Economic Development and the Effectiveness of Foreign Aid: A Historical Perspective

Yves here. Ebola is serving as a reminder that the fate of members of advanced economies isn’t necessarily divorced from those of citizens of poor, developing nations. And it isn’t as if those countries are completely neglected. They are simultaneously the recipients of foreign aid, while at the same time being de facto capital exporters. So while this study below is informative, it ignores the elephant in the room, which is the degree to which looting simply overwhelms the amount of funding provided by foreign aid.

As Nicholas Shaxson wrote in Treasure Islands (p. 157):

Global Financial Integrity (GFI) in Washington authored a study on illicit financial flows out of Africa (March 2010). Between 1970 and 2008, it concluded:

Total illicit financial outflows from Africa, conservatively estimated, were approximately $854 billion. total illicit outflows may be as high as $1.8 trillion… The GFI estimate – equivalent to just over 9 per cent of its $51 billion in oil and diamond exports during that time – simply has to be a gross underestimate of the looting. Many billions have disappeared offshore through opaque oil-backed loans channeled outside normal state budgets, many of them routed through two special trusts operating out of London.

… GFI’s shocking estimates complement the figures I mentioned – ten dollars out for every dollar of foreign aid flowing in.

… Another study emerged in April 2008 from the University of Massachusetts, Amherst … to examine capital flight from forty African countries from 1970 to 2004. Its conclusions are striking.

Real capital flight over the 35-year period amounted to about $420 billion (in 2004 dollars) for the 40 countries as a whole. Including imputed interest earnings, the accumulated stock of capital flight was about $607 billion as of end-2004.

Yet at the same time, the total external debt of these countries was ‘only’ $227 billion. So, the authors note, Africa is a net creditor to the rest of the world, with its net external assets vastly exceeding its debts. Yet there is a crucial difference between the assets and the liabilities: The subcontinent’s private external assets belong to a narrow, relatively wealthy stratum of its population, while public external debts are borne by he people through their governments.

Factoring in the impact of this pilfering might make the results of foreign intervention seem less random. It would also help if advanced economies weren’t such eager facilitators of this type of expropriation.

By Sebastian Edwards, Henry Ford II Professor of International Economics at the University of California, Los Angeles. Originally published at VoxEU

The effectiveness of official development aid is the subject of heated debate. This column argues that aid affects recipient economies in extremely complex ways and through multiple and changing channels. Moreover, this is a two-way relationship – realities in recipient countries affect the actions of aid agencies. This relationship is so intricate and time-dependent that it is not amenable to being captured by cross-country or panel regressions. Even sophisticated specifications with multiple breakpoints and nonlinearities are unlikely to explain the inner workings of the aid–performance connection.

Foreign aid is controversial in development economics. Three distinct camps may be distinguished:

• One believes that official assistance is ineffective, and has harmed poor countries throughout the years.

This views official aid as creating dependency, fostering corruption, and encouraging currency overvaluation (Easterly 2014 and Moyo 2010). It also prevents countries from taking advantage of the opportunities provided by the global economy.

• Another camp believes that aid levels have been too low, and that large increases would help reduce poverty.

This camp, however, believes we need a rethinking on the way in which aid is provided (Sachs 2009 and Stiglitz 2002). In particular, specific interventions, such as anti-malaria programmes, should be emphasised.

• The third camp is less vocal, and includes authors such as Collier (2007), who has emphasised the role of a number of ‘traps’ in perpetuating destitution, and Banerjee and Duflo (2011) who argue that the use of ‘randomised control trials’ may help devise effective and specific aid programmes in the war against poverty and underdevelopment.[1]

These schools of thought have historical precedents.

Foreign Aid Policies from a Historical Perspective

Foreign aid is a relatively new concept in economics. The classics – Smith, Ricardo, and Stuart Mill, for example – didn’t address the subject in any significant way. If anything, classical economists thought that the colonies would catch up – and even surpass – the home country quite rapidly.[2] In Chapter VII of The Wealth of Nations, Adam Smith provides a detailed discussion on the “causes of the prosperity of the new colonies.”[3]

The first legal statute dealing expressly with official aid was passed by Parliament in the UK in 1929.[4] In 1940 and 1945, new laws dealing with aid to the colonies were passed in the UK. These Acts increased the amount of funds available, and made commitments for longer periods of time – for up to ten years in the Colonial Development and Welfare Act of 1945. More important, the Act of 1945 established that aid plans had to be prepared “in consultation with representatives of the local population.”[5]

In the US the first law dealing with foreign assistance came quite late, with the adoption of the Marshall Plan in 1948.[6] In his inaugural speech on 20 January 1949 – the so-called ‘Four Point Speech’ – President Harry Truman put forward, for the first time, the idea that aid to poor nations was an important component of US foreign policy. He said that one of the goals of his administration would be to foster “growth of underdeveloped areas.”[7]

In spite of Truman’s vehement allocution, aid commitments to poor countries were considered temporary. In 1953, when Congress extended the Mutual Security Act, it explicitly stated that economic aid to US allies would end in two years; military aid was to come to a halt in three years.

In the early 1960s – and largely as a result of the escalation of the Cold War – the US revised its posture regarding bilateral assistance, and, jointly with other advanced countries, founded the Development Assistance Committee (DAC) at the newly formed Organisation for Economic Cooperation and Development (OECD). The main objective of the DAC was – and continues to be – to coordinate aid to the poorest countries.[8]

Foreign Aid Policies and Academics’ Views on Development

Academic research has helped shape international aid policies. During the 1950s and 1960s, aid agencies’ work was influenced by the Harrod–Domar growth model and by W. Arthur Lewis’ unlimited supplies of labour model. As a result, most agencies funded very large capital-intensive projects, and neglected policies, projects, and programmes related to labour, human capital, and productivity.

This changed in the late 1960s and 1970s with the ascendance of Solow’s neoclassical model of growth, and the development of the ‘basic needs’ approach to welfare economics. Aid policies changed focus, and a higher percentage of funds were devoted to social programmes (health and education), programmes aimed at directly reducing poverty, and programmes that strengthened skills and human capital.

Further changes in aid policy came with research that related openness and exports’ expansion to productivity growth. The work of Anne Krueger and Jagdish Bhagwati was particularly important. During the 1980s and 1990s, international assistance became increasingly conditioned on the recipient countries liberalising their economies through the elimination of quantitative import restrictions and the lowering of import tariffs.

The development of the ‘dependent economy’ macroeconomic model, with tradable and nontradable goods, in the late 1970s and early 1980s, helped put emphasis on the crucial role of the real exchange rate in the resource allocation process. Works by Robert Mundell, Rudi Dornbusch, and others pointed out that real exchange rate overvaluation was costly and at the heart of devastating currency crises. These works, in conjunction with research undertaken by Robert Bates and Elliot Berg, among others, influenced aid agencies’ views regarding currencies, incentives, exports, and agriculture. The very poor performance of the agricultural sector between 1965 and 1985 in most regions – and in particular in Africa – also affected thinking in the aid agencies, and contributed to a new view that emphasised ‘getting prices right’.

In the 1990s, two research lines influenced aid policy.

• Work on incentive compatibility and strategic behaviour persuaded aid officials in donor countries to become more flexible, and to incorporate recipient governments in the design and management of aid programmes.

This approach received the name of ‘programme ownership’, and has been at the heart of improved relations between donors and poor nations in the last two decades.

• New research on capital mobility and the international transmission of crises, resulted in a more nuanced and pragmatic view regarding the use of capital controls.

Many agencies – including the IMF and the World Bank – supported a limited use of capital controls (especially controls on capital inflows) and so-called macro-prudential regulations, as a way of avoiding destabilising forces and currency crises.

Econometric Studies

Academic and aid-community economists have used a battery of econometric methods to analyse whether aid is effective in the sense of generating higher growth and better economic outcomes.[9] Some of these studies have tried to tackle issues of reverse causality, and have used a series of instruments – some more convincing than others – in an attempt to deal with the fact that slower growth (in very poor countries) may attract additional aid.

Some research focused on whether aid only works under certain conditions, or whether a minimal degree of institutional development is required for international assistance to bear fruit (Burnside and Dollar 2000, 2004). Many of these studies have considered nonlinear functional forms, and have investigated if there are meaningful interactions between aid and other variables, such as the degree of literacy, the level of corruption, the extent of macroeconomic stability, institutional strength, the quality of overall economic policies, and geography.

In general, most studies have relied on cross-country or panel data, and have attempted to distinguish between short- and long-term impacts. A number of authors have used ‘Dutch disease’-related models to analyse the extent to which increased aid results in currency overvaluation, poor exports performance, and crises – see Rajan and Subramanian (2011).
Fragile and inconclusive results

Overall, the results from this large body of research have been fragile and inconclusive.

After analysing 97 studies, Doucouliagos and Paldam (2008, 2009) concluded that, in the best of cases, it was possible to say that there was a small positive, and yet statistically insignificant, relationship between official aid and growth.

This conclusion was also reached by Rajan and Subramanian (2008) in an analysis that corrected for potential endogeneity problems, and that considered a comprehensive number of covariates. In particular, according to this study there is no clear relation running from more aid to faster growth; this is true even in countries with better policy environment and stronger institutions – see also Rajan and Subramanian (2008) and Quibria (2014).

Bourguignon and Sundberg (2007) have argued that one should not be surprised by the inconclusiveness of studies that rely on aggregate data.

According to them, aid affects economic performance, directly and indirectly, through a variety of channels. Treating all aid as homogeneous – independently of whether it is emergency assistance, programme aid, or project-based aid – is misleading. In their view it is necessary to break open the ‘black box’ of international aid, and deconstruct the causality chain that goes, in intricate and non-obvious ways, from aid to policymakers, to policies, and to country outcomes. This type of analysis would explore a number of specific ways in which international assistance could impact economic performance. In particular, according to Bourguignon and Sundberg (2007) it is important that studies that try to determine the impact of aid on growth consider issues related to technical assistance, conditionality, level of understanding of the economy in question, and the government’s ability to implement specific policies.[10]

Bitter Policy Controversy

In spite of its intensity, the academic debate pales in comparison with recent policy controversies on the subject. The level of animosity in this veritable war of ideas is illustrated by the following quote from an article by Jeffrey Sachs published in 2009:

Moyo’s views [are] cruel and mistaken… [Moyo and Easterly are] trying to pull up the ladder for those still left behind.[11]

Easterly’s reply, also from 2009, was equally strong:

Jeffrey Sachs [is]… the world’s leading apologist and fund-raiser for the aid establishment… Sachs suffers from [an]… acute shortage of truthiness…[12]

The Easterly–Sachs debate has generated public attention because it has been couched in rather simple terms. These are simple narratives based on ethnographic arguments that resonate with large segments of the general public. But behind the different positions there are hundreds of academic studies – most of them based on advanced econometric techniques – that have tried to determine the extent to which foreign aid is effective. The problem, as noted, is that much of this body of empirical work has resulted in fragile and inconclusive evidence.

Intermediate Positions

For an increasing number of economists, the issue of aid effectiveness is neither black nor white. Indeed, a number of authors have taken intermediate positions. For example, in an influential book that deals with the plight of the poorest of the poor, Collier (2009) has argued that both critics and staunch supporters of official aid have greatly exaggerated their claims and distorted the empirical and historical records.

Collier’s reading of the evidence is that over the last 30 years official assistance has helped accelerate GDP growth among the poorest nations in the world – most of them in Africa – by approximately 1% per year. This is a nontrivial figure, especially when one considers that during this period the poorest countries have had an aggregate rate of per capita growth of zero. That is, in the absence of official assistance, the billion people that live in these nations – the so-called ‘bottom billion’ – would have seen their incomes retrogress year after year.

Banerjee and Duflo (2011) argue that there is need for a “radical rethinking of the way to fight poverty.” In their view, the acrimonious debate between the Easterly and Sachs factions has missed the boat. Banerjee and Duflo join a growing group of researchers in arguing that this controversy cannot be solved in the abstract, by using aggregate data and cross-country regressions. The evidence, in their view, is quite simple – some projects financed by official aid work and are effective in reducing poverty and moving the domestic populations towards self-sufficiency and prosperity, while other projects (and programmes) fail miserably. The question is not how aggregate aid programmes have fared in the past, but how to evaluate whether specific programmes are effective.

Persuasive ‘Aid Narratives’

In Edwards (2014b) I discuss the effectiveness-of-aid literature from a historical perspective, and I argue that international aid affects recipient economies in extremely complex ways and through multiple and changing channels. Moreover, this is a two-way relationship – aid agencies influence policies, and the reality in the recipient country affects the actions of aid agencies. This relationship is so intricate and time-dependent that it is not amenable to being captured by cross-country or panel regressions; in fact, even sophisticated specifications with multiple breakpoints and nonlinearities are unlikely to explain the inner workings of the aid–performance connection.

Bourguignon and Sundberg (2007) have pointed out that there is a need to go beyond econometrics, and to break open the ‘black box’ of development aid. I would go even further, and argue that we need to realise that there is a multiplicity of black boxes. Or, to put it differently, that the black box is highly elastic and keeps changing through time. Breaking these boxes open and understanding why aid works some times and not others, and why some projects are successful while other are disasters, requires analysing in great detail specific country episodes. If we want to truly understand the convoluted ways in which official aid affects different economic outcomes, we need to plunge into archives, analyse data in detail, carefully look for counterfactuals, understand the temperament of the major players, and take into account historical circumstances. This is a difficult subject that requires detective-like work.

See original post for references and footnotes