The Rise of Self-Interests: How Shifting Political Goals Changed the Effectiveness of Dutch Official Development Aid to Bangladesh From 2005 Until 2016

Aid effectiveness has been situated in the central debate of development studies and, particularly, among foreign aid policy making community. As the Netherlands was one of the unique countries which focused extensively on the needs of its recipient, it is important to investigate the Netherlands’ recent practice of its development aid, particularly to Bangladesh and the associated goal ambiguity. This article attempts to answer the question, “to what extent does goal ambiguity in Dutch development aid affect the effectiveness of Dutch development aid given to Bangladesh?” This analysis is based on independent assessments of Dutch official development aid (ODA) to Bangladesh and Dutch coalition agreements in 2005-2010 and 2011-2016. This study finds that, in the second period, goal ambiguity has increased, while the aid effectiveness has decreased. This change is unlikely to be attributed to the conditions for aid effectiveness since these remained largely stable or even in slightly in favour of the second period. While the comparison of different sectors between timeframes lacks full robustness, there is direct evidence that goal ambiguity has played a role in lower effectiveness in the water related aid in the second period. Therefore, this study suggests that aid effectiveness in Bangladesh has decreased in the period 2011 to 2016 due to an increase in goal ambiguity caused by increasing self-interests in Dutch ODA policy.


INTRODUCTION
In the past few years, foreign aid has been an important feature in the contemporary debate on foreign policy. Foreign aid has been inextricably related with how state actors instrumentalise their foreign policy to essentially reduce poverty and support developing countries' development (Ying, 2002;Virgianita, 2011). The debate goes beyond whether official development assistance (sometimes called official development aid, development aid, foreign aid, etc.) is only an extension of states' desire to politically expand their influence or soft power. Among scholars, there is no general consensus that development aid is effective (Apodaca, 2017;Easterly, 2006;Easterly, 2008;Lancaster, 2007;Sachs, 2005). The current debate, nonetheless, still sits on the problem of effectiveness of foreign aid. Assessing the effectiveness of aid is extremely difficult, according to Bigsten, Gunning and Tarp (2006). Most studies about effectiveness do not clearly define their definition of effectiveness. Although effectiveness is a highly relevant concept that is somewhat clear to most, an overall accepted definition is absent.
From a different body of literature, that of global prohibition regimes, Getz (2006, pp. 248) proposes, "that effectiveness is, first and foremost, a function of the problem being addressed". Therefore, it is significant to further the discussion on development aid's effectiveness through empirical assessment, such as examining state actor's practice on development aid. This study will focus on the Netherlands.
Historically, the Netherlands was a donor of development aid that was known for its focus on the needs of the recipient country (Olsen, 1998), but this has changed.
Since Prime Minister Rutte came into power in 2010, multiple non-governmental organisations (NGOs) have voiced their concerns about the new direction of Dutch foreign aid (Oxfam Novib, 2016;Keultjes, 2018). Oxfam Novib (2016) worries that the Dutch aid policy has focused more on self-interests since then. It fears that this 'selfinterested' policy change has created a conflict of interest that might harm the effectiveness of Dutch development aid. When goals conflict or when they are unclear, one can speak of goal ambiguity. In a different policy field-the public administration literature-it has indeed been shown that goal ambiguity might decrease the effectiveness of a policy (Leune, 2007). However, it has not been extensively empirically researched in the development aid literature. Thus, it remains to be seen if goal ambiguity has negative consequences on aid effectiveness as well. Up until now, however, the relationship between goal ambiguity and aid effectiveness has not been researched thoroughly. Nevertheless, there has already been some speculation outside the field of science. Ban Ki-moon, the former United Nations (UN) Secretary-General, addressed the Bretton Woods 1 institutions with this following statement: "All too often, aid is driven more by politics than by need, undermining its effectiveness" (Ban, 2007, pp. 1).
The policy change in 2010 included a reduction in the number of aid recipient countries. From the previous 33, 15 eventually remained (OECD, 2012). One of those countries is Bangladesh, which has recently become a middle-income country (World Bank, 2018a;World Bank, 2018b). If the previously voiced concerns about the Dutch aid policy shift hold true, then these effects could be examined by studying Bangladesh before and after 2010. Therefore, this research will focus on the period six years before the policy change and the period 6 years immediately after. By comparing those two periods, the possible causal relation between goal ambiguity and aid effectiveness can be tested.
The purpose of this article is to evaluate the relationship that goal ambiguity policy might have with the effectiveness of the aid. This is done to incorporate insights from the public administration literature into the aid effectiveness literature, by introducing the concept of goal ambiguity and its effects, which adds to the existing Policy Coherence for Development framework. This coincides with another goal of this research, namely to evaluate the effects of the policy change introduced under the Dutch People's Party for Freedom and Democracy's (VVD) Prime Minister Mark Rutte's first cabinet in 2010 and the cabinet Rutte II that followed. The main question is, therefore: "To what extent does goal ambiguity in Dutch development aid affect the effectiveness of Dutch development aid given to Bangladesh?" To answer that research question, this article utilises goal ambiguity as the main framework of analysis. This article focuses on multiple secondary sources. These mainly include scientific articles, parliamentary papers, policy reviews, and independent evaluations of Dutch aid made by the Dutch Policy and Operations Department (abbreviated as IOB in Dutch) and the OECD. The structure is as follows, first the analytical framework will be set out, then the methodology will be discussed, and followed by the results in the discussion. The article closes of with a conclusion, including remarks on further research and policy implications.

Official Development Assistance
The OECD defines Official Development Assistance (abbreviated as ODA and used interchangeably in this article with aid) as "government aid that promotes and specifically targets the economic development and welfare of developing countries." (OECD, 2018a, pp. 1). This assistance is in the form of resource flows to the recipient countries, territories, and multilateral development institutions (OECD, 2018a). The selection of eligible countries to receive ODA changes over time and it depends on their per capita income (OECD, 2018a). The OECD has also made a more specific definition: "ODA is the resource flows to countries and territories on the DAC List of ODA Recipients (…) and to multilateral development institutions that are: i. Provided by official agencies, including state and local governments, or by their executive agencies; and ii. Concessional (i.e. grants and soft loans) and administered with the promotion of the economic development and welfare of developing countries as the main objective." (OECD, 2018a, pp. 1). This definition has recently been altered. In 2014, it was agreed upon that only grants and those parts of concessional loans (loans with a lower than market interest rate or with a longer payback time) that are similar to a grant in nature are considered as ODA.
However, this was only imposed for the ODA figures from 2018 and onwards, with the earliest OECD numbers being released by this more stringent definition in early 2019 (OECD, 2018a). It is important to notice that this research will only use data from before 2018 (which was based on the old definition of ODA).
ODA can be further classified in sub-categories. ODA can differ in terms of time and the number of donor countries involved. With respect to the former, one can distinguish between aid that focuses on the short-term (emergency aid) and aid focuses on the long-term (structural aid). With respects to the latter dimension-that of the number of donor countries involved-one usually makes the distinction between bilateral aid (meaning one donor country and one recipient country) and multilateral aid (two or more donor countries and one or more recipient countries). Sometimes aid is called multilateral and bilateral at the same time, abbreviated as multi-bi aid. Aid is labelled as such when it is earmarked and given by one donor country using a multilateral aid programme . There are more reports on structural aid than on emergency aid. Therefore, this research will focus on the structural aid that the Netherlands has given to Bangladesh. In addition, the Netherlands supports Bangladesh with both multilateral and bilateral ODA and they are both included in aid assessments.
Therefore, both are examined in this article.
Lastly, one can make the distinction between tied and untied aid. Tied aid is defined as "official grants or loans that limit procurement to companies in the donor country or in a small group of countries" while untied aid is simply all other aid that is not bound by this criterion and it can thus be spent more freely (OECD, 2018c, pp. 1).
Tied aid has been under fire for many years because the costs of development projects can increase by as much as 15% to 30% when it is in the form of tied aid instead of untied aid (OECD, 2018c). The international community has made efforts to reduce the amount of tied aid because of this (OECD, 2012). Unfortunately, progress towards this goal has been slower than expected (OECD, 2012). Tied aid also seems to put selfinterests above selfless-interests, thereby contradicting a criterion of ODA, namely that helping the recipient country is the main objective (OECD, 2018a). Unfortunately, specific numbers of tied aid (to Bangladesh) are not available; therefore, this will not be a part of this research.

Effectiveness of ODA
According to the OECD (2007, pp. 9) the measure of effectiveness differs depending on whom you ask: "For countries that receive foreign aid, it means allocating resources to their policy priorities and plans to promote justice, stimulate economic growth and improve social welfare for their citizens". However, as outlined above, the goal of helping recipient countries should be the main objective of ODA (OECD, 2018a). There are various ways of obtaining this goal and different methods (and types of aid) have different effects for the recipient country. A recent literature assessment by McGillivray et al. (2006) shows that there is much debate about the effectiveness of aid and the conditions that shape it. Factors that claim to be a condition for aid are often contested.
Up to date, the following factors have been reported to influence aid effectiveness: the climatic conditions, the amount of political stability, the level of economic growth, the geographical location, and the policy environment (see the causal model below in Figure 1).
Concerning the climatic conditions, Guillaumont and Chauvet (2001) find that worse climatic conditions (i.e., the frequent occurrence of natural disasters) result in better aid effectiveness. This is somewhat unexpected. They explain this by giving the hypothesis that aid mitigates effects of natural disasters, therefore making aid more effective in a country with many natural disasters than in a country with fewer natural disasters. Political stability may also play a role in aid effectiveness. To be more precise, more political stability creates a better condition for aid to be effective (Chauvet & Guillaumont, 2002). Another condition for effective aid is economic growth. Gomanee, Girma and Morrissey (2003) have found that aid is only effective when there is a GNP growth of 2% or higher in a given year. 2 In their model, aid has no effect on development when that threshold is not reached. Others have found that the geography of a country can exert influence on the effectiveness of aid. Dalgaard et al. (2004) stated that aid given to countries with jungle (i.e., countries in the tropics) has lower effectiveness than aid given to other countries. The geography, however, will not be operationalised because, even though, Dalgaard et al. (2004) have found a statistically significant relation between geography and aid effectiveness, it seems highly unlikely that this is caused by the geography of a country itself. This is even what Dalgaard et al. (2004) admit themselves. More probably, it is not a condition for effective aid but rather a factor that strongly correlates with one that is. While it could still be potentially meaningful to include a proxy variable, like geography, into an analysis of a crosscountry comparison, for an intra-country comparison like this study where the country's borders have not significantly changed, the proxy variable would not add anything to this qualitative analysis since it would stay the same. Finally, a good policy environment can also increase aid's effectiveness (Burnside & Dollar 1997Kaufmann & Kraay, 2002). The research from Burnside and Dollar (1997; has been critiqued by Easterly, Levine and Roodman (2004) who used their dataset but extended it and found no statistical significance. On the other hand, Kaufmann and Kraay (2002) used other indicators to measure a good policy environment, namely those of the World Bank, and found a statistically significant effect that to the best of the author's knowledge has not been disproven. Thus, a country's policy environment does matter for aid effectiveness. Finally, the many factors that influence ODA's effectiveness analysed above will form an integral part of this research.

Policy Coherence for Development
Policy coherence is a term originating from the public administration literature, but it can also be applied in the field of development aid. For development aid, a specific approach has been developed called Policy Coherence for Development (PCD). There are multiple definitions of this concept. Experts in the field of aid define PCD as follows: "Policy Coherence for Development means working to ensure that the objectives and results of a government's (or institution's) development policies are not undermined by other policies of that government (or institution), which impact on developing countries, and that these other policies support development objectives, where feasible." (OECD, 2005, pp. 28). The concept was created to generate better input, output, outcome, and impact results (Picciotto, 2005b). This research will focus on improving the output level, Picciotto (2005b, pp. 148) defined this as "to minimise the negative impact on developing countries of policy initiatives taken to achieve domestic goals". In other words, in the PCD framework, non-aid policies and aid policies of the donor country should not interfere with each other in such a way that it reduces the development outcomes of the donor country's policies.
The PCD approach consists of multiple dimensions (Picciotto, 2005a). The first dimension is that of internal coherence. This is about the consistency of goals and objectives, modalities, and protocols of a donor government's aid. The second dimension is called the intra-country coherence, and this is defined as the consistency of aid and non-aid policies of a government to jointly contribute to development. The third dimension is called the inter-donor coherence and is described as the consistency of aid and non-aid policies of various donor countries in terms of their contribution to development. The fourth dimension is named donor-recipient coherence. This is about the consistency of the policies that the donor countries have with the policies of the recipient country (Picciotto, 2005a). This research will focus on a part of the internal coherence of the Dutch development policy, namely the consistency of goals and objectives. This will be taken as the independent variable. For goals to be coherent, they have to be clearly defined and they must not contradict each other as well. Another way to describe this is with the concept of goal ambiguity. From the public administration literature, one can derive that goal ambiguity comes down to leeway in interpretation of the set goals (Chun & Rainey, 2006). Internal policy coherence of Dutch development policy can thus be extracted by analysing how much leeway there is in interpreting the goals of the development aid.
There can be multiple causes for goal ambiguity. In the case of ODA, an explanation can be sought in the fact that donor countries also use their aid to help themselves to a certain extent (Burnside & Dollar, 1997;Berthélemy, 2006) and this can cause friction with the goal of helping the recipient country develop. That domestic goals play a role in giving development aid has already been proven by Burnside and Dollar (1997) and more recently by Berthélemy (2006, pp. 192) who did a study on the motives of donor countries to give aid. It was concluded that most donors "behave in a rather egoistic way." The literature on how much a donor country profits from spending on ODA varies, but a recent analysis shows that on average every euro spent on ODA increases exports of the donor country by the same amount (IOB, 2014). Moreover, donors often select recipient countries based on their trade linkages with those countries, therefore implying that domestic economic interests play a significant role in the allocation of aid (Berthélemy, 2006). The conflict between a self-interested goal and a selfless goal creates goal ambiguity.
Goal ambiguity can have multiple consequences. The selection of goals, for instance, always influences the selection of the means (WRR, 2004) and when this happens, a trade-off might be created between the different goals. In the case of aid policy, unless officials, despite the ambiguity of goals, always use their agency to select aid policy goals to take priority over other goals, goal ambiguity between aid and nonaid policies will likely hurt the effectiveness of ODA. Moreover, Leune (2007)-a public administration researcher who focused on education-states that for policy to be effective, it is a requirement to have non-conflicting goals. 3 According to him, this does not always have to be the case, but there is a high chance of policy failure when goals do conflict. Little research has been done, however, to investigate these effects in the field of development aid. If goals can be interpreted as conflicting, then there is goal ambiguity. These effects are accumulated into the causal model and hypothesis of this paper.
Consequently, the hypothesis in this research is that if there is more focus on goals focused on advancing Dutch actors' interests, then because of trade-offs caused by this goal ambiguity, Dutch development aid will become less effective. The causal model for this hypothesis is represented in Figure 1. In order to ensure that a change in effectiveness is not caused by other factors, the climatic conditions, economic growth, political stability and policy environment will be evaluated as well.

Operationalisation
The operationalisation of development aid is straightforward. It is a quantitative variable that can be further dissected into the indicators of ODA and untied aid. ODA is simply measured by the recorded ODA amount that the Dutch government provides to the OECD. Not all aid is ODA because sometimes aid does not adhere to the criteria set by the OECD Development Assistance Committee (DAC). That part of the development aid of the Netherlands will therefore not be taken into account.
The next concept to operationalise is goal ambiguity. As derived from scientific literature in the previous section, goal ambiguity is about the vagueness of goals and about conflicting goals and thus creating leeway in interpretation. This article will focus on both aspects. Moving on, conflicting goals can be found by analysing the goals mentioned in coalition agreements. One goal of ODA is always to help the recipient country. Goals will be seen as conflicting when they go against the interest of the recipient country, which is Bangladesh.
As mentioned in the previous section, the definition of effectiveness depends on the perspective one chooses. This research will focus on aid effectiveness from the perspective of Bangladesh and three indicators are used in order to do so. The first indicator is 'attainment of project-specific goals obtained from evaluation reports' and statements will be searched for as proof of this when they describe a project as successful, effective, or working as intended. The second indicator is 'added value of projects obtained from evaluation reports'. 'Added value' means that aid was necessary to create this project. Thus, when assessments of Dutch aid speak about a project that could have been done without aid, it will be marked as non-additional. The third indicator is 'successful output of aid'. This is measured by marking statements in assessments that speak of improvement of living conditions due to the projects.
Based on the literature, external conditions for aid effectiveness are indicated by the variables of the policy environment of the recipient country, climatic conditions, political stability, economic growth, and geography. The policy environment will be based on the world governance indicators: voice and accountability, government effectiveness, regulatory quality, rule of law, and control of corruption from the World Bank (2018c). The climatic conditions variable is about natural disasters (Guillaumont & Chauvet, 2001). Therefore, this will be operationalised by the number of natural disasters and the number of which are labelled as major disasters. These data are obtained from the Asian Disaster Reduction Centre (ADRC; 2018). Political stability will be operationalised by the world governance indicator called 'political stability and absence of violence/terrorism' from the World Bank (2018c). The variable 'economic growth' will be operationalised by using GNP figures from the Bangladesh Bureau of Statistics (Trading Economics, 2018). As discussed in the theory section, the geography will not be taken into account and it will therefore not be operationalised. An overview of the entire operationalisation can be found in Table 1.

RESEARCH METHOD
This section will explain how the research was conducted and why it was done in this way. This section first sets out to account for the choices made with regards to the timeframe and the research design.

Timeframe and Research Design
The timeframe of this research will encase the period ranging from 2005 to 2016. This

Research Limitations
All possible validity issues from evaluation assessments from IOB and the OECD affect the validity of this research. IOB, however, has a de jure independence. This also seems to be the de facto case since there have been instances of sharp criticism towards the Dutch government, making the cover-up of politically sensitive information or other validity issues unlikely (Hoebink, 2017). Thus, somewhere around 25% of aid has been evaluated in the first period and somewhere around 61% has been evaluated in the second period.

DISCUSSION
In this section, the results of this research will be described. This section is divided into three parts. The first part explains how the conditions for effective aid have changed between 2005 and 2016. The second part is about the changes in aid effectiveness in this period. The third part elaborates on that goal ambiguity might have had.

Conditions for Aid Effectiveness in Bangladesh
Looking at the first of the aid effectiveness conditions, one can see that the policy environment of Bangladesh has improved minimally when comparing the time periods of 2005-2010 with 2011-2016 (see Table 2 for the averages). Most change has happened in the control of corruption in which Bangladesh almost doubled its world ranking. The regulatory quality and the rule of law in Bangladesh have increased, but only to a small extent. On average, there is an increase of about two percentile ranks. Due to the high error margin, however, these results should be conservatively interpreted. Dollar (1997, 2000) state that a better policy environment leads to better aid effectiveness. Thus, while it seems that the period 2011-2016 has better conditions for aid effectiveness, it is more appropriate to state that the policy environment has remained stable and that this is not in the benefit of either two periods. Furthermore, none of these natural disasters was classified as a major natural disaster. In line with conclusions from Guillaumont and Chauvet (2001), more natural disasters leave more room for aid to counteract development disturbances and thus be more effective. Therefore, the period of 2005-2010 better fulfils this condition than the period of 2011-2016.
Political stability is somewhat higher in the second period than in the first (see Table 3 for the averages). In 2014, political stability reached its highest point and in 2005, it was at its lowest (World Bank, 2018c). However, the error margin in all of these years is quite large (World Bank, 2018c); therefore, these results should be conservatively interpreted. Labelling political stability as stable over the two periods is, therefore, more appropriate in this case. Thus, neither of these two time slots better fulfils the criterion of Chauvet and Guillaumont (2002).  Gomanee et al. (2003). In that year, there was a negative GNP change of  Table 4.

Changes in Aid Effectiveness
Overall, the report by IOB (2011)  "results achieved poorer than planned" (MFA, 2017a, pp. 8). One of the main reasons is that involvement of Dutch small and medium enterprises (SMEs) was pushed, while they were not adapted to the local market, "This makes the country not very [sic] suited for SME initiatives from the Netherlands" (MFA, 2017a, pp. 4).

Goal Ambiguity
A change in goal ambiguity, the leeway in interpretation of set goals, might be the explanation for the lower effectiveness of aid in the second period. There were three Dutch cabinets in power during the period of 2005-2010 and they all had very similar goals that were clear and not mutually conflicting to a large extent. The first cabinet is that of Balkenende II who was in power from 2003 to 2006. Its coalition agreement was a bit special since it was coined as an outline agreement (Rijksoverheid, 2003a) instead of a coalition agreement and it was later on extended by a declaration of Prime Minister Balkenende (Rijksoverheid, 2003b). Therefore, both have been examined. In the coalition agreement, the stated goal is to help poorer countries develop. This is seen as a moral duty (Rijksoverheid, 2003b). Another goal mentioned is stabilising the international environment through aid (Rijksoverheid, 2003a(Rijksoverheid, , 2003b. Although there are two main goals mentioned, they are not conflicting because stabilising the world can be in the best interest of developing countries as well. This might just be a way of legitimising the high aid budget to the electorate because Balkenende stated that the budget would stay at 0.8% of the GDP 10 during his reign (Rijksoverheid, 2003b (Rijksoverheid, 2007). Because only one goal is voiced, it is by default a non-conflicting goal. However, later on, it is also mentioned that the ORET programme (which focuses on private sector development) will be adjusted to create a larger target group for the medium and small Dutch enterprises in the developing countries (Rijksoverheid, 2007). This is a first and moderate revelation of private interests playing a role in the allocation and form of aid. Private interests have not been directly stated as a goal in the coalition agreement, but it seems to be at least of some  (Rijksoverheid, 2010a). The first goal is of special importance. It is characterised by a change 'from aid to investments' (Rijksoverheid, 2010a). According to the OECD (2012)  interests. Further evidence for this can be seen in the following translated quote 11 from the coalition agreement, "within the budget for development aid, the room for business will expand greatly". This is also noticeable in their policy according to IOB (2013a).
Because Rutte I was a minority government, there was also a parliamentary support agreement signed with the Party for Freedom (Partij voor de Vrijheid or PVV) in order to rule. In this agreement, however, development aid is only briefly mentioned in relation to diverting funds to deal with unwanted migration to the Netherlands (Rijksoverheid, 2010b With the fall of Rutte I, Rutte II came about in 2012. There were no goal statements in relation to aid in the coalition agreement from Rutte II (Rijksoverheid, 2012). Nevertheless, the year 2012 also heralded a major change in Dutch aid policy.
From that year on, a ministerial post was created, for the first time in Dutch history, combining the posts for development aid and trade (Rijksoverheid, 2012). This minister was part of cabinet Rutte II and was put in place in order to "emphasize the importance of coherence between aid and trade" (Rijksoverheid, 2012). While in addition to the numerous benefits for Dutch companies that were put in place in the new Dutch aid policy, the budget for development aid was also reduced further than under Rutte I (Rijksoverheid, 2012). These changes point towards the increase in the importance of self-interests in Dutch aid creates a conflict with the selfless goal of aid. In addition, since there are no official goals, the goal of aid is ambiguous. Thus, under Rutte II, goal ambiguity has increased further compared to Rutte I. Goal ambiguity in 2012 to 2015 was therefore high.

Causal Relationship
Finally, the above results can be summarised in Table 5. This research also demonstrates that poor effectiveness in one of the four subsections of aid for better water management from 2011 to 2016 named 'trade and development cooperation' was fully attributed to an increased focus on the involvement of Dutch SMEs. Dutch businesses were involved in aid, but they were not adapted to the Bengali business environment, and this resulted in poor aid effectiveness.
To summarise, under Rutte I and II, the goal ambiguity of the Dutch government has increased, and this coincided with lower effectiveness of aid in Bangladesh at the same time. The poorer effectiveness of aid cannot be attributed to a change in conditions since these have been stable in 2005 to 2016. If one takes in the empirical evidence from the water management aid into account as well, then there is a basic level of evidence to suggest the causal model of this research is correct, namely that goal ambiguity decrease aid effectiveness for the receiver. This has been proven in the case of Dutch aid given to Bangladesh following the self-interested policy change in 2010.
Ban Ki-moon might have been right when he stated, "all too often, aid is driven more by politics than by need, undermining its effectiveness" (Ban, 2007, pp. 1) because the Dutch aid seems to be a prime example of this.

Implications of This Research
To interpret and describe the significance of this research as well as to comment on the research difficulties, a number of reflections will be given below. Firstly, the findings in this research have been consistent with the expectations derived from the public administration literature such as those from Leune (2007) and WRR (2004) and continuing to explore the concept of goal ambiguity outside of the domestic public sector might be fruitful for other researchers. Therefore, this concept could potentially help to explain policy failure in cases where no explanation has been found as of yet.
Secondly, the comparison between the same and different sectors is not impeccable and more research on this topic is needed in more complete case studies.
The results in the previous section show that there is a discrepancy between the effectiveness in the first period and the second period. In the first period, aid for better sexual health and education has been researched and in the second period, aid for better sexual health and water management has been evaluated. However, the water management aid was the aid that was underperforming in the second period whilst the aid for sexual health was not. The water management aid has not been researched in the first period due to a lack of reports. Consequently, the comparison between the two periods is not perfect and rather than giving conclusive evidence, findings of this research need to be seen as an indication. On the other hand, this research provides an opening for future researchers to explore the connection between goal ambiguity created by self-interests and ODA effectiveness further, since there is some direct evidence that goal ambiguity played a role in the lower effectiveness of the Dutch ODA to the water sector in Bangladesh in 2011 to 2016.
Finally, policy makers and politicians could benefit of this study and those that follow in this path by recognising that introducing new goals creates goal ambiguity and that by the mechanism of trade-offs, the original goal might be compromised. The full extent of the rise of self-interests in Dutch ODA in the period 2011 to 2016 under a right leaning coalition with VVD leadership have yet to be fully researched in all ODA recipient countries. Nonetheless, the evidence from this case study of Dutch ODA to Bangladesh suggest that an increasing focus of a donor's self-interests in ODA can result in increased incidence of policy failure.
1 The Bretton Woods institutions are the International Monetary Fund and the World Bank, often mentioned together because of the resemblance in their function and approach. They were created at the same time in 1944 in a city called Bretton Woods in the USA to help Europe recover economically but they have since changed their mission and they are now heavily involved in financing developing countries.
2 GNP, gross national product, is different from GDP, gross domestic product, in the sense that GNP is about the production of nationals only (those abroad as well) and not about the production in country (including expats and other non-nationals) like GDP.