The need to align results orientation and accountibility : German bilateral aid implementation

(1)

THE NEED TO ALIGN RESULTS

ORIENTATION AND ACCOUNTABILITY

G

ERMAN

B

ILATERAL

A

ID

I

MPLEMENTATION

Research Project: Managing Public and Private Organizations – What’s the difference? Supervisor: Dr. Anne-Marie Reynaers Second reader: Prof. Dr. Mark Rutgers

Author: Leonie Neuffer Student Number: 11093048 Date: 24 June 2016

(2)

1

Abstract

Results orientation has been the dominant management approach in development cooperation since the 1990s. It is an approach that is part of the New Public Management agenda that sought to make the public sector more efficient. However, for management reforms to become accepted and effective, they need to be aligned with a corresponding practice of accountability. Traditionally, accountability emphasized compliance, rules and procedures, but, according to NPM, accountability should shift focus towards results. By using a theoretical framework that combines different dimensions of accountability and NPM prescriptions of results orientation, this study aims to uncover to what extent the practice of accountability reflects results orientation as prescribed by NPM. This is being studied on the case of German bilateral aid implementation by means of a qualitative analysis of semi-structured interviews. The main research finding is that the rhetoric of results orientation has been adopted, but the way in which accountability is practiced remains close to a traditional compliance-centered understanding. Consequently, there is a need to modernize the understanding and practice of accountability towards a focus on results.

Key words: Results orientation, accountability, German bilateral aid implementation, New Public Management

(3)

2

1. Introduction

In the 1990s, an approach called results-based management (RBM) gained popularity in the aid sector, which was a response to criticisms that donors were operating undemocratic, lacked transparency in the decision-making, and were not subject to public accountability (Missoni and Alesani 2013, p. 271). The adoption of this results orientation was intended to increase and demonstrate effectiveness of aid (Missoni and Alesani 2013, p. 271). Also, this focus on results was part of the New Public Management (NPM) agenda that aimed to make the public sector more efficient and has been promoted prominently by the Organisation for Economic Co-operation and Development (OECD).

One of the most important changes brought about by NPM is a shift away from process values, such as proper administrative process and the rule of law, to a strong emphasis on results and performance (Bemelsman-Videc 2007, p.26). As this shift has implications for all areas of public management, it also requires a rethinking of the concept of accountability. In fact, it means moving from a traditional compliance-oriented understanding of accountability to increased managerial authority in return for results-based accountability. More precisely, traditionally accountability emphasized on process, rules, and procedures, as well as a hierarchical relationship and blaming when things go wrong (Mayne 2007, p. 65). The NPM reform proposed a model in which a results-oriented culture should replace the preoccupation with rules and procedures and should ensure a focus on outcomes and results. Also, public managers should be held accountable for performance to encourage thinking in terms of results. This requires a performance management system that helps define targets, monitoring operation and assessing results. In order to be able to take appropriate action in the pursuit of performance, managers should be empowered, and the public sector deregulated (Verbeteen and Speklé 2015, pp. 955-956).

However, for management reforms, such as results orientation, to become accepted and effective, they need to be aligned with appropriate accountability practice. If accountability practices do not reinforce results orientation, administrative reforms are not likely to become rooted in organizational practice (Romzek 2000, p.37). For instance, if reform rhetoric stresses employee discretion and autonomy, but performance reviews stress rules and process, then employees will focus on rules and process in their work efforts (which in turn may jeopardize performance oriented change). Thus, it is important that the practice of accountability reflects management emphases and administrative tasks at hand (Romzek 2000, p.39). Appropriate accountability mechanisms are essential because accountability has the potential to ensure democratic control and the legitimacy of governance.

(5)

4 Germany had been known for the traditional administrative tradition strongly coined by Max Weber, yet has taken up the NPM reforms. In fact, its development sector has strongly embraced the focus on results. Germany has a separate Ministry of developmental affairs and economic cooperation (BMZ) which does not have implementing capacities and therefore, mainly relies on the two publicly owned aid organizations. These are the German Organization for Development Cooperation (GIZ), which focuses on technical cooperation, and KfW Development Bank, which concentrates on financial cooperation. Both emphasize results and impact orientation as the basis of their work in development projects and programs (KfW 2015; GIZ 2016). However, it remains to be seen in how far Germany has been able to move its practice of accountability beyond the principles of the traditional administrative model towards results-oriented accountability.

This study, therefore, aims to uncover the following question: To what extent does the

practice of accountability reflect NPM prescriptions of results orientation in German bilateral aid implementation? Firstly, answering this question involves defining these NPM

prescriptions and examining whether they are (or can be) lived up to in practice. Secondly, it means investigating how accountability is practiced and to what extent this practice exhibits the NPM prescriptions. This means to clarify, if and why accountability practice remains close to the principles of the traditional bureaucracy or to the principles of NPM.

Societal and Academic Relevance

This study seeks to contribute to the academic literature by using a theoretical framework to create a systematic understanding of the practice of accountability in a NPM results context by using the case of German bilateral aid implementation. Few studies have explored the accountability mechanisms of aid implementing organizations so far. Usually, accountability in aid concentrates on the partner government in a developing country and how to establish good governance, including accountability. Little attention is being paid, in literature and among practitioners, to the accountability regime of the donor organizations themselves. The research findings of this thesis may assist aid agencies to get a clear grip on how they practice accountability, which can serve as a basis to find ways to further increase their responsiveness to various stakeholders. The complexity of accountability relationships that aid agencies work in make it particularly important to study and get a clear and more structured understanding of how accountability is practiced and how and why it reflects NPM results orientation. Thus, by providing empirical evidence of how accountability in its different dimensions looks like and the extent to which its practice reflects results orientation, as well as by shifting attention to the

(6)

5 bilateral aid agencies themselves, this study, therefore, aims to make an important contribution to the literature.

Structure of this Study

This study starts by giving a review on the central theoretical concepts under investigation – results orientation and accountability. It continues by establishing a theoretical framework that will be the basis for the empirical study. This framework consists of different accountability dimensions that exhibit ideal-type NPM results orientation, as well as traditional compliance-based accountability serving as a counterfactual. Next, an overview of the methodological approach is given. This qualitative study is based on data raised through semi-structured interviews at GIZ and KfW Development Bank and analyzed by means of coding. The findings are structured according to the theoretical framework. Therefore, the practice of accountability is examined by assessing the focus on results in each accountability dimension, as well as by scrutinizing the practice of NPM prescriptions regarding results-based accountability. This is done for the KfW and GIZ separately. Finally, the findings are being discussed in relation to existing literature on NPM and accountability. The study concludes that the practice of accountability remains close to compliance-oriented traditional understanding. This is because NPM prescriptions, such as using performance information to hold managers accountable and giving them discretion in managing for results are difficult to put into practice. Results orientation is formally adopted, but has not been supported with corresponding accountability mechanisms. This means that, in practice, no one is held accountable for actual results, but instead for compliance with in the usage of results tools, as well as rules and procedures.

(7)

6

2. Theoretical Framework

In this chapter, the central concepts of Results-Based Management (RBM) and accountability will be discussed. The section on RBM starts with the reasons why it is a popular concept in the aid sector, and then continues to outline its origins in the context of the aid effectiveness debate and NPM reform agenda. Also, the results chain will be explained, which is the underlying logic of RBM. Besides that, the difference between RBM and results-based approaches will be made clear. With regard to the second concept, accountability, an overview of its different dimensions, as well as a definition will be given. Besides that, an analytical framework will be developed that combines different accountability dimensions with NPM results orientation and traditional compliance-based accountability. More precisely, firstly, accountability in a traditional input- and compliance-oriented Weberian public administration will be presented. This approach will serve as a counterfactual. Secondly, accountability for results according to NPM is surveyed. This requires sufficient discretion for managers to manage for results, as well as a functioning performance measurement to track progress and to have a basis for holding managers accountable. Finally, various challenges in fulfilling those two requirements are being discussed. This framework will be used to structure and discuss the findings.

2.1 Conceptualizing Results-Based Management (RBM)

Numerous reports and discussion papers by aid agencies describe a so-called aid fatigue and an emerging debate on aid effectiveness towards the end of the 1990s (Vähämäki 2011 et al., Missoni and Alesani 2013). Against the backdrop of lessons, such as difficulties of project aid, low levels of ownership, and unclear development results, results-based management (RBM) approaches gained popularity (Vähämäki et al. 2011, p.15). However, RBM not only needs to be understood in terms of a response to pressure from the public in donor countries to demonstrate the effectiveness of aid (Bester 2012, p.8) but also as part of an effort to introduce private-sector management techniques into development agencies (Obser 2007, p.2).

Since the 1980s, the New-Public-Management-reform agenda became popular in OECD countries, for example, in Germany, the Netherlands, and Great Britain, and eventually also took hold in development agencies, particularly bilateral donors (Best 2014, Obser 2007, Bester 2012). Of the ten principles to ‘reinvent government’ that Osborne (1993) puts forth, the most important NPM principle in development aid is ‘results-oriented government.’ It entails that accountability must shift from inputs to outcomes and results and requires an (entrepreneurial) government or public organization that measures the results of an agency’s actions while at the same time establishing incentives to achieve those results (p.353). RBM approaches were established as a means to check performance and to hold public agencies accountable (Obser

(8)

7 2007, p.2). According to Binnendijk (2000), the most central features of public sector reforms in the 1990s were improving performance and ensuring that governments achieve the desired results (p.6).

RBM has been preceded by Peter Drucker’s concept of Management by Objectives in the 1960s and 1970s (Bester 2012, p.8). Meier (2003) thus argues that RBM needs to be seen as an evolution not a revolution in management, with origins in management sciences and closely linked to earlier efforts to implement the Management-By-Objectives approach (p.4). Similarly, Obser (2007) claims that RBM is a combination of the traditional approach of monitoring implementation with the assessment of results (p.2). Performance-informed RBM links implementation progress with progress in achieving the desired objectives or results of development projects and programs (Obser 2007, pp.2-3). Performance measures thereby serve as a means to assess the extent results achievement.

Results-based management (RBM) in development cooperation is not a homogenous concept, as a singular RBM model does not exist (Vähämäki et al. 2001, p.10). The OECD Development Assistance Committee (DAC) defines RBM as “a management strategy focusing on performance and achievement of outputs, outcomes, and impacts” (OECD/DAC 2002, p.30). A discussion paper commissioned by the Canadian International Development Agency expanded that definition to the following (Meier 2003, p.6):

“Results-Based Management (RBM) is a management strategy aimed at achieving important changes in the way organisations operate, with improving performance in terms of results as the central orientation. RBM provides the management framework and tools for strategic planning, risk management, performance monitoring and evaluation. It’s primary purpose is to improve efficiency and effectiveness through organisational learning, and secondly to fulfil accountability obligations through performance reporting.”

With that definition, Meier (2003) aims to move towards a common understanding of RBM among development cooperation agencies. It remains the most commonly used definition of RBM in the literature and will, therefore, be adopted in the context of this study (e.g. Missoni and Alesani 2013; Vähämäki et al. 2011). Whenever the term results are used in this study, it refers to outputs, outcomes, and impacts in line with the definition of the OECD/DAC (2002) definition from above. RBM in German bilateral aid is called “wirkungsorientiertes Management” (GTZ 2008, p.2), which is why RBM in this study is mostly referred to as results orientation and focus on results. Because the terms results orientation and focus on results are used more commonly in German bilateral aid, it is those terms that will be found predominantly in the chapter of the findings.

The logic of RBM is often depicted in a results chain (table 1). Following the reasoning of planning in aid, the activities of a project and the resources invested or used (inputs) are expected to

(9)

8 produce certain outputs. The results entail the immediate use of outputs by the intended target groups, as well as the direct benefits or outcomes, and the more far reaching results, namely the long-term overarching developmental impact, which can only indirectly be attributed to the activities of a project. Those impacts are intended to contribute to achieving the highly aggregated results at development policy level, such as the Sustainable Development Goals (GIZ 2012a, p.9).1_{These different levels of results will be used in the case analysis, where it will be} investigated on which of those levels the practice of accountability has its focus on.

HOW should this be implemented? WHAT should be produced? WHAT

do we expect from this intervention?

WHY should we do this?

Inputs Activities Outputs Use of

outputs outcomes Direct overarching Long-term developmental

impact RESULTS

Table 1. Results chain (adapted from Meier 2003, p.7 and GIZ 2012, p.53)

It is important to note that this study only pays attention to the focus on results, not results-based approaches. Results-results-based approaches are an advanced version of the focus on results and is based on linking results achievement to financial resources allocation. This means payment of a development intervention depends on the achievement of a priori determined result, e.g. a certain amount of students passing a standardized performance test, a certain amount of women receiving medical care during child birth, or transparent and participatory operations of a communal administration (Blume 2012). This approach has been implemented and tested in pilot projects at KfW and GIZ, but it is too early to study its impact on the practice of accountability and therefore is not considered in this study.

2.2 Conceptualizing Accountability

As stated above in the definition of results-based management by Meier (2003), RBM in aid has the aim to fulfill accountability obligations. However, before looking into the relationship between accountability and results-based management, it is important first to grasp the concept of accountability alone.

Accountability in contemporary Public Administration is an “elusive” concept and lacks a shared definition (Sinclair 1995, p. 219). Even though there is a great interest in accountability, at the same time, there is much confusion about what accountability exactly means (Day and Klein

1_{The Sustainable Development Goals that were adopted on 25 September 2015 at the United Nations Sustainable} Development Summit include 17 targets to end poverty, fight inequality and injustice, and tackle climate change by 2030. The Sustainable Development Goals, build on the Millennium Development Goals (MDGs), eight anti-poverty targets that the world committed to achieving by 2015 (UNDP 2015).

(10)

9 1987, p.1). Accountability is invariably seen as a “good thing” and the absence of accountability as one of the greatest abuses in human history. Accountability is closely linked with the exercise of power, the legitimation of policies, and those pursuing them (Lonsdale and Bemelsman-Videc 2007, p.3). However, a lack of clarity of the concept limits the meaningfulness of the term (Perrin, Bemelsman-Videc and Lonsdale 2007, p. 241). This makes it important to study the concept and to shed light on different ways to practice accountability.

2.2.1 Dimensions of Accountability: How to Study Accountability

There are numerous ways to approach the concept: Bovens (2010) distinguishes between accountability as a virtue and accountability as a mechanism. The former refers to a normative set of standards for evaluating the behavior of public actors, while the latter describes an institutional arrangement in which an actor is held to account (Bovens 2010, p.946). Willems and Dooren (2011) identify two different interpretations of accountability – one that points to answerability and another that refers to managing expectations, whereby the latter is a more process-oriented interpretation than just the actual outcome. Other scholars, in turn, understand accountability in terms of different forms, such as political, public, managerial, professional, and personal accountability (Sinclair 1995, p. 219).

Bovens (2007) claims that there are four important questions to be asked with regard to accountability. Firstly, the question of whom is an account to be rendered to? This relates to the type of forum to which the actor is required to render account. Secondly, there is the question who should render account? Thirdly, the question is about what is an account to be rendered? This concerns the aspect of the conduct of which information is to be provided, such as e.g. financial, procedural or programmatic accountability. The fourth question concerns that of why the actor feels compelled to render account? This refers to the nature of the relationship between the actor and the forum, and the nature of the obligation, for instance, obligations arising from a hierarchical relationship, a contractual agreement or from voluntary agreements (pp.454-455). A commonly used definition of accountability is the one by Bovens (2007) who defines it as a social relation (p.452):

“A relationship qualifies as a case of accountability when: 1. there is a relationship between an actor and a forum, 2. in which the actor is obliged, 3. to explain and justify, 4. his conduct; 5. the forum can pose questions; 6. pass judgement; 7. and the actor may face consequences.”

This definition leaves open the nature of the conduct – be it following rules and procedures or be it the achievement of results, which is the reason why this definition is chosen to serve as a basis for this study.

(11)

10 2.3 Accountability Framework: Two Approaches to Accountability

In order to study the practice of accountability in German bilateral aid and the extent to which it reflects NPM prescriptions of results-based management, a theoretical framework is developed (see table 2). It consists of different accountability dimensions, which are inspired by Bovens (2007), that include: the accountable actor, the accountable forum, the nature of the conduct, the accountability instrument and sanctions and incentives. For each of the dimensions, the ideal type accountability in the context of input- and compliance-oriented traditional public administration, as well as results-oriented accountability of NPM, are presented and will be further explained in the following section. Accountability needs to be studied in its context because only the context gives meaning and life to the concept. The context of NPM is chosen, as results orientation is an NPM principle and the objective of this study is to find out the extent to which accountability reflects NPM results orientation. Accountability in traditional public administration, in this case, serves as the counterfactual. In addition to that, the framework includes two requirements that are necessary for accountability for results according to NPM and that are identified from the literature. This framework serves as a means to clarify, if and why accountability practice is close to the principles of the traditional bureaucracy or to the principles of NPM.

Table 2. Accountability dimensions in traditional public administration and NPM, as well as their elements.

Accountability

Relationship Nature of

Conduct Accountability Instrument and Sanctions Incentives Accountable

Actor Accountable Forum

Who is accountable? To whom is an account to be rendered? About what is an account to be rendered?

How is the actor held accountable by

the forum?

How are actors motivated to meet

expectations and what are consequences if they do not do so?

Traditional Accountability Hierarchical and collective accountability (control from top to bottom) Superiors in the bilateral aid agency Contracting party Tax payer Beneficiaries Partners Rules and procedures Input Audit and control Sanctions for non-compliance with organizational rules or legal commitments NPM Accountability for Results Project manager (individual managerial accountability) Output Outcome Impact Performance measurement Incentives to encourage management for results

NPM Requirements for Results Orientation

1. Performance information to allow performance measurement, which forms the basis for accountability for performance

2. Discretion and managerial authority for managers to be able to manage for results

(12)

11

2.3.1 A Traditional Understanding of Accountability

The classical model of administration is strongly influenced by the administrative tradition of Germany and the principles of bureaucracy articulated by Max Weber. Mayne (2007, p. 65) summarizes traditional accountability in the following way:

“– Focusing on process, especially on the extent to which administrative rules and procedures have been followed.

– Hierarchical, with a junior party being accountable to a senior party. – Assigning blame when things go wrong.”

What accountability means in this model for the individual dimensions is presented in the following.

A. Accountable Actor

In the traditional bureaucracy, public administrators are integrated elements of responsible, collective public bodies (Lægreid 2014, p. 326). Control from top to bottom is emphasized, which means that policy is set at the top and carried out through a series of offices, with each employee reporting to a superior and held to account by that person (Pfiffner 2004, p.443). In this system of hierarchical control, the bureaucrat is strictly subordinate to the political superior and is required to follow a set of rules and regulations flowing from the public law (Pfiffner 2004, p.443).

B. Nature of the Conduct

In a traditional Weberian control and accountability system, the dominant focus lays on inputs and procedures, as well as hierarchy and values such as legality, fairness, integrity, procedural correctness and economy (Verhoest 2005, p.138). The traditional model stresses accomplishing the mission and accountability for resources. Thereby, particularly the inputs (e.g. number of personnel, money spent, the number of vehicles, or the numbers of computers) used to accomplish that mission are emphasized. For those measures, accountability is easy to establish, and they help to assure that resources are not being stolen, but used for the intended purpose (Pfiffner 2007, p.450).

C. Accountability Instrument

The traditional public administration operates with legal programming through public law, budgetary provisions, audit controls and judicial review (Goetz 1999, p.169).

(13)

12

D. Incentive and Sanctions

The traditional approach to accountability focuses on assessing compliance with predetermined procedures, organizational rules, legal commitments, and for executing and perhaps accomplishing what has been initially agreed to (Perrin, Bemelsman-Videc and Lonsdale 2007, p.252). Such an accountability system thereby aims to protect against the abuse of power and the misuse of resources, thus avoiding the negative aspects of delegated authority. It concentrates on providing feedback through mostly negative response (sanctions), instead of a more positive response (rewards or incentives) (Zapico-Goñi 2007, p.222).

2.3.2 Accountability and a Results Focus as Prescribed by NPM

The NPM according to the OECD rests on three pillars: Firstly, it involves establishing a results-oriented culture and thereby replacing the preoccupation with rules and procedures with a focus on outcomes and results. Secondly, public managers need to be made accountable for performance, in order to encourage thinking in terms of results. This demands performance information, in order to define outcome objectives, to monitor operations and to assess performance. Besides, it requires incentives that are linked to performance. Thirdly, public organizations should accomplish internal deregulation with the aim to empower managers to take appropriate action in pursuing performance (Verbeteen and Speklé 2015, pp. 955-956). Thus, the core idea of managing for results according to NPM is using performance information to increase performance by holding managers accountable for precisely determined goals and providing them with adequate authority to achieve those goals (Moynihan 2006, p.78).

NPM and its focus on results have provoked claims about several changes in accountability relationships and mechanisms. While it does predict more or less accountability, it rather implies a transfer from one set of accountability relationships to another. Lægreid (2014) maintains that traditional Rechtsstaat accountability relations retain their importance, but have been supplemented by NPM accountability relations that produce accountability tensions and dilemmas for civil servants. Without a classical Weberian model that concentrates on rules, legality, impartiality, due process, ethical capital and public ethos, NPM reforms are problematic (p.328). Therefore, NPM should not be seen as an alternative to Weberianism, but as a shift in emphasis from rules and procedures to managing for results.

A. Accountable Actor

With regard to the question of ‘who’ is accountable, scholars have noticed a shift away from collective accountability towards individual accountability. NPM strengthens individual managerial accountability, which encompasses that each person acts for himself within strategic

(14)

13 guidelines and according to self-interest and is personally accountable for the results (Lægreid 2014, p. 325).

B. Nature of the Conduct

The first and most obvious change is that NPM and its focus on results have shaped ‘for what’ a manager is accountable. Accountability no longer focuses on processes and compliance with rules and inputs (finances) and on fairness, equity, and impartiality, but more on performance and results, increased efficiency, competition, and cost effectiveness (Lægreid 2014, p. 326). NPM rejects measuring inputs and emphasizes the use of performance measures to evaluate projects and management. According to NPM, accountability for resources plays a less important role than the accomplishment of goals at a given cost (Pfiffner 2004, p.8).

C. Accountability Instrument

According to NPM, performance measures are sought to be the tool through which accountability for results is achieved (Pfiffner 2004, p.447). The success of managers is measured by their performance with regard to accomplishing goals, rather than in the prudent accounting for the resources and inputs used (Pfiffner 2004, p.451). Performance measurement emphasizes performance indicators, which are supposed to be measurable as objectively as possible (OECD 1997, p.98).

C. Incentive and Sanctions

Performance-related incentives schemes need to be adopted to support the focus on results and outcomes (Nielsen 2014, p.433). Despite the fact that intrinsic motivation to meet performance expectations and professional pride are mentioned as motivators, the significance of extrinsic incentives cannot be exaggerated: ‘performance measures are only relevant if there are consequences for over- or underperformance against targets’ (OECD, 1997, p. 22). Such external incentives can take different forms, such as a performance-based promotion or pay-for-performance compensation schemes (OECD, 2007).

E. Two Conditions for Accountability for Results

A review of the literature reveals two prerequisites for establishing accountability with a focus on results in line with NPM. Those two requirements and the accountability tensions and dilemmas for civil servants they produce will be examined in the following part.

(15)

14

1) Performance Management to Assess Managers Progress in Managing for Results

An important reason for the attractiveness of performance measurement is the idea that performance measures can help to overcome some of the seemingly inherent characteristics of public agencies: Performance measures aim to serve as a bottom line to substitute vague and contradictory goals and could help transcend the information asymmetry of agents (Moynihan et al. 2011, p.142). Internally, performance management could contribute to specifying vague and often symbolic missions by translating them into action and by producing performance feedback to adjust goals and actions. Externally, performance regimes could offer access and transparency to government, facilitate citizen participation, improve accountability, equity, and responsiveness of government (Moynihan et al. 2011, p.153).

Limits of Performance Measurement:

Potential Reasons Why Accountability and Results Focus Are Not Aligned

Several challenges to performance management that may explain why accountability and a focus on results are difficult to align are identified. A lot of the challenges relate to the assumption that projects are simple enough for some measures to represent performance accurately. But in practice, and particularly in development aid, tasks of public managers are shaped by complexity, and multiple goals, different perspectives and interests of stakeholders, and therefore not easily represented with performance measures.

Steward and Walsh (1992) argue that performance management assumes that managers can be given clearly bounded tasks, performance targets and be held accountable for the use of resources to achieve those tasks (p.513). However, the assumption that the tasks of managers can be bounded and measured is problematic. Besides that, the extent to which adequate or complete performance measures can be found in the public domain is limited, as performance assessment is at the heart political, subject to public discourse and the measure adopted may serve a particular interest (Steward and Walsh 1992, pp.513-514). Also, performance management can potentially restrict public managers if it limits their capacity to learn and adapt (Steward and Walsh 1992, p.513).

According to Lægreid (2014), stronger managerial accountability through performance measurement often causes the burden of internal scrutiny and paperwork to increase. Stronger accountability often suggests that operational managers need to provide data on performance indicators or performance audits. This effect has been labelled as re-bureaucratization, audit explosion, increased regulation inside government, or target overload (p.327). Furthermore, Lægreid (2014) argues that by focusing on specific outputs, public managers tend to ignore outcomes and emphasize efficiency over effectiveness. It is difficult to hold managers

(16)

15 accountable for a focus on outcomes, because, on the one hand, attributing the outcome to actual behavior and actions is problematic, and on the other hand, because of difficulties of timing. Often, it is challenging to link outcomes to annual accountability assessment, since outcome usually manifests itself over longer periods of time. That is why accountability systems tend to favor outputs, as they are easier to measure, less costly to monitor and easier to hold people accountable for. NPM, in fact, excessively stresses outputs. (p.327).

Behn (2011) claims that performance audits tend to focus more on compliance than on performance, often picking out small errors and that it is biased towards punishment and hierarchy, which undermines public trust.

Also, Perrin (2015) argues that performance measures can fail to provide a meaningful account of actual performance. When meeting indicators becomes the objective, “making the number” may become more important than doing what the program is originally intended to be doing (pp.189-190). Besides, holding managers accountable to meet targets may cause a strong pressure to misrepresent and distort results (p.190). Pfiffner (2004) maintains that the difficulty of performance measurement lies in choosing the correct indicators that validly measure that the project is intended to accomplish (p.451).

In sum, according to the literature, the following assumption of performance measurement appear to be problematic: the idea that managers can have clearly bounded tasks, that complete performance measures can be found, that management can be isolated from the political process, and that outcome can be attributed to actual behavior and the annual performance assessment. Also, performance measurement can be challenging, as it often represents a bureaucratic burden, may limit manager’s capacity to learn and adapt, results in an excessive focus on outputs and compliance, and may overemphasize indicators over actual objectives, and can create pressure to misrepresent data.

2) Managers Need Discretion to Manage for Results for Which They Are Held Accountable

One requirement for results-based accountability identified in the literature is that managers receive sufficient discretion to manage for results. According to Nielsen (2014), increased managerial authority should be granted in return for stronger results-based accountability, which is a fundamental prescription of NPM (p.434). The idea is to grant public managers the flexibility necessary to engineer performance-oriented change. Without managerial authority, public managers may be aware of strategic goals and may have access to performance information, but are unable to use this information to change routines and structures to improve performance (Nielsen 2014, p.432). Results-based accountability cannot occur if public managers lack authority over the process, and ultimately, outputs (Moynihan 2006, p.84). The potential for performance

(17)

16 systems in which public managers are formally held accountable for results may be jeopardized, if financial and personnel systems continue to exist that emphasize compliance and error avoidance (Moynihan 2006, p.84).

Moynihan (2006) theorizes that high managerial authority combined with a high focus on results is the so-called managing for results ideal type, in which managers have clear goals and authority to achieve goals. This, in turn, facilitates program effectiveness, higher technical efficiency and results-based accountability (p.84). In his study, though, he finds that in practice, reformers have adopted the language and expectations of managing for results, but actually only adopted the reform partially: they implemented an increased focus on results but neglected an increase in managerial authority (p.78). Consequently, low managerial authority combined with a high focus on results leads to pressures for performance, but public managers have limited discretion to engineer change. In that case, the lack of managerial authority undermines the scope of performance improvement and potential for results-based accountability (p.84).

Besides that, Verbeteen and Speklé (2015) note that NPM distinguishes between operational and strategic autonomy. Operational autonomy involves decentralizing decision rights in operational matters, while strategic autonomy refers to defining goals and priority setting in policy affairs. In NPM thinking, operational autonomy is desirable, as it intends to provide increased performance and flexibility (Verbeteen and Speklé 2015, pp.955-956).

Challenge: Distinguishing Managerial Discretion and Political Discretion

The demand for more discretion, authority, and flexibility for managers raises the question of in which area managerial authority should be given. NPM assumes that it is possible that policy makers make policy, and then delegate the implementation to managers that are held responsible by contract (Kettle 2006, p.457). In that way, politicians work as “chief executives,” assuming a strategic role. They formulate general goals and assess results without being involved in daily implementation. On the flip side, civil servants operate as managers at arm’s length from politicians and are held accountable though performance systems and incentives (Lægreid 2014, p.326).

However, this dichotomy between policy-making and management is problematic to put into practice. This is because it assumes that a clear division between politics and administration is feasible, yet in reality, a lot of the work of public administration is political. Also, Lægreid (2014) maintains that the question whether executive politicians are willing or able to operate solely as strategic formulators is problematic. They prefer involvement in details of implementation (p.327). Civil servants tend to lose sight of the unified public service, which increases the distance towards political executives and thus reduces responsiveness and accountability (Lagreid 2014,

(18)

17 p.328). Also, civil servants can never be isolated from the political process, because their actions in the public domain can and should be subject to public criticism, to which the political process should be able to respond (Steward and Walsh 1992, p.513). Consequently, the politics-administration divide becomes blurry and makes administrative and managerial accountability difficult and demanding (Lægreid 2014, p.329).

Given the differences between both approaches – rules and procedure-oriented accountability and results-oriented accountability, it is important to apply both concepts coherently in the implementation. This means either emphasizing inputs compliance with rule and procedures like in the traditional public administration, or embracing the results focus as prescribed in NPM by fulfilling the two requirements that enable a clear accountability chain.

The following definition by the OECD/DAC (2002, p.11) presents the understanding of accountability in development aid and is, in fact, an example of how elements of the different approaches to accountability are thrown together, an example of how accountability and the results focus are not very well aligned.

“Obligation to demonstrate that work has been conducted in compliance with agreed rules and standards or to report fairly and accurately on performance results vis a vis

mandated roles and/or plans. This may require a careful, even legally defensible, demonstration that the work is consistent with the contract terms. Note: Accountability in

development may refer to the obligations of partners to act according to clearly defined responsibilities, roles and performance expectations, often with respect to the prudent use of resources. For evaluators, it connotes the responsibility to provide accurate, fair and

credible monitoring reports and performance assessments. For public sector managers and policy-makers, accountability is to taxpayers/citizens.” [emphasis added]

This definition, on the one hand, encompasses a compliance-emphasizing accountability that is based on contractual relations – reflecting a traditional understanding of accountability. On the other hand, it shows indications of an NPM-oriented accountability that focus on performance results. This mixed approached to accountability at the OECD leaves us to predict that accountability and results focus are still lacking alignment in German bilateral aid as well. This will be tested in the following case study.

(19)

18

3. Research Design

The research starts out with a theoretical framework for accountability, in order to structure and operationalize the study of the way accountability is practiced. Combined with theory on NPM prescriptions regarding results orientation, this framework guides the research process. The aim is to generate a systematic understanding of how far the practice of accountability reflects NPM results orientation in German bilateral aid implementation, as well as potential explanations for it. In other words, this study intends to (inductively) inform existing theory with new insights (regarding the practice of accountability) and to (deductively) test theoretical propositions of NPM prescriptions of results orientation.

3.1 The Case Study: German Bilateral Aid Implementation

A case study approach is chosen, as it can help to “derive richer, more contextualized, and more authentic interpretation” of a complex phenomenon like the relationship between the practice of accountability and results orientation in aid (Bhattacherjee 2012, p.93). Before looking into the findings of this, some background to the case will be provided. Firstly, the structure of German bilateral aid will be explained. Secondly, the basis for the results focus on German aid will be presented. Thirdly, light will be shed onto the complex web of accountability relationships in which bilateral aid works. Fourthly, criteria for the case selection will be considered.

The Structure of German Development Cooperation

In contrast to the majority of OECD (Organisation for Economic Co-operation and Development) countries, Germany has a separate Ministry of developmental affairs and economic cooperation, namely the Federal Ministry for Economic Cooperation and Development (BMZ). This Ministry, however, does not have implementing capacities. Therefore, it commissions publicly owned, as well as private organizations to implement projects and programs for the government.2_{This means that German bilateral aid sector consists of the} German (publicly owned) organization for development cooperation called Gesellschaft für Internationale Zusammenarbeit GmbH (GIZ) focused on technical cooperation and implementation of project and programs, and KfW Development Bank, which concentrates on financial cooperation and financing of development projects and programs. The clear division between financial and technical development cooperation is a characteristic of German aid (Hüging and Klinnert 2013, p.381).

2_{BMZ is the main commissioning or contracting party of GIZ and KfW Development Bank, but there are also} other Ministries or even private actors that commission GIZ and KfW Development Bank to implement or finance project and programs.

(20)

19 KfW Development Bank has the goal to improve social and economic conditions in developing countries by giving credits and grants to developing countries on behalf of the German government. The resources provided by the government are complemented by KfW Development Bank’s own market funds that are given to countries depending on their degree of development (Hüging and Klinnert 2013, p.385). Next to this mixed form of financing, KfW finances grants consisting exclusively of own KfW Development Bank resources to market-consistent conditions (Hüging and Klinnert 2013, p.386).

GIZ, on the other hand, focuses on capacity development and advising people and organizations on learning and change processes. Thereby GIZ aims to help organizations, public authorities and private businesses to optimize their organizational, managerial and production processes. Besides that, GIZ advises governments on how to achieve objectives and implement nationwide change processes by incorporating them into legislation and strategies (Hüging and Klinnert 2013, p.387). GIZ, working in more than 130 countries with about 17.000 employees, develops and supports project and programs related to vast amount of different topics, such as health, education, agriculture, economy and employment, promotion of democracy, peace, and security, reconstruction and civil conflict management, environment, climate protection, transportation, communication, emergency aid and migration (Hüging and Klinnert 2013, p.387).

GIZ and KfW Development Bank will be portrayed separately in the case study, as they differ slightly due to their different nature relating to technical and financial cooperation, respectively. However, despite the difference in the nature of their organizations, this study aims to find a common pattern with regard to the practice of accountability in German bilateral aid.

Results Focus in German Bilateral Aid

The focus on results in German bilateral aid was formalized with the introduction of the developmental commissioning framework called AURA (entwicklungspolitischer Auftragsrahmen) which was introduced in 2002. This result-oriented commissioning framework between GTZ (Deutsche Gesellschaft für Technische Zusammenarbeit)3_{and BMZ (Federal Ministry for Economic} Cooperation and Development) focuses on outputs and outcomes of projects and programs, which were also reflected in a new reporting system. Previously, the contact clarifying the relations between BMZ and GTZ was input-oriented which meant that in a project offer, the services needed for the implementation were listed and then executed. With the introduction of AURA, the focus was shifted to results and impacts, so that in project offers no longer the promise of delivering agreed services is central, but the “promise of impacts.” This procedure is

3_{On January 1, 2011, the organizations GTZ (Deutsche Gesellschaft für Technische Zusammenarbeit), DED} (Deutscher Entwicklungsdienst), and InWEnt (Internationale Weiterbildung und Entwicklung) were merged to GIZ (Gesellschaft für Internationale Zusammenarbeit GmbH).

(21)

20 also valid for GIZ since the merger, as well as for KfW Development Bank. It aims for a clear demarcation between the tasks of BMZ, which is responsible for developmental and policy decisions, and GIZ and KfW, which are responsible for operational activities. Furthermore, the new procedure intends a higher degree of flexibility in the implementation, results and impact orientation and shared responsibility in the implementation between the partner in the developing country and GIZ (Hüging and Klinnert 2013, p.388). This means that GIZ and KfW are not accountable for the specific procedure that they describe in their project offers, but for acting in the most optimal way towards achieving the targets.

The Complex Accountability Relations in Bilateral Aid: An Accountability Triangle Bilateral aid agencies work in a complex (at least) triangle of accountability relationships (see figure 1). The accountability triangle is a conceptual framework for the analysis of the provision of social services put forth by the World Bank in the World Development Report 2004. While exploring the details of the logic of the concept goes beyond the scope of this study, here, it suffices to say that it serves as a means to illustrate the different (and complex) accountability relations that development cooperation operates in. Usually, accountability works in a relationship in which one party is paying and the other party delivers a service in return. However, in aid, the contracting party who is paying for an intervention is not the same actor, as the one benefiting from it. In fact, poor people often lack direct access to service providers. This results in an accountability triangle, in which there is a direct relationship between the beneficiary and the service provider (GIZ or KfW), as well as a direct relationship between the contracting party (BMZ) and the service provider (GIZ and KfW), and a more indirect relationship between the contracting party (BMZ) and the beneficiary in the developing country. In this study of the practice of accountability, these main relationships – and due to limitations of scope, only these– will be examined.

When investigating the accountability practice at KfW Development Bank, next to internal accountability, the focus will lay on the relationship between KfW and the contracting party. This is because responsibilities and accountability in project management at KfW is distributed somewhat differently due to the fact that KfW is a bank. More concretely, KfW provides financing for the partner (e.g. a Ministry in Vietnam) to contract a consultant who has the technical knowledge to implement a project. This puts KfW in a more supervisory role, as it is not directly responsible for the implementation. GIZ, on the other hand, implements projects and programs itself and works more directly with the partner in a sense that they are both responsible for implementing a project or program. Thus, when studying the accountability practice at GIZ, the relationship to the contracting party (BMZ), as well as to the partner will be

(22)

21 considered. Given the complexity of this web of accountability relations, this study will look at each of the accountability relations individually when analyzing the extent to which they reflect results orientation.

Figure 1: Accountability Triangle (the size of the arrows reflects the strength of the relationship)

Criteria for Case Selection

The case of German bilateral aid implementation is chosen on the following criteria: a bilateral aid sector that has embraced result orientation, and that is from a country that has been under the influence of NPM. GIZ and KfW Development Bank are selected as a case for this research because they are the publicly owned organizations of the German aid sector that are concerned with aid implementation. Therefore, both institutions are chosen as a typical case aiming to provide insight into the complex relationship of accountability and results orientation in aid (Gerring 2007, p.91). By sketching this single-case of German bilateral aid as a two-case design consisting of the two organizations GIZ and KfW, the overall study of German bilateral can be regarded as being more robust (Yin 2014, p.67).

3.2 Data Collection

Data in this study is obtained through interviews, which serve the purpose of finding out the actual practice of accountability. More precisely, in the course of this study, six semi-structured interviews are conducted with employees at GIZ, and three interviews are conducted with employees at KfW Development Bank. Interviews are able to provide thick and in-depth data,

Implementing agency (GIZ or KfW)

Partner in the developing country (e.g.

a Ministry in Vietnam) Commissioning party

(e.g. a German Ministry

such as BMZ) Mutual

accountability

Population in the partner country (e.g.

Vietnam) Taxpayer in the

donor country (Germany)

(23)

22 which is useful to receive insights into intangible notions such as accountability. Also, because this research aims for insights of the actual practice of accountability, interviews seem as a more appropriate method than reports, which may reflect only ideal-type accountability and thus a distorted picture of reality.

Criteria for the Selection of Respondents

Interview candidates are selected by the means of purposive sampling, since the aim is to get a sample with various different perspectives on the topic, in order gain a thorough understanding of the issues at stake. This means both, corporate- and project-level views, as well as, implementation- and policy-level views, are included. Interviewees at GIZ include employees with experience at the headquarters in Bonn and Eschborn, as well as in projects, programs, and the country offices in the aid receiving countries. At KfW Development Bank, all employees, therefore also all interviewees, are based in the headquarters in Frankfurt. With regard to the organizational positions of each interviewee, staff members of GIZ at various stages of the accountability chain that deal with results-based management are interviewed. On the country level, this includes program/project managers, advisors, and mangers at the country office. At the headquarter level (Bonn and Eschborn), this includes the evaluation corporate unit and staff in the different departments. Besides attempting to maximize range, the sampling is conducted with a certain extent of pragmatism. This means that the personal network of the researcher and snowballing technique helped to attract respondents with the required experience in project management and expertise on the topic of accountability and results orientation. Despite the fact that convenience sampling may not be ideal for generalization, using this technique can still be justified to some degree (Weiss 1994, p.26). Firstly, it allows the respondents’ own assessment of generalizability. Secondly, it permits the interviewer’s own identification of others worth recruiting and thirdly, it is the idea that a certain amount of universalism with regard to the phenomenon studied, exists among groups of respondents (van der Wal 2008, p.49). There is a disadvantage of having only relatively small amount of respondents, as it allows only limited generalizability within the organization. However, this is aimed to be offset to some degree by the advantage that analytical conclusions arising from the two cases (KfW and GIZ) are more powerful than those coming from a single case (e.g. only GIZ) (Yin 2014, p.64).

(24)

23 Concerning the concrete selection, respondents had to fulfill at least one of the following selection criteria in order to be considered as an informant for this study:

- Expert with experience in project management and thereby expertise in the practical application of accountability tools,

- Expert with knowledge in reporting and evaluation, as reporting (and to a lesser extent evaluation) is often considered as an accountability instrument,

- Expert with knowledge about accountability and good governance.

In table 3, the respondents of this study and their corresponding professional background and characteristics are presented.

Table 3. Characteristics of the Interview Respondents. In-depth Semi-Structured Interviews

As a next step in the research process, an interview guide is constructed, which represents “a listing of areas to be covered in the interview along with, for each area, a listing of topics or questions that together will suggest lines of inquiry” (Weiss 1994, p.48).

The interviews always start with an introductory talk, in which the researcher lays out the purpose of the interview and a short question on the professional background and present function of the respondent. The qualitative research method is a semi-structured interview, which means that the questions do not always follow on in the way outlined on the schedule, and a few questions are included that were not in the guide, as the researcher picks up on things said by the respondent (Bryman 2008, p.438). The reason for choosing semi-structured interviewing for this research design is that it provides insights into how the research participants view and experience the

Respondent Characteristics

GIZ

A1 Male Key account manager with more than 30 years of experience, while more than half of the time was spent in project management abroad A2 Male Policy advisor in a project abroad

A3 Female Senior evaluation officer at the evaluation department

A4 Male Project manager in Germany and good governance specialist with some experience in project management abroad

A5 Male Program director abroad with almost 20 years of experience and extensive experience in project management abroad

A6 Male Senior policy advisor in Germany and specialist for policy perspective of aid

KfW Development Bank

B1 Male Project manager at the evaluation department

B2 Female Sector economist and good governance specialist with some experience as project manager

(25)

24 practice of accountability and results orientation without pigeon-holding the response of those being interviewed. Thus, the interview guide enables a degree of structure, but there is enough room for the interviewees to pursue related issues of particular importance to them (Bryman 2008, p.438). The interview guide is attached at the end of this study (see annex 1).

At the beginning of the interview, each respondent is asked to give their personal definition of accountability. This technique is utilized to get the first grip upon their understanding of accountability, which could be a first hint towards what matters in their practice of accountability in their organization. In the following first set of questions, questions are asked on how accountability functions on the organizational level and personal level using the dimensions of the accountability framework of this study. More specifically, this included questions on the accountable actor, forum, the nature of the conduct, instrument, as well as sanctions and incentives. In order to get a precise grip on the practice of accountability, respondents are also asked to give examples of a situation when they hold someone accountable or were held accountable themselves. Although many respondents initially consider accountability in abstract terms, most are able to make clear for themselves and the interviewer, how accountability in the organization was lived in practice.

The second set of questions concerns the practice of results-based management and serves the purpose to find out to what extent and how results-based management is used for accountability. Questions regarding performance measurement (indicators) and managerial freedom are aimed to get insights into how NPM prescriptions of results orientation have materialized in practice. In the last set of question, accountability and results are directly related to each other, in order to find out whether accountability for results is practiced, or whether accountability concentrates on a different type of content, e.g. inputs. Questions concerning challenges serves the purpose to find out potential reasons for why accountability has or has not kept pace with the focus on results. A final question on what the respondent would like to add, allows the respondent to address aspects that he or she finds important, but are not addressed with the questions, as well as to emphasize what he or she finds of particular importance with regard to accountability and results-based management.

It is important to note that the views expressed by the respondents are solely personal opinions and do not represent the official views of the GIZ and KfW Development Bank respectively. With regard to any internal and external rules and procedures mentioned in the findings and discussion, it needs to be emphasized that those reflect only the rules and procedures that the respondents highlighted – making not claim to be exhaustive or complete.

Each in-depth interview lasts between 40 – 70 minutes, depending on the progress made during the interview and the availability of the interviewee and was conducted either via phone, skype or

(26)

25 in person. Also, all interviews are conducted and transcribed in German. The quotes presented in the findings are translated into English ex post. The transcripts of the interviews result in the basis for this research that need to be systematically analyzed.

3.3 Data Analysis

The assembled data is analyzed by means of segmenting it into meaningful chunks, reassembling it and finally organized through coding. It is an analysis that is oriented towards themes present in the data (Boeije 2010, p.94). More precisely, the goal of this study is to portray the practice of accountability and the extent to which it reflects NPM results orientation on an organizational, rather than individual level. For that reason, the data analysis is issue-focused, instead of case-focused. This issue-focused analysis aims to “to describe what has been learned from all respondents about people in their situation” (Weiss 1994, p.153). In other words, this study paints a general, but at the same time contextual picture of German bilateral aid implementation (van der Wal 2008, p.51).

Initial coding of the interview material starts after all interviews were conducted. The interviews are converted into text documents and imported as separate “hermeneutic units” into the software tool ATLAS.ti. Two separate files for the interviews at GIZ and the interviews at KfW Development Bank are created, in order be able to analyze the results separately. The coding process begins with defining some general code families relating to accountability, such as the accountable actor, the accountability forum, the nature of the conduct, the accountability instrument, as well as sanctions and incentives. Also, two general code families with regard to results orientation, namely performance management and managerial freedom are established. On the one hand, this helps to systematize the findings from the start and on the other hand, by only pre-determining code families, but not the codes themselves, allows for topics emerge from the data. In other words, initially the data was coded by assigning these code families to text fragments. Later, in the following rounds of coding, those code families were complemented by more specific open codes that became sub-code of those families. The coding tree used for KfW and GIZ is attached in annex 2. The entire text is coded, except for the parts that have no specific relevance for answering the research question. In the process of coding and going back between data and codes, new codes are created, and old ones adapted so that gradually more definite codes are established (Klostermann 2003, p.43). The aim of the coding process is to find out how accountability and to what extent NPM prescriptions of results orientation are practiced at GIZ and KfW Development Bank. Lastly, all observations are combined in a coherently written report “as a coherent story, so that the material presented early in the report prepares the

(27)

26 reader for material that will appear later and later material draws on the earlier, and the reader, in the end, can grasp the report entire” (Weiss 1994, p. 153).

Validity and Reliability

In case-study research, the external validity of inferences tends to remain relatively weak, which means it is difficult to generalize inferences from a case study to other contexts or organizations (Bhattacherjee 2012, p.93). This study though does not aim to prescribe a universal framework for accountability in results contexts. Clearly, “what works in one country may not be appropriate in another, given differing political systems, administrative cultures, and levels of development” (Perrin 2006, p.17). Instead, this study aims to gain insights into the practice of accountability with regard to results orientation in the German case and where challenges may persist.

Due to this low generalizability of this study, validity is aimed to be achieved internally rather than externally. In order to actually measure what the research is supposed to measure, and thus increase internal validity, participants have to express their true beliefs and perceptions when answering the questions. This, in turn, means that researcher aims not to steer participants in a certain direction (asking leading questions) and to be very clear and concise in explaining the purpose of the interview (Punch 2014, p.146). Besides that, to guarantee a validity of the interview material, “careful, concrete level interviewing within the context of a good interviewing partnership” is aimed for (Weiss 1994, p.147).

Standardization of at least some of the questions by means of the interview guide is used to increase data reliability. Besides that, probing questions are an invaluable tool to ensure the reliability of the data, as it, for instance, allows for completing the information, provides opportunities to explore sensitive issues, and enable the researcher to clarify inconsistencies within the respondents’ accounts (Barriball and While 1994, p.331).

The need to align results orientation and accountibility : German bilateral aid implementation