Enhancing service delivery in humanitarian relief chains: the role of relational capital

Relational capital is a concept that is widely studied across a variety of disciplines and industrial sectors. The concept is defined as a set of all relationships established between firms, institutions and people that stem from a strong sense of belonging and a highly developed capacity of cooperation typical of culturally similar people and institutions (Capello and Faggian, 2005; Welbourne and Pardo-del-Val, 2009). Relational capital is a perquisite for inter-cluster coordination. Coordination costs are reduced when firms have relational capital (Gittell, 2006; Bijman et al., 2011). Relational capital enhances the accuracy and timing of information and reduces information exchange costs among clusters (Adler and Kwon, 2002). Provision of accurate and timely information enhances effective humanitarian responses.

Humanitarian relief environments involve a number of stakeholders that include international relief organizations, host governments, the military, local and regional relief organizations and private sector companies, each of which may have different interests, mandates, capacity and logistics expertise. Typically, no single actor has sufficient resources to respond effectively to a major disaster (Bui et al., 2000). Many factors contribute to coordination difficulties in disaster relief. Such factors include the inherently chaotic post-disaster relief environment, the large number and variety of actors involved in disaster relief and the lack of sufficient resources. Because of this, humanitarian organisations often find it too difficult to collaborate (Fenton, 2003) because of increased complexity and absence of relational capital. Although there are few coordination success stories, coordination continues to be a fundamental weakness for humanitarian action (Rey, 2001). Given the continuing challenges and previously failed initiatives, coordination is receiving increased attention, due to the increasing scarcity of resources and accountability concerns (Lindenberg and Bryant, 2001). The literature addresses various aspects of relief sector coordination (see Minear, 2002; Kehler, 2004), highlighting the complexities and challenges associated with coordinating humanitarian assistance (see Rey, 2001; Stephenson and Schnitzer, 2006); however, the influence of relational capital on inter-cluster coordination and service delivery have not been given significant attention. While a lot has been said about inter-cluster coordination, empirical studies examining the role of inter-cluster coordination are still absent. A cluster is a group of agencies that gather to work together towards common objectives within a particular sector of emergency response (Ulleland, 2013). Inter-cluster coordination is a cooperative effort among sectors/clusters (Vinbury, 2017).The cluster approach was instituted in 2006 as part of the UN Humanitarian Reform process. This was is an important step on the road to more effective humanitarian coordination. Ultimately the cluster approach aims to improve the predictability, timeliness and effectiveness of humanitarian response and pave the way for recovery (Ulleland, 2013).

A number of the humanitarian relief organisations tend to specialise in areas such as water and sanitation, health, shelter, nutrition (Jahre and Spens, 2007; IASC, 2006; Jahre and Jensen, 2010) with many having their own funding and systems. When such organisations combine, they can face a series of problems related to coordination. Such problems may include overlapping relief, poor governance structures and some populations of beneficiaries not being well served (Adinolfi et al., 2005; OCHA, 2007). An important concern is when and how the relief organisations can collaborate and be coordinated given that the humanitarian systems are fragmented. The relief organisations may collaborate through creation of relational capital among clusters. Both relational capital and inter-cluster coordination play a critical role in facilitating the development of the strategic response plans and assure a coherent and coordinated approach to planning and operationalizing the shared strategic objectives as set out in the strategic response plan.

Natural disasters and climate-related extreme weather events are increasing in scale and frequency. Developing countries especially small developing countries are significantly more exposed than developed countries to natural disaster risks. For example, the number of reported disasters in Africa has significantly increased in recent decades, putting at risk its recent development gains. Besides, natural disasters are more frequent and of higher intensity in developing countries, Africa inclusive, and their economic cost, as a proportion of GDP, is several times larger than for developed countries (European Commission, 2012; Food and Agriculture Organization, 2017). An analysis of 20 years of data shows earthquakes and tsunamis are the biggest overall killers, followed closely by climate-related disasters, which have more than doubled over the period in developing countries including those of Africa (United Nations, 2016). In Africa, the Eastern and Southern Africa regions continue to face multiple and more frequent humanitarian crises, including conflict and insecurity, economic shocks, climate change, natural hazards and disease outbreaks (Uganda Country RRP, 2019, 2020; UNICEF, 2018).

As an African country, Uganda has been prone to disasters including natural disasters and wars requiring humanitarian support. For instance, Uganda hosts the largest number of refugees in Sub-Saharan Africa and is identified as the third largest refugee-hosting country in the world after Turkey and Pakistan (Uganda country refugee response plan RRP, 2019). There are currently 1.4 million registered refugees and asylum seekers in the country, most of whom are from South Sudan (66%), the DRC (27%), Burundi (3%), Somalia (2%) and Rwanda (1%) (UNHCR, 2019; REACH Initiative Report, 2018; UNHCR, 2018). Refugees in the country include both protracted refugee populations and new arrivals. Wars, violence and persecution in the Horn of Africa and Great Lakes Region were the main drivers of forced displacement into Uganda, led by South Sudan's conflict, insecurity and ethnic violence in the Democratic Republic of the Congo (DRC) and political instability and human rights violations in Burundi (Uganda country refugee response plan, 2019; Mastercard Foundation, 2019). More than 60% of Uganda's refugees are under the age of 18, one of the most visible consequences of conflicts in neighbouring countries and with clear implications for the provision of protection services (Mastercard Foundation, 2019). While the number of refugees per 1,000 inhabitants has tripled to 35 since 2016 putting a huge pressure on local resources and services, external aid has been progressively dwindling over the years, causing major gaps in the refugee response. Disaster response partners continue to face enormous challenges in stabilising existing programmes, meeting the minimum standards of service provision and investing in long-term and more sustainable interventions.

Besides, flooding in Butaleja District and landslides in Bududa in February and March 2010 were among the 10 deadliest disasters worldwide causing 385 deaths and displacement of over 3,000 people (Doocy et al., 2013). In June 2019, a total of four major landslides occurred in Bududa leaving 6 dead, 27 injured and 480 people displaced (International Federation of Red Cross and Red Crescent Societies, 2019). Further, in Teso in Eastern Uganda, floods and landslides almost left the entire region in ruins and displaced (Uganda Red Cross Report, 2018). Many organisations including Uganda Red Cross, UNICEF, UNHCR (Uganda Red Cross report, 2018), Oxfam, International Rescue Committee, Save the Children, Plan International and World Vision among others have been involved in providing relief supplies (Care International Report, 2018). Nevertheless, service delivery for the people affected remained low. For instance, while disasters in Teso begun in April 2018, distribution of relief and food items started on 17 August 2018 in a one-week long distribution exercise (Uganda Red Cross report, 2018). In refugee camps, the average amount of water was less than the recognised UNHCR standard of 20 L per person/day, while reductions in food assistance were observed since August 2016. Most settlements lacked medical requirements and nutrition programmes have not meet UNHCR/World Food Programme standards (REACH Initiative Report, 2018). In Kyaka II and Kyangwali refugee camps, 59% (300,000) children were not receiving education (Care International Report, 2018). Further, many of the communities that host refugees still do not have secondary schools, although this is a priority for the coming years (Mastercard Foundation, 2019).

Although several clusters have developed close working relationship with host state governments, country of origin governments, the refugee community, civil society, the private sector and other United Nation (UN) agencies, the contextual evidence shows that there was a gap in services delivery. This has led to the unanswered empirical question as to whether relational capital and inter-cluster coordination in humanitarian relief chains related to service delivery. Service delivery in humanitarian relief chains refers to the provision of the required aid in the right time (Beamon and Balcik, 2008). Managing disaster risk is a global challenge requiring collective action from more than just the humanitarian organisation. Although, developing countries rampantly often experience natural disasters, the relevancy of relational capital and inter-cluster coordination in humanitarian relief chains is not empirically known. Most research relates relational capital to supply chain performance (Shujaat et al., 2019; Zhao et al., 2019a, 2019b; Zhang and Wang, 2018; Afshar and Fazli, 2018) and innovation (Thi Mai Anh et al., 2019). Therefore, the research aimed at examining the influence of relational capital on service delivery in humanitarian organisations, the influence of relational capital on inter-cluster coordination and the mediating and moderating roles of inter-cluster coordination on the relationship between relational capital and service delivery in humanitarian organisations.

The link between relational capital, inter-cluster coordination and service delivery may be explained using the Resource-Based View Theory. The Resource-Based View (RBV) by Barney (1991) identifies factors that can affect services delivery in humanitarian relief chain. RBV posits that service delivery of organisations derives from the resources that are valuable, rare, imperfectly imitable and not substitutable. Such resources include tangible and intangible assets (Barney et al., 2011) such as relational capital with other organisations that may enhance inter-cluster coordination. Hence, performance is a result of an organisation's specific resources. The essence of an organisational strategy is or should be defined by the organisation's unique resources and capabilities (Theriou et al., 2009). RBV proposes that performance results are a consequence of organisation-specific resources. These resources and capabilities can be important factors for superior services delivery (Kamasak, 2017). Therefore, this theory was the basis for relating factors namely relational capital and inter-cluster coordination to service delivery in humanitarian relief chains.

This section reviews literature on three study variables that include relational capital, inter-cluster coordination and service delivery. Literature review sub-sections discuss a review of the literature on the relationship between relational capital and service delivery in humanitarian organisations, relational capital and inter-cluster coordination, the mediating and the moderating role of inter-cluster coordination on the relationship between relational capital and service delivery. A moderator variable is a variable that affects the strength AND /OR direction of the relationship between the dependent or criterion variable (Y) and the independent or predictor (X) variables (Baron and Kenny, 1986) while a mediator variable (ME), also called “intervening or process variable”, is the variable that causes mediation in the relationship between the dependent variable (called outcome) and the independent variable (called causal variable) (Namazi and Namazi, 2016).The relationships between the variables are shown in the conceptual framework using arrows (see Figure 1 below).

Conceptualisation of the three variables under study is undertaken in this subsection. The variables include relational capital, inter-cluster coordination and service delivery.

To conceptualise relational capital, articles published on relational capital in other fields other than supply chain management were considered first. Findings on the measures for relational capital were later compared with those published on relational capital in the supply chain management context. The majority of the studies emphasised trust and commitment (see Table 1). After making a comparison of the measures, measures for relational capital relevant to the humanitarian relief chain were identified. These measures include trust, commitment, respect and reciprocity. The humanitarian relief chain in developing nations experiences challenges because there is lack of trust between the local non-government organisations (NGOs) and UN agencies. Lack of trust results in reduced commitment and resource sharing (Yao et al., 2019). Disrespect of participants may yield negative emotions. Such emotions have powerful effects on social interaction, particularly on trust-building (Weber and Weber, 2007). Many local NGOs think they are not respected thus affecting reciprocity.

The dimensions for inter cluster coordination were established through a review of the supply chain and humanitarian literature. Possible measures for inter cluster coordination were identified from the humanitarian literature while those of supply chain coordination were identified from the supply chain literature. The majority of the measures from the supply chain literature were related to commercial supply chains. A comparison of the possible measures shows communication, information exchange, partnering and joint planning as relevant measures for both setting. In order to come up with dimensions for inter-cluster coordination, measures with similar characteristics are grouped together to form dimensions. The two dimensions formed include information technology and responsibility interdependence. Information technologies include information management and analysis, information sharing and effective communication while responsibility interdependence is made up of active participation, joint problem solving, joint planning and establishment of joint priorities (see Table 2). Effective communication within or across different clusters is essential to response and recovery (Kim and Bui, 2019). However, power shortages and damage to communication equipment threaten the ability to communicate effectively with key stakeholders. Information management and sharing is a key activity for the cluster (Jahre and Jensen, 2010). Cluster stakeholders require relevant, timely and accurate information. Such information may be archived for future use. Further, information on what humanitarian organisations are doing, where they are based and when they are active may provide knowledge on imbalances in relief provision (Van de Walle and Dugdale, 2012; Camacho et al., 2019). Poor information management may result in loss of information hence affecting relief provision processes.

Responsibility interdependence enhances inter-cluster coordination when stakeholders in the different clusters are able to actively participate in relief provision, jointly plan and solve problems and establish joint priorities. Responsibility interdependence is more useful than task interdependence in highly uncertain situation such as those that involve disaster relief operations. It enables actors focus their attention on the contribution they are making to their clusters and other clusters. This promotes a sense of shared responsibility. The stronger the responsibility interdependence between and among cluster members, the more private information will be shared and jointing planning and problem solving and active participation will take place.

Measurement of performance in commercial supply chains takes into account delivery times, costs and value for the end customer. Similarly, key performance indicators in the relief chain make use of some of the key performance indicators developed for the commercial supply chain that include response time, efficiency (Cost of supply, distribution and inventory holding), effectiveness, flexibility, the number of relief works, the amount of time spent providing aid and the amount of time spent serving a recipient. Other scholars suggest the use of the balanced score card and the supply chain operations reference model as alternative approaches for measuring performance of commercial supply chains (see Table 3). These approaches have also been applied to humanitarian relief chains with both the balanced score card and supply chain operation reference model being commonly used. However, existing balanced score cards in the humanitarian relief chain are considered as in transition between generation one (early stage) and generation two (balanced score card with strategy map) categories. The balanced score card with strategy map shows interdependencies of performance indicators in humanitarian relief chains. Further, existing balanced score cards in humanitarian relief chains emphasize a balance between financial and non-financial performance indicators, the procedure on how such performance indicators should be selected is unclear (Behl and Dutta, 2019). Also decision-makers experience difficulties regarding the choice of the performance indicators and their order of relevancy. Lack of knowledge on the performance measures and their order of relevance may lead to ineffective implementation (Anjomshoae et al., 2019). Overall, measurement problems arise from the uncertainty surrounding the humanitarian relief chain. Thus making it a duty for those engaged in provision of humanitarian relief aid to determine the relevant key performance indicators based on the existing circumstances.

Service delivery is driven by relational capital. A higher level of relational capital improves planning, problem solving and troubleshooting, all of which most likely increase service delivery efficiencies (Thi Mai Anh et al., 2019; Attar et al., 2019). Relational capital is defined as the set of all relationships that include market relationships, power relationships and cooperation established between firms, institutions and people that stem from a strong sense of belonging and a highly developed capacity of cooperation typical of culturally similar people and institutions (Welbourne & Pardo-del-Val, 2009). Most research examining the impact of relational capital only considers operational improvements (Polyviou et al., 2019; Hughes et al., 2019) rather than its contribution to humanitarian service delivery. For example, previous research findings indicate that relational capital leads to improvement in cost, quality, flexibility and productivity (Xu et al., 2019; Molodchik et al., 2019; Kengatharan, 2019). Other research relates relational capital to manufacturing performance (Tumwine et al., 2012), firm performance(Datta and De, 2017; Kengatharan, 2019; Bayraktaroglu et al., 2019), product innovation (Eiteneyer et al., 2019; Ji and Ma, 2019) and the mediating effect of relational capital on the relationship between corporate reputation and competitive advantage (Wang, 2014). Relational capital provides critical information that serves as a useful external guide for improving and developing new knowledge (Zhao et al., 2019a, 2019b; Lee and Ha, 2018). Such knowledge aids in improving humanitarian service delivery through the development of innovative solutions.

Much of the research on humanitarian organisations focuses on collaboration (Clarke and Fuller, 2010), cross-sectoral partnerships coordination (Nurmala et al., 2017) and intellectual capital (Ataseven et al., 2018). The concept of relational capital in humanitarian relief chains has not been given significant attention. Although collaboration is often seen as beneficial to humanitarian organizations, humanitarian organisations also face difficulties in managing collaborative efforts, including collaboration with the private sector and the beneficiaries. The difficulties may stem from cultural conflict (Skute et al., 2019), technology barriers (Oraee et al., 2019; Agarwal and Singh, 2018), conflicting goals and mandates (Kumar and Havey, 2013; Kovács and Spens, 2007), structure conflict (Akhtar et al., 2012) or lack of performance metrics (Agarwal et al., 2019). Presence of such difficulties may have an effect on service delivery. Such difficulties may be overcome through relational ties created among the different partners. Relational ties increase accountability and flexibility, decrease duplication of services at the local level, reduce distortion of information accuracy and eliminate vulnerability and enable strategic alliances internally and externally with partners, a major challenge in humanitarian programmes (Bush et al., 2015; Kraft and Smith, 2019). Other benefits include more effective response to collective problems, improvements in service delivery, spreading of risks and increased access to resources (Lee et al., 2019). Given the above discussion, it can be hypothesised that,

The concept of relational capital has been applied to research examining buyer–supplier relationships within the supply chain literature (Chang, 2018; Handoko et al., 2018), environmental management (Yu and Huo, 2019), project management (Aaltonen and Turkulainen, 2018) and recognized in group behaviour research (Li et al., 2019a; Johnston and Lane, 2018; Paoloni and Demartini, 2018). The relevance of relationships has been over emphasised in the humanitarian literature. Because no single humanitarian agency can cover all humanitarian needs, building relationships is not an option but is a necessity. However, most of the relationships remain transactional and competitive, rather than reciprocal and collective (Bennett, 2016). Further, Streets et al. (2010) argue that for ensuring cohesiveness of the humanitarian response, the relationships between clusters are as important as the relationships within them. In addition, Luo and Ye (2019) and Liang et al. (2015) argue that relational capital is a perquisite for coordination. In addition to flexibility relational capital enhances joint problem solving and information sharing thus enhancing inter-cluster coordination (Ramadass et al., 2018; Lee and Ha, 2018). Coordination does not automatically produce good performance; it has to be properly managed to generate positive results. In fact, poor coordination among the actors involved in humanitarian disaster relief is cited as the main cause of performance gaps; conversely, good coordination is crucial to guarantee successful emergency operations (Cook et al., 2018; Nkengasong and Onyebujoh, 2018). Staff in different clusters work more efficiently due to more concise and accurate information sharing and responsive technical assistance (Najjar et al., 2019; Comes et al., 2018). Further, investment in specific machines, apparatus, or instruments may be made to meet the needs of beneficiaries. Besides, relations among clusters of humanitarian organisations may bring new logistic solutions in difficult situations, thereby improving the agility and adaptability levels of the humanitarian organisations. For example, humanitarian organizations may learn from each other on how to reach the most remote parts of the world through imaginatively employing unconventional delivery systems. Although the review of the literature shows the existence of a positive relationship between relational capital and inter-cluster coordination, empirical research testing such relationship is missing. It can therefore be hypothesised that:

Coordination is a concept that is widely covered in the supply chain management literature with more coverage in commercial supply chains (see Tang et al., 2018; Heydari et al., 2018). It involves putting existing interdependencies in order (Raju et al., 2018; Lemma et al., 2015). Given the nature of the interdependence between clusters, coordination is a necessary prerequisite for integrating their operations to achieve the mutual goal of disaster relief as a whole as well as those of the different clusters (Raju et al., 2018). Because coordination problems in humanitarian relief chains arise due to inherently chaotic post-disaster relief environment, the big number and variety of actors involved and the lack of sufficient resources, inter-cluster coordination is suggested as a remedy for overcoming coordination problems so as to improve service delivery (see Balcik et al., 2010). However, the sector-based approach to coordination, exemplified by clusters, is not the best method for addressing complex, multi-sectoral needs of affected people (Knox Clarke and Campbell, 2016). Although there may be a progressive increase in collaboration among diverse stakeholders from government, civil society and private domains of society at multiple scales, inter-cluster coordination among government agencies and departments is lacking (Food and agriculture organisation of the United Nations, 2017). Findings by MacRae and Hodgkin (2011) on the Yogyakarta earthquake show that neither cluster decision making nor inter-cluster coordination provided much space or opportunity for participation by local actors, NGOs, government representatives and local communities.

Much of the inter-cluster coordination maybe conducted through routine meetings that all clusters attend typically also including support from an inter-cluster information management working group. Although important, development of inter-clusters may be affected by differences in languages, over-arching goals, unclear roles or taking on too many roles by the individual NGOs (Jensen and Hertz, 2016). Besides, a dearth of research examines the concept of inter-cluster coordination empirically in humanitarian relief chains. Further, the mediating and moderating role of inter-cluster coordination on the relationship between relational capital and service delivery is an area that has not been given significant attention. The influence of relational capital on service delivery in humanitarian relief chains may be moderated and mediated by inter-cluster coordination among humanitarian organisations.

With inter-cluster coordination, humanitarian organisations may share relational capital (McLachlin and Larson, 2011; Franco and Esteves, 2018). With relational capital, humanitarian organisations are able to deal with emergencies (Cao et al., 2017) through influencing inter-cluster coordination. Unlike the supply chain literature where relational capital is widely recognised (Zhao et al., 2019a, 2019b; Yu and Huo, 2019), rarely do organisations recognize that they can tap into a wealth of knowledge from their own clients in humanitarian relief chains. Whereas relational capital may result into an improvement in service delivery of humanitarian organisations, inter-cluster coordination plays an important role in achieving this goal. The influential effect of relational capital on service delivery may be mediated by inter-cluster coordination. Absence and low levels of inter-cluster coordination may result in low levels of service delivery. Although import, empirical studies examining the mediating role of inter-cluster coordination are lacking. Further, relational capital is focused on commercial transactions and entrepreneurial activities (Aisyah et al., 2019; Yu and Huo, 2019). With the above, it can be hypothesized that,

The moderating role of inter-cluster coordination on the relationship between relational capital and service delivery has not been given significant attention. A dearth of research relates inter-cluster coordination and service delivery. Further, the research lacks empirical evidence as its majorly qualitative in nature (Jahre and Jensen, 2010; Jahre, M. and Spens, 2007). Over the last decade there has been considerable interest and activity in clustering. The majority of the research on clusters is focused on industrial clusters or business clusters and their relationship with innovation and knowledge management (Anokhin et al., 2019; Ganau and Rodríguez-Pose, 2018) other than inter-cluster coordination. Clusters can thus be embedded in larger networks of global relations often involving cluster actors (Schüßler et al., 2013). Clustering stimulates cooperation between companies, as well as between business, encourages knowledge flows, information exchange, technology transfer, and learning processes, as well as contributes to the development of relational capital and trust (Kowalski, 2012). However, inter-cluster relations (relations between firms) demand a high degree of trust, time to develop, consensus over goals and cognitive, cultural and institutional proximity (Bathelt and Li 2014; Schüßler et al., 2013). Differences between firms in terms of interests, cultures, organisational design, and technologies give raise to coordination problems (Späth and Rohracher, 2015; Li et al., 2019b). In such circumstances, relational capital may improve coordination problems. Relational capital includes the set of economic, political, and institutional relations developed and maintained between the organization and non-academic partners (Espinosa, 2019). In the humanitarian context, poor relational capital in war affected zones or disaster zones affects inter-cluster coordination which in turn affects service delivery. For example, host community relations may affect inter-cluster coordination. Even when the humanitarian community does try to respond, their efforts may be blocked by the government (Yoshikawa and Teff, 2011; Tronc et al., 2018) which makes aid organisations seem to be reluctant to counter government policies (Dix, 2006). However, favourable host community relations may increase inter-cluster coordination. The higher the level of inter-cluster coordination, the more service delivery levels will be improved. This means that the refugees will be able to attain the required basic services. It can therefore be hypothesised that;

A pragmatic research paradigm was employed. The pragmatic paradigm neglects the ontology and epistemology assumptions and is more concerned with application of a method that works based on the need to provide a solution to a problem (Patton, 2015). Although research on supply chain management generally and humanitarian relief chain management specifically employ both quantitative and qualitative methods, the majority of the research is within a postivist paradigm (see Vybornova and Gala, 2016; Noori and Weber, 2016; Kunz and Reiner, 2016). Similarly, this research used a quantitative method involving hypothesis testing; however, it was carried out with the pragmatic paradigm because the research dealt with perceptions and beliefs of respondents. Respondents were required to give their perceptions of what they thought was absolute truth rather than relying on the absolute truth of knowledge because absolute truth may never be established. Although previous research on inter-cluster coordination employs a qualitative approach, a quantitative approach was carried to obtain the percentage contribution of inter-cluster coordination and relational capital in enhancing relief aid. Knowing the percentage contribution may aid in devising other better means to improving the operation of humanitarian clusters and also service delivery as well.

Similar to most studies in supply chain, a cross sectional quantitative survey design was employed. The respondents for the study included staff from 60 humanitarian agencies in Northern Uganda that belonged to any of the six established clusters. The clusters include Emergency Shelter and Non-Food Items (ESNFI), Food Security and Agriculture (FSAC), Health, Nutrition, Protection and Water, Sanitation and Hygiene (WASH), as well as a Refugee and Returnee Response Plan as led by UNHCR. All these sixty humanitarian agencies are currently working in the country as resident organizations (UN report, 2015) and these include government organisations, UN Agencies, International and Local NGOs as included in the 2018 Ugandan Gazette. Potential respondents were five officers from each humanitarian organisation involved in delivering humanitarian assistance. Officers engaged in lower management and technical staff of the humanitarian organisations were the target respondents for the study.

The survey instrument included items covering relational capital, inter-cluster-coordination and service delivery. Measures for relational capital were adapted from Marqués et al. (2006) and service delivery from Kaluki (2015) and Beamon and Balcik (2008). Measures for the inter-cluster coordination variable were developed based on previous rigor research (see Inter-Agency Standing Committee (2013); United Nations Children's Fund (UNICEF), 2015a, 2015b) and are being tested for the first time. Measures for relational capital included capacity to search for information, research and development agreements, strategic alliance creation and cooperation. Measures for service delivery included efficiency, delivery time and responsiveness while those for inter-cluster coordination included information sharing, effective and inclusive consultative and feedback mechanism, cluster flexibility, predictability and accountability, integration of cross-cutting issues for example the environment and complementation of partner actions. To overcome inefficiencies in Internet and postal services in Uganda, data were collected using a hard copy survey questionnaire through a pickup and drop-off method. The survey included 123 items plotted on a seven point Likert scale, similar to other studies (Akman and Pışkın, 2013; Tsireme et al., 2012; Sarkis et al., 2010).

The analysis was conducted using the Analysis of Moment Structures software (AMOS): a covariance based structural equation modelling (CB-SEM) software, Smart PLS software and SPSS software. AMOS is structural equation modelling software that may be used for conducting tests like model fit and mediation while SPSS may be used to examine for normality reliability and descriptive statistics. Smart PLS was only used when testing for convergent validity and discriminant validity, composite reliability and Average variance extracted while Statistical Package for Social Scientists (SPSS) and Analysis of Moment Structures (AMOS) were used in the rest of the analyses. SPSS was used to test for relationships and the predictive potential of the independent variable. The Covariance-Based Software was used to test for mediation and running a confirmatory factor analysis. It also provides the p-values for measurement items as well as model fit indices. Covariance-Based Software (AMOS) was used because it has been identified as the best software for testing mediation. It also provides the p-values for measurement items as well as model fit indices. Once collected, the data were tested for normality, skewness, kurtosis, common method variance and reliability using SPSS Version 21. All variables had a Cronbach alpha higher than the recommended value of 0.70; composite reliability values higher than the minimum threshold of 0.40; average variance extracted values higher than the minimum threshold of 0.40 implying the existence of convergent validity; and common method variance values below the maximum threshold of 0.50 (see  Appendix 3). All p-values were less than or equal to 0.0001 meaning that both the questionnaire and the items were valid (see  Appendix 2). Discriminant validity was measured using the heterotrait–monotrait method. All the values were below the set threshold of 0.9 (Friman et al., 2019; Henseler et al., 2015) (see  Appendix 4). The data were tested for normality. Previous research on normality suggests that the absolute value of univariate skewness should be <2 while the absolute value for univariate kurtosis should be <7 (Curran et al., 1996; Xiong and King, 2015). Skewness values for all variables were less than 2 with a range from −0.500 to −0.609 while kurtosis values for all variables were less than 7 with a range from −0.157 to 0.421. Skewness or kurtosis values lying between +1.0 and −1 indicate the existence of a normal distribution (George and Mallery, 1999).

The measurement model was confirmed through running a factor analysis. Factor analysis is a construct-validation tool that is used to determine measurement items of a given construct. The most important factors for a dimension are identified through factor analysis (Cronbach and Meehl, 1955). Factor analysis was conducted using SPSS (version 21) and CB-SEM (version 21), so that the weaknesses for each software package are catered for. Each of the software packages uses a different approach for the factor analysis and mediation analysis. Factor analysis results were obtained using the CB-SEM Software.

Using SPSS, Kaiser–Meyer–Olkin Measure of Sampling Adequacy and Bartlett's test of Sphericity were used first to test whether the sample was adequate for conducting a factory analysis. Results for the Kaiser–Meyer–Olkin Measure of Sampling Adequacy were 0 0.844 and so acceptably close to 1 and results for the Bartlett's test of Sphericity were significant at p < 0.001. Therefore, the results supported the use of factor analysis to obtain item loadings.

All item loadings had values above 0.30. The item loadings ranged from 0.31 to 0.88 with all having p-Values less than 0.0001 and critical ratios above the minimum established threshold of 1.96 for the confirmatory factor analysis and between 0.47 and 0.842 for the case of an exploratory factor analysis (see  Appendices 1 and 2). Exploratory factor analysis results were obtained using maximum likelihood as an extraction method and varimax rotation. Maximum likelihood is a factor analytic technique that is used in circumstances where the data meet normality assumptions and where the analysis is based on sample data (Cho and Hong, 2013). For normally distributed data, maximum likelihood allows computations for a wide range of model fit indices to take place; permits testing for statistical significance of measurement item loadings and correlations among measurement items and computation of confidence intervals (Fabrigar et al., 1999). The varimax rotational procedure is commonly used in the orthogonal rotation method. The results indicate that the measurement items are relevant to the constructs they are measuring.

This section provides findings for the influence of relational capital on inter-cluster coordination and service delivery in humanitarian organisations; and the mediating and moderating role of inter-cluster coordination on the relationship between relational capital and service delivery in humanitarian organisations.

The results indicate the existence of positive relationships among the variables. A positive relationship exists between relational capital and service delivery (β = 0.515^**, p-value = 0.01); relational capital and inter-cluster coordination (β = 0.523^** p-value = 0.01); Inter-cluster coordination and service delivery (β = 0.471^**, p-value = 0.01). However, issues of gender, qualification or level of education and experience have a role to play in enhancing relational capital, inter-cluster coordination and service delivery (see  Appendix 5).

Results indicate the existence of a significant positive moderate relationship between relational capital and service delivery and inter-cluster coordination and service delivery in humanitarian organisations as per Ratner (2009). The results imply that rather than carrying out humanitarian work solely, humanitarian organisations belonging to the different clusters need to cooperate in order to ensure better service delivery. Also inter-cluster coordination leads to effective coordination of humanitarian and early recovery assistance, especially among international assistance organizations with and in support of national entities. Also easy resource mobilisation can easily be obtained as humanitarian organisations in any given cluster will be able to put together resources for a common cause. However, for achievement of better service delivery, clustering of humanitarian organisations also requires the existence of relational capital among the employees of the humanitarian organisations and members in the different clusters. Findings are supported by Franco and Esteves (2018) and Food Agricultural Organisation (FOA) (2010) who argue that inter-cluster coordination enhances knowledge sharing and learning among the actors involved, leads to easy resource mobilisation and allows improvement in activities carried out or service provided, at both the firm and regional level. Also, inter-cluster coordination may be a more resilient way to responding to disruptions caused by men and nature, and reducing the total logistics cost (Cedillo-Campos, 2012). However, relational capital remains at the core of the humanitarian relief chains' cluster strength (Swann et al., 1998; Braun et al., 2005). This is because it promotes trust, commitment and reciprocity among the participants in humanitarian relief provision (Chen et al., 2017; Belanche et al., 2019). For example, host community relations may affect effectiveness of inter-cluster coordination as humanitarian efforts may be blocked by governments even when the humanitarian community does try to respond (Rohwerder, 2014). Also given the limited resources, existence of relational capital between the host governments and humanitarian organisations may help reduce issues of resource constraints through initiation of resource mobilisation campaigns. In addition, inadequate information sharing between the beneficiaries and the UN organizations resulted in development of poor plans that culminated in delayed response to material requirements (UNCHR, 2015). Also findings by Anderson et al. (2012) show that people requiring international aid may not ask for more aid because of the poor relationships between the aid providers for example UN agencies and the beneficiaries (people requiring the aid).

The results indicate that relational capital is a significant predictor of service delivery (see  Appendix 6). According to the results, an increase in relational capital signifies an improvement in service delivery. The findings are consistent with Beaudoin, 2007; Dynes, 2006; Koh and Cadigan, 2008; Nakagawa and Shaw, 2004; Chamlee-Wright, 2010 who argue that relational capital has a strong predictive power on service delivery. Relational capital is a resource that should not be disregarded by humanitarian service organisations (Buckland and Rahman, 1999; Pelling, 1998; Aldrich and Crook, 2008; Reininger et al., 2013). Humanitarian organisations with relational capital are likely to be more successful, particularly as the rate of disaster occurrences increases. Their ability to respond to humanitarian problems will be higher compared to those humanitarian organisations without relational capital because they will be receptive to different cultures and will exchange information required for undertaking humanitarian services such as information on beneficiaries (Zhang and Wang, 2018). Further, with relational capital, humanitarian organisations are able to obtain resources required to perform the different activities during disaster discovery. For example, provision of humanitarian aid to Rwandese refugees demonstrated the need for much closer linkages between humanitarian and political policies in the principal donor countries and the UN system and also with the neighbouring countries and regional bodies such as the Organisation of African Union (OAU) (Maghsoudi, 2016). The creation of task forces or contact groups composed of key interested parties served to encourage closer linkages. Such linkages paved the way for relief provision to the Rwandese refugees through the development of more effective interventions.

The results indicate that relational capital influences inter-cluster coordination (see  Appendix 7). With relational capital, humanitarian organisations are able to deal with emergencies (Cao et al., 2017) through influencing inter-cluster coordination. Given the challenges embedded in the adoption of a cluster approach by non UN humanitarian organisations, establishment of relational linkages would overcome such challenges. For example, clusters have been found to be most effective in sectors with long-standing relationships, prior to cluster implementation, such as in Kenya (Integrated Regional Information Networks, 2008b) and the WASH cluster in the Democratic Republic of the Congo, Ethiopia and Zimbabwe (Humantarian Reform Project (HRP), 2009). However, it is a common practice that the humanitarian organisations form partnerships after the clusters have been formed. Findings by Humphries (2013) show that partnerships were rated very weak amongst the evaluations on clusters implying that a cluster approach has been weak in fostering partnerships between UN agencies, national NGOs and local NGOs. However, Luo & Ye (2019) and Liang et al. (2015) argue that relational capital is a perquisite for coordination. In addition to flexibility, relational capital enhances joint problem solving and information sharing thus enhancing inter-cluster coordination (Ramadass et al., 2018; Lee and Ha, 2018).

While some humanitarian agencies have had a positive experience with the cluster approach, others have had a negative experience implying the need for relational capital if the cluster approach is to be effective. For example, a NGO aid worker in Chad explained the Cluster Approach as “nothing more than a way for the UN to control us” (Integrated Regional Information Networks (IRIN), 2008a). Another example is after the earthquake in Haiti, the Haitian non-government organisations found themselves sidelined by international actors. Haitian NGOs' were frequently relegated to the role of implementing partners, excluding them from clusters and coordination mechanisms and diminishing their ability to contribute to conceptual issues (Street, 2011). Lack of an inclusive partnership in the early days of an emergency renders the cluster approach inadequate, resulting in coordinating mechanisms gaps, resources, knowledge and assets to expand coverage (Knudsen, 2011). Also inclusion of local NGOs on an ad hoc basis creates insecurity among these organisations. Thus limiting the amount of information that may be shared due to limited relational capital amongst the organisations (Steets et al., 2010).

The results indicate the existence of a partial mediation between relational capital and service delivery. Although relational capital remains a significant predictor of service delivery, the introduction of inter-cluster coordination in the model reduces the direct influence of relational capital on service delivery (see  Appendix 8). Findings are supported by McLachlin and Larson (2011) and Franco and Esteves (2018) who argue that with inter-cluster coordination, humanitarian organisations may share relational capital. With relational capital, humanitarian organisations are able to deal with emergencies (Cao et al., 2017) through influencing inter-cluster coordination. Inter-cluster coordination is viewed as a critical strategy for advancing global goals at the national level and International level. For example, non-governmental humanitarian organisations that cooperated closely with each other in the Rwandese genocide had highly impressive performance. However, there were numerous examples where this was not the case. Some non-governmental humanitarian organisations sent inadequately-trained and -equipped personnel, some undertook to cover a particular sector or need and failed, and others were unwilling to be coordinated (Borton et al., 1996). The major causes of the problems were lack of relational capital and inter-cluster coordination.

Humanitarian organisations with relational capital are more likely to have better coordination amongst their clusters thus improving service delivery. Improved service delivery may be as a result of having common, mutual responsibilities to reach the objective of effective and timely humanitarian response for affected people (Inter-Agency Standing Committee, 2013). Besides, information management and exchange is a core activity for the cluster. The range of information covered is potentially quite wide but mainly focuses on transport routes, infrastructure status and the availability of transport resources. Such information is often quite obvious to those in the field but difficult to obtain for those who are not. However, the success of the overall cluster seems to depend, at least in part, on finding an acceptable and more participatory style of leadership (Cosgrove et al., 2007) and creation of both strong and weak ties. In the humanitarian relief chain, ties exist between humanitarian actors that include the host government, non-governmental organisations (NGOs), UN agencies, donors/suppliers and local communities. With such ties, beneficiaries in the humanitarian relief chain receive the right aid and supplies through mobilising resources and sharing information (Maghsoudi, 2016). Also affected communities are able to send information about their problems and unmet needs or open up a discussion with aid supply providers during response. Further, organisations that act autonomously without a jointly developed plan may provide similar needs while other necessary needs might be left unmet.

Inter-cluster coordination is viewed as a critical strategy for advancing global goals at the national level and these can be complemented by multi-stakeholder collaboration inclusive of government, civil society and private sector at multiple levels. However, most governments are organised administratively within a framework of sector based ministries and agencies with resource allocations and accountability managed accordingly. This is also the case for most technical organisations at all levels and this has hampered actions that need to be undertaken across sectors-such as food security, climate change or sustainable agriculture.

The results indicate that inter-cluster coordination does not moderate the relationship between relational capital and service delivery in humanitarian organisations (see  Appendix 9). The findings are contrary to the review from literature that indicates that inter-cluster coordination may moderate the relationship between relational capital and service delivery. The contradicting results may be due to a number of factors that include minimal coping capacities, lack or ineffective coordination of the interests and activities of various stakeholders, weak governance and lack of adequate partnerships (van Niekerk et al, 2018; International Strategy for Disaster reduction, 2004). Minimal coping capacities may arise due to limited resources and financial backing. Lack of adequate partnerships may result in inadequate inter-country cooperation and coordination to address common hazards and communication and shallow natural hazard risk assessment. Further, inadequate communication may be due to under developments in technology within the African countries. Further, most African countries, Uganda inclusive have limited resources to invest in disaster risk reduction and minimal fiscal space to fund relief and recovery efforts after a major disaster. Critical infrastructure such as roads, telecommunication and dams often lag behind rapidly growing needs or are not constructed according to risk prone standards. In the immediate aftermath of disasters, they become critical infrastructure for relief and recovery operations. Therefore, transport infrastructure, schools and hospitals need to be constructed and maintained according to minimum standards to resist certain earthquakes, cyclones or flood events. Unclear specification of responsibilities and weak communications systems are the cause of infective coordination in disaster relief operations. Weak governance and institutional capacities are the problems experienced in many countries in Sub-Saharan Africa. Countries in Sub-Saharan Africa also experience significant governance challenges, including an institutional and policy framework to effectively respond to disasters and manage risk reduction measures. This includes poor staffing and skills, weak analytical and implementation capacity, an unclear institutional landscape addressing disaster risk management across various ministries and agencies, and weak partnerships with other agencies and academia, non-governmental organisations, and the private sector.

Besides, cluster coordination is labour intensive, requiring a significant amount of time and resources for effective participation (Humphries, 2013) thus creating a barrier to inclusive partnership in the cluster Approach. Other factors could be inexperienced cluster coordinators, high turnover rates of cluster coordinators and infective coordination of disaster responses between nations and the people. Besides, the presence of very weak structures for disaster management affects the interactive effect of inter-cluster coordination. Much of the disaster relief activities are usually conducted by international organisations including UN Agencies. Although the organisations may be mandated to manage disasters, the funding sources, interests and priorities create gaps and overlaps which indicate the need for improved coordination (Penuel and Statler, 2011).

The research aimed at examining the influence of relational capital on inter-cluster coordination and service delivery, and the mediating and moderating role of inter-cluster coordination on the relationship between relational capital and service delivery. Findings indicate that relational capital influences service delivery, relational capital influences inter-cluster coordination, inter-cluster coordination partially mediates the relationship between relational capital and service delivery while no moderation effect. Although a positive influence was obtained, the results indicate that relational capital influences inter-cluster coordination by 56.4% and service delivery by 65.4%. Based on the results there are other factors other than relational capital that need to be considered such as the communication mechanisms. Relational capital aids information management and exchange. Information management and exchange are core activities for a cluster approach. However, information is often quite obvious to those in the field but difficult to obtain for those who are not. Also, with relational ties, beneficiaries in the humanitarian relief chain receive the right aid and supplies through mobilising resources and sharing information and willingness of humanitarian organisations to accept a cluster approach in humanitarian relief provision increases. Affected communities are also able to send information about their problems and unmet needs or open up a discussion with aid supply providers during response.

Partial mediation results may be due to the existence infective coordination of stakeholders' interests and activities, unclear specification of responsibilities and weak communications systems for informing stakeholders. Also weak governance and institutional capacities in Sub-Saharan Africa causes significant governance challenges, including an institutional and policy framework to effectively respond to disasters and manage risk reduction measures. Institutional challenges include poor staffing and skills, weak analytical and implementation capacity, an unclear institutional landscape addressing disaster risk management across various ministries and agencies, and weak partnerships with other agencies and academia, non-governmental organisations and the private sector.

The non-moderation effect is due to cluster coordination being labour intensive, requiring a significant amount of time and resources for effective participation thus creating a barrier to inclusive partnership in the cluster approach. Other factors are inexperienced cluster coordinators, non-inclusive cluster partnerships, high turnover rates of cluster coordinators and infective coordination of disaster responses between nations and the people. Also weak structures for disaster management affect the interactive effect of inter-cluster coordination. Much of the disaster relief activities are usually conducted by international organisations including UN Agencies. Although the organisations are mandated to manage disasters, the funding sources, interests and priorities create gaps and overlaps which indicate the need for improved coordination.

The research has scholarly or theory, policy and managerial implications. To theory, the research makes a contribution to literature through examining the influence of relational capital on inter-cluster coordination and the mediating and moderating role of inter-cluster coordination. The humanitarian literature addresses various aspects of relief sector coordination, highlighting the complexities and challenges associated with coordinating humanitarian assistance. Further, a group of studies describe coordination efforts observed during previous disaster relief operations and evaluate the factors leading to the success or failure of these efforts. There are also non-academic resources, such as practitioner reports, handbooks, training documents, agency websites and blogs that describe current practices and emerging initiatives in relief chain coordination. However, the literature lacks empirical studies that broadly and systematically address relief chain inter-cluster-coordination, relational capital and service delivery in humanitarian relief chains. Besides, the majority of the research is qualitative rather than quantitative with a few meta-analytical studies. Managerial implications include, humanitarian organisations being able to devise ways of working together to identify and reduce gaps and duplication, establish joint priorities and address cross-cutting issues in order to improve humanitarian response. They will also work together with partners thus pooling technical knowledge and even improving approaches covering the various strategic areas of procurement, such as forecasting, quality assurance and product innovation. Further, humanitarian organisations may also build relational capital with host governments to avoid restrictions and other humanitarian organisations prior to their inclusion in the different clusters.

To policy, governments and humanitarian organisations will develop policies that enhance effective coordination of the interests and activities of various stakeholders; adequate inter-country cooperation and coordination to address common hazards, adequate communication and creation of adequate partnerships with communities. Also governments may establish strong political leadership, strong governance and adequate institutional capacities in order to promote effective inter-cluster coordination and relational capital that will the end improve humanitarian service delivery.

Areas of further research include the influence of cross border policies, sourcing and procurement practices on inter-cluster coordination and service delivery in humanitarian relief chains. Effective inter-cluster coordination may require the consent of the nation state in whose territory relief operations are to be implemented, consent of the neighbouring nation state from which any transit operations will take place and consent of any personnel in control of territory through which the relief goods must transit. Countries have varying cross border policies which may at times be driven by inter-country political conflicts. Differences in policies imply variations in the provision of relief operations. Further, timely availability of relief supplies is relevant for the success of relief operations. However, timely availability of relief supplies requires presence of effective sourcing and procurement policies. Whereas institutionalisation and execution of cluster responsibilities including procurement policies is a duty of the UN agencies, the influence of established sourcing and procurement policies on inter-cluster coordination has not been examined. Also a multi-dimension approach may be taken when studying inter-cluster coordination and relational capital given that this study used a unidimensional approach when measuring inter-cluster coordination and relational capital. Lastly, given that the study was quantitative, an interview approach may be employed in future to obtain an in depth understanding of the role inter-cluster coordination in humanitarian relief chains.