The UK Bilateral Aid Review

The UK Bilateral Aid Review

Analysis and response

March 2011

1. Introduction

CAFOD’s public response to the government’s Bilateral and Multilateral Aid Reviews was very positive and welcoming[1]. We defended aid and backed the government’s commitment to honouring its pledgeson BBC 5Live, LBC radio, Premier Radio and Vatican Radio. We also briefed journalists from the Financial Times and the Independent, both of whom subsequently led with positive analysis pieces.

Our top-line message is that we are pleased with the reviews. The outcomes,however, encompass a complex set of issues which require careful consideration. The purpose of this paper is to analyse these issues and give a more detailed policy response.

We comment on the Bilateral Aid Review (BAR), rather the Multilateral Aid Review (MAR), as this has been on the process that CAFOD’s engagement has been particularly targeted. Our focus here is issues around the allocation of bilateral aid between countries. Analysis of other issues – for example, on aid for trade – may follow in forthcoming work.

1. The BAR process

Whilst there was room for improvement, in the end, CAFOD were relatively happy with the process of decision-making for the Bilateral Aid Review.

From ‘top down’ to ‘bottom up’?

The government describe the Bilateral Aid Review (BAR) as an entirely new approach to allocating aid funds between countries. Instead of allocating resources through a “top-down” process from the centre, it centred on “’identifying and scrutinising from the bottom-up the results that UK assistance could achieve in country”. A full description of the process is given in the BAR Technical Report.

At the heart of the approach was the development of “Results Offers” – where every DFID country office set out what results they could realistically achieve, how much this would cost, and what these results represented in terms of value for money. These offers were reviewed through a series of evaluation processes – and formed the basis for ministerial decisions on which offers should be taken up.

There is genuine innovation in what the government has done here, and DFID should be commended for designing such a rigorous system for ensuring that results are at the centre of UK aid. The layers of peer-review and scrutiny built into the process add significant credibility to the review, and key concerns about the results agenda (for example, the risk of incentivising easy-wins over more important challenges) appear to have been thought through. CAFODagrees in principle with the government’s drive towards results, so long as this does not lead to a focus on short-term wins over addressing the structural causes of poverty. Increasing the emphasis on the MDGs could be seen as risky on these terms[2]. It is critical that the results-focus does not lead to an over-reliance on quantifiable interventions where attributing the causality of positive change is straight-forward. Our scrutiny of DFID’s detailed country-plans will focus particularly on this issue.

The review is, however, “bottom-up” only if you count the “bottom” to be DFID Country Offices. The timelines of the review limited the extent to which partner governments or broader civil society could be engaged in the process, and CAFOD’s partners have reported a lack of consultation at the country-level. As operational plans are rolled out it engagement with civil society in-country will remain important. CAFOD would be happy to help facilitate this if helpful.

Lacking more meaningful country-level engagement, this remains a system for allocating aid, rather than a two-way negotiation of the part aid will play in strategies for development. There is an assumption that whatever aid is allocated from the UK will be received by countries with both hands. This is reflected in the overarching tone of the BAR documents which describe what UK aid is going to do and achieve, and leave partner governments and Southern civil society actors almost entirely invisible[3].

The ‘poverty focus’ modelling

CAFOD had been concerned that on its own, the Offers system would risk giving unfair advantage to countries which already receive high levels of UK aid. The larger, more highly funded DFID offices would have greater capacity to prepare strong Offer documents – and the resulting evaluation would therefore become a function of the existing British aid system. We know that patterns of global poverty look very different than they did a decade ago[4] – and it is important that UK aid distributions are informed by the reality of how poverty and injustice are manifest in our world today.

We were therefore very pleased to see that DFID combined their Offer analysis with a ‘poverty-focus model’, which ranked countries according to development need and the potential for aid to be effective. These types of models are always subject to debate – but they provide a vital means to assess the need for aid in a systematic, principle-based way, which can then held up to scrutiny. This was a key issue in CAFOD’s advocacy around the aid reviews, and we are delighted that our concern here has been addressed.

In terms of the model itself, there are similarities between the government’s formula and one that CAFOD had been experimenting with ourselves to assess the need for aid. We too, had been using a compound Index of HDI (Human Development Index), together with a source for the number of people living on under $2 per day (weighted so as not to unduly skew the Index towards populous middle-income countries with high numbers of poor people).

There are questions however. Firstly, as other commentators have noted[5], there is a curious contradiction between the Fragility Index, which is used to score fragile states with a greater need for aid – and the World Bank’s CPIA score, which is used to give countries with weak institutional and policy frameworks a lower score for aid. These two measures are of very similar things, and yet are set work in opposing ways. The technical note says that the use of the CPIA “balances out” the fragility part of the Index – but it does not explain the thinking behind this or justify the apparent contradiction. There are also questions around the legitimacy of the CPIA index as an indicator of good governance[6].

The contradiction in the poverty focus model is indicative of a broader difficulty with using ‘fragility’ as a category for Indexing: It encompasses such a range of development issues that the rationale for using fragility in aid assessments tends to lack clarity. We know that fragile states lag behind on the MDGs – and we know that fragile states often have weak human rights, poor governance, and high levels of poverty generally. However, there has been a nagging concern that it is not these development issues, but the security interests of the UK which has made these countries of special interest to the government. A logical way to alleviate this concern would have been to disaggregate ‘fragility’ as a category, and allow the multiple facets of development need that it encapsulates to be accounted for independently. This would have made for a more complex conversation with aid sceptics, who respond to the argument that aid is in UK interests because it bolsters security – but it would have provided critical reassurance on the ‘militarisation of aid’ issue. Arguably, the government’s attempt to use fragility to ‘have their cake and eat it’is a weak point in the BAR, which has served neither of its audiences particularly well.

Nevertheless, overall the Bilateral Aid Review has undertaken a very challenging and complicated task in a systematic and principle-led way, and DFID deserve significant recognition for this.

1. The decision to reduce the number of DFID programmes

CAFOD supports the decision to focus DFID bilateral programmes on a smaller number of countries – so long as the withdrawal is carefully managed. Historically, a major cause of inefficiency in the aid system has been that too many donors have tried to run small aid programmes in the same country. Such fragmentation causes coordination problems, and undermines aid effectiveness. We also know that understanding the political economy of developing countries is critical to running aid programmes successfully. It makes sense to focus resources, so that a small number of high-capacity DFID offices can build deep country-knowledge and forge strong relationships with partners.

It would be interesting to knowwhy 27 (rather than, say, 22 or 32) DFID country programmes were fully funded. Presumably assessments were made on the absorptive capacityof programmes and whether the ‘rate of development return’ would diminish with higher budget lines. This question was not well addressed in the technical note accompanying the BAR.

We are very pleased that DFID will phase out their programmes slowly, and honour their existing commitments. It is important that withdrawal processes are managed in co-ordination with other donors, to avoid ‘donor orphan’ issues and ensure successful programmes can transition to new funding sources. The budget lines given in the BAR are “subject to performance and sensitive to political and economic circumstances”[7]. Does this mean that if other donors do not step into gaps being left by DFID, there might be some revision to graduation plans?

Theseissues will be a key point for scrutiny when DFID release their detailed country plans.

1. Are the BAR results biased towards the UK’s own interests?

There has been widespread concern prior to the BAR over whether the government’s focus on ‘fragile states’ would lead to aid allocations being biased in favour of the UK’s security interests – rather than given where the needs are greatest[8]. Here, we test whether there is evidence of such a bias in the BAR results.

Avoiding presumptions around ‘security interests’

The nature of security and foreign policy interestsmeans that they tend not made explicit in the public domain. This makes for it challenging to assess the extent to which security interests may have influenced aid allocations in an objective and systematic way. We may think we ‘know’ which countries are of interest to UK security – and it has become common-sense in development circles to identify Afghanistan, Iraq, Pakistan, Somalia and Yemen as the key countries to watch. Our perception of interests can be highly influenced by media reports, hear-say, and by our own prejudices – and a different view might well be taken by those from a military, anti-terrorism or diplomatic background. It is important to work from an empirical basis when assessing whether aid allocations are biased towards UK security interests – and avoid jumping to conclusions.

An obvious option for empirically mapping security interests is to look at troop deployment. The UK’s only current ‘high intensity’ operation is in Afghanistan (around 10,000 troops) – but it has small operations in Cyprus, the Falklands, Gibraltar, Sierra Leone and Pakistan; and permanent posts in Belize, Brunei, Canada, Germany and Kenya[9]. NATO have troops stationed in Afghanistan, Iraq, the Horn of Africa/ Gulf of Aden, Kosovo, Sudan, and the Mediterranean Sea[10]. The UN (to which Britain is a major contributor of personnel) has large operations in DRC, Haiti and Sudan, and smaller operations in Western Sahara, Chad, Cote d’Ivoire, India and Pakistan, Lebanon, Liberia, Morocco, Syria and Timor-Lest[11]. These patterns of deployment would back-up some elements of the common-sense view where UK security interests lie, but they would also raise some questions.

If we take the UK security interests to be primarily concerned with tackling terrorism, a useful source is the University of Maryland’s Global Terrorism Database[12]. This shows on a country-by-country basis how many violent incidents have been counted as ‘terrorist’ around the world.

An alternative indicator of security interests might be the budgets of the UK diplomatic mission to different countries, or the number of civil servants allocated to different country desks in the Foreign Office and Ministry of Defence. Unfortunately, this data is not available.

Given these information gaps, CAFOD explored a number of alternative sources for data. We have looked at the Foreign and Commonwealth Office security advice[13], which provides a standardised assessment of the level of risk from terrorism in different countries around the world. Giving points for the various threat levels “high” (100), “general” (50), “underlying” (30) and “low” (10) we constructed an Index that ranks the extent of terrorism risk.

This reveals some interesting patterns. Our “common sense” countries – Afghanistan, Iraq, Pakistan, Somalia and Yemen – are indeed listed as “high” terror threats; but so are countries where DFID are withdrawing or have no presence – such as Colombia, Niger and Indonesia. India, which tends to be discussed by development experts in terms of its MIC status, is also a “high” threat.

Thesevariousindicators of UK securityinterests have their weaknesses – and it would be a great help if the government applied their commitment to transparency to providing better data on these issues. They do, however, enable us to analyse the issue of whether aid allocations are biased in a more empirical way than by relying on ‘common sense’ alone.

Has there been a shift towards security interests?

We can address this question by using correlations to map the overall patterns in aid allocations. A correlation is a single number between -1 and 1, which describes the degree of relationship between two variables. A correlation of 1 means that two variables are perfectly positively correlated (the more of one thing, the more of something else). A correlation of -1 means they are perfectly negatively correlated (the more of one thing, the less of something else).

For example, there is a strong positive correlation (0.8) between flows of remittances and patterns of immigration. This can be shown by plotting the value of remittances to different countries against the number of UK residents who were born in other countries (see Figure 1).

Figure 1: A strong positive correlation (0.8) between flows of remittances and immigration patterns

On their own, correlation statistics cannot be read too deeply – but they are valuable when considering the relative strength of different correlations to each other. The value of correlations is to explore the changes in patterns. It should be noted that we do not imply any causation from these patterns.

We can use the sources outlined in the previous section to calculate the correlations between indicators of security interests and UK aid allocations, using a large group of low and middle-income countries (126 for which consistent data could be found).

The BAR results allocations were compared with the pre-BAR bilateral spending, from 2009[14]. There arelimitations to comparing this data with the BAR results, but it seems to be the best that is available. The 2009 bilateral spending data has been collected on a calendar year basis, whilst the BAR results cover financial years. 2009 is the most recent year for which full data is available. An alternative would be to use the 2010/11 figures in the Technical Report of the BAR (Annex F), butonly gives budget lines for countries that will continue to receive some amount of bilateral funding, and these are themselves only provisional figures.

Table 1: Correlations between security interests and UK bilateral aid, before and after the BAR

UK troops / UN peacekeeping deployment / NATO troops / FCO terror threat level / Terrorist incidents globally
Bilateral allocations in 2009 / 0.30 / 0.29 / 0.39 / 0.35 / 0.07
P-value / 0.0000 / 0.0000 / 0.0000 / 0.0000 / 0.0000
Bilateral allocations 2011/12-2014/15 (after the BAR) / 0.36 / 0.19 / 0.31 / 0.34 / 0.12
P-value / 0.0000 / 0.0271 / 0.0003 / 0.0648 / 0.1456

Running a two-tailed t-test, we tested the likelihood that these correlations are not due to chance (see worksheet). All the correlations are significant at a confidence interval of 95% (p-values are less than 0.05), apart from the correlation between FCO terror threat level and bilateral aid post-BAR, and terrorist incidents globally and bilateral aid post-BAR (shaded grey).

For most of the indicators there is a weak positive correlation between ‘security interests’ and aid allocations. This would support the hypothesis thatcountries scoring higher levels of ‘security interest’ for the UK do have slightly higher levels of aid. These correlations do not necessarily indicate, however, that security interests have caused aid to be allocated more highly in these countries[15].

The strength of correlation between security interests and UK aid allocations has stayed roughly the same, following the Bilateral Aid Review results. These figures reflect the fact that on the one hand, DFID are withdrawing from Iraq and Indonesia which are both ‘high terror’ threats according to the FCO, and pulling out of Kosovo, which has deployment of NATO troops; whilst on the other hand, DFID are doubling their spending in Pakistan, which is a ‘high’ terror risk, has a small UK troop operation and UN troops in place; and increasing aid to Yemen and Somalia (both ‘high’ terror threat countries).

Countries considered ‘high’ terror threat are getting larger budgets on average – however, they are getting a significantly lower amount of aid on a per capita basis. The average per capita spend for countries judged to be a ‘high’ terror threat by the FCO is £9.43; whilst the average per capita spend for countries not judged to be a high terror threat by the FCO is £16.43[16]. This is partly due to the very low per capita spending in countries like India (£0.94 per capita, ‘high terror threat’ country) – and the decision to pull out of Iraq, Niger and Indonesia (all ‘high terror threat’ countries). The average overall budget for ‘high’ terror threat countries is £558m; whilst for non-‘high’ threat countries it is £343 (2011-2015).