-
Notifications
You must be signed in to change notification settings - Fork 0
/
paper.qmd
127 lines (77 loc) · 25.7 KB
/
paper.qmd
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
---
title: "Foreign Aid, Democracy, and Poverty: An Inquiry into the Impact of European Union Development Assistance"
author: "Amanda Nascimento Dial"
format:
cmc-article-pdf:
fontsize: 12pt
margin-left: 1.5in
margin-right: 1.5in
margin-bottom: 1.2in
margin-top: 1.2in
keep-tex: true
fig-pos: 'tbp'
tbl-pos: 'tbp'
include-in-header: header.tex
bibliography: bibliography.bib
abstract: |
With the adoption of the Sustainable Development Goals (SDGs), the international community committed itself to eradicating poverty by 2030, a feat thought possible through the use of foreign aid. However, assessing aid’s efficacy for poverty alleviation is a challenging endeavor due to endogeneity problems. I harness a unique source of exogeneity in the allocation of foreign aid first identified by @carnegie_foreign_2017 in an attempt to address this problem. Using an instrumental variables methodology, I assess the impact of aid from the European Union (EU) on both monetary and non-monetary metrics of poverty, controlling for various measures of democracy, institutions, and economic conditions. While I find no compelling evidence of an effect of EU aid on poverty, my results are consistent with theoretical expectations of the intertemporal effect of aid and provide a novel framework for future studies with higher quality data to examine this critically important relationship. ^[All code and data necessary for replication of this study are available at the following GitHub repository: <https://github.com/amanda-dial/foreign-aid-poverty-paper>]
---
# Introduction
Despite monumental progress in the alleviation of global poverty in the last century, roughly half of the world's population today still lives on less than \$6.25 a day, with almost 700 million living in extreme poverty (less than \$2.15 a day).[^1] The adoption of the Sustainable Development Goals (SDGs) in 2015 signaled the international community's intent to eradicate extreme poverty and lift half of the world's impoverished population out of all forms of poverty by 2030. Of the available means of accomplishing this goal, foreign aid is one of the preferred choices for OECD nations: in 2022 alone, \$204 billion in official development assistance (ODA) was allocated to the world's poorest nations.[^2] However, aid is also among the most contentious methods of development support.
[^1]: "Poverty." 2023. *worldbank.org*. <https://www.worldbank.org/en/topic/poverty/overview>
[^2]: "Official development assistance (ODA)." 2023. *OECD.org*. <https://www.oecd.org/dac/financing-sustainable-development/development-finance-standards/official-development-assistance.htm>
Prior empirical studies on the effectiveness of aid for alleviating poverty yield conflicting results, with some offering support and others critiques of aid's ability to reach the poor. @bahmani-oskooee_poverty_2009 and @hirano_aid_2014 both find unequivocal and significant support for the poverty-reducing impact of aid. These studies suggest that aid fuels economic growth while promoting more equitable distributions of income, effects that directly benefit the poor. Contrarily, @chong_can_2009 and @boone_politics_1996 argue that aid has questionable effects on poverty. On its own, aid neither improves human development nor reduces commonly reported poverty metrics such as the poverty gap. Although disconcerting, such contradictory results raise a nuanced point: additional mechanisms may be at work beyond aid that impact its effectiveness. Increasingly, studies such as @kosack_effective_2003 focus on examining how institutional contexts in recipient countries influence aid efficacy. As @kosack_effective_2003 demonstrates, aid has a significant, positive relationship on quality of life in democracies, but a negative effect in autocracies.
The inconclusive evidence in the literature regarding aid's impact on poverty demonstrates the abundant quantitative challenges scholars face in their attempts to discern the relationship between aid and poverty. A lack of high-quality data on poverty and common poverty covariates such as democracy and institutions are likely contributing to biased or mistaken conclusions. Empirical studies on aid also face endogeneity problems. Crucially, donor countries allocate aid non-randomly based on existing levels of poverty in target countries, complicating efforts to extract a causal effect of aid on poverty. Many prior studies have adopted a wide array of methods to address this limitation, with varying degrees of success. Therefore, this study is poised to attempt a new causal inference strategy to adjudicate the debate on aid and poverty.
Applying a novel approach of addressing the issue of endogenous aid allocation first introduced by @carnegie_foreign_2017, I harness an exogenous source of variation in foreign aid---the rotating presidency of the Council of the European Union---to instrument for foreign aid and purge it of endogeneity. Using two-stage least squares estimation, I investigate the effect of ODA provided by the European Union (EU) between 1987 and 2011 on various metrics of poverty in an effort to mediate the debate regarding aid's efficacy. Building upon Carnegie and Marinov's discovery of positive effects of EU aid on democracy, I hypothesize that the positive conditionality of EU aid, which incentivizes recipient countries to enact democratic reforms in exchange for aid, can potentially amplify the direct effect of aid on poverty. I collect data on two additional measures of democracy---electoral democracy and egalitarian democracy--- and two measures of institutional quality---rule of law and corruption---to parse out the effect of existing democratic structures in recipient countries and test whether democracy and aid affect poverty complementarily.
Although I find little compelling evidence of a statistically significant effect of EU aid on poverty, I argue that my results are of crucial economic relevance. I find that, in the immediate term, EU aid potentially worsens the poverty gap but lowers the infant mortality rate. However, five years after receipt, aid may improve both the poverty gap and infant mortality rate. I find inconclusive results for the theory that democracy and institutions tend to work in favor of poverty alleviation. Despite my null effects, I shed light on the potential of a non-trivial impact of foreign aid on poverty, but caution that future research is necessary. In addition to addressing a key empirical challenge in aid analysis, I pave the way for and urge future studies to leverage higher quality data as it becomes available to investigate the critical relationship between aid and poverty.
# The Theory of Aid and Poverty
While aid proponents argue that aid is a wholly effective way of reducing poverty, many aid skeptics question its efficacy, arguing that it is ineffective or, in some instances, counterproductive. Such arguments rely on the phenomenon known as the *Samaritan's Dilemma*, which, in the context of foreign aid, stipulates that international monetary support perpetuates dependency, degrading domestic institutional capacity and limiting self-sustainability. While this argument may indeed hold water, it is important not to lose sight of an often overlooked middle-ground proposed by aid optimists: under the right institutional conditions, foreign aid can instigate positive feedback loops that amplify its effects in the long-run.
Since the 1950s, development economists have theorized the existence of poverty traps, vicious cycles of continuously ill-met needs that perpetuate the impoverished state of developing countries. An often proposed mechanism for breaking through poverty traps is to give developing countries a "big push", or provide them with the monetary support they need to kickstart virtuous cycles of growth and development on their own. Aid can provide countries with capital and investment opportunities, thereby spurring sustainable economic growth that, in turn, alleviates poverty [@mahembe_foreign_2019; @riddel_does_2009]. Aid can also impact poverty directly by expanding government expenditure on public services like education and health or through funding of grassroots development projects.
However, since ODA is disbursed to governments or multilateral institutions, its effect on poverty is conditional upon it actually reaching the poor, which is not independent of existing institutional structures, especially democracy, rule of law, and corruption, in recipient countries. Democracy is a crucial condition for aid to reach the poor, since, as @ross_is_2006 emphasizes, democracies face greater government accountability through the electoral process and tend to spend more on public goods and income redistribution than non-democracies. As @fig-1 shows, among countries analyzed in this study, those with higher levels of electoral democracy also tend to have below median infant mortality rates and poverty gaps, both of which are common measures of poverty.
{#fig-1 height="80%" fig-align="center"}
The direct implication of such theories is that aid is more likely to be effective at reducing poverty in more democratic states. However, they do not necessarily imply that aid is ineffective in non-democracies. Indeed, as @carnegie_foreign_2017 demonstrate, the positive conditionality of EU aid, whereby recipient countries receive aid conditional upon enacting reforms in the realms of human rights and democracy, leads to short-term improvements in rights and democracy (@tbl-cm replicates this main result from @carnegie_foreign_2017). Thus, it is intuitive to expect that aid allocated with conditions to enact policy reforms that foster the conditions necessary for aid to reach the poor---even if only temporarily---may improve aid's efficacy for poverty reduction, putting into action positive aid feedback loops.
# Data and Methods
Using poverty and economic data collected from the World Bank World Development Indicators database and Poverty and Inequality Platform, democracy data collected from V-Dem, and Carnegie and Marinov's original data on EU aid disbursements in 115 countries between 1987 and 2011, I estimate the relationship between foreign aid and poverty, implementing controls for democracy, institutions, and economic conditions to examine whether democracy and aid are complementary forces.
I focus on two frequently reported measures of poverty: the poverty gap at \$6.25 a day, which measures the depth of poverty by the average distance from the poverty line as a percentage of the poverty line (a zero value means no one lives below \$6.25 a day), and the infant mortality rate, the number of infant deaths per 1,000 live births. Since poverty is a multidimensional concept, it encapsulates both monetary and non-monetary determinants of well-being. Thus, including both measures in my analysis allows for differences in the way aid affects either dimension of poverty to be detected. Although the poverty gap and infant mortality rate are limited in that they only capture a small range of deprivations experienced by the poor, they are among the most widely available and highest-quality measurements of poverty.
In addition to the Polity IV combined score used by Carnegie and Marinov to measure democracy, I introduce four new conceptualizations of democracy and institutional quality from V-Dem: electoral democracy, egalitarian democracy, rule of law, and corruption. Each metric ranges from zero to one, with a one indicating most democratic or most adherence for democracy and rule of law, but most corrupt for corruption. Electoral democracy captures how accountable government leaders are to their citizens while egalitarian democracy measures the level of equality with respect to rights and access to resources or power. Rule of law and corruption are by no means exhaustive of the complete range of institutional metrics, but are likely among the most influential in terms of impacting poverty. Finally, since poverty is likely to be affected by economic conditions, I incorporate various economic controls for gross domestic product, trade openness, government expenditure, and inflation, among others. @tbl-ds provides summary statistics of my variables of interest and controls.
As in @carnegie_foreign_2017, I am interested in estimating the following model:
$$POV_{it}= \alpha + \beta_{1}log(ODA)_{i(t-1)} + \beta_{2}Z_{it} + \eta_{i} + \theta_{t} + u_{it},$$
where $POV_{it}$ is the dependent variable of interest (the poverty gap or infant mortality rate) in country $i$ for year $t$ and $log(ODA)_{i(t-1)}$ is the independent variable of interest (the total logged EU official development assistance) received by country $i$ in year $t-1$. As aid may take time to have effect, the foreign aid independent variable is lagged by one year to capture the impact of aid receipt in the previous year. $Z_{it}$ is a vector of controls for democracy, institutions, and economic conditions that varies across model specifications. The terms $\eta_{i}$ and $\theta_{t}$ are country and year fixed effects, respectively. Country fixed effects are included to control for country-specific, time-invariant characteristics, while year fixed effects control for time-specific, country-invariant characteristics. Finally, $u_{it}$ is the error term.
I adopt the instrumental variables framework utilized in @carnegie_foreign_2017, leveraging the assignment of the Council presidency to EU member countries to purge foreign aid of endogeneity. The Council presidency is an ideal instrument for foreign aid as it is determined by a known and as-if-random rotating selection process that is independent of existing democracy levels in recipient countries. Furthermore, countries that hold the presidency during the budget-making period can impact the amount of aid for which a country is eligible, and tend to allocate more aid to countries with which they have cultural or economic connections---particularly former colonies. Thus, before estimating my model of interest, I estimate the following first-stage regression:
$$log(ODA)_{i(t-1)} = \alpha + \gamma_{0}Colony_{i(t-2)2} + \eta_{i} + \theta_{t} + \epsilon_{it},$$
where $Colony_{i(t-2)2}$ takes a value of one if country $i$ is a former colony of the country holding the Council presidency during the budget-making period [@carnegie_foreign_2017]. $\eta_{i}$ and $\theta_{t}$are country and time fixed effects, respectively, and $\epsilon_{it}$ is the error term.
In order for former colony status of the Council presidency to be a valid instrument, it must be significantly and positively correlated with foreign aid and must satisfy the exclusion restriction assumption. As shown in @tbl-fs, I find a statistically significant and positive impact of colony status on aid receipt for both the poverty gap and infant mortality rate regressions. Substantively, former colonies receive, on average, between 14.5% (p \< 0.05) and 15.9% (p \< 0.05) more foreign aid than non-colonies of the Council presidency. Additionally, being a former colony of the Council presidency cannot impact poverty except through foreign aid. Although the enduring economic effects of colonial origins certainly impact poverty today, the appointment of the presidency of an EU governing body will have no impact on poverty in former colonies, thereby satisfying the exclusion restriction assumption.
# Results
@tbl-main shows the two-stage least squares estimates for the effect of EU aid on the poverty gap and infant mortality rates across multiple model specifications. Although I find no statistically significant effects of foreign aid on poverty, my estimates are consistent across different model specifications. Furthermore, as recommended by @visentin_earnest_2020 for the interpretation of null effects, examining the economic significance of my results using 95% confidence intervals yields nonetheless important conclusions.
Despite theoretical evidence suggesting that aid should decrease monetary measures of poverty, I find a positive relationship between aid and the poverty gap at \$6.25 a day. Estimates for the effect of aid on the poverty gap range from 0.033 to 0.112, which means that, all else equal, a 1% increase in EU aid is associated with between a 3.3 and 11.2 percentage point increase in the poverty gap. Although zero is within the confidence intervals of the full range of estimates, examining their upper- and lower-bounds are fruitful exercises. For example, the 95% confidence interval for regression (2), which controls for both electoral democracy and economic conditions, is (-0.77, 0.24). These bounds are crucially economically significant: since the poverty gap measures how far the average poor household falls below the poverty line, a household initially at the poverty line that experiences a 24% increase in their poverty gap would see their daily income fall to \$4.75. Conversely, a household consuming \$4.75 a day that experiences a 77% reduction in their poverty gap would see their income rise to \$5.91 a day. For low-income households whose livelihood can be significantly altered by marginal changes in income, these are non-trivial income shocks. Therefore, while I cannot provide definitive evidence of an effect of aid on the poverty gap due to the statistical insignificance of my results, they are nonetheless economically meaning-
| |
|:-------------------------|
|  |
: Two-Stage Least Squares Estimates for the Effect of Logged EU Foreign Aid on the Poverty Gap and Infant Mortality Rates {#tbl-main}
ful, indicating that the effect of aid (either positive or negative) is non-ignorable and worthy of further inquiry.
Unlike the poverty gap, my estimates for the effect of aid on non-monetary poverty as measured by the infant mortality rate is consistent with theoretical predictions. Estimates for the effect of aid on the infant mortality rate range from -0.417 to -1.366, such that a 1% increase in EU aid is associated with a decrease in the infant mortality rate by roughly 0.4 to 1.4 deaths per year, other things constant. Conducting a similar examination of the 95% confidence interval for regression (8), which is (-10.9, 8.2), demonstrates that estimates for a reduction in infant mortality by 11 deaths per year or an increase by 8 deaths per year are consistent with my results. As was the case for the effect of aid on the poverty gap, my findings are statistically insignificant, but nonetheless reveal that aid has a non-trivial impact on human livelihood.
The sign of the estimates for electoral democracy, egalitarian democracy, and the Polity IV score suggest that higher levels of democracy may be associated with lower levels of poverty across monetary and non-monetary dimensions, albeit statistically insignificantly. The estimate for the effect of electoral democracy on poverty is the only democracy variable to be consistently negative: all else equal, an electoral democracy score of one, symbolizing the highest level of democracy, is associated with a 24.1 percentage point lower poverty gap and about one fewer infant death as compared to countries with the lowest level of democracy. Perplexingly, my results suggest that egalitarian democracy is positively related to the infant mortality rate, such that a score of one is associated with roughly 24 more infant deaths per 1,000 births as compared to countries with a score of zero, a difficult result to comport with theory.
Estimates for the effect of rule of law and corruption on poverty are also insignificant and the most inconsistent. I find mixed evidence that better adherence to the rule of law, which is an indicator of higher quality institutions, is poverty-alleviating. The effect of rule of law on poverty varies from positive (increasing poverty) to negative (decreasing poverty) across different model specifications. Furthermore, my estimates for the impact of corruption on poverty suggest that more corruption is associated with less poverty, contrary to predictions that countries with less corrupt governments should have lower poverty as they are less likely to sequester aid funds for the elite and powerful. These results may suggest that there is a potential confounder not captured in my model. As a result of these confounding results, it is paramount for further studies armed with higher quality data and additional controls to reassess the impact of rule of law and corruption on poverty.
{#fig-2 height="90%" fig-align="center"}
Although, in the immediate term, my findings suggest that EU aid (insignificantly) worsens the poverty gap, over time, aid may eventually improve the poverty gap. @fig-2 depicts the intertemporal effect of aid from year $t$ (when aid is first received) to year $t+5$ (five years after receipt) on both the poverty gap and the infant mortality rate using the models specified in columns (1) and (7) of @tbl-main, respectively. EU aid initially increases the poverty gap in recipient countries and has a destabilizing effect through year $t+2$, at which point it exerts consistent downward pressure on monetary poverty. The infant mortality rate peaks in year $t+1$ before falling consistently through year $t+5$ after aid receipt. Despite their statistical insignificance, such patterns are insightful and should be investigated further. If future studies find statistically significant support for my results, suggesting that EU aid may have different effects on monetary and non-monetary poverty in the immediate term but eventually serves to decrease both measures of poverty as it "trickles down" to the poor, then aid may indeed be a viable long-term solution to alleviate poverty.
# Discussion and Conclusion
The alleviation of global poverty is a monumental feat, but an undeniably necessary undertaking. Though solutions for reducing poverty vary across contexts, foreign aid is one of the most ubiquitous approaches. However, its effects are still not widely understood, casting a shroud over the efforts of altruistic nations seeking to support the economic and human development of the developing world. In this study, I sought to bring clarity to the debate over aid efficacy by building upon the methodology introduced by @carnegie_foreign_2017 to address the problem of endogenous aid allocation. Using the rotating presidency of the Council of the European Union to instrument for foreign aid, I assessed the impact of EU foreign aid on two measures of poverty: the poverty gap and the infant mortality rate. The positive conditionality of EU aid, which @carnegie_foreign_2017 demonstrate effectively incentivizes recipient countries to enact democratic reforms, is a mechanism that can potentially amplify the impact of aid on poverty. Thus, I explored the relationship between democracy and poverty in an effort to understand whether there indeed exists a complementarity between aid and democracy for poverty alleviation.
I find that EU aid potentially worsens the poverty gap but improves the infant mortality rate in the immediate term, but may improve both measures of poverty after five years. My analysis of the role of democracy on poverty yielded inconsistent support for the theory that democracy is poverty alleviating. While my results are statistically insignificant, thereby precluding any conclusive evidence for an effect of EU aid on poverty, they are nonetheless economically relevant. Examining the confidence intervals for my estimates reveals that, should there be an effect of foreign aid on poverty, this effect could be large enough to have a critical impact on the livelihood of the poor (either positively or negatively).
As reflected by the lack of statistically significant results, this study is subject to some of the common challenges facing analysis of foreign aid and poverty. A lack of high-quality data on poverty is likely a contributing factor to my inconclusive results. Although the World Bank provides copious data on global poverty, these data are often collected by countries using their own resources; in developing contexts, countries may simply not have adequate means of collecting accurate data on poverty on a basis consistent enough to facilitate panel analysis [@riddel_does_2009]. Consequently, my sample sizes may be too small, leading my analysis to be underpowered to detect an effect.
By adopting a novel technique for addressing the aid endogeneity problem, my study paves the way for future studies utilizing a similar framework to investigate the relationship between foreign aid and poverty. Despite not finding strong enough evidence of an impact of aid on poverty, my results motivate a need for future studies armed with higher quality data to reassess this relationship, in addition to the role that democracy plays in promoting higher standards of living.
::: {#refs}
:::
# Appendix {.appendix}
| |
|:-------------------------------------|
|  |
: Replication of Carnegie and Marinov (2017) Table 1: Estimates for the Effect of Logged EU Foreign Aid on Human Rights and Democracy {#tbl-cm}
{#fig-cm1 height="80%" fig-align="center"}
| |
|:----------------------------------|
|  |
: Descriptive Statistics Table of Variables of Interest and Controls {#tbl-ds}
| |
|:----------------------------------|
|  |
: Estimates for the Effect of Former Colony Status of the Council Presidency on Aid Receipt: First-Stage Regression Results for Infant Mortality Rate and Poverty Gap Regressions {#tbl-fs}