University of Birmingham ‘Does it work?’ – work for whom? Britain and political conditionality since the Cold War

Summary. — Evaluations of the political conditionality (PC) phenomenon have long focused on the question of instrumental eﬃcacy – whether PC promotes policy reform in developing states. Evidence from the UK nevertheless suggests that this emphasis is misplaced and that donor oﬃcials increasingly use PC for ‘expressive’ reasons – to signal their putative commitment to delivering ‘value for money’ in a diﬃcult international economic climate. This shift in rationale raises important questions; not least, what do we know about the eﬀects of PC on public perceptions of aid and to what extent, within this dispensation, can contemporary PC be viewed as a ‘success’? (cid:2) 2014TheAuthor.PublishedbyElsevierLtd.This is an open access article under the CC BY license (http://creativecommons.org/licenses/ by/4.0/).


INTRODUCTION
Analyses of 'political conditionality' since the early 1990s have focused heavily on identifying trends in its application and answering one central question -'does it work'? A less common, but no less important, question, however, is 'do policy-makers expect it to work?' and, indeed, why do they use the instrument? The apparent return of political conditionality (PC) since the mid-2000s -particularly linked to the disbursement of General Budget Support (GBS) -provides a suitable opportunity to re-focus discussion of this phenomenon in a manner that addresses this analytical and empirical gap.
This article will therefore attempt to explore donor motivations for imposing PC. In doing so, two major categories of motivations for applying PC will be delineated. Developed from conceptual debates in the literature on economic sanctions these are: instrumental (where PC is applied to force aid recipients to implement political reforms) and expressive (where PC is applied to signal disapproval of the recipient's actions -for domestic or international audiences -without the expectation that actual reforms will follow). Establishing whether a particular PC decision is based primarily on instrumental or expressive rationales has important implications for assessing the effectiveness of the instrument; if donors do not intend for a particular suspension to result in political change, should they be criticized when it does not? Moreover, how should scholars and practitioners view and approach the instrument within this dispensation?
In investigating the key rationales for PC decisions the policies and perspectives of one major donor -the United Kingdom (UK) -will be reviewed and critiqued. Though single-case-based analyses rarely provide a solid foundation for developing comprehensive general theories on issues, they nevertheless allow for a deep and nuanced consideration of a phenomenon -and how it has changed over time. Since this article aims at filling both a theoretical and empirical gap in the PC literature a single-case study therefore represents a more appealing option. Many findings associated with the British case also, naturally, have parallels elsewhere in the international aid system and some of these will be highlighted in the conclusion.
A further set of reasons for focusing on the UK relate to Britain's particularly prominent role in promoting and employing PC in the past (particularly between c.1991-95) and in its revival since c.2004. Britain's Department for International Development (DFID) has also -more than many other donors -consciously sought to explain and rationalize its usage of PC since 2005 with a view to influencing how other Western donors employ it, sometimes explicitly (Benn, 2005). This has been a reasonable ambition given the widespread acceptance among many in the Western development community today that DFID represents a 'thought leader' second only to the World Bank. 1 The UK's position as one of the leading aid donors -particularly in Sub-Saharan Africaand providers of GBS also provide it with a degree of influence in international development fora and within 'in-country' cross-donor groupings that few other donors possess.
The article ultimately argues that while expressive motivations have always played a role in the UK's PC decisions, they have become increasingly central since the mid-2000s. Likewise, where the expressive dimensions of PC impositions in the 1990s were mainly aimed at the international level, those of today are focused far more clearly on domestic, British audiences. Finally, while instrumental rationales were highly influential in the PC decisions of the 1990s they have become largely irrelevant in the contemporary era. DFID officials do not believe that PC can force political change on aid recipients but continue to use it as a means of communicating with an increasingly aid-skeptical domestic audience. The final part of the article will suggest a range of factors which have led to this state of affairs and explore the implications of these findings for scholars and practitioners.
This analysis draws upon a range of semi-structured interviews carried out with current and former DFID staff -and former staff of DFID's predecessor, the Overseas Development Administration (ODA) -during 2007-13. Interviewees included current and former staff at varying levels based in both London/East Kilbride (henceforth 'HQ') 2 and regional and country offices. Internal DFID and ODA documents released to the author under the 2000 UK Freedom of Information Act (FOI) are also analyzed together with public statements, ODA/DFID documents, and media reporting. The author also held an Honorary Research Fellowship in the Foreign and Commonwealth Office (FCO) during 2013-14. While no information or data collected during this Fellowship has been directly cited in this article the experience has nevertheless informed its analysis.
The article is structured as follows: Section 2 reviews and critiques the existing literature on PC, highlighting in particular its reluctance to explore the intentions and perspectives of donor officials involved in imposing the modality. The literature on the purpose of economic sanctions, however, does speak to this debate -albeit indirectly -and is therefore engaged with as a means to develop a set of categories for conceptualizing donor PC rationales. Section 3 introduces the British case and analyzes the UK's employment of PC since the Cold War highlighting a significant shift in UK thinking on the modality's efficacy from instrumental to expressive. Section 4 delineates a number of reasons for this shift in the last decade linked to the domestic and international political economy of aid management in the contemporary era before returning to the question posed in the title of this piece -does [expressive PC] "work"? The article concludes by exploring the implications of these findings for development scholars and policy-makers more broadly.

POLITICAL CONDITIONALITY UNDER SCRUTINY
The term 'political conditionality' (PC) has been defined and employed in several different ways by scholars and policymakers since its coining in the early 1990s. In essence, it refers to the placing by donors of 'political' conditions upon their aid disbursements with many scholars dividing this behavior into 'positive' conditionality (where aid is released once conditions have been fulfilled) and 'negative' or 'restrictive' conditionality or 'aid sanctions' (where aid is suspended when conditions have not been, or are no longer being, fulfilled) (Crawford, 1997, pp. 69-70;Stokke, 1995, pp. 11-13;Waller, 1995, pp. 401-408). Much of the PC literature has, however, tended to focus upon the latter -as this study will also do. In so doing, the emphasis will be on exploring the reasons why 'negative' PC (subsequently simply PC) has been imposed rather than why it has also, sometimes, not.
Within this literature, however, there have been few attempts to clearly outline what makes a condition 'political' as opposed to purely 'economic' (a version of conditionality contrasted with PC by many commentators). While aid suspensions relating to democratization and civil and political liberties have invariably been included in PC analyses (Brown, 2005;Crawford, 1997Crawford, , 2000Stokke, 1995), those linked to high-level corruption and economic mismanagement have been less systematically incorporated. This division is somewhat artificial, however, since donors have frequently suspended aid based on 'governance concerns' which -when unpacked -clearly refer to interlinked democratization and corruption-related issues (De Felice, 2013). This article will therefore understand 'PC' to refer to aid suspensions linked to governance-related matters, broadly defined to include corruption as well as democratic backsliding, human rights abuses and fomenting regional insecurity. 3 Though isolated instances of PC being applied in the Cold War era can be found (Bratton and van de Walle, 1997, pp. 27-30), the instrument has become a common feature of the aid landscape only since c.1990-91. Never entirely abandoned, its usage decreased significantly by the mid-1990s only to be revived -linked particularly to disbursements of GBS -in the mid-2000s (Hayman, 2011;Molenaers, 2012;Molenaers, Cepinskas, & Jacobs, 2010). Scholarly interest in PC has tended to follow the interest of donors themselves with a substantial array of studies produced during c.1990-97 andagain, linked to GBS -since c.2008. Though much of the latter body of literature has, of course, built upon the former, the two 'generations' can be distinguished to some extent. Early 1990s scholarship, for example, focused more on normative debates surrounding PC ('should it be used?') than that of today (Adam, Chambas, Guillaumont, Guillaumont Jeanneney, & Gunning, 2004, pp. 1059-1060Barya, 1993). Likewise, contemporary PC scholarship has generally focused more on European donor behavior than that of other aid providers, particularly the United States owing to the absence of GBS in Washington's range of aid modalities (Hayman, 2011;Molenaers, 2012;Molenaers et al., 2010).
The two generations of PC commentary, however, share at least one common feature. Both focus substantially on questions of PC's ultimate instrumentality and effectiveness; 'does it work? That is, does the threat of, or actual withholding [of aid] sway recipient governments into a different course of action?' (Hayman, 2011, p. 683; see also Crawford, 1997Crawford, , 2000Levitsky and Way, 2006;Molenaers et al., 2010;Moore and Robinson, 1994;Smith, 1998;Stokke, 1995;Uvin, 1993). Exploring whether PC 'works', however, fails to investigate the extent to which donors themselves actually apply it in the hope -or belief -that it will work, whatever 'work' is understood to mean. Likewise, interpreting donor motivations for using PC only in terms of a rational state actor weighing-up its balance of interests overlooks other more arbitrary, short-term, or mundane influences on the application of PC including organizational factors, personalities of policymakers, and domestic political pressures (Lancaster, 2007;Lindblom, 1959;Natsios, 2010).
The purpose of this article is not to reject these approaches to understanding PC but, rather to offer an alternative perspective -one which focuses on exploring and comprehending the motivations of officials within donor agencies for actually applying the instrument (Molenaers, 2012). For while development practitioners often privately acknowledge that this thinking has changed in recent decades, this has not yet been explored or established in scholarly literature. An exception to this generalization can be found in the work of Collier et al. where five donor objectives for imposing conditionality are delineated. These span from incentivizing recipients to undertake reforms to 'signaling' to 'private agents' that the donor perceives a recipient country's policy and economic environment to be safe or unsafe for private investment (Collier, Guillaumont, Guillaumont, & Gunning, 1997, pp. 1400-1401. This conceptual exposition is useful in its highlighting of the broader functions of conditionality. Its value for this study is somewhat limited, however, by its framing of conditionality decisions rather narrowly in terms of principal-agent theory; influences, concerns, and drivers outside the donor-recipient relationship are largely not considered in this setup. Salient questions on motivations have, nevertheless, been addressed within the literature on economic sanctions. As long ago as 1967, Galtung argued -in a study of economic sanctions applied to Rhodesia -that sanctions can have 'expressive functions' as well as 'instrumental' ones and that 'there is the value [on the part of an interested state] of at least doing something. . .of being busy in time of crisis' -that is, in imposing sanctions to make a political statement and not just to force reform (Galtung, 1967, p. 411). Several scholars have since sought to develop this instrumental/expressive distinction at the conceptual level. Barber, for example, argues that there are three levels of 'objectives of sanctions': primary (aimed at prompting change in the 'target state'), secondary (aimed at persuading domestic and international audiences that the sanctioning state is willing to act or deflecting criticism of inaction), and tertiary (aimed at promoting a particular international norm; Barber, 1979). Likewise, Lindsay delineates five 'aims' of imposing trade sanctions -compliance (by the target with a stated policy), subversion (forcing the overthrow of the target government), deterrence (discouraging the target from repeating an action), international symbolism (similar to Barber's tertiary objective), and domestic symbolism (similar to Barber's secondary objective; Lindsay, 1986).
These scholars all, of course, acknowledge that no decision to impose sanctions is taken with only one objective or aim in mind; the same is true of a PC or any other policy decision. Their distinction between sets of motivations -particularly that between instrumental (Barber's primary objective and Lindsay's compliance, subversion, and possibly deterrence) and expressive (Barber's secondary and tertiary objectives and Lindsay's international and domestic symbolism) -can nevertheless be helpfully applied to analyses of PC decisions, which might be considered 'aid sanctions'. The intellectual and practical merits of focusing on this question of purpose are also articulated clearly in the economic sanctions literature. As Portela notes, acknowledging the 'multiplicity of goals of sanctions' also means acknowledging 'that different goals should be measured separately' (Portela, 2010, p. 11). Understanding why donors impose PC is not just useful for comprehending the phenomenon itself, therefore, it is also important for establishing a meaningful set of indicators of the instrument's effectiveness.
In applying this framework to PC motivations it is useful to consider the extent to which Lindsay's variety of aims actually applies in reality. Thus, instrumental-based aid cuts are usually premised on compliance and deterrence but rarely, at least explicitly, on subversion. The case of Malawi in 1992, however, reveals that PC impositions can -in partlead to subversion even if this was not part of the original rationale for the aid suspension. The imposition of PC by most donors in that year, for example, forced the aid-dependent Hastings Banda government to hold a referendum on abolishing Malawi's one-party state -a poll it lost heralding multi-party elections in 1994 which saw the autocratic Banda lose office (Brown, 2004). Likewise, expressive-based cuts can rarely be said to have only domestic symbolism in mind and many such impositions clearly seek to make a point to domestic and international audiences simultaneously.
Moreover, contextual factors must also be taken into account to gain a fuller understanding of the background to individual PC impositions. This includes the state of relations between the donor and recipient, the latter's perceived 'value' to the former in a range of areas (most notably security and trade) as well as social and cultural ties between the two. It also, however, includes a range of influences on the donor side including the role of individual personalities and office-holders in the decision-making process, the relative influence of government departments -and other donors -on aid policy and the perceived profile of a particular aid 'crisis' in the domestic sphere in the minds of these actors.
The purpose of this article, nonetheless, is to focus on one dimension of this context: whether or not the UK's motivations when applying PC since the Cold War, have been based primarily on instrumental rationales or on expressive ones. This is not to say that decisions have not been premised on a large range of rationales but rather that this study is interested in identifying the purpose through this particular lens. To avoid stark and unrealistic dichotomies between the two in this regard, it is perhaps best to see each PC decision as being plotted somewhere on a spectrum between the two extremes. At one end, the key focus is on the recipient state while at the other the focus is on domestic audiences and managing their perceptions of the donor. Such a spectrum might look like this (see Figure 1).
Ascertaining the motivation behind PC decisions is, nonetheless, a methodologically challenging task. In part this is because of the scarcity of quality and comparative data on decisions -particularly those in the more distant past. More generally, the separation of DFID from the FCO in 1997 and the gradual devolution of authority to regional and country offices since this time render it even more complicated to pinpoint the locus of decision-making in certain cases. While senior ministers have largely retained the lead voice over PC decisions since the Cold War (see below), the relative influence of other Whitehall departments and 'in-country' staff in this regard has fluctuated. Likewise, while DFID HQ has published clear guidelines on PC since the mid-2000s these documents cannot necessarily be read as the de facto perspectives of all DFID staff.
In addressing these challenges, the strategy taken in this research has been to focus first on gathering data through interviews. This material has then been triangulated with (in order of preference) internal ODA, FCO, and DFID documents released to the author under the 2000 FOI Act, public statements by ODA and DFID officials and publicly available ODA and DFID documentation. In recognition of the above challenges, interviewees have been sought from both headquarters (including those in ministerial offices and a number of directorates and divisions) and regional/country office level as well as from counterparts in the FCO at equivalent levels for the entire period under study. The same is true of internal ODA, FCO, and DFID material accessed through FOI requests.

Instrumental Expressive
Subversion Compliance Deterrence International symbolism Domestic symbolism This triangulation of data allows for the drawing of general conclusions on the changing motivations behind UK PC decisions since the Cold War. Greater detail is available, though, on a number of cases throughout this period and these will be explored in more depth in the subsequent analysis. These include decisions on Kenya (1991), Malawi (1992), Uganda (2006), Malawi (2011), and Rwanda (2013. Background information on interviewees is provided where possible -though it is important to note that interviews were carried out anonymously and often candid perspectives were expressed on the understanding that identities would not be disclosed; the location of interviews is also not disclosed as this may compromise anonymity, particularly in the case of CO staff.
It is first necessary, however, to introduce the UK case and clarify how the management of PC by London has changed institutionally since the end of the Cold War. It is also important to delineate the 'universe' of UK PC cases during this period.

THE UK'S USE OF POLITICAL CONDITIONALITY SINCE THE COLD WAR: CHANGES AND CONTINUITIES (a) From ODA to DFID: effects on PC policy
During the 1979-97 Conservative government, Britain's aid program was managed by the ODA -a semi-autonomous organization within the FCO. ODA had its own minister of state (Lynda -later Baroness -Chalker during 1989-97) although was represented in Cabinet by this minister's immediate superior, the Foreign Secretary. During this period, British aid policy was frequently designed to support and defend UK foreign policy interests, particularly those linked to trade (Killick, 2005, pp. 68-83). Following the election of a Labor administration in May 1997, however, ODA was separated from the FCO becoming an independent department of state -DFID -with its own secretary of state (Clare Short during 1997-2003) attending as a full member of Cabinet.
DFID rapidly established itself, strongly supported by Short's forceful dynamism, as the antithesis of its predecessor (Vereker, 2002, pp. 136-138). 4 The Department defined its sole objective as being the 'elimination of global poverty' with the linking of British aid programs to other objectives not only discouraged but effectively criminalized in the 2002 International Development Act. 5 The reluctance of many DFID officials from Short downward to maintain a comprehensive dialog with FCO colleagues over DFID policy in developing states led to longstanding tension between the two departments, not least because DFID's large budget (consistently and significantly higher than the FCO's since 1997) made it the de facto lead representative of the British government in many capitals (Porteous, 2005, pp. 281-284). 6 Though the DFID-FCO relationship has improved considerably in the last decade, DFID officials continue to guard their budgets against FCO attempts to influence spending decisions.
As well as seeking to change the intellectual culture of the Department since 1997 through the hiring of staff with closer links to development (see below), DFID managers have also attempted to decentralize development policy-making since 1997. 7 Thus ODA's 'regional office' model was phased-out in the early 2000s and replaced with the establishment of DFID Country Offices (COs) in almost every state receiving UK aid. Staff in COs are the key decision-makers on how aid money is allocated in 'their' countries and this has not changed since the election of a conservative-leaning coalition government in the UK in May 2010. 8 Indeed, the Conservative-Liberal Democrat Coalition has not only protected DFID's independence -to the chagrin of some Conservative Party grandees -but has also 'ring-fenced' its budget in spite of substantial cuts to government spending linked to the Global Financial Crisis of 2007-08 (Guardian, October 20, 2010).
An important complication within this decentralization of authority, however, concerns the imposition of PC, a decision which has remained in the hands of DFID's secretary of state and his or her senior staff (composed of senior civil servants, private office staff, and politically-appointed 'Special Advisers' (de Renzio et al., 2005, p. 30)). 9 Indeed, the dominance of PC decision-making by ministers represents an important consistency in the UK's employment of the instrument since 1991 -although prior to 1997 it was the foreign secretary in consultation with their junior minister, who was the primary actor in this field. Since the later 1990s, however, country offices (COs) have become increasingly significant in the earlier part of this process with assessments of crises and deteriorations in the aid relationship being ordinarily identified at this level before being referred to HQ for consideration and possible action. DFID officials close to the latter end of this process nevertheless acknowledge that sometimes the impetus for such discussions has come from the secretary of state. 10 Indeed, while PC decisions in the UK have long been taken by ministers the progression from identification of a 'problem' through to aid cut has rarely followed a linear pattern from CO to minister's office. Instead, PC decisions are best seen as evolving from a set of discussions within and between three core groups, all with slightly different interests and perspectives: 11 (a) The CO: the DFID team closest to the crisis and most familiar with the context. Often events interpreted as 'crises' in London are viewed as part of a more gradual deterioration in 'direction of travel' by these officials and this, together with their often close day-to-day working relationships with partner government officials, can lead COs to discourage London from 'pressing the nuclear button' and suspending an entire aid portfolio over one particular transgression. (b) Regional Directors (and relevant department heads): UK-based line managers of CO teams across particular regions or specialities who must collate information on particular crises and put together lists of 'options' for ministers to take which balance DFID's longer term programmatic and strategic objectives with short-term political and reputational concerns. (c) Ministers and their offices: the final decision-making team which includes those personally loyal to the minister and mindful of his/her future career prospects as well as senior civil servants. As well as considering their Departments' advice, ministers must also often take into account counsel from other ministries (most notably the FCO, Ministry of Defence and Department of Business or Trade, their party's political interests and, indeed, their own personal political ambitions. The positions and decisions of other donors may also enter into consideration at each of these levels to varying degrees.
This article will continue to refer to PC decisions of the 'UK', therefore, while acknowledging that the locus of decision-making on British aid suspensions has not been wholly static since the end of the Cold War. Distinctions between these groups, where relevant, will nevertheless be highlighted where necessary and will be emphasized particularly in the final section of this piece where the effectiveness of expressive PC will be considered.

(b) The UK's use of PC since 1987: Universe of cases
Though the UK had imposed what might be called PC in a number of isolated cases during the height of the Cold War, those applied to Fiji in 1987, Burma in 1988, and -particularly -Kenya in 1991 represent the first examples of a more systematic and enduring policy shift. Since 1987, the UK has suspended, delayed, or re-directed aid away from recipient governments based on governance concerns in around 31 instances (see below) with the withholding of £9 million of GBS to Malawi linked to a high-level corruption scandal in November 2013 being the most recent example at the time of writing (Daily Telegraph, November 18, 2013). These c.31 instances represent the universe of cases under scrutiny in this section.
Threats to apply PC have also been made on a number of occasions but not followed through for a variety of reasons (Crawford, 1997, pp. 75-77;de Renzio et al., 2005, p. 12). There have also been a number of 'non-cases' of PC -where aid or GBS continued in spite of political transgressionsthough a delineation of such cases ultimately depends on a somewhat normative judgment on what ODA or DFID 'should' have done, particularly as some of those cases identified by scholars were the subject of debate within DFID/ODA while others were not (Brown, 2005;. Though exploring these instances of 'non-PC' would provide a valuable insight into the political economy of PC decisionmaking this article is ultimately concerned with understanding the thinking behind decisions when they have been taken as opposed to when they have not. There is also already a rich literature which engages with the issue of PC non-cases (Brown, 2005;Crawford, 1997Crawford, , 2000Cumming, 2001;Fisher, 2014;Porteous, 2005).
It is not possible to be wholly certain that these 31 cases identified below represent the entire universe of casesalthough they include all examples cited in existing secondary literature (including systematic comparisons produced in 1997 and 2005), in media sources and ODA/DFID statements consulted and in interviews undertaken for this article. This lack of certainty is the result of the sometimes secretive or arbitrary manner in which PC decisions have been communicated -at one end of the spectrum, formal, public announcements of aid suspensions have not always been made while at the other end, some decisions (such as on Uganda in 2006) have been communicated virtually 'through the press' (FCO, 2011). Neither ODA nor DFID themselves have kept lists or records of instances of PC being applied before 2005 and DFID claims that it does not retain comprehensive records of its own 'non-essential' activities earlier than the mid-2000s. 12 The Department's lack of institutional memory in this regard has meant that it has sometimes even needed to commission external consultants to inform it of its own past practice on PC (see Mokoro, 2005).
The tables below, which delineate instances of UK aid suspensions, delays, or diversions linked to concerns on democratization, corruption, human rights abuses, and security under ODA (Table 1) and DFID (Table 2) -and the stated rationales for these decisions (where available) -since 1987, represent as complete a list as possible. It should be noted, however, that while Table 1 captures PC imposed on a range of aid modalities (albeit mostly programmatic in nature), Table 2 focuses only on PC applied to GBS disbursements -reflecting DFID (and other European GBS donors)'s almost total restriction of PC decisions to this modality in the contemporary era (Hayman, 2011). A comparison of the two tables highlights a number of basic trends. Firstly, while UK PC applied during the 1990s tended to focus primarily on instances of democratic backsliding or political space (10 of 12 -83%), PC in more recent years has been directed more toward corruption concerns (11 of 19 -58%). Secondly, while some states have been consistently targeted by UK PC since 1987 -notably Malawi -others have received greater attention in either the 1990s (e.g., Kenya, Zambia and Tanzania) or the 2000s (e.g., Uganda and Rwanda); all, however, have been African since 1992. This is perhaps not surprising given both ODA and DFID's very heavy focus on Africa in their programs.
Furthermore, while there have been some chronological gaps in the imposition of PC -particularly during 1996-2000 and during 2006-10 -the instrument's popularity does not fall quite as clearly into two separate 'generations ' (1990s and post-2005) as some commentators, notably Hayman, have argued (Hayman, 2011, pp. 678-681). The linking of GBS to one specific modality -GBS -does, as Hayman rightly points out, nevertheless distinguish recent PC impositions from their predecessors which focused on a range of programmatic instruments.
Several additional trends are also worth noting at this stage. Where PC impositions under ODA often involved the wholesale suspension of 'all non-humanitarian aid' (as, for example, those during 1991-92) those under DFID have more often entailed only partial suspensions of GBS portfolios with these funds being diverted to other modalities still within the country itself. Moreover, while formal donor coordination architectures have grown exponentially in most developing states in recent decades the UK has seldom sought to use these structures to coordinate PC impositions with other aid partners. This stands in contrast to UK approaches in the early 1990s especially where donor Consultative Group fora became the de facto setting for announcing UK PC decisions in tandem with other donors. 13 It is crucial, however, to interrogate and unpack these -and other -trends identified in greater depth in order to better understand the UK's evolving employment of PC. The following section will therefore analyze how far instrumental and expressive rationales have lain behind PC decisions by London across these 31 instances.
(c) Changes in UK PC policy rationales since the Cold War: from instrumental to expressive The UK's use of PC has always been motivated in part by expressive objectives -particularly a desire to be seen to be 'acting' in the face of high-profile governance abuses by aid recipient governments. For while some applications of the instrument -such as Kenya 1991, Malawi 1992, Ethiopia 2005 Uganda 2006 -have been linked to longer term concerns among donors over democratic backsliding they were nonetheless triggered by immediate, controversial actions by the regimes in question: the arrest and harassment of opposition leaders in Kenya and Malawi, the shooting of protestors in Ethiopia and the arrest of an opposition leader in Uganda (Aalen and Tronvoll, 2008, p. 117;Brown, 2007, pp. 307-313;Cammack, 1998, pp. 183-205;Fisher, 2012, pp.421-422).
ODA and DFID officials involved in these PC decisions discuss these events in terms of their being 'the straw that broke the camel's back', the stepping over of a 'red line' by the recipient which made PC necessary and justifiable. 14 One DFID official has emphasized the degree to which PC cuts are ultimately based on 'judgment calls' at local and HQ level (see above). 15 It is notable, in this regard, that in DFID's most recent staff guidance on conditionality the notion of a 'breach' in relations leading to an aid suspension has been replaced by the much less binary concept of a 'spectrum' of options to allow DFID staff flexibility in different circumstances (DFID, 2014, pp. 3-4).
In almost every case of PC, these 'trigger' or -to use Adam and Gunning's terminology -'show-stopper' events have simply been a symptom of a much longer term problem in the states concerned: Kenya and Malawi had been one-party states long before the 1990s, Nigeria a military dictatorship before 1993, Uganda and Ethiopia semi-authoritarian states (which, in Uganda's case, outlawed opposition parties during 1986-2005) before 2006, Rwanda a military 'spoiler' in Congo before 2004-12 and Uganda a home to major regime-sanctioned highlevel corruption long before 2010 (Adam and Gunning, 2002, pp. 2052-2054. The PC trigger in each case, therefore, has rarely been a sudden decline in governance standards (with the possible exceptions of the PC cases linked to military coups during 1991-94) but instead a high profile event which shines a light on longstanding problems in a state. Indeed, almost all of the instances of PC being applied by the UK since the Cold War occurred against the backdrop of widespread editorials and media campaigns in the British press highlighting and publicizing the trigger events concerned. 16 ODA, for example, initially opposed the idea of applying PC to Kenya with both Foreign Secretary Douglas Hurd and Overseas Development Minister Lynda Chalker publicly ruling out an aid suspension linked to poor governance in Kenya as late as September 1991 (Weekly Review, September 20, 1991). Following the arrest of Kenyan opposition leaders in November 1991 -an event covered in depth by several major UK newspapers -the two ministers reversed this position and both became leading players in persuading other donors to impose PC upon Kenya (Brown, 2007, p. 310;Weekly Review, November 22 and December 6, 1991). 17 Moreover, the decision to suspend GBS to Ethiopia after a 2005 postelection government crackdown followed a lively debate within DFID partly prompted, as a senior official from the time has noted, by the high profile of the crisis in the UK media and NGO community. 18 Other major PC applications have come in the midst of international media storms including Tanzania  It is also clear, nonetheless, that ODA officials in the 1990sand their political masters -viewed PC as a credible and powerful means to force political reform upon intransigent, semiauthoritarian regimes. Senior decision-makers involved in the aid suspensions to Kenya, Sudan, Somalia (all 1991), Malawi (1992), and Nigeria (1993) have privately stressed their 'genuine belief' at the time that PC could 'bring democracy' to Africa. 20 These statements are supported by the public and private behavior of ODA personnel during this period. PC was applied with specific guidance and recommendations as to how the aid relationship might be restored -usually the abolition of one-party states and/or a return to constitutional government (Crawford, 1997;Cumming, 2001).
Internal ODA discussions on the re-establishment of aid programs in states subjected to PC focused heavily on assessments of whether the regime in question had responded in a 'positive' or 'satisfactory' manner to the specific conditions demanded by London (ODA, 1992). Likewise, the publicizing of PC decisions to the British media appear to have been frequently opposed in order, as one internal document on Malawi produced in June 1992 puts it, 'to encourage as positive a response as possible [on political reform] from the Malawi Government' (ODA, 1992).
Early impositions of PC by the UK, therefore, appear to have been informed, at least in part, by an 'instrumental' understanding of the tool's efficacy; the belief that it could indeed be used effectively to promote political reform abroad.
Overall, this appears to have been mainly based on Lindsay's logic of compliance and policymakers often expressed clear guidance on what such reforms should entail. This sentiment on the instrumental efficacy of PC appears to have been echoed in contemporary scholarly analysis of the phenomenon which largely focused on reviewing the circumstances under which PC 'could be effective' (Crawford, 1997(Crawford, , 2000Levitsky & Way, 2006;Moore & Robinson, 1994;Smith, 1998;Stokke, 1995). Expressive objectives were also clearly at play in these suspensions although they appear to have been less central than instrumental ones and focused mainly on international symbolism -upholding and promoting the norm of multi-party democracy in the post-Cold War world.
More recent applications of PC by London, however, do not seem to have been informed by quite the same logic. With some exceptions, contemporary impositions of PC have rarely been accompanied by any clear guidance on what reforms the recipient government in question should make to restore the aid relationship (de Renzio et al., 2005, pp. 42-44). Indeed, in some instances, -such as Uganda in 2006 -recipient governments have not even been clearly informed that a decision to suspend aid has been made beforehand (FCO, 2011). Most impositions of PC since the early 2000s have been declared as a response to a transgression (such as a corruption scandal) but not linked to any recommendations on what policy changes DFID would like to see introduced.
This lack of clarity over the objectives behind more recent PC applications by DFID has occasionally led to significant confusion within Whitehall over the rationales for certain decisions. Internal UK government discussions between desk-level staff and more senior civil servants (both in London and in Lilongwe -and between the two) on suspending GBS to Malawi during 2011, for example, veered between concerns regarding 'governance issues', the closing of political space, human rights concerns, financial mismanagement, risks to DFID's own reputation domestically, and the expulsion of UK High Commissioner Fergus Cochrane Dyet in April (DFID & FCO, 2011). DFID's statement announcing the cut in July referred to concerns over 'economic mismanagement and governance' (DFID, 2011) but followed an April announcement by the FCO that London's relationship with Malawi would be 'rapidly reviewed' in response to Cochrane Dyet's expulsion (FCO, 2011). 21 The range of explanations provided for the cut by British and Malawian newspapers underlines the unclear reasoning put forward by London for the decision. 22 Significantly, however, the Malawi decision -like many others in the later 2000s -was formally announced and promoted by the Department to the British media, following discussion between DFID and the FCO over what 'press lines' to take on the matter (FCO, 2011). Unlike those few statements made to explain PC decisions by ODA/FCO in the early 1990s, however, more recent such statements, including those on Malawi in 2011 and 2013, have emphasized the extent to which protecting 'UK taxpayers' money' (DFID Uganda, 2012), ensuring that 'taxpayers' money is safe' (Daily Telegraph, November 18, 2013) and satisfying the expectations of 'British taxpayers' (DFID Malawi, 2011) have influenced the decisions of senior policy-makers. These sentiments have been echoed by former ministers themselves with one reportedly framing the 2005 Ethiopia GBS suspension in terms of his department needing to be able to 'account to the British people'. 23 DFID officials have also followed how their spending decisions have been portrayed in the British media increasingly closely since the mid-2000s. DFID officials contacted one Daily Telegraph journalist in November 2012, for example, to ask him to amend an article which argued that the UK should suspend aid to Uganda requesting that he point out that 'DFID. . .have already frozen aid to the Ugandan government earlier this month' (Daily Telegraph, November 30, 2012). This emphasis on managing how its PC decisions are framed in the British media indicates a much higher level of focus on the domestic expressive nature of the instrument among UK officials in recent years.
It is crucial to note, nonetheless, that this change in tone has been accompanied by a clear loss of faith on the part of most DFID officials in PC's instrumental effectiveness. One DFID official has claimed, for example, that the 2010 and 2011 GBS suspensions to Uganda were not expected 'to make any actual difference to the situation on the ground'. 24 Likewise, in defending his decision to partly restore GBS to Rwanda in September 2012 after suspending it in July of the same year following allegations of support by the Rwandan regime for Congolese rebels, former DFID Secretary Andrew Mitchell (in office 2010-12) told a parliamentary committee that in his view 'taking away budget support [from Rwanda] would have no effect on the [government and its policies] in Kigali' (BBC News, November 8, 2012).
Indeed, DFID officials have held quite negative views of conditionality as an instrument since the early years of the Department, describing it as the 'traditional' way of engaging with developing states -echoing an early official departmental critique of the 'old conditionalities' to be contrasted with a new approach based around donor-recipient 'partnership' (DFID, 1997, pp. 22-49). 25 In its recent literature on conditionality, DFID has reflected on past experience and scholarly analysis in declaring that 'attempts to "buy" policy reform from an unwilling partner [through conditionality] have rarely worked' (DFID, 2009, p. 19) and that 'specific conditions should not be used to impose specific policy choices on partner governments' (DFID, 2006(DFID, , p. 10, 2009(DFID, , p. 19, 2014. Though these are 'formal' positions of the department they nonetheless accurately reflect the perspectives of interviewees consulted during this research. This revised position on the purpose of PC emerged from a departmental reassessment of the role, purpose, and effectiveness of the instrument following Short's 2003 departure. During 2004-05 DFID strategists in the Department's Policy Unit carried out a comprehensive review of PC policy. Owing to the Department's limited institutional memory, this process was largely placed in the hands of several consultancy groups who produced several critiques of ODA/ DFID PC current and past practice Mokoro, 2005). These analyses ultimately fed-into a 2005 policy paper -Rethinking Conditionality -which formally laid out DFID's approach to PC and the 'circumstances in which the 'DOES IT WORK?' -WORK FOR WHOM? BRITAIN AND POLITICAL CONDITIONALITY SINCE THE COLD WAR 7 UK will consider reducing or interrupting aid' (DFID, 2005, p. 3). The 2005 paper, while acknowledging PC's 'mixed' record in successfully promoting a range of reforms, clearly sets out three circumstances whereby DFID will consider suspending aid -when countries 'veer significantly away from their agreed poverty reduction objectives or. . .of a particular aid commitment'; when countries 'are in significant violation of human rights' and when 'there is a significant breakdown in the performance of partner government financial management. . .leading to. . .corruption' (DFID, 2005, p. 9). Importantly, however, the paper stresses that PC decisions 'will be based on an assessment of the long-term impact on poverty of the overall program of the [recipient] government. . .not on failure to implement any specific policy' and that 'the UK will seek to talk the issues through with partner governments before taking a decision' (DFID, 2005, p. 10). These sentiments and guidelines have been repeated and expandedupon in several other more recent DFID practice papers where, once again, the three 'breaches' of aid relationships meriting PC are outlined (DFID, 2006, pp. 10-11).
Nearly a decade later, however, many of DFID's PC targets appear to have been selected in a far less systematic manner than these guidelines suggest they should have been (NAO, 2008, p. 5). Instead, the application of PC, as some DFID officials have acknowledged privately, 26 often reflect the prominence of a particular scandal or crisis in the British media rather than the extent to which a crisis highlights a declining governance trajectory. GBS was suspended to Malawi in 2011, for example, in part because of London's concerns over Lilongwe's intimidation and suppression of opposition and civil society organizations (DFID, 2011). Much more oppressive regimes have continued to receive GBS but never have PC imposed as a result -however, such as that in Rwanda. Kigali did, nevertheless, see its GBS cut in July 2012 following allegations that it supported rebels in eastern Congo, claims which received heavy interest from British media houses (Daily Telegraph, July 27, 2012). Likewise, Uganda has, since the mid-2000s, been a frequent target of British PC linked to corruption concerns while Mozambique, another serial offender in this regard but one which enjoys much less attention in the UK press, has never been subjected to the instrument (a discrepancy some DFID officials have privately acknowledged 27 ).
Furthermore, PC decisions continue to be made much more suddenly and unilaterally by DFID officials than the guidelines recommend. Officials in Uganda in 2006 appear to have been informed about London's GBS cuts through reading local media reports having clearly not been afforded an opportunity for 'discussion' with DFID beforehand (FCO, 2011). The circumstances in which PC has been imposed by DFID in recent years, therefore, provides further evidence of the primarily expressive -rather than instrumental -rationale that now informs London's use of the tool.
The last decade has therefore seen an explicit rejection of compliance-based, instrumental PC among UK officials. It has also seen a growing focus by DFID on the domestic symbolism aspect of applying PC -assuring taxpayers that their money is not being 'wasted' on corrupt governments. Indeed several recent DFID publications make explicitly clear that PC can and should play an important role in ensuring 'accountability to [UK] taxpayers' in the use of aid funds (DFID, 2005(DFID, , p. iii, 2009. This simultaneous rejection of PC as a compliance mechanism and increased use of it for expressive reasons is paralleled in the statements and activities of a range of other European donors (see below).

DRIVERS AND 'EFFECTIVENESS' OF EXPRESSIVE PC
In explaining this shift in thinking on PC, intervieweesdirectly or implicitly -consistently highlighted two 'drivers' which have informed the process over the last two decades. 28 The first of these -the rejection of instrumental PC -appears to have come about through the fundamental re-focusing of DFID's role, purpose, approach to recruitment, and engagement with research and NGOs in the aftermath of its 1997 creation. This has been a conscious effort directed from the top aimed, in part, at building a Department whose personnel reject any presumed efficacy of instrumental PC. On the other hand, DFID staff have come to embrace expressive PC in more recent years through a more reactionary set of responses to perceived reputational risks to the Department -and to GBS as a modality -among British citizens themselves in the context of an expanding aid budget during a time of domestic economic crisis. Interestingly, while the impetus for this move has also come from ministers' offices to a considerable degree, it has also been driven by CO staff based on the assumption (in some cases) that this is 'what ministers want'. The final part of this article will therefore briefly (for reasons of space) outline these two drivers before concluding with a broader consideration of whether DFID's employment of PC can, in fact, be considered to have 'succeeded' in its main objectives.
(a) The rise of ownership and de-campaigning of the 'old conditionalities' within DFID UK enthusiasm for instrumental PC impositions clearly declined during ODA's final years, with more suspensions occurring during 1991-92 alone than in the subsequent 5 years. Former ODA staff who later became DFID staff, however, emphasize that while there were some misgivings within ODA on the instrument's rather meager returns (notably in Nigeria and Kenya) 29 a more wholesale rethink on this issue began only after DFID's creation in May 1997. In part this was driven personally by Short, whose tenure at DFID had profound and enduring consequences for British development policy. According to some of her former key advisers, Short viewed PC as a relic of an era of UK aid policy dominated by geostrategic and economic interests rather than concern for eliminating global poverty. 30 Short's determination to instill a wholly pro-poor culture at DFID -in contrast to its ODA predecessor -lead to a specific denouncing of 'the old conditionalities' of the past and an ambition to build 'genuine partnerships' with recipient governments and societies in its foundational 1997 White Paper (DFID, 1997, p. 37).
More generally, under Short's leadership DFID re-focused its internal structures and recruitment criteria in order to attract more candidates from NGO backgrounds in particular (notably through the Social Development Advisers cadre), most of whom held similarly skeptical views of conditionality. The vast expansion of DFID's staffing numbers from c.1,080 in 1997 to c.2,600 by 2013 (an increase of 240%) -focused particularly in the 'advisers' cadres -allowed for a much more rapid installment of a broad constituency of personnel within the Department of the same mind on this issue than might be feasible in more established bureaucracies. This process was complemented and enhanced under the tenure of Hilary Benn (2003-07) who sought to incorporate a greater awareness of political economy issues into DFID's work and oversaw a major recruitment drive of political scientists and the creation of a 'Governance Head of Profession' cadre in the mid-2000s which contributed heavily to a White Paper entitled Making Governance Work for the Poor (Benn, 2006;. 31 This process ultimately contributed to the move toward expressive PC in the Department through the creation of a body of HQ and CO policy staff focused on the political economy, and risks, of development interventions (see below). 32 This cadre also, however, viewed instrumental PC as fundamentally misguided, informed by a deeper understanding of the complex relationship between external intervention and political reform in developing states than their ODA predecessors.
The undermining of confidence or support for instrumental PC within DFID should also be seen in the context of the Department's enthusiastic investment in, and engagement with, academic development research (on PC and many other areas) since 1997 as well as its determination to become a global development 'thought leader' through closer cooperation with, and contribution to debates within, the World Bank and OECD in particular. The building of closer personal and institutional relationships between DFID staff and individuals from these communities and institutions since this time has meant that DFID staff have become particularly familiar with academic critiques of instrumental PC (Adam et al., 2004(Adam et al., , pp.1059(Adam et al., -1060Killick, 2004, pp. 12-14;Nelson, 1996Nelson, , pp. 1551Nelson, -1554 and supportive of the idea of 'ownership' -a concept which has dominated discussion at High-Level Fora on Aid Effectiveness in Rome (2003), Paris (2005), Accra (2008), and Busan (2011). Within these various contexts, conditionality has been rejected as highly damaging and problematic for fostering continued partnership and recipient ownership of policies and 'dialogue' is to be the preferred method for resolving donor-recipient disagreements on issues of democratization, corruption, and human rights. This is in spite of the fact that many of these 'partnerships' have in fact been imbued, at least implicitly, with conditionalities, as a number of scholars have acknowledged from 1997 onward (Hayman, 2011, p. 679;Maxwell & Riddell, 1998, pp. 262-267;Nelson, 1996).
(b) A changing international economic climate and the move to expressive PC The embrace of expressive PC by DFID since the mid-2000s has, however, been driven by more reactionary thinking and processes -both in ministerial offices and in COs. The 2007-08 Global Financial Crisis and its consequences have placed heavy pressure upon most European governments to curtail public spending where possible and to better justify 'nonessential' expenditure, such as development assistance, to constituents. The impact of this dispensation has been felt particularly heavily within DFID given the Department's rising budget under the Labor government (1997)(1998)(1999)(2000)(2001)(2002)(2003)(2004)(2005)(2006)(2007)(2008)(2009)(2010) and protection against cuts ('ring-fencing') under the current Coalition government. DFID -consequently -became the first donor to build 'value for money' considerations into its programs comprehensively and its approach has been adopted by a range of other governments since this time. 33 Protecting the Department from accusations that funds are being 'wasted' has become a central preoccupation of DFID's bureaucratic culture and DFID officials interviewed for this study, more so within COs than UK directorates, privately bemoaned how acutely conscious they are of the -assumed -negative consequences that will befall their careers should they 'take their eye off the ball' on particular aid interventions and end up being perceived as responsible for continuing support to an abusive or corrupt regime. 34 Given the difficulties of tracking GBS expenditure this modality has been a particular victim of this dynamic and has frequently been cut by the UK on the basis of these concerns, albeit usually being diverted to other modalities which by-pass the state (Hayman, 2011;Maxwell, 2011).
This has led to a culture of growing conservatism and blame avoidance in the Department (see Hood, 2013) which has even reached up to the top; Andrew Mitchell waited until the final moments of his time in office to part-restore GBS to Rwanda in September 2012 (BBC News, November 8, 2012). The use of PC has therefore flourished in this environment given its ease of application (at least in relation to GBS) as well as its high profile and immediate impact; DFID officials note that they feel increasingly incentivized to use PC to demonstrate to their superiors -and, indirectly, to domestic audiences -that they are firmly committed to 'protecting UK taxpayers' money' from abuse. 35 (c) "Does it work?" The efficacy of expressive PC It is therefore clear that the use of PC by UK officials since the Cold War has been increasingly premised upon expressive as opposed to instrumental rationales. This is a consequence of a range of influences and pressures on these individuals operating at different levels over time. To return to the title of this article, though, whether PC can be considered as 'working' within this new dispensation, however, remains an open question. In closing, then, it is important to consider whether or not DFID officials in the contemporary era believe that their impositions of PC are delivering success vis-à-vis their expressive function.
Few DFID officials interviewed for this piece, however, were able to clarify what such 'success' might look like or indeed who the 'audience' of PC impositions was specifically -or how their reactions to aid cuts might be captured and assessed. Many CO and ministerial/senior staff linked to PC decisions since the mid-2000s emphasized the role of PC cuts in addressing 'public' or 'taxpayer' concerns on 'wasted aid' in the taking of these decisions but were unable to break-down this category of actors further or describe what mechanism might be used to establish whether, and to what extent, any such cuts had allayed these concerns among the general populace. What exactly is being 'protected' in PC impositions beyond 'UK funds' was also somewhat unclear with ministerial careers, DFID's institutional autonomy, 'the integrity of GBS as a modality,' and the reputation of the development enterprise itself all being cited by officials at different levels in the UK and in COs. 36 Furthermore, DFID has significantly reduced its investment in public opinion surveys on aid since 2010 with the most recent work conducted on this by the Department dating from September 2011. These surveys examine public opinion on aid wastage and overall support for aid but shed little light on the effects of PC cuts on such perceptions (COI/DFID, 2010). Indeed, a number of UK-based DFID staff interviewed suggested that their perceptions on public support for -or opposition to -particular aid decisions and approaches are largely based on careful readings of media reporting, with the Times, Daily Telegraph, and Daily Mail most frequently cited in this regard. 37 Several vividly described the impact of such publications' headlines in determining day-to-day activities in London starting from weekly Monday morning strategy meetings but were unclear as to whether they believed that such stories were representative of public opinion more broadly. Moreover, the fact that some CO staff have felt it necessary to be more vocal in 'DOES IT WORK?' -WORK FOR WHOM? BRITAIN AND POLITICAL CONDITIONALITY SINCE THE COLD WAR 9 highlighting potential crises and 'breaches' because of a perception that this is 'what the minister and managers want' (see above) also suggests that for some in DFID actually precipitating or the taking of a PC decision is the end goal. These observations are particularly interesting when considered alongside the findings of Adam and Gunning -published in this journal -on the use of performance indicators in the allocation of aid. These scholars argue that while performance indicators played an important role in shaping donor-recipient discussions in Uganda by the early 2000s few were actually designed to measure policy outcomes specifically, remaining 'vague' and concerned primarily 'with input and process' (Adam & Gunning, 2002, p. 2052-2054. This is not to suggest that the rise of expressive conditionality at DFID is a product of cynical careerism in the Department -civil servants are obliged to respond to the demands (perceived or directly communicated) of their political masters in a democratic system. Likewise, DFID staff have consistently -and to date successfully -sought to ensure that PC impositions result in a diversion of funds to alternative modalities within the same country rather than a withdrawal of aid itself, in line with their genuine commitment to providing assistance to the world's poorest. Nor can DFID be accused of being secretive in setting-out its position on PC and in communicating decisions on the same (at least within the UK) since the mid-2000s -in stark contrast to many other donors employing the instrument or to DFID's ODA predecessor.
What DFID has not yet successfully communicated -to UK citizens or to those in the countries it provides aid to -however, is the extent to which PC and other decisions are a trade-off between short-term political and reputational positioning and longer-term developmental programing. DFID's March 2014 guidance note on conditionality hints at these ambiguities more clearly than any previous publication and intimates, perhaps, at a greater recognition within DFID's management class that PC is a question, ultimately, of staff 'judgment' than of implementing an applied formula. To some extent, this resonates with Adam and Gunning's discussion on the value of 'vagueness' for donors in providing space to 'not have to follow through on ex ante threats' (Adam & Gunning, 2002, p. 2052. If expressive PC is ultimately, however, to be assessed as effective or otherwise then these complexities and uncertainties must not only be tacitly and privately acknowledged by DFID but also defended, both in the UK and in recipient states.

CONCLUSIONS AND BROADER IMPLICATIONS
This article makes an important contribution to the literature on PC by turning around the oft-asked question 'does PC work?' to investigate whether aid officials themselves believe it can work and, if not, why they continue to use it. Drawing on concepts discussed in the literature on economic sanctions, two broad sets of motivations have been delineated as a framework for understanding applications of PC: instrumental (focused on compliance and deterrence) and expressive (focused on domestic and international symbolism). The article has focused in depth on the UK's imposition of PC since the Cold War and argues that while expressive rationales have always informed the UK's use of PC, they are far more central now than they were a decade ago and are largely concerned with domestic rather than international perceptions. Confidence in instrumental PC among UK officials, it is suggested, was fatally undermined by the later 1990s through a ministerial-lead re-balancing of the country's approach to development and through the rapid recruitment of new staff skeptical of the tool. The move toward expressive PC, however, has been driven by more reactionary concerns among ministerial and CO staff relating to the protection of reputations -of DFID, GBS, and the development enterprise itself -in the context of domestic economic austerity and rising aid spending.
The findings of this study raise a number of questions for development scholars, as well as for donor officials themselves. Most critically, how should the effectiveness of PC be assessed and measured to take into account its expressive purposes? If PC's instrumental efficacy is no longer a major factor in its employment then clearly analyses which focus only on this dimension (such as whether PC results, for example, in greater democratization) require reconsideration. Moreover, if donor usage of the instrument is increasingly premised on managing the perceptions of domestic audiences then should donors themselves look more closely at how PC impositions are interpreted by Western citizens and how, if at all, this effects general opinions on the value and integrity of the development enterprise? It is clear, as argued in the final part of this study, that these are not questions currently being explored in any depth within DFID itself.
Though this article has focused on the UK case, similar trends can be discerned elsewhere in the European aid network. Following a re-evaluation of its GBS programs in 2011, for example, the European Commission argued that the modality should be suspended in instances where 'the EU's. . .reputation needs to be protected' (European Commission, 2011, p. 4). EU officials have also increasingly framed their objections to recipient government actions in terms of the negative impact such behavior will have on European perceptions of the Commission's development policies (Daily Monitor, December 4, 2012). German development officials have also privately characterized European aid cuts linked to anti-homosexuality laws and policies in Africa as based at least partly on a donor desire to 'communicate with domestic constituencies'. 38 Similar sentiments have been expressed to the author by representatives of a range of Scandinavian donors during discussions in Uganda (during 2009-12) and the OECD (2013). This is not to say, of course, that this is enough evidence to support a clear generalization of the findings of this study beyond the UK. To do so a far more systematic analysis would be required -one which considered the recent bureaucratic and architectural histories of other aid donors as well as the salience of aid policy and crises in different national media's and political cultures. It is apparent, nonetheless, that the UK is not alone in using PC for expressive purposes and thus the questions raised in this article have resonance outside this single case. NOTES 1. DFID as a 'trailblazer' and 'innovator' in the global development community has been a major theme in the public and private discourse of departmental officials since the late 1990s (see Vereker, 2002); then DFID Secretary, Andrew Mitchell, even described the UK as a 'development superpower', claiming it was 'in the lead' in this field in June 2011 (BBC News, June 12, 2011). This has also, however, been a common sentiment expressed by a range of donor officials in meetings attended by the author and in interviews conducted on a range of themes (security, justice, political economy analysis, governance, etc) in several European, African and North American capitals during 2009-13 including the World Bank, Swedish International Development Agency, Dutch Ministry of Cooperation, Norwegian Agency for International Development, Organization for Economic Cooperation and Development, and US Agency for International Development. DFID is also the leading funder of research among bilateral donors.
2. Between 500-600 ODA/DFID staff have been based in East Kilbride, South Lanarkshire, since 1981. These personnel were primarily administrative until 1998 when a range of policy positions and departments (including Finance, Performance and Impact -where conditionality policy is, at the time of writing, housed) were transferred there from London. A range of senior DFID staff (directors and directors-general) now also divide their time between the two offices.
3. Aid suspensions or reductions linked purely to delays in meeting public financial management targets or fiscal indiscipline not immediately linked to corruption -such as those imposed by DFID on Ghana in 2006 or Malawi in 2008 -will not, however, be considered.