ADVOCACY COALITIONS IN THE SPECTRUM MANAGEMENT POLICY SUBSYSTEM IN INDIA Rishabh, Dara Doctoral Student, Indian Institute of Management Ahmedabad, India ABSTRACT This paper reviews the public policy process in the spectrum management policy subsystem in India from the lens of the Advocacy Coalition Framework (ACF). The ACF is a theoretical lens that is used to bring structure to the inherent complexity of the public policy process. The ACF considers advocacy coalitions, operationalised using common beliefs, as the appropriate unit to deal with the multiplicity of actors in the policy subsystem. In this paper, we perform a content analysis of 144 testimonies submitted by various actors in response to public consultations by the telecommunications regulator. We use a coding frame to identify the stated beliefs of elite actors on contentious policy issues. We then perform a cluster analysis to operationalise and identify the advocacy coalitions operating in the subsystem in two time periods: from 2008 to 2011 and from 2012 to 2015. The resulting analysis finds strong evidence for existence of advocacy coalitions and a high degree of conflict on contentious policy issues between these coalitions. The evolution of these coalitions as a result of change in underlying technologies is discussed. The ability of competing coalitions to dominate and convert their beliefs into policy outputs by using various instruments and resources has been reviewed. The paper provides significant insights for policy advocates to influence the policy process in order to achieve their policy objectives. The paper also presents a critique of the ACF as a theoretical lens to analyse the policy process. KEYWORDS: advocacy coalition framework, public policy, spectrum management, india. 1.

INTRODUCTION

Theories of the policy process refers to the body of theories that may be used as a lens to develop a structured view of the policy process. Each theoretical lens has a set of assumptions, on the basis of which it provides structure to the inherent complexity of the policy process. For example, the lens may provide structure by breaking down the policy process into different functions or stages as done by the Stages Heuristic (Lasswell, 1956); or by providing an explanation for a pattern of policy change as done by the Punctuated Equilibrium Theory (Baumgartner & Jones, 2010); or by grouping policy actors on the basis of resource dependency as done by the Policy Network Approach (Adam & Kriesi, 2007); or by grouping processes and institutions into streams as done by the Multiple Streams Approach (Kingdon, 2003). In this paper, we adopt Sabatier's (1988) Advocacy Coalition Framework, a theory that provides structure to the policy process by grouping actors in a policy subsystem on the basis of common policy core beliefs. We use the Advocacy Coalition Framework (ACF) as a lens to review the policy process in the spectrum management policy subsystem in India. Spectrum management refers to the policy issues arising from the use and allocation of spectrum for provision of wireless telecommunications services. Spectrum management includes policy issues such as spectrum allocation, valuation, harmonisation, trading, sharing, liberalisation, refarming etc. The paper is structured as follows. First, the Advocacy Coalition Framework is introduced and its underlying constructs are discussed. Next, we introduce the methodology adopted for reviewing the spectrum management policy subsystem from the lens of the ACF. We then operationalise the important constructs identified in the framework and present our analysis. Finally, we discuss the limitations of the ACF and our suggestions for improving it.

2.

ADVOCACY COALITION FRAMEWORK

The ACF was initially proposed in 1988 and since then has seen major revisions in 1993, 1999 and 2007. Since its publication, the framework has received widespread academic attention and scrutiny with over 80 applications of ACF in different policy subsystems across the globe (Weible, Sabatier, & McQueen, 2009). At the macro-level, ACF considers the policy subsystem as the most useful aggregate level of analysis. At the meso-level, ACF considers advocacy coalitions as the best unit to deal with the multiplicity of actors in the policy subsystem. At the micro-level, ACF considers a model of the individual that does not necessarily maximise utility but may also have altruistic behaviour. The ACF considers the policy subsystem, and not any particular government actor or institution, as the most useful aggregate unit of analysis because over time policy issues have become so complex that they require specialisation by actors for influencing policy. A policy subsystem is comprised of all such actors from private and public organisations, including administrative agencies, legislative committees, journalists, researchers, policy analysts, that are specialised with a policy problem. The ACF recognises three dimensions of a subsystem namely (i) policy issues, (ii) policy actors and (iii) territorial scope. ACF also conceptualises nested subsystems wherein a larger subsystem has many smaller subsystems within its fold.

Fig.1 Advocacy Coalition Framework (Weible & Sabatier, 2007)

ACF argues that beliefs are translated into policies, or that people get into politics to translate their beliefs into policies. ACF conceptualises a three tiered structure for beliefs. The first tier is deep core beliefs. These are beliefs that apply to all policy subsystems. These beliefs cater to general normative and ontological axioms or assumptions about human nature such as liberty, equality, distributive justice etc. The next tier is policy core beliefs. These are contextualised applications of deep core beliefs to a particular policy subsystem and are generally resistant to change. Operationalisation of policy core beliefs helps identify advocacy coalitions. The final tier is secondary beliefs. These beliefs are necessary for implementation of policy core beliefs in the subsystem. These are beliefs that coalitions are willing to change as a result of policy oriented learnings. Advocacy coalitions are composed of people who share policy core beliefs and who act in concert. There are typically two to four conflicting advocacy coalitions in every subsystem (Weible et al., 2009). Policy brokers in the policy subsystem are concerned towards finding a reasonable compromise between conflicting coalitions. The end result of the compromise is a policy output. On the basis of the effects of the output, as well as new information arising from external factors, the coalitions may alter their strategies or revise their beliefs. This alteration of strategy on the basis of learnings and new information acts like a feedback loop as a result of which the ACF regards the policy process as a continuous process with no beginning or end. The ACF is based on the premise that an understanding of policy change requires a time perspective of at least a decade. With respect to the policy subsystem, there are two categories of exogenous variables. The first category is relatively stable, which remains stable over decades, and the second category is relatively dynamic, which gets significant fluctuations over the course of a few years. As a result, the relatively stable variables are rarely made a part strategising behaviour by the coalitions. For example, changes in socio-economic conditions and changes in technology are dynamic events whereas changes in the basic legal structure is a relatively stable parameter. Accordingly, coalitions must learn how to respond to dynamic changes in a manner consistent with their beliefs. Policy oriented learning refers to the internal feedback loops in the policy subsystem and includes increased knowledge of external factors. ACF assumes that such learning is essential for furthering one's policy objectives. Coalitions will usually resist learnings that challenge their beliefs and elaborate those learnings that challenge their opponents beliefs. A major policy change is defined as a change in the dominant policy core beliefs of a policy subsystem while a minor policy change is defined as a change in the secondary beliefs of a policy subsystem. There are four paths that lead to a major policy change: external events or shocks; policy oriented learning; internal shocks; and negotiated agreements. Policy oriented learnings will usually only be able to change the coalitions secondary beliefs. The core beliefs usually change only as a result of non-cognitive factors external to the subsystem. Smith (2000) in his critique of the ACF argues that major policy changes are better defined as a result of multiple cascading external events rather than a single external event. ACF adopts a neo-positivist outlook and outlines three categories of hypotheses that emerge from the framework. The first category of hypotheses is based on policy beliefs, the second category is based on policy change, and the third category is based on policy learnings. The hypotheses relevant to the present paper have been reproduced below. On advocacy coalitions, the authors hypothesise that H1: “On major controversies within a policy subsystem when policy core beliefs are in dispute, the lineup of allies and opponents tends to be rather stable over periods of a decade or so.”, and H2: “Actors within an advocacy coalition will show substantial consensus on issues pertaining to the policy core, although less so on secondary aspects”. ACF has been applied in more than 80 studies (Weible et al., 2009). After conducting a review of these studies, Weible et al. (2009) conclude that although the hypotheses on policy beliefs tend to be

confirmed, the hypotheses on policy learnings and the hypothesis on external events have rarely been tested. The ACF is often criticised for making beliefs central to the coalitions and thus overrating knowledge and values while explicitly rejecting short term interests as the primary motivation for policy actors (Ladi, 2005). There are numerous examples of coalitions that share interests but do not share policy core beliefs (Cairney, 1997; Hann, 1995). Researchers also identify situations in which coalitions may trade beliefs for short term interests (Nohrstedt, 2005). The authors counter this criticism by arguing that self-interest or rent-seeking may be one of the beliefs of the coalition (Sabatier & Jenkins-Smith, 1993). Hajer's (1995) discourse coalition framework facilitates a discourse-analytic critique of ACF. The author argues that ACF takes refuge in generalised statements by neglecting the social and historical context in which the policy change takes place. As a result, important explanatory factors are put in black boxes, thus ignoring the multiplicity of variables that constitute it. Hajer also finds that narrative storylines, rather than policy beliefs, need to be operationalised in the social contexts to identify coalitions. ACF has been discussed and applied mostly in the context of western democracies (Weible et al., 2009). It has found rare applications in the developing country context and only one systematic application in the Indian context. Lintelo (2009) applies ACF in the food safety and street vending policy subsystems in India. In general, the study finds ACF to be a useful lens for studying the policy process in India, while suggesting that its relevance is diminished in the analysis of informal sectors. The study finds evidence of advocacy coalitions operating in the food safety subsystem. The study identifies the need for modifying ACF to incorporate policy changes caused by coalitions of convenience that seek to influence policies in pursuit of short-term interests. The study also finds that the judiciary and senior civil servants are insulated from being part of any advocacy coalition due to the formal institutional structure. 3.

METHODOLOGY AND DATA

In order to view the spectrum management policy subsystem in India from the lens of the Advocacy Coalition Framework, we follow a modified version of the methodology adopted by Jenkins-Smith & Clair (1993). This methodology has been recommended by the authors of the framework for quantitatively assessing and applying the Advocacy Coalition Framework (Sabatier & Jenkins-Smith, 1993). The methodology relies on the public testimonies of different actors. It uses these testimonies to identify the competing advocacy coalitions operating in a policy subsystem. In this section we explain how the methodology was applied in the present study. 3.1.

DATA COLLECTION

The methodology employs content analysis of 144 comments submitted by prominent telecommunications service providers over a period of eight years, from 2008 to 2015, in response to public consultations held by the telecommunications regulator of India. The public consultation process involves a consultation paper being released by the regulator in which numerous policy issues are raised in the form of questions. Any interested person can submit comments in response to the consultation paper; and submit counter-comments in response to comments sent by any actor. In the comments, actors respond to questions raised by the telecommunications regulator and express their policy preferences and suggestions along with reasons where possible. In the counter comments, actors usually state their agreements and disagreements to the comments submitted by other actors.

Comments and counter-comments (collectively referred to as comments) for most public consultations are published on the regulator's website. For the consultations for which comments were missing, a request for information application was filed under the Right To Information Act 2005; in response to which the regulator diligently uploaded the comments on its website. The actors, whose comments have been reviewed, and the consultations they have participated in are captured in the following table: Year Consultation

Vodaf Bharti one Airtel

2009 Consultation Paper on Overall Spectrum * Management and review of license terms and conditions 2010 Pre-Consultation paper on IMT-Advanced (4G) Mobile wireless broadband

*

*

AUS COAI Idea BSNL MTNL Tata R Jio Aircel PI Tele Comm *

*

*

*

*

2011 Consultation Paper on IMT – Advanced Mobile Wireless * Broadband Service

*

*

2012 Pre-Consultation Paper on Allocation of Spectrum in 2G * band in 22 Service Areas by auction

*

*

*

2012 Consultation Paper on Auction of Spectrum

*

*

*

*

2013 Consultation Paper on Valuation and Reserve Price * of Spectrum

*

*

2013 Consultation Paper on Reserve Price for Auction of * Spectrum in the 800 MHz Band

*

2014 Consultation Paper on Allocation and Pricing of Microwave RF carriers

*

*

*

2014 Consultation Paper on Valuation and Reserve Price * of Spectrum: Licences expiring in 2015–16

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

* *

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

Video Uninor / MTS con Telenor SSTL

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

*

Table 1

These comments contain a wealth of information that helps capture the stated beliefs of elite policy actors. A coding frame comprising of 40 variables was developed and the response for each variable was captured on a five point scale. Each variable represented a contentious policy issue and was used to capture the stated belief of an actor in a comment. Codes of 1 and 5 represented extreme opposing positions of the contentious policy issue. Variables that were not addressed in a comment were coded as a 0 to represent missing or not mentioned data. The coding frame was developed through an iterative process and each time the coding frame was modified, all comments previously coded had to be recoded. The issues in the coding frame were first identified through a review of consultation papers. However, many times, actors raised new issues or new positions in comments or counter-comments that required an existing variable to be modified or new variables to be added. The public consultation process is an open process in which comments can be received from any interested person. In this paper, we do not attempt to capture the beliefs of every tom, dick and harry

who participated in the process. We instead focus the content analysis on the submissions of the elite actors who have substantial potential to influence the policy process. Specifically, we focus on the submissions of large telecommunications service providers who regularly participate in the consultation process. There was a large amount of missing data as comments are not expected to address all prevailing policy issues and be largely restricted to the questions asked in a particular consultation. At an average, every comment only addressed 13% of the variables in the coding frame. To deal with missing data, two strategies were adopted. First, variable values over a four year time period were aggregated. Second, composite scales were developed using conceptually similar variables with high correlations. Since the coding focused on consultations for a period of eight years from 2008 to 2015, the variables were aggregated within two different time periods, that is from 2008 to 2011 and from 2012 to 2015. 3.2.

DATA VALIDITY & RELIABILITY

Most comments have been submitted by the official representatives of the organisations participating in the consultation process. It is recognised that such comments may not always be a true representation of one's actual beliefs. One's stated or expressed beliefs may be different from one's true beliefs either due to organisational constraints, or due to the context in which the comment is embedded. Often, the comments are an effort to persuade the regulator and to promote the interests of the organisation. Therefore, in this study, we focus on stated policy positions or expressed beliefs. In content analysis of this sort, reliability of the coding is assessed on the basis of two dimensions. The first is the decision of the coder regarding whether or not to code a particular variable for a comment. The second is the value coded for a particular variable. Usually, in such content analysis using coding frames, reliability tests are conducted at regular intervals by checking inter-coder reliability between two or more coders. As reference, Jenkins-Smith & Clair (1993) had 15% splits, in which, one coder chose to code a particular variable for a testimony, while the second coder did not code that variable for that testimony. However, once both coders chose to code, the attributed values were less than 1 point apart 98% of the times. Since, the journey of a doctoral researcher is a lonely one, and since the option of a two coder reliability test was not feasible, the researcher chose to recode 50 comments that were randomly selected. In this, the split was 17%, and the values were less than 1 point apart 96% of the times. 3.3.

VARIABLES

The analysis presented in this paper relies on four composite scales. Each composite scale represents a policy core belief and each comprising variable represents a secondary belief. Extreme values of 1 and 5 represent the extreme policy positions for that belief. These composite scales were derived from conceptually similar variables that were used in the coding process. These conceptually similar variables displayed a high degree of correlation (>0.6). The composite scales and their comprising variables are discussed below: 1) Level Playing Field The first composite scale is titled “Level Playing Field”. This composite scale relates to concerns that the present regulations are favourable to the incumbent GSM telecommunications service providers. A value of 1 represents the position that the present playing field is distorted in favour of the incumbents. A value of 5 represents the position that the regulations do not favour the incumbents. Incumbent refers to early entrants who were awarded licenses as early as 1994. These providers use the 3GPP family of

technologies, that is GSM for 2G and UMTS for 3G. Since the 900 MHz band was vacant when these incumbents entered the market, they were awarded spectrum in this band. Subsequent entrants received spectrum in the comparatively inefficient 1800 MHz band, which requires relatively higher capital and operating expenditure to achieve the same quality of service and coverage, especially in rural areas. Most of the concerns about level playing field arise from this unfavourable allocation to the late entrants. Additional concerns arise from the argument that the quantity of spectrum allocated to incumbent operators is far higher than new entrants; while the new entrants have failed to receive adequate spectrum to launch services or provide efficient services. The variables that comprise this composite scale are as follows: • Refarming of 900 MHz Band – Some late entrants have argued that the 900 MHz band should be “refarmed” by redistributing the 900 MHz band equally amongst all service providers. Some have also argued that the band should be refarmed by taking it back from incumbents and auctioning it in blocks of 5 MHz so that the band may be used for provision of UMTS/LTE services. In lieu of spectrum taken back, alternate spectrum in the 1800 MHz band was proposed to be given to the incumbents. Other operators have argued that such “refarming”, which involves taking spectrum back from incumbent operators, will be disruptive and will affect the continuity of services. • Returning Excess of Contracted Spectrum – Some actors have argued that incumbents holding spectrum in excess of the contracted spectrum (6.2 MHz), allocated to them administratively through the subscriber linked criteria, should either be returned to the government or be paid for by the incumbents at market prices prospectively or retrospectively. • Extension of Expiring Licenses – Some late entrants have argued that when the 20 year license term of the incumbent operators ends, then spectrum linked to these licences should be put for auction. In contrast, incumbents have argued that these licenses should be administratively extended for a period of 10 years. • Valuation of 900 MHz Band – Some late entrants have argued that the reserve price for the 900 MHz band should be kept twice the valuation of 1800 MHz band due to its technical and economic efficiency, and its harmonised market ecosystem for provision of IMT services. In contrast, incumbents currently using the 900 MHz band have argued for keeping the valuation as low as possible since the band is only expected to be used for provision of legacy 2G services in the foreseable future, and also for keeping costs low in rural and remote areas. • Un-Level Allocation of Spectrum – Some operators have argued that while incumbent operators hold excess of contracted spectrum, many new entrants have failed to even receive start-up spectrum, which is the minimum required for starting proving wireless services. In contrast, incumbents have argued that they are using all spectrum efficiently and such spectrum is necessary for continued provision to current subscribers. The values for the composite scale are provided in the table below for the two time periods. The value outside the bracket is the aggregate of the values attributed to the comprising variables in the coding process. The value in brackets is the number of comments that have addressed that variable. Vodafone Bharti Airtel

AUSPI / ABTO

COAI

Idea BSNL Cellular

MTNL

Tata Tele

20082011

4.6 (5)

1.4 (5)

5 (3)

4.33 (3) 5 (2)

4 (1)

1.16 (6) 1.12 (8)

5 (2)

20122015

4.84 (13) 4.72 (11)

1.23 (13)

4.83 (6) 4.5 (6)

4 (6)

1.8 (10) 1.25 (12) 1.62 (8)

3.5 (4) 1.42 (7) 1.85 (7) 1.69 (13)

Table 2

5 (4)

Reliance Comm

Jio / Aircel Infotel

Videocon Uninor / MTS Telenor SSTL

1.5 (2)

1 (3)

2) Market Based Practices The second composite scale is titled “Market Based Practices”. This composite scale relates to the idea that spectrum should be managed through market oriented practices in contrast to command-andcontrol or administrative practices. A value of 1 represents the position that spectrum should be managed through market oriented practices such as auctions in the primary market and trading in the secondary market. A value of 5 represents the position that spectrum should be allocated and managed administratively. Historically, spectrum has been linked to licenses and has been allocated through a mix of beauty contests, auctions and first-come-first-serve. Licensees have been awarded additional spectrum mostly through administrative means such as the Subscriber Linked Criteria (SLC). To prevent hoarding of spectrum received through administrative means, there have been spectrum usage charges that escalate with the quantity of spectrum held. The variables that comprise this composite scale are as follows: • Auctions in Primary Market – Some operators have argued that spectrum should only be auctioned in the primary market, while other operators have argued that spectrum should be allocated administratively by way of first-come-first-serve or subscriber linked criteria. Actors in favour of auctions may be seeking the transparency and efficiencies of market practices, or revenue maximisation for the government, or a level playing field, or a financial barrier for challengers. Actors in favour of administrative allocation primarily seek allocation of contracted spectrum that, in their opinion, the licensor is contractually obliged to provide so that a level playing field may be established with incumbent operators. • Delinking of Spectrum and Licenses – Historically, spectrum and licenses have been linked to each other. However, for reallocation of spectrum in the secondary market, it is would be required that these two be delinked so that operators can trade only a portion of their spectrum. Opponents have argued that a license without spectrum or spectrum without a license is of no use and therefore should not be allowed. • Spectrum Trading – Spectrum trading refers to the assignment or leasing of excess spectrum to other operators in the secondary market. This allows for efficient reallocation of spectrum for most efficient use. Opposition to trading is derived from fear of hoarding and speculation, and the possibility that operators may make windfall gains if allocated spectrum at administrative prices in the primary market but allowed to trade at market prices in the secondary market. • Spectrum Sharing – Similar to trading, spectrum sharing refers to pooling of spectrum and related active infrastructure. This is emphasised given that spectrum allocation in the primary market in India has been fragmented and non-contiguous, and thus non-suitable for efficient provision of existing services or for switching to IMT services. • Spectrum Liberalisation – Liberalisation refers to the idea that spectrum can be used for any technology and that the government should not define the service or technology to be used. In this, while some support liberalisation of all spectrum, some others only support the liberalisation of spectrum that is auctioned. Some others argue that spectrum and licenses are already liberalised in India and that the current form of liberalisation is a guise for refarming. Others oppose liberalisation with the argument that if spectrum is liberalised in the hands of the incumbents, it would worsen the already distorted playing field. • Abolishment of Subscriber Linked Criteria – Historically, spectrum beyond the start-up spectrum has been administratively allocated on the basis of the number of subscribers that an operator holds. If the operator reached a threshold number of subscribers, then additional spectrum would be administratively allocated to it. • Flat Spectrum Usage Charges – In order to prevent hoarding of spectrum, the government has administratively set spectrum usage charges that escalate with quantity of spectrum held. It has been argued that such escalating charges are not required in market oriented practices wherein the licensee pays market prices for spectrum and thus has no reason to hoard it or not use it

efficiently. In contrast, others argue that flat charges is just a guise to benefit those licensees that already hold substantial quantity of spectrum. Allocation of Contracted Spectrum – Some actors argue that the Central Government is obliged to administratively allocate spectrum to those providers who have previously been granted licenses. Not all these providers have received spectrum linked to such licenses for provision of wireless services.

•

Vodafone Bharti Airtel

AUSPI / COAI ABTO

Idea BSNL Cellular

20082011

1.18 (11) 1.33 (15)

2.77 (9) 1.11 (9)

20122015

1.53 (15) 1.2 (10) 4.5 (8)

1 (3)

MTNL

Tata Tele

Reliance Comm

1.28 (7) 1.42 (7) 4 (5)

4 (13)

3.72 (11) 2 (1)

2.12 (8) 1 (2)

3.44 (9) 4.55 (9)

2.5 (4)

Jio / Aircel Infotel

3.33 (6)

Videocon Uninor / MTS Telenor SSTL

2.12 (8)

1.55 (9) 2.77 (9)

1.5 (2) 1 (4)

2.25 (4) 1.66 (6)

Table 3

3) Preferential Treatment of Public Sector Units The third composite scale is titled “Preferential Treatment of Public Sector Units”. The composite scale relates to the regulatory treatment of government owned public sector units (PSUs). A value of 1 represents the position that service provision should be left to market forces and that government owned providers should compete at an equal footing as private players. A value of 5 represents the position that the public sector units should receive preferential treatment in spectrum allotment and regulation, as their objective and focus is substantially different from that of private players. Preferential treatment may take tangible shape in the form of regulations which mandate that PSUs don't need to participate in spectrum auctions, or PSUs should not be required to pay for administratively granted excess spectrum, or PSUs should be compensated for service provision in rural or remote areas etc. The variables that comprise this composite scale are as follows: • Vacation of Underutilised Spectrum – Certain actors have argued that PSUs should return the excess spectrum with them (such as that in the 800 MHz band) which was allocated to them administratively so that it may be used more efficiently by private players. • Need to Participate in Auctions – Certain actors have argued that PSUs should not be required to participate in auctions and should be granted spectrum administratively at market determined prices. Other actors have argued that PSUs should compete at an equal footing with with private players in auctions. • Access Deficit Charges (ADC) – These are charges that were paid to incumbent fixed line operators (dominated by PSUs) to compensate them for continuing to subsidise access services. Since the principle beneficiary of the ADC has been BSNL, ADC has often been treated as a proxy for whether the public sector should be subsidised by the private sector. Vodafone Bharti Airtel 20082011 20122015

2 (2)

AUSPI / COAI ABTO

1 (1)

1.66 (3) 1 (1)

1 (1)

4 (1)

2 (2)

Idea BSNL Cellular

5 (3) 1 (1)

MTNL

Tata Tele

Reliance Comm

5 (1)

1 (1)

1 (1)

5 (3)

4 (3)

Jio / Aircel Infotel

2 (2)

Videocon Uninor / MTS Telenor SSTL

1 (2)

5 (1)

Table 4

4) Growth Path for CDMA Providers The fourth composite scale is titled “Growth Path for CDMA Providers”. This composite scale relates to concerns that the outlook for CDMA is limited and suffering from a poor ecosystem with an oblique future. A value of 1 represents the position that 800 MHz band should continue to serve as a growth

path for incumbent CDMA providers and have low valuation. A value of 5 represents the position that the 800 MHz band should should be refarmed for LTE or E-GSM services and have a high valuation. The variables that comprise this composite scale are as follows: • Valuation of 800 MHz band – Certain operators feel that 800 MHz band has an increasingly degrading ecosystem of devices and networks as a result of which its valuation should be kept low – lower than or equal to the 1800 MHz band. Other operators feel that the valuation of the 800 MHz should be kept high as its propagation characteristics are comparable to the 900 MHz band, and when liberalised, it can be used for provision of LTE services or can be harmonsied with the E-GSM band. • E-GSM – Certain operators feel that the 800 MHz band should be harmonised with the E-GSM band, effectively making 10 MHz of the 800 MHz band as part of the overall 900 MHz band ecosystem. On the other hand, other operators feel that harmonisation with the E-GSM band will negatively impact the continuity of their services and adversely affect the investments made by CDMA operators. • Refarming of 800 MHz band – Some actors have argued that CDMA services should be provided alternate spectrum in the 1900 MHz band so that 800 MHz can be refarmed for provision of LTE services. While some actors have argued that such refarming would be disruptive, other actors have argued that such refarming is currently not feasible as the 1900 MHz band is occupied. • Mobility in Wireless Local Loop & Dual Technology – Historically, CDMA operators were only allowed to provide fixed services with limited mobility even though full mobility was technically feasible. This anomaly led to the transition of the licensing regime to the technology neutral unified access services license, wherein CDMA operators could provide full mobility, and the eventual transition of some operators to provision of services using two technologies (CDMA and GSM), commonly referred to as dual technology licenses. Despite these issues being decided on by the government prior to 2008, there continue to be differences between operators on the legitimacy of dual technology licenses and the spectrum allocations to be made to such operators. Vodafone 20082011

2 (1)

20122015

1.81 (11)

Bharti AUSPI / Airtel ABTO

1.71 (7)

COAI

Idea BSNL Cellular

2 (1)

2 (1)

4.87 (8) 1.37 (8) 2 (5)

MTNL

4 (2)

Tata Tele

Reliance Comm

4 (1)

4 (1)

Jio / Aircel Infotel

4.33 (6) 4.83 (12) 2.33 (3)

Videocon Uninor / MTS Telenor SSTL

2 (1)

2 (1)

1.5 (2) 1 (1)

1.75 (4) 4.63 (11)

Table 5

4.

ANALYSIS AND DISCUSSION

4.1.

HOMOGENEITY OF BELIEFS

Over Time According to the framework, the policy core beliefs of various actors are expected to be resistant to change over time. For the scales “Level Playing Field” and “Market Based Practices”, sufficient data was available to check if there was statistically significant difference in the beliefs of the various actors between the two time periods. The results of the one-way Anova, as displayed in the following tables, demonstrate that the difference in beliefs of any given actor was not statistically significant between the two time periods. Thus, the beliefs of these telecommunications service providers did not change from one time period to the next. This is consistent with what the authors of the Advocacy Coalition Framework have argued about the obstinacy of policy core beliefs.

Table 6: Level Playing Field – Anova Across Time Vodafone Bharti Airtel

AUSPI / COAI ABTO

Idea BSNL Cellular

MTNL

Tata Tele

20082011

4.6 (5)

1.4 (5)

5 (3)

4.33 (3) 5 (2)

4 (1)

1.16 (6) 1.12 (8)

5 (2)

20122015

4.84 (13) 4.72 (11)

1.23 (13)

4.83 (6)

4.5 (6)

4 (6)

1.8 (10) 1.25 (12) 1.62 (8)

3.5 (4) 1.42 (7) 1.85 (7) 1.69 (13)

NS

NS

NS

NS

NS

NS

Anova NS

5 (4)

NS

-

Reliance Comm

NS

Jio / Aircel Infotel

-

Videocon Uninor / MTS Telenor SSTL

1.5 (2)

-

NS

1 (3)

#

NS = Not Significant (>0.1); # = Significant at 0.1; * = Significant at 0.05; ** = Significant at 0.01; *** = Significant at 0.001 Table 7: Market Based Practices – Anova Across Time Vodafone Bharti Airtel

AUSPI / COAI ABTO

Idea BSNL Cellular

MTNL

Tata Tele

Reliance Comm

3.72 (11) 2 (1)

20082011

1.18 (11) 1.33 (15) 2.77 (9) 1.11 (9)

1.28 (7) 1.42 (7) 4 (5)

4 (13)

20122015

1.53 (15) 1.2 (10) 4.5 (8)

1 (3)

2.12 (8) 1 (2)

2.5 (4)

3.44 (9) 4.55 (9)

NS

NS

NS

NS

Anova NS

NS

*

NS

NS

Jio / Aircel Infotel

Videocon Uninor / MTS Telenor SSTL

2.12 (8)

1.55 (9) 2.77 (9)

3.33 (6)

1.5 (2) 1 (4)

2.25 (4) 1.66 (6)

-

NS

NS

-

NS

NS = Not Significant (>0.1); # = Significant at 0.1; * = Significant at 0.05; ** = Significant at 0.01; *** = Significant at 0.001

Between Actors According to the framework, actors are expected to take sides on contentious policy issues concerning the policy subsystem. We checked if there was a statistically significant difference between the positions of competing actors in a given time period. Subsequently, if there was a difference, we identified the subsets of actors with homogeneous beliefs. For the scale “Level Playing Field”, there was statistically significant difference between groups as determined by one-way Anova for both time periods. A Tukey post-hoc test was performed to identify homogeneous subsets. Table 8: Level Playing Field – Anova Across Actors Vodafone Bharti AUSPI / COAI Airtel ABTO

Idea BSNL Cellular

MTNL

Tata Tele

Reliance Jio / Aircel Comm Infotel

20082011

4.6 (5)

1.4 (5) 5 (3)

4.33 (3)

4 (1)

1.16 (6)

1.12 (8)

20122015

4.84 (13) 4.72 (11)

1.23 (13)

4.5 (6)

4 (6)

1.8 (10)

1.25 (12)

5 (4)

4.83 (6)

5 (2)

5 (2) 1.62 (8)

Videocon Uninor / MTS Telenor SSTL

1.5 (2) 1 (3)

3.5 (4) 1.42 (7) 1.85 (7)

1.69 (13)

Anova

*** ***

NS = Not Significant (>0.1); # = Significant at 0.1; * = Significant at 0.05; ** = Significant at 0.01; *** = Significant at 0.001 Table 9: Homogeneous Subsets for 2008-2011 Tukey HSD Actor N Subset for alpha = 0.05 1 2 MTS SSTL 3 1.0000 Reliance Comm 8 1.1250 Tata Tele 6 1.1667 AUSPI 5 1.4000 Uninor 2 1.5000 Idea 3 4.3333 Vodafone 5 4.6000 Airtel 4 5.0000 COAI 3 5.0000 Aircel 2 5.0000 BSNL 2 5.0000 Sig. .888 .601 Means for groups in homogeneous subsets are displayed. a. Uses Harmonic Mean Sample Size = 3.196.

Table 10: Homogeneous Subsets for 2012-2015 Tukey HSD Actor N Subset for alpha = 0.05 1 2 3 AUSPI 13 1.2308 Reliance 12 1.2500 Videocon 7 1.4286 Jio 8 1.6250 MTS 13 1.6923 Tata Tele 10 1.8000 Uninor 7 1.8571 Aircel 4 3.5000 MTNL 6 4.0000 4.0000 Idea 6 4.5000 4.5000 Bharti 11 4.7273 COAI 6 4.8333 Vodafone 13 4.8462 Sig. .749 .097 .288 Means for groups in homogeneous subsets are displayed.

b. The group sizes are unequal. The harmonic mean of the group sizes is used. Type I error levels are not guaranteed.

a. Uses Harmonic Mean Sample Size = 7.804. b. The group sizes are unequal. The harmonic mean of the group sizes is used. Type I error levels are not guaranteed.

The results reveal two major homogeneous subsets of actors for both time periods. The established GSM incumbent actors like Airtel, Vodafone, Idea and their association COAI comprise one subset in both time periods. The new entrants, CDMA operators and dual technology operators comprise the other subset. This shows that the issue of level playing field has been a contentious policy issue in which actors have taken sides. For the scale “Market Based Practices”, there was statistically significant difference between groups as determined by one-way Anova for both time periods. A Tukey post-hoc test was performed to identify homogeneous subsets. Table 11: Market Based Practices – Anova Across Actors 20082011

Vodafone Bharti Airtel

AUSPI / COAI ABTO

Idea BSNL Cellular

MTNL

Tata Tele

Reliance Jio / Aircel Comm Infotel

1.18 (11) 1.33 (15)

2.77 (9)

1.28 (7)

4 (5)

4 (13)

3.72 (11)

1.11 (9)

1.42 (7)

2 (1)

2.12 (8)

Videoco Uninor / MTS n Telenor SSTL

1.55 (9)

Anova

2.77 (9) ***

2012- 1.53 (15) 1.2 4.5 (8) 1 (3) 2.12 1 (2) 2.5 (4) 3.44 4.55 (9) 3.33 1.5 (2) 1 (4) 2.25 1.66 (6) *** 2015 (10) (8) (9) (6) (4) NS = Not Significant (>0.1); # = Significant at 0.1; * = Significant at 0.05; ** = Significant at 0.01; *** = Significant at 0.001

Table 12: Homogeneous Subsets for 2008-2011 Tukey HSD Actors

N

Subset for alpha = 0.05 1 2 1.1111 1.1818 1.2857 1.3333 1.4286 1.5556 2.1250 2.1250 2.7778 2.7778 2.7778 2.7778 3.7273

3 COAI 9 Vodafone 11 Idea 7 Bharti 15 BSNL 7 Uninor 9 Aircel 8 AUSPI 9 2.7778 MTS 9 2.7778 Reliance 11 3.7273 MTNL 5 4.0000 Tata Tele 13 4.0000 Sig. .124 .162 .553 Means for groups in homogeneous subsets are displayed. a. Uses Harmonic Mean Sample Size = 8.692. b. The group sizes are unequal. The harmonic mean of the group sizes is used. Type I error levels are not guaranteed.

Table 13: Homogeneous Subsets for 2012-2015 Tukey HSD Actors

N

Subset for alpha = 0.05 1 2 COAI 3 1.0000 BSNL 2 1.0000 Videocon 4 1.0000 Bharti 10 1.2000 Aircel 2 1.5000 Vodafone 15 1.5333 MTS 6 1.6667 1.6667 Idea 8 2.1250 2.1250 Uninor 4 2.2500 2.2500 MTNL 4 2.5000 2.5000 Jio 6 3.3333 3.3333 Tata Tele 9 3.4444 3.4444 AUSPI 8 4.5000 Reliance 9 4.5556 Sig. .210 .059 Means for groups in homogeneous subsets are displayed. a. Uses Harmonic Mean Sample Size = 4.582. b. The group sizes are unequal. The harmonic mean of the group sizes is used. Type I error levels are not guaranteed.

Like “Level Playing Field”, the results show that “Market Based Practices” has also been a contentious policy issue in which sides are divided. However, for 2012-2015, many actors fall on both sides indicating that these actors possibly take issue based stands on secondary beliefs that relate to market based practices. For the scale “Growth Path for CDMA Providers”, there was statistically significant difference between groups as determined by one-way Anova for both time periods. A Tukey post-hoc test was performed to identify homogeneous subsets.

Table 14: Growth Path For CDMA Providers – Anova Across Actors 20122015

Vodafone Bharti AUSPI / COAI Airtel ABTO

Idea BSNL Cellular

MTNL

Tata Tele

Reliance Jio / Comm Infotel

Aircel

1.81 (11) 1.71 (7)

2 (5)

4 (2)

4.33 (6)

4.83 (12)

1.5 (2) 1 (1)

4.87 (8)

1.37 (8)

2.33 (3)

Videoco Uninor / MTS n Telenor SSTL

1.75 (4)

4.63 (11)

Anova

***

NS = Not Significant (>0.1); # = Significant at 0.1; * = Significant at 0.05; ** = Significant at 0.01; *** = Significant at 0.001 Table 15: Homogeneous Subsets for 2012-2015 Tukey HSD CDMA2012Actor

N

Subset for alpha = 0.05 1 2 1.3750 1.5000 1.7143 1.7500 1.8182 2.0000 2.0000 2.3333 2.3333 4.0000

3 4 COAI 8 Aircel 2 Bharti 7 Uninor 4 Vodafone 11 Idea 5 Jio 3 2.3333 MTNL 2 4.0000 4.0000 Tata Tele 6 4.3333 4.3333 MTS 11 4.6364 Reliance 12 4.8333 AUSPI 8 4.8750 Sig. .913 .066 .066 .952 Means for groups in homogeneous subsets are displayed. a. Uses Harmonic Mean Sample Size = 4.601. b. The group sizes are unequal. The harmonic mean of the group sizes is used. Type I error levels are not guaranteed.

The results show that “Growth Path for CDMA Providers” has also been a contentious policy issue in which actors have taken sides. Actors with spectrum in the 800 MHz band (MTNL, Tata, MTS and Reliance) have opposed any efforts to thwart their growth path. In contrast, actors not holding spectrum in the 800 MHz band have supported efforts to change the growth trajectory of the band. 4.2.

OPERATIONALISING ADVOCACY COALITIONS

In order to operationalise and identify advocacy coalitions, a cluster analysis was performed using the policy core beliefs captured by the composite scales. A cluster analysis displays the proximity of different actors depending on their expressed beliefs. Time period of 2008 to 2011 Not all policy issues are contentious in a given time period. Some issues are put on the back burner and remain dormant while other issues gain primacy in the pecking order. In a highly technical policy subsystem, such as the present one under study, change in underlying technology may result in previously contentious issues becoming irrelevant and some other previously non-contentious issues suddenly becoming divisive. The data reflects that issues of level playing field and market based practices were the issues in vogue during this time period. From the amount of missing data for “Preferential Treatment of Public Sector Units” and “Growth Path for CDMA Providers” from 2008 to 2011, it is apparent that these issues were not hotly debated before the regulator in this period. For this time period, a hierarchical cluster analysis was performed using the composite scales capturing “Level Playing Field” and “Market Based Practices”. The dendrogram in the figure below is a graphical representation of the cluster analysis.

Fig. 2 Cluster Analysis (2008-2011)

From the dendrogram, there is clear evidence of at least two advocacy coalitions at loggerheads during this time period. The first advocacy coalition (Coalition of Incumbent GSM Operators) comprises of incumbent GSM operators with spectrum in the 900 MHz band including Airtel, Vodafone, Idea, Aircel and MTNL/BSNL. These operators have collectively opposed any moves to create parity in the present playing field, such as by refarming of the 900 MHz band, or by returning spectrum in excess of contracted spectrum etc., which may negatively affect the advantage that they presently hold. These operators also lean towards market based practices for spectrum allocation. The second advocacy coalition (Coalition of Unified Service Operators) comprises of pure CDMA operators like MTS and dual technology licensees (CDMA and GSM) like Tata Tele and Reliance Communications. These operators have argued in favour of creating a level playing field with the incumbent GSM operators. They have also favoured administrative practices for allocation of spectrum as they fear that market based practices will worsen the already distorted playing field. COAI, the association that represents the interests of incumbent private GSM operators falls in first coalition. Likewise, AUSPI, the association that represents the interests of CDMA operators and dual technology licensees, falls in the second coalition. The second coalition also weakly comprises of Uninor, which is a late entrant in the GSM market. Unlike other members of the second coalition, it does not hold spectrum in the 800 MHz band for provision of CDMA services. Similarly, unlike members of the first coalition, it does not hold spectrum in

the 900 MHz band. These characteristics of Uninor put it in a unique position in comparison to other operators and thus explain its lack of collective action with either coalition. Its policy positions represent its alignment with the second coalition against the incumbent operators who hold spectrum in the 900 MHz band. However, its distance from the second coalition can be explained by the fact that it does not hold spectrum in the 800 MHz band to provide CDMA services, unlike all other actors in the second coalition. Table 16: Beliefs of Coalitions Coalition of Incumbent GSM Operators Level Playing Field

Market Based Practices

Bharti Airtel

5 (4)

1.33 (15)

COAI

5 (3)

1.11 (9)

Vodafone

4.6 (5)

1.18 (11)

Idea

4.33 (3)

1.28 (7)

BSNL

5 (2)

1.42 (7)

MTNL

4 (1)

4 (5)

Aircel

5 (2)

2.12 (8)

Anova

NS

*** (Including MTNL) NS (Excluding MTNL) Coalition of Unified Service Operators

Level Playing Field

Market Based Practices

Tata Tele

1.16 (6)

4 (13)

Reliance Comm

1.12 (8)

3.72 (11)

AUSPI

1.4 (5)

2.77 (9)

MTS SSTL

1 (3)

2.77 (9)

Uninor

1.5 (2)

1.55 (9)

Anova

NS

** (Including Uninor) NS (Excluding Uninor)

NS = Not Significant (>0.1); # = Significant at 0.1; * = Significant at 0.05; ** = Significant at 0.01; *** = Significant at 0.001

Time Period of 2012-2015 The year 2012 has been chosen as the starting of the new time period because of a subsystem altering event in which the Supreme Court of India, while deciding on a matter related to irregularities in spectrum allocation, ruled that spectrum should only be allocated through fair and transparent mechanisms such as auctions. The Supreme Court also cancelled a number of previous licenses such as those held by Uninor, MTS and Videocon, amongst others. As discussed previously, in a highly technical policy subsystem, such as the present one under consideration, policy issues evolve quickly due to the progress in underlying technology. This time period reflects rising concerns about the degradation of the CDMA ecosystem. Like the previous time period, in this time period, issues related to level playing field and market based practices were also hotly debated. The coding analysis reflects that the degradation of the CDMA ecosystem put concerns about its growth path in the forefront along with the other two issues. For this time period, a hierarchical cluster analysis was performed using the composite scales capturing “Level Playing Field”, “Market Based Practices” and “Growth Path for CDMA Providers”. The dendrogram in the figure below is a graphical representation of the cluster analysis.

Fig. 3 Cluster Analysis (2012-2015)

In this time period, we can see how the advocacy coalitions have evolved from the previous time period. The dendrogram shows evidence of sub-coalitions emerging in the policy subsystem. It can be seen that the Coalition of Incumbent GSM Operators, as identified in the previous time period, comprising of Airtel, Vodafone, Idea, Aircel and COAI remains strong and cohesively bound. These private operators continue to support market based practices for spectrum allocation and argue that the 800 MHz band should be harmonised with E-GSM and valued at par with the 900 MHz band. These operators continue to oppose any efforts to put them at a disadvantage to establish a level playing field. We can see evidence of a new coalition (Coalition of New Entrants) emerging comprising of Reliance Jio, Uninor and Videocon. Of these three, Uninor and Videocon had their spectrum cancelled pursuant to the 2G spectrum scam judgement and had to repurchase their spectrum in the subsequent auctions for the 1800 MHz band. Reliance Jio is a new entrant which has bought additional spectrum in the 1800 MHz band for complementing its spectrum in the 2300 MHz band. The spectrum held by these operators is liberalised and can be used for provision of LTE based services. They form a distinct

coalition because their position is neither aligned with the incumbent operators nor aligned with the dual technology operators. These operators seek to establish a level playing field with the incumbent operators and do not share concerns about the growth of CDMA with the dual technology operators. Further, since these operators have purchased spectrum in auctions, they support market based practices. It can be seen that the second coalition (Coalition of Unified Service Operators), as identified in the previous time period, compromising of dual technology operators and pure CDMA operators, has also seen widening rifts. The dendrogram reflects that differences in policy positions have emerged between MTS and the dual technology operators Tata Tele and Reliance Communications. These operators continue to share concerns about the distorted playing field favouring the incumbent and share a new concern about the future growth path of CDMA services. While all members of the coalition remain united against the un-level playing field favouring the incumbent GSM operators, MTS has begun to support market based practices after having its licenses cancelled and having purchased spectrum in the open market. In contrast, the dual technology operators continue to demand and support administrative forms of allocation of spectrum. The position of the public sector operators (MTNL and BSNL) is entirely unique in the subsystem. They, like the incumbent private operators, hold spectrum in the 900 MHz band. However, they share concerns with the CDMA operators about the growth path for CDMA operations since they also hold spectrum in the 800 MHz band. This unique position puts them out of sync with other coalitions operating in the policy subsystem. Coincidently, the members of the Coalition of Incumbent GSM and the members of the Coalition of New Entrants are all collectively organised as a prominent association known as the Cellular Operators Association of India (COAI). Similarly, members of the Coalition of Unified Service Operators are organised as the Association of Unified Telecom Service Providers of India (AUSPI). In the following table, we perform a one-way Anova to check which beliefs are homogeneous across the constituent members of COAI and AUSPI.

Table 17: Advocacy Coalitions Mapped to Industry Associations COAI and COAI Members – Coalition of Incumbent GSM Operators and Coalition of New Entrants Level Playing Field

Market Based Practices

Growth Path for CDMA Providers

Vodafone

4.84 (13)

1.53 (15)

1.81 (11)

Bharti Airtel

4.72 (11)

1.2 (10)

1.71 (7)

COAI

4.83 (6)

1 (3)

1.37 (8)

Idea

4.5 (6)

2.12 (8)

2 (5)

Aircel

3.5 (4)

1.5 (2)

1.5 (2)

Reliance Jio

1.62 (8)

3.33 (6)

2.33 (3)

Uninor

1.85 (7)

2.25 (4)

1.75 (4)

Videocon

1.42 (7)

1 (4)

1 (1)

***

* (Including Reliance Jio) NS (Excluding Reliance Jio)

NS

Post-hoc test using Tukey provided below

Post-hoc showed entire set as homogeneous subset using Tukey

AUSPI and AUSPI Members – Coalitions of Unified Service Operators Level Playing Field

Market Based Practices

Growth Path for CDMA Providers

AUSPI

1.23 (13)

4.5 (8)

4.87 (8)

Reliance Comm

1.25 (12)

4.55 (9)

4.83 (12)

Tata Tele

1.8 (10)

3.44 (9)

4.33 (6)

MTS

1.69 (13)

1.66 (6)

4.63 (11)

#

** (Including MTS)

NS

Post-hoc showed entire set as homogeneous subset using Tukey

NS (Excluding MTS)

NS = Not Significant (>0.1); # = Significant at 0.1; * = Significant at 0.05; ** = Significant at 0.01; *** = Significant at 0.001 Table 18: Homogeneous Subsets for Level Playing Field for COAI Tukey HSD Coalition1Level2012Actor

N

Subset for alpha = 0.05 1

2

3

Videocon

7

1.4286

Reliance Jio

8

1.6250

Uninor

7

1.8571

Aircel

4

3.5000

Idea

6

4.5000

Bharti

11

4.7273

COAI

6

4.8333

Vodafone

13

4.8462

Sig.

.911

.086

4.5000

.970

The results show that COAI members are united in support of market based practices for allocation of spectrum and are also united in support of changing the growth path for CDMA services. However, COAI members are divided on the issue of level playing field. Specifically, the Coalition of New Entrants is opposed to Coalition of Incumbent GSM Operators on the issue of level playing field and this tussle between the two coalitions takes place within COAI itself. Therefore, the internal rules within COAI for resolving these differences and reaching consensus would have been under strain in this time period. Till now, the official submissions of COAI in public consultations have clearly aligned with the Coalition of Incumbent GSM Operators highlighting the possible dissidence of the Coalition of New Entrants within COAI. The results also show that AUSPI members are united on issues of level playing field and in maintaining the current growth path for CDMA services. However, AUSPI members are divided on the issue of market based practices. Specifically, unlike other members, MTS has been leaning towards market based practices for allocation of spectrum after having purchased spectrum in open auctions pursuant to the cancellation of its licenses as a result of the 2G spectrum scam.

4.3.

DOMINATION OF COALITIONS

According to the framework, a dominating coalition is able to convert its beliefs into policy outputs. In our analysis, we find that different coalitions dominated on different policy issues concerning the subsystem. The policy broker, which in the present case is the telecommunications regulator, played an important role in reaching a compromise between competing coalitions. In addition to the open consultation process, the regulator also held open houses, wherein it pushed coalitions to reach consensus on technical issues. On issues where competing coalitions were not able to reach consensus, the regulator would give reasoned decisions, thus grounding policy recommendations in transparency and logic, which enhanced trust in the regulator as the policy broker. Here we recognise the multiplicity of venues and the interaction between them. Specifically, we recognise that different coalitions may dominate in different venues. In the present case, the telecommunications regulator is a distinct venue from executive bodies known as the Department of Telecommunications and Telecom Commission. Procedurally, the regulator can only give non-binding recommendations, which are usually considered seriously by the executive but not always accepted. Therefore, dominating at the venue of the regulator need not necessarily imply that the belief will be translated into a policy output by the executive. The executive provides an additional venue wherein an aggrieved coalition may be able to get the regulators decision overturned. For example, the Coalition of Incumbent GSM Operators dominated the venue of the regulator with respect to the question of harmonising the 800 MHz band with E-GSM in 2012-2013. However, the Coalition of Unified Service Operators was able to overturn that decision at the venue of the executive decision-making. Notably, effectiveness of coalitions is different at different venues. The nature of resources available to coalitions and the type of advocacy adopted by these coalitions may also differ from venue to venue. Attributes of the venue, such as the level of transparency, also determine the effectiveness of coalitions. For example, the executive as a venue of decision-making, is not required to give reasoned decision and lacks the transparency of the regulator. On the issue of level playing field, the telecommunications regulator supported the positions of the Coalition of Unified Service Operators. In both time periods under consideration, the regulator released multiple recommendations (2010, 2012 and 2013) in support of refarming of the 900 MHz band, auction of spectrum linked to expiring licenses, and a fee for spectrum held in excess of contracted spectrum. On the issue of market based practices, the telecommunications regulator supported the Coalition of

Incumbent GSM Operators. The regulator supported delinking of spectrum and licenses, auction of spectrum in the primary market, liberalisation of spectrum, and spectrum trading and sharing in the secondary market. On both issues, it took the shock/event of the 2G Spectrum Scam to make the executive imbibe the recommendations of the telecommunications regulator a part of its official policy. This new policy came to be known as the National Telecom Policy 2012. The effect of events, and the ability of competing coalitions to channelise these events to create major policy changes is discussed in the following section. 4.4.

COALITION RESOURCES AND EVENTS

According to the framework, external and internal events or shocks are recognised as important paths to policy change. However, it is not expected that all events will translate into policy change. The ability of coalitions to translate events into policy changes depends on the constraints and resources of coalitions. Sabatier and Weible (2007, pp. 201–2) identify six categories of coalition resources: formal legal authority to make policy decisions, public opinion, information, mobilizable troops, financial resources, and skillful leadership. Coalitions mobilise these resources in order convert their common policy core beliefs into policy outputs. For example, to counter the issue of spectrum refarming, the Coalition of Incumbent GSM Operators hired the services of Analysys Mason to prepare a report on the costs of refarming and its effect on consumer tariffs.1 Similarly, the Coalition of Incumbent GSM Operators also hired the services of PwC to prepare a report on the implications of high reserve prices on consumer tariffs.2 These reports translated complex recommendations into tangible and simplistic conclusions about the amount of rupees by which the tariffs would rise in case adverse recommendations by TRAI were accepted. These reports were used by the Coalition of Incumbent GSM Operators to generate media attention and build public opinion on TRAI recommendations. These reports were also used as advocacy material before the executive and cited in future submissions to TRAI. Amid pressure to justify its recommendations, TRAI conducted its own independent study on the implications of these recommendations on tariffs.3 This study showed that the rise in tariff would not be as substantial as reported by those opposed to the recommendations. The Coalition of Incumbent GSM Operators then released a press release rejecting the calculations of TRAI stating that the exercise appears to be a pre-determined answer to a pre-determined conclusion.4 Eventually, two internal shocks helped the Coalition of Incumbent GSM Operators convince the telecommunications regulator to maintain low reserve prices. The first was the failure of the auction in 2012 wherein no bidders expressed interest in the 800 MHz band and substantial spectrum in the 1800 MHz band remained unsold. The second was the failure of the auction in 2013 wherein no bidder expressed interest in the 900 MHz and 1800 MHz bands. The Coalition of Incumbent GSM Operators was able to successfully channelise the second shock to convince the telecommunications regulator to reduce the reserve price. Similarly, the Coalition of Unified Service Operators was able to successfully channelise the first shock to reduce reserve price for the 800 MHz band. The foremost and most significant event captured in the time period under study in this paper is the 2G spectrum scam and the Supreme Court judgement of 2012. This event is also selected as the event for 1 2 3 4

http://timesofindia.indiatimes.com/home/India-The/articleshow/13846802.cms www.pwc.in/press-releases/coai-and-pwc.jhtml http://www.trai.gov.in/WriteReadData/Recommendation/Documents/Final%20reply%20to%20DOT.pdf http://www.sify.com/news/coai-slams-trai-mobile-tariff-calculations-news-default-mhocd4cgfdhsi.html

dividing the two time periods (2008-2011 and 2012-2015) under consideration. In this event, administrative allocations made in 2008 through first-come-first-serve were found to be irregular by the Supreme Court. The court cancelled the allocations made in 2008 and mandated the government to follow auctions as a fair and transparent method for allocation of natural resources. The media attention around the scam pre-dated the judgement amid allegations by the CAG of substantial revenue loss due to the 2G spectrum scam. Due to the nature of the scam, the event was channelised by the Coalition of Incumbent GSM Operators to influence beliefs surrounding market based practices for allocation of spectrum. This event was used to advocate for auctions in the primary market, discontinuation of subscriber linked criteria, delinking of spectrum from licenses etc. In contrast, the Coalition of Unified Service Operators changed their narrative only slightly pursuant to the scam. While conceding that auctions are the best method for allocation of spectrum in the primary market, the coalition remained opposed to other tangible forms of market based practices such as discontinuation of subscriber linked criteria, trading of spectrum in the secondary market, flat spectrum usage charges, delinking of spectrum and licenses etc. The government was criticised by the Supreme Court for disregarding and not implementing previous recommendations of the regulator. As a result, the event also strengthened the role of the telecommunications regulator. The executive went in an overdrive pursuant to media pressure surrounding the spectrum scam. Notably, the government released a new telecommunications policy in 2012 that mostly codified the previous recommendations of the telecommunications regulator. 4.5.

CHANGES IN SECONDARY BELIEFS

As explained earlier, the ACF visualises a tiered structure for beliefs. Policy core beliefs are common across a coalition and are resistant to change. Effectively, these are used for operationalising advocacy coalitions. In comparison, secondary beliefs may vary within a coalition and are comparatively less resistant to change. The analysis in the previous section showed that the policy core beliefs (as captured by the composite scales) were not significantly different between the two time periods. While the composite scales have shown resistance to change, a qualitative analysis of changes in their comprising variables, which represent secondary beliefs, shed light on the following aspects of secondary beliefs: (i) deviations in secondary beliefs over time; (ii) deviations in secondary beliefs between actors belonging to the same coalition; (iii) deviations in secondary beliefs due to selective application. Deviations Over Time To analyse deviations in secondary beliefs over time, we identified cases wherein the actor changed positions on a policy issue by at least 3 points on the coding scale. Additionally, counter-comments were found to be a useful reference wherein actors pointed out deviations in positions of other actors. This helped shed light on how actors take positions on issues and how changes in circumstances and technology affect the positions that they take. Reversals in policy position show that in a technological policy subsystem, it is relatively easy to develop a technical narrative to justify an opposing position. Such reversals also show that the position of actors is generally flexible on secondary beliefs. A few examples are discussed below: • MTS on Spectrum Usage Charges: On the issue of whether spectrum usage charges should be graded or flat, MTS reversed its position twice. MTS had argued in 2009 that spectrum usage charges should be graded. The argument was as follows: “As per the laws of Spectral efficiency: as the amount of deployed spectrum increases, the capacity of a network to carry traffic increases in a greater proportion than the proportion of increase in spectrum. Therefore, as subsequent spectrum is allotted it will provide higher spectral efficiency opportunities for the

•

•

•

5 6 7 8 9 10 11 12 13 14

operators. Accordingly, the rate of increase in the annual spectrum usage charges should be higher as the quantum of spectrum allocation increases with any operator”5. However, 2012, MTS flipped its argument and took the position that spectrum usage charges should be flat. The argument made was as follows “Currently per the system present in India, operators was not paying upfront for the right to use spectrum, hence the graded spectrum charge was logical. Since the right to use spectrum now will be won at market price, using the current methodology may not be logical anymore.”6. In 2013, MTS reversed its position again and argued that spectrum usage charges should continue to be graded. It argued that “As the amount of spectrum holding increases due to increased trunking efficiency, the benefit derived from the spectrum also are higher as with larger chunks of spectrum, there will be larger SUC percentage. Graded system also creates a barrier to an operator hoarding excessively large amount of spectrum that it does not really need... Further the graded SUC provides level playing field to all operators - existing as well as new set of operators... Our submission is as follows:... ii. SUC should escalate (slab-wise) with the amount of spectrum holdings.”7. The arguments are reproduced to show the relative ease with which the narrative was changed to justify the belief. Similarly, MTNL also reversed its position on spectrum usage charges from graded in 20098 to flat in 20139. Idea on Spectrum Trading: On the issue of whether spectrum trading should be allowed, Idea reversed its position twice. In 2009, Idea argued that spectrum trading should be allowed to encourage spectrum consolidation and to improve spectrum utilisation efficiency.10 However, in 2012, Idea reversed its position and argued that spectrum trading should not be permitted.11 Idea argued that the government should instead focus on the MVNO policy, exit policy, sharing policy and merger policy, and only then consider trading. Further along, in 2013, Idea argued that spectrum trading should be allowed as it will ensure optimal allocative efficiency of a limited natural resource as long as spectrum caps are imposed.12 Similarly, Reliance and AUSPI and Uninor changed their positions on spectrum trading. Tata Tele on Auctions for Spectrum Allocation: In 2009, Tata Tele argued that auctions are not the appropriate method for allocation and pricing of spectrum. It argued that “Some of the incumbent operators have suggested that spectrum can be only priced through auction. We do not recommend Auctions as methodology for pricing spectrum, as this will bring disparity between existing players hoarding so much of free spectrum from times when call rates were so high and whereas new players who have actually helped bring prices down including new and unique schemes of Pay Per Second etc would end up paying large upfront fee which also could be tacked to unrealistic prices due to cash availability with existing players.”13 However, after the Supreme Court judgement on the 2G spectrum scam, Tata Tele came out in support of auctions “We support the auction as suggested by the Hon'ble Supreme Court. Auction rules must be fair and give equal opportunity to all.”14 Bharti Airtel and COAI on Subscriber Linked Criteria: A testimony of Reliance is reproduced below from a counter-comment. This testimony alleges that Bharti Airtel and COAI reversed their positions to suit their material interests: “Earlier, incumbents cornered spectrum without any guidelines, without payment of entry fee for spectrum beyond the contracted value but now when that route is closing, they have started pitching for more spectrum through auction,

Pg42 Q45 http://www.trai.gov.in/WriteReadData/ConsultationPaper/Document/201505190550513411980Sistema%20Shyam.pdf Pg31 Q27 http://www.trai.gov.in/WriteReadData/ConsultationPaper/Document/MTS-Auction-spect.pdf Pg10 Q18 http://www.trai.gov.in/WriteReadData/ConsultationPaper/Document/201308140635067531516MTS.pdf Pg10 Q55 http://trai.gov.in/WriteReadData/ConsultationPaper/Document/201505190549266224401MTNL.pdf Pg5 Q18 http://www.trai.gov.in/WriteReadData/ConsultationPaper/Document/201308160616364905866mtnl-2013.pdf Pg9 Q24 http://trai.gov.in/WriteReadData/ConsultationPaper/Document/201505190548575600466Idea.pdf Pf35 Q33 http://www.trai.gov.in/WriteReadData/ConsultationPaper/Document/Idea21-03.pdf Pf25 Q8 http://www.trai.gov.in/WriteReadData/ConsultationPaper/Document/201308150408211067177Idea.pdf Pg4 Point2 http://www.trai.gov.in/WriteReadData/ConsultationPaper/Document/TATAC.pdf Pg1 Point1 http://www.trai.gov.in/WriteReadData/ConsultationPaper/Document/TTSL.pdf

secondary market, M&A etc. Proponents of delinking spectrum and allocation through auction had submitted following views till two years back when the SLC was suiting them, however they have now changed colours and trying to hoard spectrum through M&A and auction route.” Deviations Between Actors of a Coalition Most advocacy coalitions are organised as material associations that have internal rules for developing a coordinated response to consultations. However, we found evidence for deviations in the secondary beliefs of actors within a coalition. To analyse deviations in secondary beliefs within a coalition, we identified cases wherein the difference between the policy positions of constituent actors was at least 3 points on the coding scale for submissions to the same consultation. A few examples are discussed below: • MTS and AUSPI (from Coalition of Unified Service Operators) on Spectrum Trading: In response to a question regarding spectrum trading in 2012, MTS argued that trading should be allowed and offered the following reasoning “Yes, spectrum trading should be allowed in India in line with liberalization of spectrum holdings. Spectrum is a scarce resource and most efficient use of spectrum is in the best national interest. Spectrum trading will allow the natural migration of spectrum to the hand of operators who would be able to generate maximum value from the spectrum. It will also be the most efficient approach that will allow the migration of spectrum to best possible use in the shortest possible time.”15 In contrast, for the same consultation paper, AUSPI argued that spectrum trading should not be allowed. It argued that “The spectrum trading will only encourage spectrum hoarding so that it can be traded at a premium. ... Spectrum is a national asset with Govt having a sovereign right over it. Natural resource is allowed for use for a certain period and should not be allowed to be traded during that period. The TRAI should specify spectrum cap equivalent to the Prescribed limit so that there is no excess spectrum, no hoarding and no possibility of Trading”16. • Aircel and Vodafone (from the Coalition of Incumbent GSM Operators) on Subscriber Linked Criteria (SLC): In comments submitted in 2009, Vodafone argued for the discontinuation of the subscriber linked criteria for allocation of spectrum by asserting that “The SLC criterion is not an optimum approach for determining spectrum allocations and is not used in any other market of which we are aware.”17 In response to the same consultation, Aircel supported SLC for creation of a level playing field “In order to ensure level playing field amongst existing operators, SLC based criteria should be kept in vogue till they arrive at a threshold of 6.2 MHz, beyond which auction mechanism be adopted.”18. • Idea and Vodafone (from the Coalition of Incumbent GSM Operators) on Spectrum Trading: In 2012, Vodafone strongly supported introduction of Spectrum Trading in India “Vodafone fully support the introduction of spectrum trading in India... we believe that ideally, a trading regime which allowed changing both allocation and assignment rights would be most beneficial in terms of promoting an economically efficient outcome.”19. In response to the same consultation, Idea opposed introduction of Spectrum Trading20 and suggested that the government should instead focus on prerequisites such as the MVNO policy, exit policy, sharing policy and merger policy, and only then consider trading. Deviations Due to Selective Application

15 16 17 18 19 20

Pg35 Q32 http://www.trai.gov.in/WriteReadData/ConsultationPaper/Document/MTS-Auction-spect.pdf Pg19 Q32 http://www.trai.gov.in/WriteReadData/ConsultationPaper/Document/AUSPISPECTRUM.pdf Pg26 Q50 http://trai.gov.in/WriteReadData/ConsultationPaper/Document/201505190552539036947Vodafone.pdf Pg22 Q46 http://trai.gov.in/WriteReadData/ConsultationPaper/Document/201505190548258881680Aircel.pdf Pg26 Q32 http://www.trai.gov.in/WriteReadData/ConsultationPaper/Document/Vodafone-26-03.pdf Pf35 Q33 http://www.trai.gov.in/WriteReadData/ConsultationPaper/Document/Idea21-03.pdf

There were numerous instances wherein the actor made a submission in which a position was selectively applied to a subset of possible applications; and the opposing position was applied to the remaining subset. This selective application of a belief to a subset was often a choice of convenience for promotion of self-interest. Examples of such selective applications and the arguments made for the selective application have been reviewed here: • MTS, Reliance and Uninor on Delinking of Spectrum and Licenses: In 2009, these actors argued that spectrum and licenses should be delinked for new licensees; but remain linked for old licensees. Similarly, while they argued that new entrants should be allocated spectrum through auctions, they also argued that existing licensees should continue to be allocated spectrum administratively. These actors argued that existing licensees should be excluded from market based practices else it will worsen the already distorted playing field with incumbent GSM operators, who have already received spectrum through administrative allocation. The narrative shows that these actors were not concerned about the level playing field for new operators. This selective applications of beliefs by creating classes of licensees was clearly an effort towards self-preservation and for promotion of self-interest. This highlights the need to balance beliefs with interests in the framework. • Tata Tele on Refarming for 800 and 900 MHz: The regulator had been considering refarming of the 900 MHz band by providing alternate spectrum to incumbent users in the 1800 MHz band, and refarming of the 800 MHz band by proving alternate spectrum in the 1900 MHz band. Tata Tele strongly supported and advocated for refarming of the 900 MHz band with the objective of creating a level playing field with incumbent GSM operators. However, since Tata Tele had spectrum in the 800 MHz band, it argued that presently refarming of the 800 MHz band is not possible given that alternate spectrum in the 1900 MHz band is not available. Thus, it generated a narrative that supported refarming of the 800 MHz band knowing that refarming of the 800 MHz band is not feasible given the non-availablility of alternate spectrum in the 1900 MHz band. Therefore, this is a case where a belief was selectively applied for one band and not for the other. • MTNL on Allocation to PSUs: MTNL and BSNL have made numerous submissions wherein separate positions are applied to private and public actors. The following extract highlights such selective application: “Each new licensee needs spectrum to rollout its network, accordingly in our opinion the initial spectrum should be linked with license. Later on depending upon subscriber base criteria i.e. after attainment of certain subscriber base, additional spectrum may be allotted on the payment of upfront charges decided based on any of the methods i.e. either by beauty contest method, lottery or auction. The spectrum beyond threshold may be allotted through a combination of beauty contest and auction in steps of 2x1 MHz. However, PSU operators should be allotted spectrum based on the subscriber base criterion on the rates as decided through auction process.” • Telenor on Spectrum Refarming: In a counter-comment, Vodafone pointed out that Telenor selectively applied a position to India that deviated from the position that it applied to other countries. The following extract highlights how beliefs of an actor may deviate across venues and jurisdictions: “We also caution the Authority against placing emphasis on views submitted to this consultation which are very different from the views submitted by the same company in other jurisdictions. Specifically, we are concerned that Unitech (which is majority owned by Telenor) has submitted views in this consultation which are the opposite of those it submitted on very similar issues in Sweden in February 2009 with respect to re-farming of 900MHz spectrum. Telenor recommended to the Swedish regulator that all existing 900MHz operators be reallocated their 900 MHz spectrum. In contrast, presumably because it suits Unitech’s commercial interests in India, the opposite has been proposed to the Authority – that 900MHz licenses should not be renewed or it would put new entrants at an 'undue disadvantage'. We

note that Telenor in Sweden requested (and received) a renewal of its 900 MHz licence without any competitive process or redistribution.”21 5.

CONCLUSION

5.1.

INFERENCES AND APPLICATIONS OF ANALYSIS

This paper reviews the public policy process in the spectrum management policy subsystem in India from the lens of the Advocacy Coalition Framework (ACF). The ACF is a theoretical lens that is used to bring structure to the inherent complexity of the public policy process. The ACF considers advocacy coalitions, operationalised using common “beliefs”, as the appropriate unit to deal with the multiplicity of actors in the policy subsystem. In this paper, we performed a content analysis of 144 testimonies submitted by various actors in response to public consultations by the telecommunications regulator. A coding frame was used to identify the stated beliefs of elite actors on contentious policy issues. A cluster analysis was performed to operationalise and identify the advocacy coalitions operating in the subsystem in two time periods: from 2008 to 2011 and from 2012 to 2015. The resulting analysis demonstrated that actors were divided into opposing camps on contentious policy issues. We found strong evidence for the existence of two advocacy coalitions in 2008-2011 and three advocacy coalitions in 2012-2015. The policy core beliefs of actors were found to be resistant to change over time. The evolution of advocacy coalitions was attributed to the change in underlying technology. Different advocacy coalitions dominated on different policy issues - no single advocacy coalition was found to dominate the entire policy subsystem. The ability of competing coalitions to use various instruments and resources to dominate and convert their beliefs into policy outputs was reviewed. The paper provides significant insight for policy advocates by bringing structure to the policy process and by providing insights on how to benefit from collective action. 5.2.

A CRITIQUE OF THE ADVOCACY COALITION FRAMEWORK

The Advocacy Coalition Framework helped bring structure to the inherent complexity of the policy process through multiple means. First, it helped define the scope of the policy subsystem and the actors participating in the process. Second, it helped focus the unit of analysis on the collective behaviour of actors and the beliefs uniting these actors. Third, it helped understand how these actors collectively try to influence the policy process to convert their common beliefs into policy outputs. Fourth, it helped understand how coalitions may use different resources and instruments to capitalise on events and shocks to influence the policy process. Finally, the framework helps define constructs in a structured and definite manner, thus allowing extrapolation of insights from one subsystem to another. However, the framework has many gaps that need further study. The framework is yet to evolve into a coherent framework that can provide structure to the specific needs and attributes of a technical subsystem in a developing country context. We identify a few of these problems and suggest ways to improve the framework to overcome these problems. The primary weakness of the framework and its methodological applications appears to be that it takes a simplistic approach to the inherent complexity of a policy issue by visualising that opposing beliefs exist as a dichotomy. The framework needs to instead recognise that beliefs actually exist as a 21

Pg13 http://www.trai.gov.in/WriteReadData/ConsultationPaper/Document/VodafoneCC.pdf

polychotomy; and that dichotomous beliefs are only a special case. By reducing beliefs to a dichotomy, the framework ignores the complex technical and legal narrative that surrounds the issue and the multiplicity of variables that comprise a belief. Beliefs are often a function of time, legacy, technology, audience and circumstances, and definitely far more complicated than what the authors of the framework have set it out to be. For example, on a question about whether spectrum should be auctioned or administratively allocated, an actor stated that it supports auctions for new licensees providing liberalised services but supports administrative processes for allocation of additional spectrum to existing licensees; further auctions should only be conducted if reserve prices are set low, spectrum blocks are contiguous, and secondary markets are allowed. Similarly, other actors also supported or opposed spectrum auctions while attaching new conditions and creating permutations of conditions stated by others. To reduce the complex variables that comprise this technical narrative into a black box wherein one simply opposes or supports auctions would indeed ignore the deeper issues that play a part in the policy process. Further, the framework appears to greatly downplay the role of short term interests in influencing long term policy. Beliefs, in a subsystem dominated by private players, are often characterised by the organisationsal objectives, or the interests of the organisation, and the personal ambitions of the people within that organisation. The narrative surrounding stated beliefs was found to be conveniently flavoured and polished to suit the short-term interests of these organisations. For example, while Tata Tele supported refarming of the 900 MHz band to hurt competing incumbent GSM operators, it also supported refarming of the 800 MHz band (the band in which itself operates) knowing that refarming of the 800 MHz band is not feasible given the non-availablility of alternate spectrum in the 1900 MHz band. This was a clear moulding of stated beliefs to suit the short-term interests for creating a level playing field. Similarly, there were examples wherein existing licenses supported delinking for new licensees and opposed delinking for existing licensees. These examples highlight how stated beliefs of actors are often flavoured by short term interests, and how the technical narrative developed to support a belief can easily be moulded to support such interests. Additionally, it was found that relevance of issues is short-lived when technology evolves at a fast pace, and thus beliefs, as a function of these technologies, also evolve at an extremely fast pace. In a highly technical policy subsystem, such as the present one under study, change in underlying technology may result in previously contentious issues becoming irrelevant and some other previously non-contentious issues suddenly becoming divisive. The framework needs to thus recognise the possibility of fast evolving nature of beliefs. Operationally, the framework needs to recognise the need for using different beliefs for operationalising advocacy coalitions across different time periods. As a result, the advocacy coalitions themselves can not be expected to remain stable and should be expected to continuously evolve with technology. The assumption that beliefs persist creates the assumption that coalitions persist and remain static. Focusing the unit of analysis on such static coalitions, operationalised using multiple policy core beliefs, undermines single-issue based dynamic coalitions that emerge and dissolve with that issue. For example, important actors may have beliefs only on a single issue in the subsystem, such as Ministry of Finance on revenue maximisation and Ministry of Defence on captive spectrum. These actors participate in subsystem on specific issues and remain silent on most other issues. However, they play a substantial role in influencing the policy process while remaining outside of any time-persistant static coalition. Additionally, it is assumed in the framework that actors can be a part of just one coalition at one time. This static and compartmentalised nature of advocacy coalitions does not reflect ground realities wherein actors take issue based stands and may not owe allegiance to any coalition in the form of collective action.

Another critique exists around the use of policy core beliefs for operationalising and identifying advocacy coalitions. There are two dimensions for measuring a belief: (i) the position that an actor supports; and (ii) the relevance of that belief to that actor. While it may feasible to group actors on the basis of their positions, the beliefs itself may be of varying relevance for different actors within a coalition, and thus the intensity of the same belief may vary for different actors. As a result, some actors within a coalition may be willing to compromise with a belief while others may not be willing to do so. These intra-coalition dynamics due to varying relevance of a belief need further exploration. The framework also needs to focus on legacy and legal issues that play an important role in policymaking but are not captured in the beliefs of elite actors. It is difficult to capture the beliefs of actors who “may” possibly file a legal case thus thwarting policy change, but are not otherwise active participants in the subsystem. For example, this may take shape in the form of failed bidders opposing changes in license terms and conditions for successful bidders.

6.

REFERENCES

Adam, S., & Kriesi, H. (2007). The network approach. Theories of the Policy Process, 2, 129–54. Baumgartner, F. R., & Jones, B. D. (2010). Agendas and Instability in American Politics, Second Edition. University of Chicago Press. Cairney, P. (1997). Advocacy coalitions and policy change. Contemporary Political Studies, 1, 884–894. Hann, A. (1995). Sharpening up Sabatier: belief systems and public policy. Politics, 15(1), 19–26. Kingdon, J. W. (2003). Agendas, Alternatives and Public Policies (Second). Longman. Ladi, S. (2005). Globalisation, Policy Transfer and Policy Research Institutes. Edward Elgar Publishing. Lasswell, H. D. (1956). The decision process: seven categories of functional analysis. Bureau of Governmental Research, College of Business and Public Administration, University of Maryland. Sabatier, P. A. (1988). An Advocacy Coalition Framework of Policy Change and the Role of Policy-Oriented Learning Therein. Policy Sciences, 21(2/3), 129–168. Smith, A. (2000). Policy networks and advocacy coalitions: explaining policy change and stability in UK industrial pollution policy? Environment and Planning C, 18(1), 95–114. Weible, C. M., & Sabatier, P. A. (2007). Chapter 9 - A Guide to the Advocacy Coalition Framework. Handbook of Public Policy Analysis, 123. Weible, C. M., Sabatier, P. A., & McQueen, K. (2009). Themes and Variations: Taking Stock of the Advocacy Coalition Framework. Policy Studies Journal, 37(1), 121–140. http://doi.org/10.1111/j.1541-0072.2008.00299.x