Centre of Government and Delivery

Shoulder to Shoulder

Posted on: 22nd November 2016
Dan Hymowitz
Former Senior Advisor
Download PDF



    For a few months in 2010, one of our Governance Advisors worked with the head of a presidential delivery unit to help establish the office’s processes: staff roles and responsibilities; reporting templates; and meeting routines. After weeks of our adviser being excluded from the unit’s daily co-ordination meeting, it became clear that while the President was eager for us to work with this official, the official didn’t want to work with our adviser. Only after several months did we realise that we needed to assign the adviser to a different role.

    It may sound obvious, but the lesson from that experience, as an external actor, is to recognise when government officials want to work with you and when they don’t. Sometimes being government- led means not working in situations where you can’t contribute.

    The first three papers in the Art of Delivery series have focused on lessons for governments on policy implementation from the centre. This final piece focuses on lessons for external actors – typically international partners – on how to support government delivery. What does it take to work effectively with governments to build new roads or improve the health system while also developing government capability to do those things better in the future?

    We’ve structured this paper around four elements that we’ve found, through our work, to be crucial for partners seeking to support delivery.

    1. Putting government in the lead: working with government on a limited number of priorities that it is genuinely committed to achieving.
    2. Navigating local politics: working with the grain of local politics and helping government to harness political opportunities.
    3. Integrating implementation and capacity development: supporting government to drive results while simultaneously developing government skills, systems and structures.
    4. Embedding for the long term: working shoulder to shoulder with government colleagues over months and years enables the successful delivery of the first three elements.

    None of these four elements are revolutionary: development experts have argued their importance for many years. Despite agreement in principle, however, in practice these elements are often done poorly because they can be difficult to execute and because many traditional development actors – both funding and implementing organisations – operate under constraints that restrict their ability to carry them out. We have been able to test, learn and adapt our approach to these elements over the past eight years, in part because much of our work has been financed through flexible sources such as philanthropists. Often we’ve learned ‘the hard way’, as with the example cited at the beginning of this paper.

    The first part of this paper offers reflections on how to carry out these four elements in practice: for example, what does it mean to navigate local politics to support delivery successfully? The final section provides people in funding and implementing organisations with recommendations on how they can navigate their organisation’s constraints to work in these ways.

    A few clarifications and caveats before we begin: first, this piece is about supporting government delivery, which has commonalities with, but also distinctions from, ways international partners often work with governments. For example, many international organizations advise on policy which, as David Booth points out, is “a fundamentally different activity” from supporting implementation.1 In Table 1, we differentiate between working on delivery and other common ways international partners support governments.

    Second, the next section of this piece looks at the role and approach of implementing actors or in-country staff, rather than staff based in donor countries. However, both donors and implementers are essential to working successfully in these ways, as we’ll discuss in the recommendations section.

    How supporting delivery differs from other partners' approaches

    The Shoulder to Shoulder Approach in Practice

    The Shoulder to Shoulder Approach in Practice

    Putting government in the lead

    The importance of government ownership is so widely accepted in the development effectiveness community that it borders on cliché. For decades, there’s been consensus that when donors create “aid programmes around a series of stand-alone projects that are designed, implemented and managed by the donor, they don’t result in sustainable development.” Aid effectiveness meetings in Paris, Accra, Busan and Addis Ababa over the last 11 years have each reaffirmed international commitment to country ownership principles.

    However, ‘government ownership’ is too often praised at high-level conferences and missing in reality. Governments and international partners both contribute to this phenomenon. External partners can find it difficult to determine a government’s priorities because different government ministries and agencies are not on the same page (something we discuss in our paper on prioritization, ‘Choosing the first among equals’4). Furthermore international partners often don’t trust government to use support well, or come with their own policy agendas. Today most international partners talk of ‘local’ or ‘country’ rather than government ownership and increasingly work through non-state actors. When international partners do seek to work with government, they find ways to focus on their own agenda, for example taking a passing reference to an issue in a government’s Poverty Reduction Strategy (PRS) or a minister’s vague expression of interest in a meeting, as an indication that their favourite issue is a real priority.

    In our experience, there are several keys to providing genuinely government-led support. First, it’s important for partners to invest time and energy in understanding the government’s priorities.

    Reading the PRS isn’t enough because they tend to be aspirational rather than realistic, well-prioritized plans. As a Cambodian government official once quipped, the “donors would need to be planning to build hotels on the moon not to align” with the country’s national planning paper.5

    Instead, international partners need to focus on people, getting to know key individuals throughout the political system. This can take time; we’ve spent months scoping work in new countries in part to understand which areas government cared about most. As we discuss in ‘Choosing the first among equals’, this may also mean working with government to define its priorities in cases where that government is not yet committed to the idea that it should prioritise or where officials aren’t on the same page.

    Effective government-led delivery also entails international partners being open to working on unexpected issues. For example, as we scoped a project in Ethiopia, it became clear that the government’s agenda was focused on an ambitious expansion of its manufacturing sector aimed at economic growth and job creation. It was not initially clear how (or whether) we might support this. We’ve always worked with the president’s office or national government ministries and agencies. However, in this instance we decided to change our approach: we placed advisers within the Industrial Parks Development Corporation (IPDC), a state-owned enterprise leading the development and operation of industrial parks throughout the country. Why? Because we came to understand that setting up new industrial parks was the central delivery challenge that would make or break the government’s industrialisation strategy.

    At IPDC, our team has contributed to the adoption of best practices in the development of eco-industrial parks and supported planning for future parks to ensure access to power and finance. Combined with our support to the Ethiopian Investment Commission (EIC), which aims to attract investors for the parks, this work is contributing to progress in Ethiopia’s industrialisation agenda. For example, a dozen leading Asian apparel suppliers are investing in Hawassa Industrial Park, which is expected to create over 60,000 jobs and generate US$1billion in exports. Two more parks are scheduled to be finished by the end of 2016, with similar investment and job-creation targets.

    Finally, being government-led means shifting gears when circumstances change. Sometimes this is about being responsive to changes in government priorities. During the Ebola crisis, for example, we stopped working in the sectors we’d been supporting in Guinea, Liberia and Sierra Leone to work alongside those governments on their responses to the epidemic. In each case, the government made clear to us that existing priorities had been placed on hold and that they preferred us to work with them on Ebola.

    Navigating local politics

    The development community now widely accepts that understanding and working within the constraints of local politics is essential to effective assistance.7 However, as with government ownership, partners “have found it hard to move from thinking politically to working differently”.8 For one thing, navigating political environments requires working flexibly, which is a challenge for most donor-funded projects given donor rules and procedures.9 It’s that much harder because, while staff for most donor-financed technical- assistance projects have the necessary technical expertise, they don’t always have the skills needed for working in politically smart ways.

    Meanwhile, although many international organisations conduct political economy analyses at the start of projects, too often this is an afterthought, and anyway, “formal guidelines for political analysis can help but only to a limited degree.” Working politically “is a social activity that combines ideas with relationships”; in other words, being able to understand and engage effectively with people.

    What does it mean to be politically smart in supporting delivery? Being government-led is an important starting point. That’s because if you’re working on the issues to which political leadership is most committed, it’s more likely that you will be able to work through delivery challenges when they come up.

    However, being politically smart is not just about following government’s lead; it also requires working proactively with government colleagues to navigate local politics in ways that are helpful for delivery. One example is our work with the Liberian government on its private-sector development strategy. The government had been struggling to implement its private-sector development plans because different government institutions, including the National Investment Commission (NIC), the Ministry of Finance (MoF), the Ministry of Commerce and Industry (MoCI) and the Ministry of Agriculture (MoA), had each developed their own strategies. This meant the government kept hitting roadblocks because these strategies weren’t aligned. For instance, the MoCI strategy required the MoA to pivot its focus from food crops to cash crops, which was out of sync with the MoA strategy.

    We worked alongside these key institutions to shape the Liberia Agriculture Transformation Agenda, which incorporated these various strategies into one coherent document with a common focus on six priority value chains. We also supported the creation of a Presidential Taskforce, which used the convening power of the President to bring together these ministries and agencies to coordinate implementation of the plan. Because our advisers were working with each of these focal institutions, the advisers were able to understand the priorities and concerns of the institutions, and to help work towards compromises on contentious issues.

    Progress is starting to show. A model for oil-palm outgrower schemes has been agreed, after four years of deadlock, and a pilot project will roll out this year. Liberia has launched an alternative form of rubber-processing, aided by support from the different ministries. The country has also restarted fish-processing and exporting, which had been stalled since Liberia’s civil conflict. And ministries have agreed how to reform the rice and cocoa sectors, which has led donors to re-programme projects to this new plan.

    For international organisations to work in these ways, they need staff with the right skills and approaches: the ability to build trusted relationships, to be resilient and adaptive, and to work within local culture. How do you make sure you have staff with these subtle but essential abilities and ways of working? This is hard to teach, so we put a lot of effort into hiring people who already have these capabilities and approaches; for example, through interactive case studies during interviews.

    We’ve also developed a programme called ‘CraftWork’, which aims to improve our staff’s ability to understand and navigate political environments through activities ranging from coaching to team-based problem-solving sessions.


    How do we become more politically savvy?

    To help our staff better understand the skills and approaches that are essential to navigating local politics and facilitating change in partner governments, we’ve developed ‘CraftWork’.

    CraftWork, which is summarised in this diagram, is a basic model for our staff to think about these abilities and ways of working. It emphasises the importance of individuals understanding themselves, the country they’re working in and their team so that they can build shared intent with partner  governments.

    We have also shaped a programme of activities to help our staff improve at CraftWork. These activities range from coaching to build self-awareness, to finding ways to solicit richer feedback from our government colleagues on our work. Our Africa-based staff also developed innovative activities to contribute to this programme such as a ‘CraftWork audit’, which one of our teams used to identify the specific CraftWork skills they needed to develop further.

    Integrating implementation and capacity-development

    Development organisations often work in ways that focus too much either on implementation or on capacity-building, at the expense of the other.

    One problem, as highlighted in Table 1, is when international- development actors step in to directly implement without attempting to develop local capacity. This might make sense in emergency situations where health facilities need to administer rapid treatment to the sick or when displacement camps have to be built quickly to provide shelter. Capacity-building is more time-consuming for an international organisation than implementation, and when lives are at risk the urgency of the need to implement may trump the importance of strengthening capacity.

    However, development actors frequently default to this implementation focus even when not working in crises. One good example is Haiti, dubbed ‘the Republic of NGOs’ because of the thousands of international organisations operating long-term programmes there focused on service-delivery. Critics have pointed out that while there are benefits to citizens from the services delivered by these organisations, this model has exacerbated problems of weak government capacity and institutions. Unless national and local systems get the chance to develop through implementation, institutions can’t be strengthened. That means trying out different approaches, making mistakes, and learning along the way. This model is not incompatible with NGO service-delivery, but at a minimum that delivery needs to be in partnership with local- government institutions and staff.

    The converse is also a problem: capacity-development initiatives that are disconnected from government delivery. A common example are training programmes for individuals, often held abroad. This is unlikely to be the right method for building the capability of government staff. Training may well be useful if “integrated with the organisational and institutional changes necessary to put new skills to work effectively”. However, individual capacity is developed mainly while learning on the job: by making mistakes, learning from those mistakes, then incorporating lessons and feedback into future actions.

    In terms of strengthening institutions and systems, reforms often fail because they don’t start with a problem that the government itself was trying to fix, as Pritchett, Andrews and Woolcock have argued. This results in “isomorphic mimicry”: institutional reforms that look good on the surface but fail to result in effective functions. Government capability won’t be strengthened unless institutions develop capabilities in line with an agenda on which the wider system is focused.

    That’s why you need to integrate both sides. This means starting with the government’s priorities and supporting the simultaneous implementation of those priorities and the development of the individual and systemic capability needed to drive future implementation. This is hard to do without working shoulder to shoulder with government colleagues, typically over a long period of time, as we discuss further in the next section.

    Table 2 is a simplified model of how we engages with government colleagues at the initial, middle and later stages of our work with a government institution.

    Rwanda’s Strategic Capacity Building Initiative (SCBI), which we helped to establish and supported, is one example of successfully balancing these twin aims. The programme is structured around delivery objectives in four Rwandan government priority areas – investment, agriculture, energy and mining. SCBI has specific implementation deliverables, and has simultaneously invested in the long-term potential of young talent by pairing Rwandan counterparts with embedded international consultants. The consultants and counterparts work together on those delivery priorities.

    SCBI has advanced Rwanda’s development agenda and built capacity, with many of the young professionals who were mentored through the programme now leading critical government work. For example, Rwandans trained through SCBI led negotiations for a US$30 million solar-energy deal that will increase the country’s energy-generation capacity by 10%. Other programme participants negotiated a US$22 million mining concession that will more than double Rwanda’s production of refined minerals. And a recent Department for International Development (DFID)-funded evaluation found that SCBI has led to growth in smallholder agriculture output and incomes, “critical capacity for research” and “knowledge and skills for smallholder farmers across the country for research and extension services, soil and water conservation and biosafety of agricultural inputs and outputs”.

    Embedding for the long term

    The first three elements are enabled by, and unlikely to succeed without, working daily alongside government colleagues, sometimes for well over a year.

    You need to take time – often months – to understand the government’s agenda, and how you can support it effectively, in order to be government-led. Being embedded for a long period is also important for navigating local politics. In our support for Liberia’s private-sector development strategy, we couldn’t have helped find common ground between the different institutions without understanding the varying priorities of, and political dynamics between, the different ministers – all of which we gained through relationships. SCBI was inspired in part by the Rwandan government’s frustration with the trend of international ‘fly-in, fly-out’ consultants, which the government felt was failing to develop Rwandan capacity.

    This is not to argue that all development technical assistance needs to be embedded and long-term. There are many situations where external expertise is needed on a technical issue that does not require an understanding of local political dynamics and therefore can be delivered more quickly – particularly, to borrow Lant Pritchett’s taxonomy, with “implementation light” problems.18

    However, if a government is looking for support on both delivery and improving government capability, embedded and long-term is essential. As David Booth has argued in ‘Arm’s length aid’, “effective facilitation of change calls for more local knowledge than donor staff can hope to acquire during a short posting, and much more continuous contact with local stakeholders than the typical office routines of donor agencies permit.”

    Towards a Shoulder to Shoulder Approach

    Towards a Shoulder to Shoulder Approach

    In our view, putting government in the lead, navigating local politics effectively, integrating delivery and capacity-development, and being embedded for the long term will lead to better delivery results. So how can development actors produce more of this type of work? This paper has focused so far on the role of implementing organisations. However, in order for organisations to work shoulder to shoulder with governments effectively, both funders and implementers would need to do things differently.

    Table 3 illustrates the logic behind the recommendations in this section. We start with the four elements of supporting delivery (the first column) and what is essential for doing them well (the second column). We then try to identify what prevents organisations from working in these ways (the third column). The final column describes what funding and implementing organisations might do differently in the future, recognising that much of the detail will vary by organisation. We conclude this section by distilling the final column of Table 3 into recommendations.


    Overcoming constraints on working shoulder to shoulder with governments

    Recommendations for Funders and Implementers

    Recommendations for Funders and Implementers

    To better support government delivery, implementing organisations must practice more of the effective versions of the four elements discussed in this paper. The analysis within Table 3 points to ways international funding and implementing organisations might contribute to this.

    1. Donors and implementing organisations need to employ a mix of strategies to deliver smart, government-led support.

    They need to:

    • Find ways to work with governments, even in tough circumstances. Donors prefer to avoid partnering directly with governments they view as untrustworthy or low capacity. However, if you don’t work with government, you can’t build systemic capability. There are ways for donors to work with governments that minimise risk. First, they can focus on technical assistance rather than financial support. Second, even within government systems with high levels of corruption, there are usually leaders – whether a deputy agriculture minister or the mayor of the capital city – striving to deliver results for citizens. Donors need to make the effort to identify these individuals and structure projects around them.
    • Take the time to understand government’s priorities. Funders need to create space for their in-country staff to understand the priorities of presidencies and ministries. To do this well, in-country staff must go beyond reading government planning documents and interact extensively with key political leaders within government. Meanwhile the staff of implementing organisations need to make a continuous effort to work collaboratively with government colleagues. Implementing organisations may then have to revise their plans in response to changing circumstances and government demand.
    • Navigate their own internal systems. The previous two recommendations can only be delivered if funding and implementing organisations work out how to execute these approaches within the constraints of their own strategies and systems. Within any donor agency there tend to be people who have figured out how to do this. As Dan Honig has written, some of the most effective development-partner employees “spend much of their time in conflict with their own organisations over procedures and process”.

    2. Implementing organisations need to recruit staff who can build strong relationships and navigate political environments, and funders need to incentivise organisations to do this.

    Working in government-led ways and navigating politics successfully requires staff who can build close, trusted relationships with government officials and who have the skills and approaches to work effectively in political environments. This means that, in the words of the International Rescue Committee (IRC) and Mercy Corps’ recent report on adaptive approaches to development, implementing organisations should “enable flexible competency prioritization in hiring key positions, placing value in soft skills, adaptive mindset, and change management experience”.21 Funders can play a role in enabling this by writing terms of reference that emphasise these capabilities and by assessing the bids of implementing organisations accordingly.

    3. Donor organisations should finance more capability- strengthening with the ‘right’ ingredients.

    Donor-funded projects focused on delivery too often ignore capacity- building, and capacity-building initiatives overemphasise limited approaches such as training. There is an integrated solution: donors should finance projects that provide technical assistance both to support delivery and on-the-job government capacity-strengthening. We acknowledge that it’s difficult for funders to finance capacity- building for learning on the job because it’s hard to describe this approach in project proposals and difficult to measure. However, there are ways for funders to incentivise this better: for example, requiring implementers to develop specific plans for shoulder to shoulder capacity-building that include timelines for ‘stepping back’ to allow government colleagues to take on the responsibility for implementation.

    4. Donors need to create space for adaptive ways of working.

    Staff must adapt to changing circumstances in order to work in government-led, politically smart ways. For this to happen, donors must continue to open up space for adaptive approaches, and implementing organisations need to seize opportunities created by that space.

    The good news is that progress is happening, with donors developing new approaches and mechanisms. For example, DFID’s Smart Rules were created after an internal DFID review found that “programmes need to be flexible and responsive to changing political realities and conflict dynamics on the ground” and that “to achieve this we need to improve our ability to commission and manage adaptive, flexible programmes”. These new rules aim to change processes, capabilities, incentives and culture within DFID to allow for flexible approaches. Meanwhile, a new Global Learning for Adaptive Management programme will establish a centre for learning on adaptive management for DFID and USAID programmes.24 Other funders are taking similar steps.

    This is a promising start, but it’s not enough. Donors can do more to push implementing organisations, as IRC and Mercy Corps have suggested, by providing “funding for activities focused on context analysis and ongoing learning”. As with any development trend, without additional incentives there’s the risk that implementing organisations will pay lip service to the fashionable language of adaptive management in their bids, and fail to deliver it in practice. At the same time, changes such as DFID’s Smart Rules open up the space for motivated implementing organisations to experiment and innovate.



    We’ve published our Art of Delivery series to share reflections from our work on government implementation in 10 African countries over the past eight years. The timing seemed right, in part because we felt our experience could build a bridge between two emerging international-development trends that sometimes fail to connect: delivery, and doing development differently (DDD).

    Delivery units are springing up around the world26 because governments and international organisations are rightly keen to see better development results for citizens, recognizing that picking the ‘right’ policies doesn’t automatically lead to good outcomes. However, external actors supporting delivery sometimes fail to operate effectively within local politics and are too wedded to best-practice solutions. In other words, we fail to absorb the lessons from the DDD movement, which emphasises local context, politics and adaptation.

    Organisations seeking to do adaptive programming can also learn from the delivery community. Many organisations that operate in locally-led and politically savvy ways do so by working with civil society and influencing government politics from the outside. This is important work. But more organisations should also apply these key DDD principles to the challenge of working with governments to deliver results and build systems.

    In our first Art of Delivery paper, we argued for a greater focus on the more ‘artistic’ building blocks of delivery – power and incentives – when developing delivery systems.27 The second piece made the case that prioritisation is a pre-requisite for effective delivery, and presented ideas on how to do it better.28 Our third paper argued for customising delivery systems to local contexts and for adapting them to changing circumstances over time.

    Each of these three papers featured lessons for both governments and external organisations supporting implementation. This fourth and final paper, in contrast, focuses on the role of external actors. Delivery systems can’t succeed unless they’re owned and operated by governments, but the reality is that many delivery initiatives are funded and supported by outside actors; so external organisations are playing a central role. That makes the delicate balance of the four elements described in this paper all the more important in order for delivery to live up to its promise.

    The work described here was carried out by the Tony Blair Africa Governance Initiative, it is now being continued by the Tony Blair Institute for Global Change.