NPG reforms. The mechanisms advocated by NPG for democratic renewal and public value creation-including participatory budgeting, citizens' and neighbourhood forums, and citizens' juries-have also experienced increasing resistance, and have failed to foster trust between citizens and politicians (Hyndman & Liguori, 2016). For example, mechanisms such as participatory budgeting emphasises promoting citizen empowerment rather than individual learnings (Van Helden & Uddin, 2016). Moreover, the fact that the motives of citizens' participation can be very different from the propagated ones has largely been neglected when promoting such governance reforms. For instance, in their study of "austerity localism," Ahrens and Ferry (2015) demonstrate how the grassroots' groups at Newcastle City Council (United Kingdom) were offered greater scope to engage in determining local policies and services, to produce a consensus on local spending cuts. In many emerging economies, citizens' engagements and participatory budgeting has become a means for politicians to promote patronage and maintain their tenure, rather than facilitating social and output-oriented accountability and engendering public value (Célérier & Botey, 2015).
Learning aspects of accountability are rarely acknowledged when discussing the accountability arrangements within both NPM and NPG reforms. Instead, scholars have continued their debate on the accountability crisis in the public sector (Bovens, 2010). The literature has reflected on this crisis by applying terms such as "the web of accountability arrangements" (Bovens et al., 2008), "accountability dilemma" (Behn, 2001), and "accountability trap" (Van Thiel & Leeuw, 2003). For instance, Hagbjer et al. (2017) have highlighted various strategies that those accountable can adopt, including decoupling, structural differentiation, role attribution, and compromising, to deal with perceived ambiguities in public sector accountability. Similarly, Ahrens and Ferry (2015) discuss the efforts of a local authority to pursue "dispersed accountability," which implies the process of spreading accountability across different groups.
Ongoing debates, however, signal a shift in discussions from accountability deficits to accountability overload (Schillemans, 2011;Willems & Van Dooren, 2011). In practice, both NPM and NPG public sector reforms have simply entangled administrators in the web of accountability arrangements, which at times overlap and conflict with each other (Bovens, 2007;Bovens et al., 2008;Ferry, Eckersley, & Zakaria, 2015), without equipping them with skills and competencies required to handle such tasks (Lapsley, 2009). Thus, learning aspects of accountability are largely silent in the discussion of both the NPM-and NPG-based accountability arrangements. Messner (2009) argues that too much accountability, as experienced by contemporary public administrators, can become an ethically problematic burden, damaging both their motivation and morale. Furthermore, the accountability overload can stifle innovative and entrepreneurial behaviour of public administrators, and undermine the very objective of reforms itself (Bovens et al., 2008;Koppell, 2005;Schillemans, 2011;Virtanen et al., 2018).
Conflicts, tensions, and resistance to reforms are not surprising when public administrators are exposed to multiple accountability relationships (Luke, 2010;Narayan et al., 2017). For instance, Adhikari and Jayasinghe (2017) have delineated how the externally purposed public sector accounting reforms, with a rhetoric of improved governance and accountability, contradicted the existing operating contexts; they were thus resisted by multiple stakeholders, including government accountants, officials, and auditors. Such tensions and resistance, resulted from a demand for multiple accountabilities, can potentially be mitigated by ensuring that each individual public administrator is equipped with required skills and competencies that are identifiable and measurable. This demands greater attention to learning accountability, in which the central focus is on continuous improvements (Crawford, Morgan, & Cordery, 2018;Ferry et al., 2015;Horton, 2000;Luke, 2010;Nybø, 2004). For instance, Virtanen et al. (2018) have discussed a service systems perspective, placing front line managers and users at the centre of service provision and accountability. They emphasise the need for building intelligent organisations that align the service environment with staff capacity-building, and embedding a culture of continuous learning and adoption.
Although organisational learning, including the application of management control systems in facilitating learning, has been extensively covered in prior work (Bol & Moers, 2010;Levitt & March, 1988;Pant, 2001), the role of accountability in promoting learning in pursuit of continuous improvement has received less attention in the literature (Van Acker & Bouckaert, 2018). Brodtrick (1998) argues that learning tends to be stifled once an organisation institutionalises its own culture. This is perhaps more striking in the public sector, as its entities are alleged to have a lack of flexibility, dynamism, and innovativeness, due to their enduring preference for maintaining budget stewardship. For instance, in their study of local authorities in the United Kingdom,  show how the factors such as austerity, centralised budgetary restrictions, and pressures to deliver a balanced revenue budget have restricted management control innovations, and thwarted the learning attitude of public sector entities.
For learning to take place in the public sector context, entities are required to identify their core competencies, rethink the processes, and set up a dynamic learning culture. This requires the enforcement of an appropriate learning tool, the CF being an example (Brodtrick, 1998).
Scholars such as Argyris and Schon (1978) and Bovens et al. (2008) have emphasised the role of individual learning within organisations by introducing the learning aspects of accountability, in which the latter is connected to "deutero learning," that is, an institutionalised capacity to learn (p. 232). Bovens (2007) states that public accountability also concerns motivating individual administrators to search for more intelligent ways of discharging their assigned roles and responsibilities at multiple levels. Searching for intelligent and innovative ways of delivering tasks is paramount, as the clients (citizens) and other stakeholders are provided with the mechanisms to sanction public administrators in the event of errors and shortcomings. In intelligent organisations, public administrators are aware of the issues, such as what is expected of them, what works, and what does not (Bovens, 2007, p. 463;Bovens et al., 2008). Public administrators have therefore the opportunity to review and readjust the effectiveness of their performance in achieving both social and functional goals, and maintaining a balanced relationship with heterogeneous stakeholders (Crawford et al., 2018). Willems and Van Dooren (2011) state that the learning aspects are central to the discharging of public accountability in intelligent organisations. In fact, ensuring learning accountability is claimed to be the only way to promote acceptance of governments' authority and citizens' confidence in contemporary societies, in which the citizens and other interest groups are becoming increasingly critical towards the government, given the ever-growing need to provide public value (Bovens, 2007;Bracci et al., 2014;Cooper & Lapsley, 2019;Ferry et al., 2019). Lindblom (1965) has developed a broader understanding of learning accountability, referring to it as "the intelligence of democracy," the superiority of which is predicated on its ability to promote pluralistic democracy, which offers a great number of incentives to encourage intelligence and individual learning in the process of policymaking. Bovens (2007) mentions that this aspect of accountability has great potential to articulate a pluralistic (deliberate) form of democracy and social and output-oriented accountability more effectively, by offering the required skills and competencies to individual public administrators. In an ideal situation, learning accountability needs to generate information about the administrators' functioning, and to improve performance by reflecting on the achievements and challenges of their past activities. Bovens et al. (2008) state that the very notion of accountability can only be realised in the public sector when public administrators start generating and acting upon external feedback regarding their performance. It is therefore vital that public administrators be provided with certain tools that could stimulate their continuous learning and self-evaluation of their performance. However, little regard has been given to specifying such tools through which learning accountability can be enacted and practised in the public sector. Even less is known about how public administrators respond to such tools in practice, particularly in emerging economy contexts. Therefore, we argue in this paper that the CF can be an integral component of learning accountability in the public sector, given its focus on enhancing the capacity and effectiveness of public administrators by equipping them with competencies, in order to become more responsive to the needs of public services. By examining the dissemination of the CF in the state of Kerala, the study extends the research on the challenges involved in the institutionalisation of learning accountability in the public sector of emerging economies.

RESEARCH METHOD
Data for the study were derived from three main sources: document analysis, semistructured interviews, and observations of participants in the dissemination workshops held in Kerala. We commenced our research by reviewing the documents on the CF and accountability that have been published by the IMG and the Government of Kerala since 2010. The purpose of document analysis was to generate preliminary insights into the competency definitions applied; reasons for enforcing the CF in state departments; identifying the expected benefits of the CF for state departments; and gaining an understanding of the methods used in determining duties, responsibilities, and competencies for public administrators. Given that many of the documents and competencies have been developed in consultation with staff members, document analysis proved valuable in terms of identifying and approaching the key stakeholders and administrators involved in disseminating the CF in four state-level departments. The key documents we reviewed included: () handbooks on duties and responsibilities of the Social Justice Department, the Scheduled Tribes Development Department, the Police Department, and the Panchayat (Local Self-Government) Department; (2) the CF for civil servants in these four departments; (3) the civil services competency dictionary; and (4) the CF for functionaries (see Appendix 1a).
Forty semistructured interviews, which included three follow-up interviews, were conducted over a period of 3 years (2016-2018). This method was useful for collecting in-depth information on stakeholders' opinions, thoughts, experiences, and feelings on the dissemination of competencies. Most of the interviews were conducted collectively in the presence of all coauthors. In some instances where not all coauthors were involved in undertaking interviews, we ensured that there was sufficient collaboration in developing interview formats before the interviews, and exchanged the interview transcripts on the same day, to ensure the continuity and shared knowledge aspects of the process.
To contact potential informants, we mobilised both the networks of two native coauthors and snowball sampling.
In addition to the informants from the IMG, we also interviewed former and current senior civil servants of the Government of Kerala, and officers at different levels in four state-level departments that were piloting the CF, namely: Scheduled Tribes Development, Local Self-Government, Police, and Social Justice (see Appendix 1b). We started our interviews by asking the informants open-ended questions about their knowledge on the CF. We continued by discussing their engagement in the process of determining and developing competencies for administrators with varied roles and responsibilities; learning that occurred in the process of disseminating competencies; the use of competencies in continuous learning and discharging accountability; and the challenges experienced in disseminating the CF at different administrative levels. We aimed to generate a comprehensive understanding of how learning took place in the process of disseminating the CF both at the individual and organisational levels, and how the CF functioned as a tool for developing learning accountability. The interviews, conducted mostly in English and Malayalam (regional language), lasted between 30 and 90 min. Before the interviews, we assured the respondents that their anonymity would be preserved when we present and analyse their views and observations. With participants' consent, all interviews were recorded and subsequently transcribed in English by the authors. The verbatim transcription took place immediately after the interviews, followed promptly by discussions that included all authors. This helped us share a common understanding of all the relevant issues in a time-bound manner.
We had the opportunity to review the same set of stakeholders over a period of time by attending the subsequent workshops in 2017, 2018, and 2019. Although a greater momentum towards the CF was generated during these 3 years, the process faced challenges at different levels. Our representative sample therefore included different categories of public administrators in different state departments and at different administrative echelons. We observed that at the individual level, all participants attending the workshops were enthusiastic about the CF. An element of learning was evident, as the administrators were involved in determining competencies for positions within their respective departments, and many of them had become change agents. These administrators perceived the CF as a learning tool that would enable them to deliver the assigned duties more efficiently and effectively. However, at the organisational level, the existing bureaucratic system appeared to be a key impediment to disseminating the CF as intended. Many of the interviewed administrators at the top level were relatively unconvinced, compared with middle-level administrators, about the CF's applicability across the full range of departments. This might be partly due to a lack of clear commitment by top-level government to pursue the CF project further. These observations provided us with additional insights into the bureaucratic structure of the state, and the influence of trade unions and other associations in the day-to-day running of public administration, all of which could have an adverse impact on the process of learning and discharging accountability. Over the duration of 3 years, we have had the opportunity to ensure the reliability and validity of our interview data by undertaking three follow-up interviews (see Appendix 1b), as well as to clarify emerging issues.
Based on the transcriptions and recorded data, we identified the most valuable data for the discussed issues.
Firstly, we listed the views and issues frequently expressed by our interviewees in the process of disseminating the CF. Following this, we identified the learning elements present in this process, and the way the CF served as a tool for continuous learning and accountability. Based on these responses, we developed a number of themes, such as selecting departments for piloting the CF; the learning process involved in the CF's dissemination; responses to the CF by public administrators at different levels; and the factors influencing the dissemination process, amongst others. At the last stage, the data representing the themes were clustered, and we endeavoured to establish a link between the themes, so as to create narratives of the process involved, and the challenges faced in disseminating the CF project in four departments of the state of Kerala.

COMPETENCY FRAMEWORK: THE EXPEIENCE OF KERALA
Public administration in India has experienced a series of changes in the last decade, which resemble the NPM trend (Budhwar & Boyne, 2004). As part of the reforms, several financial measures, including performance budgeting, zerobased budgeting, and more recently, output-based budgeting, have been introduced across public entities (Gomes, Sahadev, Glaister, & Demirbag, 2015;Haq, 2012). An increasing emphasis is being placed on result-based rather than rule-based approaches, in delivering public services and discharging public accountability. A striking aspect of the reforms taking place in India, as compared to other emerging economies, is perhaps an extended focus on the professionalisation of the civil service and the capacity development of civil servants, mainly at the state level. The 12th five- year plan and the report by the Administrative Reforms Commission (ARC), both issued in 2008, as well as the 2012 Right to Service Charter, had all urged state governments to search for initiatives to professionalise the civil service, as part of establishing more service-oriented and citizen-centric civil servants (IMG, 2018). Professionalisation has been envisaged as a way of making public administrators more accountable, and of delivering better public services.
In 2014 Amongst Indian states that are experimenting with CF, Kerala has become the first to maintain high performance in public service delivery and accountability. Historically, the state has earned a reputation for being a frontrunner in introducing a number of social transformations from the 19th century onwards, which have resulted in instigating a range of administrative initiatives. A high rate of literacy, advanced social indicators, and welfare provisions reflecting those prevailing in western countries have encouraged the state to introduce new measures for improving accountability and service provision. However, the drawback of the state has been its mediocre public administration driven by a historically established administrative culture, which tends to stifle rather than stimulate the changes introduced. The IMG was therefore quick to adopt the CF, as the latter appeared to be a tool that would professionalise HRM and pave the way for setting up citizen-centric governance and accountability (IMG, 2014). Several benefits of the CF for departments, superior officers, employees, state governments, and citizens were propagated, which promoted its dissemination (IMG, 2016a(IMG, , 2016b(IMG, , 2016c(IMG, , 2018. For instance, departments can link organisational and personal objectives, and make employees aware of their expected performance. The officers can use the framework to structure the employees' development and training needs; consequently, employees can gain a better insight into the department's strategy, learn new skills in discharging multiple accountability relationships, and increase their career progression. In a similar vein, governments can evaluate public service delivery more transparently, and the citizens would receive high-quality services promptly.

EMPIRICAL FINDINGS
This section consists of two parts. First, we analyse the process of disseminating the CF and the learning activities that took place in this process. Next, we present the factors that impeded the dissemination of the CF, thus shedding light on the challenges that the CF and learning accountability can face in an emerging economy context, the case of Kerala being an example.

Learning in the process of disseminating the CF
In India, the colonial legacy still dominates HRM in the public sector (Budhwar, 2009). Nevertheless, the needs of a modern democracy such as India are very different from those in the colonial era; this requires innovative thinking in the recruitment and training of public administrators, based on their skills and demonstrated ability. While academic achievement has continually become a central element in recruitment, the promotion of public administrators is largely determined based on their seniority. Trainings for public administrators are focused on making them aware of the existing rules and regulations, and helping them to discharge hierarchical accountability. Key concerns in discharging public sector accountability have been to control the misuse of authority, and to provide assurances regarding the appropriate use of both authority and resources. There is insufficient scope for fresh ideas and innovative ways of enhancing downward accountability. Furthermore, in an absence of continuous learning, there has been an element of complacency in the public sector, which has hindered the performance of public administrators and innovation. For instance, during our interviews, a retired public administrator stated: Even today, recruitment is not seen as a means to address performance and functional gaps in service delivery and administration. The dissemination process itself was articulated as a learning exercise for the IMG, the adopting departments, and the administrators representing these departments. For instance, at the outset, task forces were formed by identifying the individuals who were knowledgeable about the departments; these groups incorporated both the internal and external stakeholders. Indeed, the task force members were the crux of the activity to develop the CF.
The systems were implemented to ensure that the members were appropriately incentivised. In the next stage, all available literature that referred to orders, communications, standing instructions, and oral traditions concerning each cadre/category of personnel, was studied and analysed. This, in turn, led to the identification of desirable competencies through a series of consultative workshops; here, the task force members debated and identified the vital competencies required for executing the key duties and responsibilities that were entrusted to and expected of the employees/cadres. One erstwhile member of the task force commented: We were told during the task force discussions to elucidate to our colleagues in the department the importance of having competencies, setting up a learning culture which could help us obtain such competencies, and executing them in service delivery and discharging accountability. The departmental approaches were followed in disseminating the CF with a view to inducing learning. The process started by identifying all duties and responsibilities assigned to each cadre in the department, and determining the competencies needed to discharge each of these duties and responsibilities. The competencies needed for performing the duties were evaluated and assigned weightage by an engaging process with stakeholders consisting of task force members, peer groups, and by client reviews to ascertain and validate their significance. For instance, while the CF for the Scheduled Tribes Development Department was being disseminated, two new competencies (hitherto absent in the competency dictionary issued by the Government of India) were additionally developed and added to the framework during the process of consultation. In a similar vein, the CF was developed for 17 categories of the Local Self-Government (Panchayats) Department; for 20 categories of the Social Justice Department; and for six categories, including cadres of women, in the Police Department.
The dissemination of the CF was more deliberative in the Local Self-Government Department, reflecting on lessons learned from previous departments where the CF was initially piloted. A professor at the IMG explained: We also selected retired officers in the task force, given their wealth of professional experience and availability.
Next, we realised that a one-full-day workshop was not effective, as the participants were distracted after lunch concerned more about return to their homes and confirming transportation. We have started organising workshops of at least two days' duration, offering the participants accommodation for an overnight stay.
An underlying idea was that the dissemination of the CF would require administrators to continually learn, in order to achieve the required competencies for delivering public services. The CF was envisaged as a tool with which to redefine public sector accountability, prioritising learning and catering to the requirements of public services. However, a significant number of state departments are yet to be convinced of the pertinence of the CF, and various forms of resistance have been experienced in the process of disseminating the CF across state departments. There was a lack of momentum amongst the Personnel Department of the state government to capitalise upon this pioneering initiative, and to promote the CF as "the intelligence of democracy" (Bovens, 2007;Lindblom, 1965) in a comprehensive manner across the departments.

Factors impeding learning in the process of dissemination
Despite the initial support at the central and state levels, the established organisational culture appears to be a major impediment in disseminating the CF across state departments. This is perhaps unsurprising, as learning is more effective in young organisations, because they have to be both proactive and responsive to their milieu, to ensure survival and progression (Brodtrick, 1998). Setting up a dynamic learning environment while introducing reforms is rather arduous in public sector entities, as they are bound by rigid rules, regulations, and bureaucratic procedures.
Public sector entities need to make the citizens aware of the underlying meanings, purposes, and procedures of the reforms before their dissemination. Ambiguities in understanding and interpreting the term "competency," and the limited role that the central and state-level governments have played in addressing such complexities, have certainly limited the dissemination of the CF. For instance, a project consultant at the IMG commented on the failure of public administrators to understand the very essence of the CF: Public administrators had different ideas about competencies, and they kept on shifting. For some people, a competency turned out to be academic degrees, and others envisaged it as having specific and professional skills, including IT and accountancy. Some police officers even understood it as a skill required to identify suspects.
Both the central and state governments were more concerned with the adoption of the CF being focused on the policy level. The lack of guidelines on relevant practices has limited the process of creating a conducive learning environment in which learning becomes a part of an ongoing process. For instance, the central government has signalled its support for the CF by incorporating it in its national human resource strategy and allocating a budget for training. A total budget of ₹ 1.14 crore (approximately 1.6 million USD) was allocated based on the IMG's proposal to the Depart- This does reinforce the fact that the government is convinced of the potential of the CF to re-engineer the human resource management of the country. However, the allocated budget is nothing compared to the millions invested in administrative reforms over the years. Changes such as the CF should be viewed as an ongoing process to be institutionalised and be part of an annual budget.
At the state level, the support offered at the outset was confined to translating the competency dictionary into local Malayalam. However, in the process, the then chief secretary emerged as a learning champion, by encouraging the extension of the CF to other departments, in particular the Local Self-Government (Panchayats) Department.
Commenting on the efforts of the Chief Secretary, a deputy director of the Local Self-Government Department stated: We were lucky that the incumbent Chief Secretary enabled us to pilot the CF. Otherwise, it is hard to imagine that the innovative HR mechanisms such as the CF would be approved, let alone appreciated by our department, which is largely driven by traditional values and local accountability relationships.
Such initiatives by the central and state governments were, however, inadequate in terms of disseminating the CF across departments and setting up a learning culture. For instance, the institutional structure in public administration has persisted, in which public administrators continue to be evaluated based on their adherence to rules and their commitment to avoiding risks. No mechanisms have been introduced that incentivise public administrators to generate bottom-up feedback regarding their performance, or to rethink accountability based more on the learning aspects.
Interviewees mentioned that those who attended the dissemination workshops represented a minority group in their respective departments; the following statement from a Local Self-Government Department participant serves as an example: You will find very few administrators willing to embed new ideas, listen to mechanisms for improvements, and learn new ways of discharging accountability. When I return to my department after participating in the disseminated workshops, I am ignored and discouraged by others, as they have no desire to learn about the CF.
Public administrators lacked both motivation and the necessity to act proactively and respond to the CF. Instead, there were questions about the significance of introducing the CF without any attempt at altering the existing structure of public administration. A programme officer at the Social Justice Department stated: There is no dearth of policies and prescriptions regarding the professionalising and improving accountability of public administrators. What we need is not new measures, but the political and administrative will for improvements and a more engaging environment.
Limited interactions and motivation imply that the CF continued to appear as an elusive concept for a vast majority of public administrators, mainly the middle and lower level groups. This limited awareness of the CF and its accountability implications was reflected in the statement by a supervisor at the Social Justice Department: The limited awareness of the CF was more striking in the Local Self-Government Department, which regulates about 1,000 "panchayats" (villages) scattered across the state. We also found evidence of this while observing the CF dissemination workshops for the Local Self-Government officers, held in March 2018 and 2019. Each panchayat has developed and institutionalised its practices in terms of delivering public services and settling public grievances. Such practices have remained static, even though society is continually evolving. Conceding that the CF has failed to replace such practices and to introduce a learning culture, a lead consultant at the IMG stated: A secretary at the Local Self-Government Department added: Everyone is influenced by some organisations, ranging from trade unions to religious associations. The supervisors, peers, actually no one, really appreciate the fresh inputs. I am not saying that the measures such as the CF would never be implemented, but it will not happen in the next few years.
A lack of consistency in transfer policies has served as another obstacle to acceptance of the CF. Although both the Code of Ethics 2013 and several other regulations-for instance, the pathways for an Inclusive Indian Administration (PIA) 2014-state that there should be a gap of up to 5 years after each transfer is made, such provisions are rarely put into practice. The prevalence of patronage and party-based politics has enabled public administrators to influence the higher level officers and politicians; furthermore, and they get frequent transfers to offices where more spaces are available to cater for their interests; for instance, by working closer to home and family members. Commenting on the challenges caused by the frequent transfers of public administrators, an associate fellow at the IMG, under deputation from the state secretariat observed: In the department studied, we initiated communication with a certain set of people, but after six months, when we returned to the department, we found a completely different team. Administrators who had attended our workshops and with whom we had established a relationship had all been dispersed and transferred to different places, making the task more challenging.
A key issue, however, concerns the institutionalised mindset of public administrators in assuming their superiority in the society. Public administrators, once they are appointed to their jobs, tend to perceive themselves as superior in terms of knowledge and the positions they occupy. The prevailing system of recruiting public administrators is based on the performance in the civil service exam, which is reckoned to be one of most stringent in the region (Budhwar, 2009;Haq, 2012); this makes them behave more as an elite, exclusive group, rather than as public servants. The accountability relationships with citizens, and learning attitudes, are therefore almost nonexistent within the vast majority of administrators. An assistant professor from the IMG remarked: These bureaucrats are like the cream of the cream (academically), so they often perceive that they are in a better position to guide people rather than listen to them. For the CF to be a success, serving the public (accountability) should be made a key criterion that the stakeholders could evaluate objectively and purposively.
A civil police officer from the Police Department added: It is just a waste of time and resources trying to change the existing cadre through the CF, unless there is this kind of backing from the polity. The best alternative could be to amend the recruitment criteria, and try to adopt the CF in stages by roping in the new recruits rather than the serving administrators/police officers.
We identified that the perceived stability in government employment has further limited the desire to learn, and contributed to complacency amongst public administrators. In the state of Kerala, as well as in other states, government employment has continued to represent a form of stability and social prestige. It has also become a pathway for administrators to gain better career opportunities in the private sector after their retirement; the following statement by an assistant director at the Local-Self Government Department provides an example: In the IMG's reports, several benefits of the CF have been outlined for departments, higher level officers, employees, different levels of governments, and citizens (IMG, 2016a(IMG, , 2016b(IMG, , 2016c(IMG, , 2018. However, following our conversations with public administrators and officers who represent different departments, such benefits are far from being realised, and have remained elusive. We observed that although the CF is central to learning accountability, this notion of accountability is virtually non-existent in the public sector of India in general, and Kerala in particular. The existing structure of public administration itself prevents any initiatives that call for learning and engagement with the citizens.
The fact that public administrators regard themselves as superior to the citizens, both in terms of knowledge and authority, means that establishing and practising learning accountability is not in demand. Political patronage, along with issues such as the influence of trade unions and different associations, an absence of communication flows, and a lack of consistency in managing public administrators, have further limited the significance of any reforms and innovations, the materialisation of which requires a dynamic culture of continuous learning-of which the CF is just one element.

DISCUSSION AND CONCLUSION
Drawing on the theoretical insights from learning accountability, we have in this paper studied the development of the CF in the Indian state of Kerala, and the challenges encountered in its dissemination in four state departments.
The CF has widely been applied by private enterprises for the effective implementation of HR bundles and to improve organisational performance (Gooderham et al., 2008;Vos et al., 2015). Its application is also linked to wider cultural changes in HRM and the institutionalisation of learning accountability (Brans & Hondeghem, 2005;Horton, 2000;Nybø, 2004). In the public sector, the CF is reckoned to be an alternative to traditional job-based approaches (Horton, 2000;Hyndman & Lapsley, 2016;Nybø, 2004) and hierarchal accountability. The CF has, however, drawn relatively little attention in the public sector, compared to other managerial and governance reforms prescribed by the NPM and NPG. Public sector reforms, in general, and in emerging economies, in particular, have often resulted in unintended consequences, leading to an accountability crisis (Hagbjer et al., 2017;Van Helden & Uddin, 2017;Wiesel & Modell, 2014). While the earlier crisis was mainly caused by accountability deficits, more recently, public sector entities and administrators have been encountering a situation of accountability overload, where different accountability arrangements are layered upon one another (Bovens et al., 2008;Ferry et al., 2015;Virtanen et al., 2018). This has put public administrators in a situation of conflicts and tensions, and made them resistant to learning and reforms. By examining the CF as a tool for inducing learning accountability, we have provided in-depth explanations for the accountability crisis in emerging economy contexts.
In the Indian state of Kerala, four state departments have experimented with the CF as an important learning and accountability tool, which has the potential to transform the adopting departments into intelligent organisations (Bovens et al., 2008). The aim was to use the CF to determine competencies needed to perform the assigned duties, and help the administrators acquire such competencies by setting up a culture of continuous learning (IMG, 2014(IMG, , 2018. Insights into self-performance and the opportunity to self-evaluate the performance through external feedback were envisaged as central benefits of the CF, which would encourage administrators to search for intelligent ways of delivering services to the citizens (Bovens, 2007). We have identified that the dissemination and implementation of the CF across the departments has itself become a learning process, which engages a wide range of stakeholders, including public administrators, at different levels, and involves collectively identifying competencies for each assigned role and responsibility. This, however, has not been effective in altering the existing administrative structure and culture, including the civil society's reluctance to accept reforms, in which learning was insignificant.
An important finding of the paper is perhaps the manifestation of the ambiguities in the acceptance of learning as a form of accountability in India, as well as in other contexts (Bovens, 2007(Bovens, , 2010Bovens et al., 2008;Hagbjer et al., 2017). Learning was more adaptive than deliberate at the central and state levels, as the focus was on the adoption of the CF, rather than on creating a conducive environment that could motive public administrators to embrace the CF and engage in obtaining competencies. The learning initiatives instigated by the IMG in the process of disseminating the CF were stifled, as these had minimal impact on diverting administrators from the existing structure of accountability, which emphasised compliance and the control by authorities. The CF continued to be an elusive concept, while local practices driven by political patronage, red tape, and trade unions and associations persisted, and the public administrators' mindset that privileged their superiority remained intact. Rather than being a learning tool, the CF appeared to be more of a threat to the employment and tenure of public administrators, and it was perceived as another layer of burdensome reforms.
The paper contributes to the public sector accountability literature in several ways. First, the paper provides an in-depth explanation of how administrative contexts and structures interact with the accountability reforms and result in unintended consequences, mainly in emerging economy contexts. The reforms undertaken in the public sector across countries are often justified in the name of improving governance and accountability (Hagbjer et al., 2017;Hyndman & Liguori, 2016). Different accountability arrangements are discussed when proposing NPM and NPG reforms that range from managerial and output-oriented to social accountability-based (Ahrens & Ferry, 2015; Alawattage & Azure, 2019; Hyndman & Liguori, 2016). However, a key element that has a substantive impact on improving accountability, namely, continuous learning, has been given limited attention. The Kerala experience shows that despite acceptance at the policy level, reforms such as the CF, which are intended to instigate learning, in fact, generate fear, tension, and resistance when they are cascaded down to administrators in the form of policies and directives, particularly in the absence of collective understanding and a lack of a learning culture (Brodtrick, 1998).
We argue that there is a need to establish a learning environment prior to the dissemination and implementation of reforms that offer public administrators the opportunity to learn and exchange ideas in multiple ways-such as through networking, dialoguing, informal training, and sharing insights within the cluster of administrators who are working in the same department and doing similar work.
Next, we argue that imposing public administrator accountability reforms such as the CF would be meaningless unless an emphasis is placed on learning accountability. The latter has been relatively marginalised within wider discussions of NPM and NPG in the last decades. As stated by Lindblom (1965), learning accountability serves as "the intelligence of democracy," a fact that has largely been ignored when pursuing reforms in the public sector, mainly in emerging economies. There is still a widespread belief in many emerging economies, India serving as an example, that public sector reforms are a top-down process, and that the middle and lower level administrators are seen more as obstacles rather than facilitators of reforms. In fact, learning at the middle and lower levels of administration is restricted to obeying rules and regulations, and accountability is discharged by complying with these rules and regulations. As stated by Brodtrick (1998, p. 85), there is a need to "unlearn" or "discard" such traditional complianceoriented accountability practices, by elucidating to public administrators the fact that public accountability also implies continuous learning and institutionalising a culture of learning.
The fact that traditional attitudes, behaviours, and culture demotivate public sector entities to pursue learning and innovations when they face the situation of funding constraints and austerity is discussed in prior work . Such studies, however, have been conducted in the context of developed countries. A few studies conducted in the United Kingdom also demonstrate the fact that austerity and funding cuts may also on some occasions induce learning in the public sector (Ahrens & Ferry, 2015;Ferry & Ahrens, 2017;Ferry et al., , 2019. For instance, public sector entities may learn to focus more on generating public value, promoting grassroots participation in policymaking, and instigating corporate governance mechanisms. However, the continuity of such learning efforts has raised doubts, as they appear to be concurrently driven by the motives of obtaining grassroots legitimacy by transferring the blame to dwindling public services, and by the citizenry's anger at funding cuts by the central government (Ferry & Ahrens, 2017). It should be noted that the context of our research setting, Kerala, is unique even in the context of emerging economies, given the provision of high levels of social development and literacy. However, even in Kerala, we noticed the challenges faced in establishing a culture of learning accountability; this topic warrants further research in diverse contexts, to broaden our understanding of the observed accountability crisis in the public sector.

NOTES
1 The Institute of Management in Government is the Apex Administrative Training Institute in Kerala. This institute also has a mandate to act as the think tank of the state government. The institute has the chief secretary of the state as its president.

DATA AVAILABILITY STATEMENT
Data sharing is not applicable to this article as no new data were created or analysed in this study.