Investment for development: The UK’s strategy towards Development Finance Institutions

Our response to Call for Evidence.

Mustapha Douch[1], T.Huw Edwards[2], Todd Landman[3] and Sushanta K. Mallick[4]

Human Rights provision and the effectiveness of UK Aid programmes.

In the Terms of Reference[5], we refer in particular to the following:

• What criteria does BII use to determine investment decisions and how are financial returns balanced with achieving impact?

In addition, from the Background section

in light of diminishing Government budgets and the impact of the war in Ukraine, it is important for projects to deliver maximum impact overseas in lifting people out of poverty.

Also, the following from Chair of the Select Committee, Sarah Champion MP:

As we await the Autumn Statement, we will be asking whether BII delivers impact and value for the UK taxpayer, do the funds reach those most in need and is there transparency and accountability over where this money is invested.

More specifically, our interest is particularly in the criteria for selection of countries for BII projects. There is an announced shift in emphasis towards two regions: i) the Indo-Pacific (in particular, the Philippines and Indonesia plus the Mekong countries of Laos, Cambodia and Vietnam) and ii) The Caribbean. However, this is an expansion from the existing core of supported countries (e.g. Africa and South Asia). Much of this is outlined in the BII’s 2022-26 Technical Strategy Document. Appendix A of this document lists the current recipient countries (split into Mature, Powerhouse, Stable and Poorest and Most Fragile groups) and newer members – Indo-Pacific (climate finance) and Caribbean.

Brief summary of our findings

1. We emphasise the importance of aid conditionality – the effectiveness of aid in achieving socio-economic goals depends to a large degree upon the socioeconomic – and particularly governance environment of the recipient country.
2. We have investigated in detail the effects of human rights (especially personal integrity) provision on the effectiveness of aid, as measured by the contribution to economic growth, and find a strong positive association.
3. On the basis of this, we advocate that human rights (HR) measures should 4be used in a basket of conditions in assessing aid recipient countries, alongside more standard economic measures.
4. There is also evidence that better HR provision tends to be associated with lower levels of socioeconomic inequality. While we have not specifically examined other social outcomes, this can be seen as supporting an a priori case that better HR will also improve the outcome of wider social indicators (other than economic growth) in response to aid.
5. Countries receiving aid usually tend to be below average in terms of HR provision. This is the case with DFID recipients, in particular, possibly because the countries with highest need often tend to be those experiencing poor governance.
6. While other UK aid providers, including BII, tend to have a less bad HR record on average than DFID, there are reasons why caution should be taken in providing, and monitoring aid to countries at the lower end of the HR scale.

Our study: Human Rights and Aid Conditionality .

We have recently published the following article, on which we draw:

Douch, M., Edwards, T.H., Landman, T. and Mallick, S.K. (2022). “Aid Effectiveness: Human Rights as a Conditionality Measure”. World Development 158, October.

We note that there is a broad consensus around the ‘conditionality hypothesis’, as introduced by Dollar and Pritchett (1998) and Burnside and Dollar (2000); where the policy environment meets certain conditions, aid has been generally effective in producing economic growth and/or social/environmental benefits. For reasons of tractability and consistency with the literature, we focus primarily on the growth impacts of aid, but we regard wider impacts in Environmental, Social and Governmental (ESG) indicators (see BII(2022) Section 3.5) as being highly relevant, and we revisit this question in the Discussion section.

The conclusion of the conditionality literature is that various policy and governance indicators are useful, not just as performance measures for development assistance, but as pre-selection indicators, to ensure that aid is spent in those political/governmental environments where it will be effective. The original Burnside and Dollar (2000) article used a policy variable composed of three policy dimensions: the government’s budget surplus for fiscal policy, the inflation rate for monetary policy and the Sachs Warner index for openness or trade policy, and we accept that these variables are positively related to growth (Arndt et al. (2015)). However, drawing on research in political science, development studies, and human rights, we argue that there is a strong case for the use of more political and economic indicators as well (Szirmai, 2012). In particular, we investigate the effectiveness of using measures of human rights protection (Landman and Larizza, 2009) as a potential screening variable in the allocation of aid.

Our primary chosen Governance Indicator (human rights index).

Our human rights measure (HR Factor) follows Landman and Larizza (2009) and produces a summary index of civil and personal integrity rights using four ‘standards-based’ (Jabine and Claude (1992)) human rights (HR) scales: (1) the Amnesty International version of the Political Terror Scale, (2) the U.S. State Department version of the Political Terror Scale, (3) the Cingranelli and Richards (1999) Index of Personal Integrity Rights (http://www.humanrightsdata.com), and (4) the Freedom House civil liberties scale (Landman and Larizza (2009)). Since these different scales are significantly positively correlated, we derive a single summary measure using principal components factor analysis, producing an aggregate ‘personal integrity rights scale’. To make it intuitive, we then invert this, so that low values of the factor score correspond to a low protection of human rights (high violations) and high values correspond to a high protection of human rights (low violations). This variable has a mean of 0, a minimum value is -2.7 and the maximum value is 1.97.

Figure 1 reports the HR factor distribution across the globe. This clearly emphasizes the low level of HR protection in some African countries as well as in Asia and Latin America with respect to Europe, Australia and North America.[6]

Figure 1: Human Rights Distribution, this graph represents the average inverse HR factor for all countries including Least Developed Countries (LDCs). Note, positive HR Factor indicates better human rights protection. Average for 1989-2012.

To summarize – our HR indicator (IHRFACTOR2) is positively correlated with:

1. The level of development of a country.
2. Other governance indicators (such as the POLITY4 democracy indicator, measures of macroeconomic stability, or the KOF globalization indicator).
3. Most indicators of human needs.
4. Economic equality.

We note that causality is not simple in any of these cases. More advanced countries generally demand higher levels of both HR provision and democratic decision-making, as well as the rule of law and provision of economic freedoms and basic human needs protection. At the same time, most of the governance indicators listed tend to produce better economic and social outcomes. In this regard, we note that HR provisions are inversely correlated with measures of corruption, such as the corruption perception indices produced by Transparency International.

Figure 2: relationship between HR and corruption perception (Source: Transparency International) 2010.

There are strong theoretical reasons, which we discuss in our paper, to believe that better HR provisions (along with democracy and the rule of law) not only help countries to contain problems of corruption, but that as a result they produce economies which function better in terms of spending resources well, both in terms of economic growth and other social objectives (Douch et al, 2022). Indeed, good standards of HR, particularly personal integrity rights, are important for the functioning of the legal and democratic systems.

The effect of Human Rights on Aid effectiveness

We examine the relationship between HR provision, net aid receipts (as percent of GDP) and the growth of GDP for a panel of 42 Least Developed Countries (as listed by the UN) between 1989-2012 inclusive. In this analysis, we also take account of series of other control variables, such as the KOF index for economic globalization, income inequality, child mortality and inflation rates. Subsequently, we also construct a policy index, based upon the classic Burnside and Dollar (2000) study.

The most important, and consistent conclusion, of our studies is that, while aid receipts benefit GDP across almost all countries, the estimated effect of aid is consistently and significantly better for countries with better human rights than those with worse human rights.

We also find positive effects from good macroeconomic policy (control of inflation, good budget, openness to trade), but the interaction with aid is more variable.

We also plot the relationship between the effect of increasing aid (‘marginal effect’) on GDP at various human rights levels. An example is shown in Fig 3.             

Figure 3: marginal effect of increasing aid on GDP, at various HR levels.

The various points on Figure 3 correspond to representative countries: the worst (Human rights -2.631 average over the period) corresponds to Myanmar, where the simulated effect of aid would probably be slightly negative, to the best (Kiribati, 1.0053 and Vanuatu, 0.638), where aid would be expected to be clearly beneficial. In between, aid is estimated to has appositive but varying effect on countries such as Bangladesh, Nepal, Mozambique and Lesotho.

We also carry out a comparable analysis on middle income aid recipients, and confirm the human rights provision also improves the effectiveness of aid in these cases.

Implications for the UK’s choice of aid recipients

In the Appendix, we list countries from two lists: UK official AID recipients in 2019 from FCDO statistics and countries listed in Appendix A of the BII’s Technical Report as being currently supported, or part of the post-2022 expansion into the Indo-Pacific and Caribbean regions. A few countries have to be dropped, as we do not have human rights data. The remaining countries are ranked in terms of average human rights provision over the period of our study (1989-2012).[7]

Taking 0 as the mean HR score over all countries, we see that many of the countries on this list are relatively poor in terms of HR. Aid recipients tend to be towards the poorer end of the spectrum for most governance indicators. Even though DFID does not provide aid to China, North Korea or Iran, in 2019 it supported Syria, Sudan, Pakistan, Eritrea, Somalia, Zimbabwe, Afghanistan and Iraq. Of course, this may to some extent reflect humanitarian, rather than economic objectives.

Taking numerical averages of HR scores for DFID aid recipients, all UK aid recipients or those named in the BII pre-2021 and post-2022 strategies, we find the following.

This indicates that, indeed, DFID recipients tend to be at the poorer end of the spectrum in terms of HR provision, although it is perhaps only those at the very bottom end of the scale (listed above) where there should be a question of whether aid in fact makes a positive contribution to growth.

Non-DFID aid is reported as going also (in small amounts) to China and North Korea. However, on the whole, non-DFID aid seems mostly to go to countries with better HR records.

This issue is even clearer where we weight DFID and other UK aid recipients by reported spending.

Syria, Somalia, Zimbabwe and Afghanistan were reported in 2019 as being large aid recipients, which may be a factor in producing this poor weighted average.

Regarding BII provision: the countries listed in the BII Appendix have relatively better average HR scores than those covered by DFID. Nevertheless, many countries at the bottom of the HR league table are on the BII list, and we should urge caution in the provision of support to these countries. We do not, unfortunately, have data on how much money BII spends in each country.

Discussion

Our analysis above suggests that analysis of countries’ human rights provision provides a potentially useful indicator of the likely prospect of positive outcomes from aid, at least in terms of economic growth. We discuss here a few points which may be of relevance to the Committee.

1. When examining aid conditionality, the literature has included several measures, some of which we examine. We acknowledge that many of the better governance indicators are, in fact, correlated, so that it is hard to say that any individual measure is the main driver. In general, as we do with inclusion of the more economic measures in Burnside and Dollar (2000)’s paper, we would advocate studies using various baskets of governance indicators: however, the evidence we have is that HR provision should be taken seriously as one of these indicators.
2. The BII (2022) Technical Report gives plenty of emphasis to wider ESG aims, rather than simply economic growth. We have not specifically investigated the relationship between HR, aid and these other indicators. However, previous studies, such as Landman and Larizza (2009) show a strong positive correlation between better HR performance and greater economic equality (without confirming the direction of causality), and our own paper indicates that one of the potential routes by which HR provision affects economic performance is by curtailing the misappropriation of resources by oligarchs. This would indicate that it is likely that aid will produce better social, as well as economic outcomes, where HR provision is better, although more research is merited.
3. We do not specifically examine the relationship between the type of aid provision and its effectiveness. We acknowledge that the BII is primarily aimed at supporting private (or public/private) schemes, and that projects will be monitored for ESG performance. While we cannot specifically comment on how effective these measures might be, we agree that it would be a reasonable precaution that, where aid is given to a country with poorer governance, these other measures should be more strictly enforced.

References

Our recent paper:

Douch, M., Edwards, T.H., Landman, T. and Mallick, S.K. (2022). “Aid Effectiveness: Human Rights as a Conditionality Measure”. World Development 158, October.

Other references

Arndt, C., Jones, S., and Tarp, F. (2015). Assessing foreign aid’s long-run contribution to growth and develop ment. World Development, 69:6–18.

Burnside, C. and Dollar, D. (2000). Aid, policies, and growth. The American Economic Review, pages 847–868.

Cingranelli, D. L. and Richards, D. L. (1999). Measuring the level, pattern, and sequence of government respect for physical integrity rights. International Studies Quarterly, 43(2):407–417.

Dollar, D. and Pritchett, L. (1998). Assessing aid: A world bank policy research report. Washington DC.

Jabine, T. B. and Claude, R. P. (1992). Human rights and statistics: getting the record straight. University of Pennsylvania Press.

Landman, T. and Larizza, M. (2009). Inequality and human rights: who controls what, when, and how. Inter- national Studies Quarterly, 53(3):715–736.

Szirmai, A. (2012). The Dynamics of Socio-Economic Development: An Introduction. Cambridge University Press, Cambridge

Online documents

BII(2022) https://assets.bii.co.uk/wp-content/uploads/2022/01/06170001/2022-2026-technical-strategy-2.pdf

APPENDIX: Average Human Rights Scores (1989-2012) of current UK AID recipients and BII supported countries.