Unmasking the open secret of posting and transfer practices in the health sector.

This article focuses on processes of granting posts and transfers to health care workers and administrators that can be described as 'mission inconsistent (MI)', meaning that they are not conducted in a way that maximizes health outcomes or that respects the norms of health care worker professionalism. We synthesize relevant literature from several different disciplinary perspectives to describe what is known about the problem of MI posting and transfer in the health sector, to critically engage and interrogate these literatures, and to briefly discuss efforts that have been made to maximize mission consistency. The article concludes by suggesting principles for future research that would foster a more complete understanding of posting and transfer practices. Published by Oxford University Press in association with The London School of Hygiene and Tropical Medicine

Practices of assigning posts and transfers in many countries undermine the health system’s functioning as a core social institution.

Despite widespread acknowledgement of the problem, there is little peer-reviewed research on posting and transfer.

Existing theoretical frames that have shaped research and policy do not capture the full complexity of posting and transfer.

Introduction

The traditional tools of public health research and analysis often disappoint in their ability to provide a cogent basis for crafting policy interventions to address effectively the deep dysfunction that plagues the health systems of many countries, preventing us from achieving universal coverage and other global health goals. Those who work on the ground, within struggling health systems, often see and experience challenges that are all but ignored in the health literature. One such challenge is the way in which health workers and administrators are posted and transferred in public systems. Although human resources for health has become a priority topic in health initiatives at all levels, most attention has been devoted to calculating deficits, organizing training in clinical and managerial skills, and designing incentives for retention and performance. The actual dynamics of posting and transfer practice, which regularly sabotage the best laid plans of health officials, and which often pre-occupy the workers in the system, are treated as almost taboo, raising as they do does uncomfortable questions for everyone whose life chances are determined by their operation.

This article focuses on posting and transfer practice that can be described as ‘mission inconsistent (MI)’, meaning that it does not advance population health goals and that it is incompatible with prevalent professional ethics regarding civil servant and health care worker rights and responsibilities (United Nations 1996). Although we recognize that the definition of consistency may vary by context, the label of MI is a helpful global construct as it enables us to focus on the practices with which we are most concerned—those that undermine organizational justice and service delivery.

What form does MI posting and transfer take? In general terms, a MI post or transfer assignment can be driven by the wishes of the transferee or the transferor. Transferees may employ clientelism, bribery, or other means to avoid serving in undesirable areas, or to obtain a post in particularly desirable areas. Transferors may respond to these requests, and they may also effect a transfer for their own reasons, including displeasure with health care worker or administrator performance, a desire to free up a post, or as part of an effort to build or maintain a local political constituency. Actual posting and transfer practice is likely a negotiated outcome of transferee and transferor preferences and objectives, embedded in larger political and social dynamics of the health system and beyond. Academics and policy-makers have stated that MI posting and transfer contributes to health worker shortages in the most deprived areas; health care worker and administrator demotivation; poor co-operation among and within health care cadres; inefficiency; delays in reform and public health programme implementation; lack of accountable relationships between health care workers and administrators and the communities they serve as well as among health care system cadres; and inability to accurately monitor health care worker and administrator performance and to develop rational training, recruitment and scale-up plans (Collins ; McPake and Koblinsky 2009; Second Administrative Reforms Commission 2008; Tjoa ; Diarra 2012).

This article synthesizes relevant literature from several different disciplinary perspectives to describe what is known about the problem of MI posting and transfer in the health sector, to critically engage and interrogate these literatures, and to briefly discuss efforts to maximize mission consistency. It concludes by suggesting some principles that we believe should guide further research in the field.

MI posting and transfer practices in the health sector: what do we know?

We conducted a review of the academic and grey literature on posting and transfer in the health sector specifically, as well as in the public sector overall. We searched for the Boolean terms ‘transfer AND health’, ‘corruption AND health’, ‘transfer AND civil serv*’, ‘merit-based AND health’, ‘turnover AND health’ and ‘bureaucratic neutrality’. The last two terms were searched as a result of findings in the first round of review. We searched several different peer-reviewed databases, including LexusNexis Academic, in order to see if the issue appeared in the print media contained in the database. We also searched the main library database at Columbia University, and specific journals, as well as both Google and Google Scholar. Abstracts were read to determine whether not the paper was relevant to our areas of interest. We used the references of the initial sources to identify further articles as well as books. We conducted limited (Google Scholar) literature reviews for ancillary topics that are discussed briefly in this article, namely human resource retention in low and middle income countries, research on locational preferences among health care workers and new public management.

Overall, the plurality of articles and books found related to India. It is not clear if this relates to the comparatively high number of academics studying public health and health systems in India (vs other low and middle income countries), the prevalence of MI posting and transfer practice, widespread interest in India’s colonial legacy and its impact on post-independence governance, or other factors.

An assessment of the peer-reviewed and grey literature revealed a small number of references to MI posting and transfer in the health sector. Some studies and reports merely mention that a particular health project was hindered by frequent transfers (Bossert ; Maiga ; Barker ; Datta 2009; DFID 2011). A handful of papers report the prevalence and/or the drivers of MI posting and transfer among frontline health care workers and administrators. For example, a qualitative study conducted in the Dominican Republic found that health personnel rolls were largely driven by patronage. Promotions were reportedly based on recommendations from politicians, political parties, military officials and high-level government officials (La Forgia ), findings that were echoed by a qualitative study in Niger (Diarra 2012). Similarly, Collins et al. conducted a literature review and a pilot qualitative study (eight semi-structured interviews) to assess the frequency and the rationale of transfers in the public sector in Balochistan, Pakistan. They found that transfers could be quite frequent, with the cadres involved in health administration more affected than those involved in service delivery (Collins ). Blunt et al. provide one of the most comprehensive overviews of MI posting and transfer in the health, infrastructure and education sectors in a given locale; in this case, in Indonesia. Semi-structured interviews and focus group discussions conducted among 211 civil servants at different levels of government found that midwives and nurses frequently paid fees to be accepted into a nursing academy, to pass the civil service test, to be appointed to a desirable position, and to gain favourable transfers or promotion (Blunt ).

Several articles explore the rationale for MI posting and transfer among transferees. Opportunities for private financial gain are among the most frequently cited motives. A 1985 paper by a political economist noted that transfers to superintendent posts at hospitals in big cities in India cost more than other transfers, as the potential for making money through the allocation of beds or the black market sale of stolen medical supplies is so high (Wade 1985). This market for transfers with earning potential has been termed the ‘transfer trade’ (Chambers 1983) with particularly lucrative posts functioning as ‘earning centres’ (de Zwart 2000). The interviewed civil servants employed in Indonesia’s health sector explained that desirable transfers were to urban areas and to places where there were development projects that provided further opportunities for (illicit) income (Blunt ).

Desire to live in certain areas for reasons other than income is also a common explanation for MI posting and transfer. An ethnographic study of district health services in Nepal observed that most health staff who are posted to isolated rural areas try to avoid going by lobbying regional and central authorities for a transfer. If they are unsuccessful, they might begin their work, and then go on an extended vacation as part of their continued efforts to negotiate a transfer; posting and transfer practices thus may contribute to absenteeism in some contexts (Aitken 1994). Similarly, a study of a tertiary hospital in Northern Ghana cited reasons why doctors did not want to be posted there, including the fact that the hospital is in a rural area with a bad climate, and lacks resources and staffing, accommodation and good schools (Andersen 2004).

Indeed, within the human resources for health literature, a body of work describes the factors health care providers report shape their locational preferences. For example, 90% of medical students and two-thirds of nursing students surveyed in Ethiopia reported that they would prefer postings in urban areas, primarily because these areas offer access to good education for one’s children, opportunities for promotion, and access to further training (Serneels ). Those who were more willing to work in a rural area reported higher intrinsic motivation to help the poor (Serneels ). Discrete choice experiments are frequently used to quantify health care worker preferences and the likely impact of policy interventions, such as higher salaries in rural areas. A 2009 review of such experiments found that, in low and middle income countries, health care workers in general preferred urban areas, and they also placed high value on characteristics such as well-equipped facilities, good facility management and housing. Urban preference was not universal, and some health care workers reported that they were more concerned about salaries than others (Lagarde and Blaauw 2009). These studies offer important insight into what post attributes health care providers report as being important to them, but they offer less insight into the how these preferences are negotiated in the larger context of policies guiding provider distribution, and the power relationships among transferees, transferors and other stakeholders that underlie actual practice.

As noted, transferor-driven decisions can be motivated by objectives related to the consolidation of political power. This dynamic can also shape the system on a more macro-level. It could be that those with decision-making power over postings and transfers intentionally undermine the bureaucracy. For example, a study conducted by the Institute for Development Studies (and authored by a former civil servant in India) concluded that between 1990 and 2005, elected officials in Bihar, India purposely fostered state incapacity by not filling key civil service posts, as the only qualified personnel were from historically dominant groups. The Chief Minister of Bihar eviscerated the public sector, concentrated decision-making authority within the political system, as opposed to the bureaucracy, and maintained support among a key constituency of marginalized groups by marshalling their anger at the wealthy and elite (Mathew and Moore 2011).

The academic and grey literature shows widely shared agreement in the global health, development and anti-corruption communities that health care worker and administrator posting and transfer practices are problematic (McPake and Koblinsky 2009; Nugroho 2011; Olowu 1999; World Bank 1997). Media coverage suggests that these concerns are shared by the public in at least some affected countries (The Statesman 2002; The Statesman 2007; United News of India 2011; Press Trust of India 2011). For their part, governments have at least rhetorically acknowledged the need for transparent, rule-based, posting and transfer systems (Adegoroye 2005; United Nations 1998).

The above represents a summary of almost all of the literature identified regarding MI posting and transfer in the health sector. We have enough evidence to confirm that MI practices undercut the equitable and effective delivery of health care, but not nearly enough information to understand the factors shaping posting and transfer in different contexts; how these play out in practice; and how much they contribute to maldistribution, absenteeism and inefficient use of health administrator time. Other significant health system costs may be incurred as well. Evidence from the Indian Administrative Service suggests that the need to ingratiate oneself to obtain desired transfers means that public sector employees may be unwilling to address important but politically sensitive issues (Banik 2001). In brief, although it is clear that MI posting and transfer merits attention, the existing evidence is a thin foundation on which to base policy recommendations and interventions; posting and transfer is ripe for further enquiry.

The following sections highlight insights from multiple disciplines. These approaches provide further background to the systems of posting and transfer in the public sector. At the same time, synthesis of the literature surfaces several blank spots, including how transferor and transferee preferences are negotiated in different contexts, how posting and transfer is experienced by health care workers and administrators, the salience of formal and informal norms in different contexts, and the ways that posting and transfer reflect and reinforce the workings of power in the health sector.

Historical insights

We do not endorse a fatalistic interpretation of path dependence, but recognize that what theoreticians of public administration call the ‘administrative heritage’ of a particular country shapes current practices (Kearney 1986, p. 144; LaPalombara 1963). Most of the historical analyses identified describe the continuity and interplay of the formal and informal norms under-girding the transfer system within the civil service overall.

India, Pakistan, Swaziland and others provide examples of continuity of formal policy. They continue to use the old British system of ‘confidential reports’ of civil servant performance written by supervising officers; these reports are not shared with supervisees (McCourt 2006). However, the widespread persistence of this practice need not imply significant similarities among the countries that practice confidential reporting. For example, a case study on human resources management in Mauritius found that there was very close attention paid to following the rules associated with implementing formal policy, but little attention paid to the strategic intent of the policy (McCourt and Ramgutty-Wong 2003). Confidential reports may not actually play a role in decision-making about promotions. Knowing that certain policies are carried out may offer little insight into how public sector management really operates; the rules may tell only part of the story.

On the other hand, sometimes a particular policy or practice (formal or informal) becomes so ingrained that it is a defining feature of the system. For example, the purchase of posts—legally or illegally—is a tradition in many contexts. Some civil posts were formally available for purchase in the Mughal Empire, a practice that continues illicitly in some contexts in India. Legal purchase of posts was also prevalent in seventeenth and eighteenth century Europe, China and the Ottoman Empire (Allen 2005; de Zwart 1994; Hoselitz 1963).

Transfers, too, have assumed a central role in some systems, with evolving incentives and conditions justifying their use. In colonial India, the British effected frequent transfers in order to minimize opportunities for corruption. Over time, corruption diminished and the justification for the transfers weakened, particularly given that frequent transfers undermined administrator effectiveness. However, during this same period, colonial service in India became less prestigious, and the British instituted a generous leave policy in order to entice more talented professionals to join the colonial service. The staffing gaps associated with generous leave in turn fed the transfer machine, reinvigorating the transfer system (de Zwart, 1994).

Particularly in the post-colonial context, administrative heritage is important to mission inconsistency on a more fundamental level—upstream from posting and transfer practice. Colonial administrations prioritized territorial control and profit over service delivery; they were by design accountable to the colonial powers, not to the people (Adamolekun 2002; Banik 2001; Dwivedi 1989; Haque 2007). In the case of India, colonial medicine and sanitation favoured areas populated by the colonial administration and the ‘sanitary cordon’ established around them (Ramasubban 2008). Post-colonial governments have failed to rectify some of these gaps. For example, medical and public health staffing needs are often determined according to population size, rather than disease burden and poverty level, or median distance to a health facility (the average person living in a far flung area may have to travel quite far to reach a facility) (Ramasubban 2008). This is relevant for two reasons. First, it is important to acknowledge that the structure of the health system itself can in some ways be ‘MI’. Second, any staffing gaps caused by MI posting and transfer practices are overlaid onto a human resources situation that was already insufficient to cover need in some areas.

In brief, two key points emerge from the historical literature. First, continuity of practice should be understood in context. A particular practice such as confidential reporting may be relevant as one of many elements of the system but not particularly determinative. On the other hand, some practices, such as buying posts or frequent transfers can be such deeply embedded norms that they persist, despite rules to the contrary. To be sure, a MI transfer could be made despite a confidential report that suggests otherwise. Second, MI practice may be overlaid on a system that is itself MI. The earlier example of ‘state incapacity by design’ in Bihar illustrates that the mission consistency of the system as a whole can evolve or change.

Corruption and health care

When MI posting and transfer is broached in academic and grey literature, it is often framed as one of many types of corruption widespread in the health care sector. Corruption pervades some health systems, constraining health care worker and administrator decisions about whether or not to participate, and challenging efforts to transform the system. The corruption nomenclature is fitting in some—but not all—cases of MI posting and transfer. Befriending your supervisor in order to convince her to transfer you to the city where your children are in school is probably not corrupt. On the other hand, most would agree that paying your supervisor in order to be transferred to another post where there are more opportunities to extort unofficial payments from your patients is corrupt behaviour.

Corruption: When used here, the word ‘corruption’ refers to the ‘abuse of trusted authority for private gain’ (as cited in: Harrison 2007). We do not endorse this definition as describing the full range of corrupt practices (see, e.g. the World Bank’s Africa Development Indicators 2010, on ‘quiet corruption’, which is not captured by the definition above). However, as we are examining the traditional corruption literature and how this lens has been applied to posting and transfer, we opted to use a conventional definition.

An archetypal corruption lens detects countless cases of MI posting and transfer in the public sector. The academic and grey literature includes many examples of a market for posts and transfers across many sectors in different countries (Wade 1985; Davis 2003; Kristiansen and Ramli 2006; Kim ; International Crisis Group 2009). Almost 20 years apart, development researchers Robert Wade and Jennifer Davis described a market for transferee-driven transfers in India (Wade 1985; Davis 2003). Wade created a model for the ‘cost’ of particular posts in the irrigation sector, finding that the cost was determined by the potential for earnings associated with the post, with an adjustment for the cost of living (Wade 1985). Davis, too, described a cash market for particular posts. Transferees paid local political leaders in order to obtain the transfer they desired (Davis 2003). In contrast to Wade, however, Davis concluded that opportunities for corrupt tendering, procurement, and so on were not assets of a post for which one paid. Instead, these corrupt practices were driven in part by the need to pay for desirable transfers (Davis 2003). In his research in India, de Zwart found that both reasons pertain; individuals may engage in corrupt behaviour in order to pay for a transfer, or they may seek a transfer in order to have greater opportunities for corruption (de Zwart 1994). These differences highlight the importance of developing an emic understanding to posting and transfer; the differing motivations behind the common behaviour of purchasing a post reveal very different values and priorities, and demand different approaches to reform.

The market for posts and transfers is by no means unique to India. As explained, researchers have described such a market in Indonesia. Kristiansen and Ramli interviewed 60 civil servants, all of whom admitted to paying for their current posts (2006). Building on these findings and experience managing a civil service reform project in Indonesia, Peter Blunt and colleagues explain that development projects can foster and even exacerbate patronage-based posting and transfer through their collusion in a narrative that policy changes and reforms are leading to improved governance, rather than acknowledging and addressing the fact that the funds associated with these programmes can create opportunities for illicit income (Blunt ).

The literature on corruption also offers relevant theoretical approaches regarding the nature of posting and transfer systems. Most corruption frameworks rely on a principal–agent definition of accountability, wherein the principal (health care users and the voting public) hold an agent (health policy-makers and providers) to account for service delivery. In the case of corrupt posting and transfer practice, accountability between the health care system and the people it should serve is almost absent. The corruption literature offers several reasons for this. First, the principal–agent relationship in the health care sector is characterized by asymmetric information (Savedoff and Hussmann 2005; Vian 2007; Bloom ); health care users lack the data and the capacity to assess the frequency and impact of MI posting and transfer. They thus do not mobilize to demand change.

Second, actors in the health system are dispersed (Savedoff and Hussmann 2005). Although posting and transfer decisions can be effected in order to control the transferee, there is usually no one individual or institution that controls posting and transfer system-wide. In part because the actors influencing posting and transfer are dispersed and have dispersed interests, it has proven difficult to marshal the near-universal recognition that current practice is MI and transforms the system (de Zwart 1994).

Finally, as noted, the initial, and, in some cases, continued, rationale for transfers is to ‘prevent corruption’. To be sure, the belief that transfers prevent corruption has become a stable ‘mental frame’ that continues to justify transfers, regardless of whether or not frequent transfers do, on balance, lessen the frequency of corruption (de Zwart 2000). Civil servants persistently cite this justification, despite their contemporaneous acknowledgment that posting and transfers may feed corruption (de Zwart 2000). De Zwart explains this seeming contradiction by describing two types of corruption associated with posting and transfer—market corruption and parochial corruption. Transfers are intended to prevent ‘parochial’ corruption, or corruption that results from the development of personal relationships. However, transfers promote ‘market’ corruption, when civil servants charge clients for services that should be free (de Zwart 2000), or otherwise earn money illicitly. Market corruption may be easier when civil servants stay for such brief periods that they do not feel obligated by social ties, or that documenting and adjudicating a case of corruption are not feasible (de Zwart 2000; Second Administrative Reforms Commission 2008).

As described, corruption does not encompass the full range of MI posting and transfer practice. Nor does the language of corruption necessarily aid reform. It may alienate participants, who in fact are acting rationally given the nature of the system. Were they to refrain from advocating for better postings, they may risk a lifetime of undesirable postings (Wade 1985; Banik 2001). Moreover, labelling and prosecuting as corrupt those who pay for posts for reasons such as maintaining a standard of living for their family likely fail to address the underlying problems. Finally, it is important to consider who defines corruption, and which actors are implicated. Although MI posting and transfer and its relationship to corruption in the health sector are often described in the corruption literature, the ‘symbiotic relationships’ that Blunt et al. describe as existing between ‘donors’ and patronage systems are rarely subject to interrogation in the corruption literature (Blunt , p. 78).

In conclusion, common approaches used in corruption theory and research can be used to explain some factors that underlie MI practice, and to catalogue some types of MI posting and transfer. These approaches can also illuminate the persistence of the system of mission inconsistency; it is difficult for individuals who wish to avoid corruption to buck the trend. Corruption researchers increasingly use qualitative methods to understand behaviour (Vian 2007). These methods could be applied productively to the study of MI posting and transfer, fleshing out the constellation of motives, interests and norms that shape posting and transfer practice. At the same time, we remain cognizant of the research limitations of a corruption lens; it does not capture the full range of behaviour that can be described as MI. Moreover, the lens may have programmatic limitations as well, insofar as individuals may adopt an unhelpful defensive posture when the concept of corruption is evoked. The costs and benefits of explicitly naming and denouncing behaviour we seek to change need to be weighed in the given context.

Public administration approaches to research and reform

Much of the corruption literature is part of the larger universe of the study of public administration. Studies and debates in public administration further explain posting and transfer. This literature is crucial to understanding posting and transfer as it describes many approaches that have been applied to public sector reform since the end of the colonial period.

Empirical assessments of MI posting and transfer reveal the salience of politics in shaping the posting and transfer system. Control over transfers is an important political asset that can be used to consolidate power. Following study of the Indian Administrative Service, political scientist Dan Banik identified several types of politically motivated transfers. Politically motivated transfers initiated by the transferor can include transfers before an election to get rid of politically unpopular administrators, after an election to install loyal administrators, and to bring in a friend or exile an enemy (Banik 2001). Election-related transfers or appointments based on political affiliation have been described elsewhere in South Asia (Patil 2008), Ethiopia (Mengistu and Vogel 2006), Indonesia (Blunt ) and Nigeria (Dlakwa 1998). Politicized decision-making may conflict with two key goals of public sector reform—promoting a meritocracy and political neutrality.

Much of the public sector reform that has occurred over the last 40 years aimed to institutionalize merit-based bureaucracies. Though some disagree (Marx 1963), a merit-based bureaucracy is widely perceived to be an important indicator and requisite of economic and political development (Kearney 1986; World Bank 1997; Evans and Rauch 1999; Daddah 2000; Cameron 2010). Merit-based systems employ civil servants according to their qualifications. However, the ideal extent of the meritocracy has become a subject of debate. Should merit-based criteria apply only to the initial recruitment or to all promotions? Some argue that ‘promotions’ based entirely on seniority are fairer, as they lessen the role of possibly subjective performance evaluations and decrease opportunities for corrupt practices. Importantly, a purely merit-based ‘recruitment’ system precludes reservations and other types of affirmative action that are entrenched in some countries, including India. Thus, from a national policy perspective, the appropriate degree of meritocracy depends on health system governance preferences, as well as broader state goals, such as increasing minority representation in the bureaucracy.

Political neutrality, too, is often posited to foster public administration effectiveness, though many qualify this by saying that politicization at the top of the public sector might aid the process of state consolidation and facilitate the implementation of decisions taken by the executive branch (Adamolekun 2002; Matheson ; Cameron 2010). In some contexts, MI posting and transfer is a way that nominal merit-based systems maintain attributes of a political system (LaPalombara 1963; Levitan 1942; Daddah 2000; Haque 2007). Thus, here too, there is no ‘ideal’ level of neutrality. Though most would agree that a bureaucracy that is changed from top to bottom every time the party in power changes cannot be effective, the proportion of the positions that are politicized depends on state objectives.

What can be done to promote mission consistency?

The following section provides a brief summary of posting and transfer-specific and broader policies that were intended to influence posting and transfer practice. This discussion is relevant insofar as it illustrates the most common responses to MI posting and transfer, and the challenges that many efforts encounter.

Many governments have demonstrated rhetorical commitment to reforming posting and transfer (United Nations 1998; UNDESA 2001), and their degree of success speaks to the potency of policy resistance—when the system reacts to neutralize the intent of a policy. Since becoming independent, Bangladesh, Cambodia, India, Nigeria, Sri Lanka, Swaziland and Tanzania, among others, have undergone several reviews and commissions related to recruitment and transfers that were subsequently overturned, partially (and perhaps poorly) implemented, or just ignored (Zafarullah and Khan 1983; National Commission to Review the Working of the Constitution 2000; McCourt 2006; Kim ; Second Administrative Reforms Commission 2007; Jacobs 2009; Kim and Monem 2009; Gundu 2011; Nunberg and Taliercio 2012). In some cases, public service commissions, the very institutions mandated to ensure transparent, fair, and rules-based posting and transfer, are themselves an integral part of the system of MI posting and transfer (McCourt 2006; Kim ).

These governmental commissions, as well as academics, non-governmental organizations (NGOs) and think tanks have proposed ways of lessening MI posting and transfer. Suggested policy and practice interventions include making transfers subject to committee review, strengthening outside watch bodies, (Wade 1985; Second Administrative Reform Commission 2008), creating shared ethical standards for the cadres involved (McCourt 2007; Transparency International 2010), addressing civil servant concerns regarding their standard of living in certain areas (Davis 2003), making posting and transfer decisions more transparent (Davis 2003; Kristiansen and Ramli 2006; DFID 2010), attempting to bring a greater sense of pride and service mentality to the workplace (Davis 2003; Kristiansen and Ramli 2006), eliminating the constitutional protections that civil servants enjoy, developing detailed transfer guidelines, establishing minimum civil servant tenures (Second Administrative Reforms Commission 2007), instituting a system of asset declaration for public servants, publicizing the educational background and qualifications of key personnel, using the private sector to make recruitment and posting more efficient (DFID 2010), randomly distributing transfers to civil servants who have served a certain amount of time (Transparency International India 2005), instituting competitive examinations as the basis for recruitment and promotion, and separating the political and administrative domains (McCourt 2007). In addition, the Organization for Economic Cooperation and Development (OECD), among others, has created broader frameworks to guide the development of legislation governing the civil service (OECD 1996).

Government agencies have initiated mechanisms to lessen MI posting and transfer. For example, the South African Public Service Commission, which oversees recruitment, transfers, promotions and dismissals, maintains a National Anti-Corruption Hotline for Public Service that addresses complaints related to nepotism and favouritism in filling posts, including in the health sector (Public Service Commission 2011). The Indian state of Tamil Nadu created a transparent ‘Counselling’ process, whereby multiple people must be involved in a decision to transfer a health care worker and justification must be documented. Promotions are granted according to seniority (Raha ).

In India, civil servants and others have used existing official mechanisms to expose the MI posting and transfer problem. Indian courts have adjudicated complaints by some servants who felt their transfers were inconsistent with the public interest (de Zwart 1994). However, these cases are difficult to prove, and comparatively few civil servants lodge official complaints. Moreover, the courts are ineffective in the cases of mass transfers. Others have turned to India’s recently passed right to information law. Individual civil servants and activists have successfully filed right to information requests regarding the reasons for particular transfers, the rules for transfers, and the location and function of offices involved in posting and transfer decision-making (www.rtiindia.org).

Besides posting and transfer-specific efforts, governments and NGOs have attempted to reduce MI posting and transfer in the health sector within the framework of human resources for health interventions, as opposed to reforming the civil service system wholesale. Kenya has tried a programme to recruit health care workers through contracts that tie them to specific locations (McPake and Koblinsky 2009), and several countries have compulsory rural service and/or cadres that are exclusively posted to rural areas. Rather than creating mechanisms to prevent transfers, others have looked at ways to make rural posts more attractive, such as providing better housing for health care workers in rural areas, better prospects for career advancement and greater financial incentives to work in rural areas; and enabling health care administrators and workers to choose in which rural area to work (Mullei ; WHO 2010). A key objective of the choice experiments described earlier is to assess the likely influence these inducements would have on shaping health care worker preferences.

However, merely implementing discrete remedies, such as these, can be insufficient, as they fail to sufficiently engage the sub-systems comprising the larger system of posting and transfer (United Nations 1997). Examples of policy resistance proliferate. Despite the creation of the Counselling system in Tamil Nadu, the state High Court reportedly recently mandated that the government redo the Counselling process, as maldistribution of physicians continued (IBN Live 2012). In Indonesia, local governments created ‘fit and proper tests’ to ensure that candidates were selected based on competence. Heads of regions manipulated the tests to ensure that their chosen candidates were promoted (Turner ). These failures suggest that the constellation of power was hostile to the interventions, so they were not sustainable, regardless of whether or not they were initially effective at achieving the stated goal. Although the formal rules regarding transfers had changed, the informal rules remained the same.

Experiences from the efforts described earlier can inform future interventions in posting and transfer practice. However, the scattered literature describing these interventions is not sufficient to guide action. Much of the research and policy interventions assume that posting and transfer practice is the result of individual preferences and incentives. These preferences are certainly important to understanding posting and transfer practice—but in conjunction with a host of other factors. Health worker and administrator preferences are mediated through the preferences of others with influence in the system; indeed, there is a host of preferences that must be negotiated. These negotiations occur in a particular policy context as well as in a larger political economy of power relations. Research should capture the richness of this negotiation. Then, we can design interventions that are disruptive and adaptive.

Research principles

Having a complete view of MI posting and transfer—unmasking the open secret—would aid action. Our proposed research principles should yield more ample, thick descriptions of posting and transfer practice, surfacing the many interests and norms implicated and pointing to ways to promote mission consistency.

Seek an emic understanding of how posting and transfer is negotiated. Emic research will illuminate: the motives of those (transferees and transferors) acting in the system, how MI posting and transfer is experienced by health care workers at all levels, and the informal norms governing the system of posting and transfer. This understanding will help policy-makers to avoid a priori application of simplistic, ideal types, such as whether or not a bureaucracy is merit-based. Emic research will inform efforts that can work with the grain, by understanding the nuances of a particular social, political and economic context, and identifying appropriate avenues for meaningful change. Moreover, it will allow for in-depth exploration of normative concerns that relate to individual health care worker and administrator satisfaction, workplace morale, and esprit de corps. To be sure, very little of the discussion related to informal norms in the literature describes pro-social informal norms, such as altruism or professionalism. Such exploration is crucial for efforts to promote organizational justice and the goal of mission consistency.

Make informal institutions central. Too often, assessments focus on ascertaining existing policy, detecting gaps and then recommending new, better, policies. Understanding the unwritten rules of the game is crucial for understanding why the written rules may not be respected. Moreover, if we seek health systems that are more than mechanistic delivery systems, but that are a means to reaffirm the citizenship and entitlements of those who work within them and those who are served by them, then health systems ‘software’ is at the core of our work (Sheikh ).

Interrogate the diversity of interests and forces that comprise ‘the state’. Theories and policies that guide public sector reform often assume the state is a unified actor. Consistent with systems thinking, research should instead seek to surface the many—often conflicting—interests and norms that shape individual and organizational behaviour within the public sector and the political domain.

Do not avoid politics. Although pragmatism may be necessary in some cases, systematically rejecting the role of politics is naïve. Indeed, explicating the workings of power shaping posting and transfer in different contexts will inevitably raise difficult political issues. MI posting and transfer is a weapon both wielded by and used against the weak; mapping these workings of power is germane to altering them. Posting and transfer can be one of the few areas where health care workers and administrators, particularly those at the bottom of the hierarchy, exercise agency, but it can also be yet another institution that makes them feel powerless. Thus, although research on posting and transfer would undoubtedly touch on many political sensitivities, understanding and addressing these may be essential to the project of organizational justice in the health system.

Acknowledge and accommodate complexity and promote remedial action, by avoiding over-reliance on uni-dimensional theoretical approaches. Models such as principal–agent or an archetypical corruption lens help when they simplify, but not when they are so reductive that they misrepresent reality and offer ineffective, or even harmful, programme interventions. Corruption fails to capture the full range of practices, and risks alienating those most affected by the posting and transfer system. Similarly, principal–agent approaches can foster adversarial rather than co-operative relationships, and inappropriately simplify relationships and interests. If we wish to achieve just workplaces, then we cannot paper over the workings of power that thwart justice. A tightly linked and self-organizing posting and transfer system cannot be changed by separately addressing the two designated sides (transferee and transferor) of the system. Instead, our task is to address the interests and norms shaping the system, with an eye towards ending the health care worker maldistribution that is necessary to reach universal coverage.