SEARCH

SEARCH BY CITATION

Keywords:

  • Rent seeking;
  • Corruption;
  • Foreign aid;
  • Uncertainty
  • C53;
  • F35;
  • F47;
  • O11

Abstract

  1. Top of page
  2. Abstract
  3. I. Introduction
  4. II. Aid Flow Uncertainty and Rent Extraction
  5. III. Empirical Evidence
  6. IV. Concluding Remarks and Policy Implications
  7. References
  8. Appendix

This paper examines the effects of aid on governance from a different perspective by asserting that aid unpredictability can potentially increase corruption in recipient countries by providing incentives to risk-averse and corrupt political leaders to engage in rent-seeking activities. Analyses of data from 80 developing countries over the period 1984–2004 offer evidence that higher aid unpredictability is associated with more corruption as measured by a synthetic index. We also find further evidence that this latter impact is more severe in countries with weak initial institutional conditions. These findings are a supplementary advocacy for the need for better management and better predictability of aid flow in developing countries.

I. Introduction

  1. Top of page
  2. Abstract
  3. I. Introduction
  4. II. Aid Flow Uncertainty and Rent Extraction
  5. III. Empirical Evidence
  6. IV. Concluding Remarks and Policy Implications
  7. References
  8. Appendix

Recent international initiatives have called on the aid donor community to urgently increase Official Development Assistance (ODA) flows to allow poor countries to reach their Millennium Development Goals (MDGs). In 2007 (midway through the 15-year-long process of achieving these MDGs), mid-term reviews of the goals stressed that a significant number of countries were way off achieving the expected results and there was an absolute necessity to bring aid flows to higher levels. Also, through international commitments like the Paris Declaration on aid effectiveness and the Accra Agenda for Action (recently adopted at the Accra High Level Forum on Aid), donor countries pledged to make aid more effective through better coordination of donors, better ownership by recipient countries, better alignment of aid interventions with national development strategies, better results-based management of aid, and better mutual accountability for the results achieved. The Accra Agenda for Action also importantly stressed the need to improve other aspects of the quality of the management of aid, aiming at increasing the medium-term predictability of aid. Aid volatility and unpredictability issues are of crucial importance for the MDGs. Till now, aid flows toward developing countries have been widely volatile and unpredictable.1 In comprehensive reviews of aid volatility, Bulir and Hamann (2001, 2003) and Bulir and Lane (2002) provided strong evidence that aid is highly volatile, with the coefficients of variation exceeding those of fiscal revenues. Vargas Hill (2005) provided the evidence for sub-Saharan Africa that aid flows are five times more volatile than GDP and seven times more volatile than OECD countries' GDP. The work of Pallage and Robe (2001) also showed that aid is more volatile than the revenue of developing countries, while Fielding and Mavrotas (2005), disaggregating aid flows into program aid and project aid, provided support for the fact that the former is more volatile than the latter.

According to a growing body of research examining aid flow instability issues, predictability of overall aid and other various types of aid are significant and potentially costly problems in aid-dependent countries. In general, aid commitments exceed aid disbursements and the former are known to be bad predictors of the latter (Bulir and Hamann 2001). In recent years, aid predictability issues have attracted some research that documents the extent of its implication for development programs and macroeconomic management in recipient countries. According to OECD (2008), on average, less than 50% of committed aid is delivered on schedule. Celasun and Walliser (2008) found significant absolute deviations between commitments and disbursements. They also provided evidence that aid flows are less predictable in countries that are weakly covered by IMF programs (a proxy of a stable country environment). The work of Fielding and Mavrotas (2005) confirmed that aid flows are unpredictable and that this lack of predictability is related to the type of aid provided, program aid being more unpredictable than project aid. From a macroeconomic perspective, the lack of aid predictability2 can have some adverse consequences in aid-dependent countries. One of the main consequences of aid unpredictability is that it makes fiscal planning and the implementation of a recipient country's development agenda extremely difficult, since aid commitments have shorter terms than governments' development planning. Aid unpredictability also makes the ownership of development programs by recipients much more difficult since they are relying on funds that are uncertain. Elsewhere, the lack of predictability of aid may increase the likelihood of fiscal and monetary instability (Bulir and Lane 2002). Aid unpredictability associated with aid pro-cyclicality also increases output volatility and hence reduces growth (Ramey and Ramey 1995; Lensink and Morrissey 2000). Lensink and Morrissey (2000) found that the effect of aid on growth is insignificant unless some measure of aid uncertainty is included in the regression, and that uncertainty about aid is detrimental to growth.

This paper switches the attention from the macroeconomic effects of aid unpredictability to a more “political economy” approach by linking aid flow uncertainty to rent-seeking behaviors in recipient countries. Institutional issues have recently returned to the foreground in debates on economic development. Academic research has extensively investigated the impact of aid on the quality of institutions in aid-recipient countries and has focused on aid intensity ratios (aid/GDP, aid/GNP, aid/exports, aid/public expenditures, and aid per capita) as measures of aid dependence. A number of studies have empirically demonstrated that aid is, on average, associated with more corruption and more rent-seeking activities in aid-recipient countries (Alesina and Weder 2002; Svensson 2000),3 while others have come to the opposite conclusion (Tavares 2003).

To our knowledge, no work has focused on the effects of aid flow uncertainty on recipient countries' institutions. Looking for new evidence on the effect of aid on institutions in recipient countries, this paper switches from traditional measures of aid dependency to one feature of its delivery: its unpredictability. Does aid unpredictability lead to more corruption in aid-recipient countries? Through this core research question, the paper focuses on aid-dependent countries and investigates whether higher aid flow uncertainty is associated with a higher level of corruption. The basic political economy rationale is that aid flow uncertainty reduces the temporal horizon of the aid rent capture. Ventelou (2001), investigating the effect of political survival on rent capture, concluded that the shorter the probability of political survival, the greater the incentive for leaders (assumed to be kleptocrats) to engage in rent capture. The paper uses a similar theoretical framework and explains that the greater the uncertainty of future aid flows, the greater the incentives of kleptocrat leaders to engage in rent seeking in countries where institutions are weak. The paper then provides an empirical evaluation of these theoretical arguments with supportive results. Rent seeking is proxied by an index of corruption. Corruption is, of course, an extreme form of rent seeking. Even if rent seeking can take forms other than corruption (costs of ensuring protection, costs of seizing rents, and costs of dealing with competition), the sparse availability of such data leads us to use this proxy. Fixed effects estimations with a sample of 80 developing countries over the period 1984–2002 confirm that aid dependency is associated with less corruption, while aid unpredictability leads to more corruption. We find further evidence that this latter impact is more severe in countries with weak initial institutions. Sensitivity analyses then show that the type of aid matters for the nature and the size of the effect and that the impact is more severe in countries with weak initial institutions.

The remainder of the paper is organized as follows. Section II discusses the link between aid unpredictability and rent-seeking behaviors. Section III deals with the empirical evaluation. Section IV concludes the paper.

II. Aid Flow Uncertainty and Rent Extraction

  1. Top of page
  2. Abstract
  3. I. Introduction
  4. II. Aid Flow Uncertainty and Rent Extraction
  5. III. Empirical Evidence
  6. IV. Concluding Remarks and Policy Implications
  7. References
  8. Appendix

How does aid flow uncertainty explain rent-seeking behaviors and corruption by “kleptocratic” leaders? This question, while not seemingly new, has not yet been explicitly addressed in the political economy literature on aid. There is a huge body of political economy research on aid and endogenous political leaders' behavior. Elsewhere, a substantial number of studies have investigated the impact of aid on institutions in recipient countries, fueling a controversy about the nature of the impact. Yet this literature has focused only on the level of aid flows, investigating its impact on rent-seeking behaviors. This paper incorporates a risk factor in the analysis, that is, the effect of aid flow uncertainty on political elites' behavior. From a theoretical perspective, the expectation would be that high aid flow uncertainty (under some assumptions) should generate a higher level of corruption and rent seeking.

We consider a theoretical reasoning framework in which the political leaders (the elites in and around the government) are rent seekers (this assumption is strengthened by the considerable evidence of rent-seeking activity in many developing economies) and where aid transfers can be the subject of predation and dissipation. Moreover, we assume (from the evidence of aid flow uncertainty—see subsection III.B.2) that leaders do face uncertainty about the future of the rents they extract. So we relax the assumption of a benevolent government found in the political economy literature of government spending and assume that, as a rule, aid-recipient countries are managed by politicians who draw partial utility from rents and who face uncertainty about future aid flows.

The intuition of this paper is as follows: in a theoretical setting where politicians aim to maximize the amount of the rent they capture and where they have intertemporal smoothing considerations, greater unpredictability of aid can lead them to engage more than proportionally (compared with the optimal path) in rent seeking since they face a shortfall risk of aid. Investigating the political foundations of the negative impact of resources booms on the economy with a political economy model, Robinson, Torvik, and Verdier (2006) showed that politicians have the incentive to over-extract natural resources (generating rents) compared to the most efficient extraction path. This is led by the probability of their staying in power, which is a discount factor of the future. In other words, the less certain they are of staying in power, the more likely they are to over-extract the resource and benefit from the rents. Robinson, Torvik, and Verdier (2006) explained that the future stock of resources (and therefore rents) only matter if the politicians are in power.

The work of Ventelou (2001) also supported the notion that political risk determines the tendency of politicians to over-engage in rent seeking. Considering a government that has the choice either to invest public resources in productive goods or to appropriate them to finance private consumption, he showed that as the probability of political survival4 decreases, the level of politicians' rent capture increases. The less the government in office has the chance to remain in power in the next period, the more it will be inclined to capture the maximum rents in the current period, since the returns from productive investments will benefit the next government.

We rely on a similar theoretical reasoning, that the probability of receiving transfers from which rent is extracted determines the behavior of leaders insofar as one can predict that they will tend to be more engaged in rent seeking when this probability is low or unknown. Anticipations and expectations about the future can also be affected, not only by the probability that the leaders will stay in power in the future (determining their ability to capture the rent) but also by the probability of receiving the income (foreign aid) from which the rent is extracted (determining their ability to capture the rent as well). So, it is as if the rent-seeking leaders are risk averse, over-extracting the current rent from aid instead of waiting for an uncertain amount of future rent. Contrary to Svensson (2000), who showed that the mere expectation of foreign aid provides incentives to increase rent dissipation, we suggest that the mere anticipation of an aid shortfall provides incentives to rent seekers to increase rent dissipation. Acemoglu, Robinson, and Verdier (2004) showed that aid provides kleptocratic rulers with greater resources to finance their tenure of power by buying off opponents. The greater uncertainty of such a resource for such leaders would increase their incentive to over-extract the rent.

This paper's theoretical reasoning is inspired by the political economy literature describing the behavior of governments facing economic and political risk. Battaglini and Coate (2008) studied the relationship between politicians' rent-seeking incentives and public debt and deficits, and found that in the presence of (political) risk, rent-seeking governments over-extract the rent and therefore hold a level of debt that exceeds that of benevolent governments. Myopic politicians facing a risk prefer to extract the rents as early as possible.5

III. Empirical Evidence

  1. Top of page
  2. Abstract
  3. I. Introduction
  4. II. Aid Flow Uncertainty and Rent Extraction
  5. III. Empirical Evidence
  6. IV. Concluding Remarks and Policy Implications
  7. References
  8. Appendix

A. Data and Base Specification

The data6 for this paper come mainly from World Bank statistics,7 the International Country Risk Guide (ICRG), Development Committee Assistance (DAC) statistics, and the Global Development Network Growth Database. Our sample is comprised of 80 developing countries: eight are from East Asia and the Pacific, one is from Eastern Europe and Central Asia, 14 are from the Middle East and North Africa, three are from South Asia, 30 are from sub-Saharan Africa, and 24 are from Latin America and the Caribbean. The relatively low number of countries is due to our desire to have the most balanced possible database for our main variables. The ICRG index of corruption is on a scale from 0 (worst situation of corruption) to 6 (best situation of corruption). Lower scores indicate that “high government officials are likely to demand special payments,” “illegal payments are generally expected throughout lower levels of government” in the form of “bribes connected with import and export licenses, exchange controls, tax assessment, police protection, or loans.” We chose not to rescale the index so that an increase means a reduction of corruption. Our main measure of aid intensity is “official development assistance”8 measured as a percentage of GDP, which is standard in the empirical aid literature. Even though this measure of aid may reflect changes in GDP with aid constant, rather than changes in aid, it does capture the importance of aid. As indicated in Appendix A, aid data are available from the World Development Indicators (WDI), based on aid data provided by the OECD's Development Assistance Committee. We also make use of one other measure of aid intensity for robustness, which is aid as a percentage of Gross National Income (GNI) (also available from the WDI).

Table 1, which presents some basic descriptive statistics for the main variables used in the empirical analysis (data are averaged over 1984–2004), shows that values of aid as a percentage of GDP range up to 47.1% (Somalia). For aid as a percentage of GNI and imports, the figures are higher. The table also shows that no country in the sample reaches the maximum score that indicates the best corruption situation (6). The average score of corruption is also relatively low (2.60, on a scale from 0 to 6), indicating that the developing countries we focus on in our analysis are, on average, corrupt.

Table 1. Descriptive Statistics (1984–2004)
 MeanStd. Dev.Min.Max.N
Source: Author's computations.
A. Corruption measure:     
Corruption (ICRG)2.600.740.404.6380
B. Aid variables:     
Aid (%GDP)7.319.78−0.0147.180
Aid (%GNI)7.6710.13−0.0150.3578
C. Countries characteristics:     
Log(init. income)1,764.622,151.14123.8810,453.1976
Population3.54e+071.05e+08372903.69.16e+0880
Rents1.072.64016.2179
Eth. frac.47.9730.2209373
Legal orig. (British)0.320.470179
Accountability3.021.020.745.0680
Landlock0.140.350180
Africa0.430.500180

Table 1 also reveals considerable variation in initial income (initial GDP per capita, constant 2000 US dollars) with a standard deviation of 2,151. We rely on the literature on the determinants of corruption to select the remaining control variables, namely ethnolinguistic fractionalization (eth. frac.), GDP per capita (initial income), the share of mineral exports in GDP (rents), total population (population), British legal origin dummy (British), democratic accountability (accountability), and geography (landlock).

B. Aid Unpredictability and Corruption

1. Measuring aid unpredictability

From a statistical viewpoint, uncertainty over an economic variable is, in most of the studies, proxied by unconditional measures such as the standard deviation or the variance of the variable's movements. It is worth noting that simply using such proxies is questionable on both economic and statistical grounds. Variability does not necessarily imply unpredictability. As underlined by Knack (2001), when estimating the effect of aid volatility (using the coefficient of variation of aid) on the quality of governance, a high variability of aid should not be likened to uncertainty, since it could be the product of a strong and steady upward or downward trend in aid levels over time. Dehn (2000) also pointed out that simply using the standard deviation of a series to proxy its uncertainty leads to overestimation of the unpredictable part and underestimation of the predictable part, since the variable's trend is not taken into account. A high volatility can be anticipated or not, and what matters from a political economy perspective is unpredictability. A political leader facing a badly contained risk is not expected to exhibit the same behavior as one who has much more information on the future movement of a variable. So conditional measures of volatility are better proxies of the uncertainty faced by economic agents. While GARCH-based approaches9 are well suited to estimating uncertainty, they require high frequency data, which were not available to us. Following Aizenman and Marion (1993) and Lensink and Morrissey (2000), we therefore make use of alternative measures of uncertainty, which consist of two steps. First, we estimate the following forecasting equation, specified as a second-order, auto-regressive process and extended with a time trend:10

  • display math(1)

where Aid is total ODA net disbursements, v is the forecast error, T is a time trend, and t stands for years. We then measure aid uncertainty by calculating for each country and each subperiod11 in our sample, the standard deviations of the residuals of equation (1). This measure of aid unpredictability is intended to separate simple variation from uncertainty and thus capture unanticipated changes in aid.

2. How unpredictable is aid?

Figure 1 presents the time evolution of the aid forecast errors over the period 1982–2001, for a set of 12 countries in our sample that are the most aid dependent (the dependency ratio used is net ODA/GDP). These countries are Comoros, Guyana, Honduras, Liberia, Lesotho, Mali, Mauritania, Malawi, Rwanda, Zambia, Uganda, and Chad. The error forecasts are computed from equation (2). The x axis represents the year and the y axis represents the residuals, the variability of which is considered as a proxy of aid uncertainty. The scatters show that for all the selected countries, the residuals vary greatly around zero, and computations indicate that the mean standard deviation of the residuals for this subset of countries is relatively high at about 4.31.

figure

Figure 1. High Aid-Dependent Countries' Net ODA Errors Forecasts

Source: Author's computations.

Download figure to PowerPoint

3. The identification strategy
(a) Baseline estimations and results

In order to investigate the effect of aid unpredictability on corruption, we specify the following equation:

  • display math(2)

where corruptionit is the average level of rent seeking (proxied by the ICRG index of corruption) for country i in period t,12 uncert is a proxy variable of aid unpredictability (log value is considered to account for the likely nonlinear relationship), X is a vector of controls including initial income, mineral exports, population size, accountability, ethnolinguistic fractionalization, legal origin, geography, and an Africa dummy. τ is a time trend.

Table 2 presents results (which are meant to give a first broad sense of the relationships) using the ICRG-averaged index of corruption as the dependent variable. All columns report random effects regression results, testing the effects of aid unpredictability on corruption. As most of the controls are time invariant and need to be included in the regressions, we do not rely on fixed effects estimations that would automatically drop them. However, except for column (1), which assesses the cross-sectional and bivariate evidence of the impact of aid unpredictability on rent seeking, all our regressions include regional fixed effects to take into account the unobserved heterogeneity across regions. As expected, we found that aid unpredictability increases rent seeking proxied by corruption in a statistically significant way. A 1% increase in the aid uncertainty log measure is associated with a 0.12% increase in the corruption index (column 2).13 This result is confirmed by the cross-section regressions in column (1), yet with a smaller coefficient of 0.09. From column (3) to column (6), we gradually include in the regression additional controls that are an Africa dummy, ethnolinguistic fractionalization, and geography, which are assumed to have an effect on corruption.

Table 2. Aid Unpredictability and Corruption (cross-section and panel regressions, three-year period averages, 1984–2004)
 Coefficient (std. err.)
(1)(2)(3)(4)(5)(6)(7)
  1. Notes: 1. Random effects-based results.

  2. 2. Beside the coefficient value, the standard errors, which are computed using heteroskedastic-consistent standard deviations, are reported in parentheses.

  3. 3. All the estimations include regional controls, which are not reported for reasons of space.

  4. ***, **, and * represent statistical significance at the 1%, 5%, and 10% levels, respectively.

Log(uncertainty)−0.09* (0.05)−0.12*** (0.06)−0.14* (−1.79)−0.14* (0.08)−0.15* (0.09)−0.17*** (0.08)−0.19** (0.09)
Log(init. income)  −0.15 (0.17)−0.15 (0.17)−0.23 (0.20)−0.28 (0.20)0.30* (0.2)
British  −0.02 (0.18)−0.02 (0.18)−0.03 (0.21)−0.01 (0.21)−0.02 (0.2)
Rents  −0.04 (0.03)−0.04 (0.03)−0.03 (0.03)−0.04 (0.03)−0.04 (0.03)
Log(pop.)  −0.14* (0.07)−0.14* (0.08)−0.14 (0.09)−0.16* (0.08)−0.17*** (0.08)
Accountability  0.29*** (0.06)0.29*** (0.06)0.31*** (0.07)0.30*** (0.07)0.30*** (0.07)
Africa   −0.03 (0.37)−0.06 (0.39)−0.06 (0.38)
Eth. frac.    −0.007*** (0.003)−0.007*** (.0003)−0.007** (0.003)
Landlock     −0.37 (0.27)−0.38 (0.27)
SS Africa      1.17 (1.57)
Uncertainty*SS Africa      0.06 (0.12)
Intercept1.35* (0.75)0.86 (1.02)3.21** (1.54)3.15* (1.70)4.02*** (1.82)4.57*** (1.74)4.09*** (1.82)
Reg. dummiesNoYesYesYesYesYesYes
Obs.80158149149137137137
R20.050.050.270.270.320.340.34

Some of the findings do not match the previous findings in the literature while some do; that could be due to differences in the sampling design and model specifications. For example, coherently with the literature (Svensson 2000), only ethnolinguistic fractionalization enters significantly, while not changing the aid and uncertainty coefficients' size and significance. In column (7), we test whether the impact of aid uncertainty may depend on regional clustering. Since sub-Saharan African countries exhibit the lowest corruption scores,14 we focus on this sub-region dummy that we interact with uncertainty. No significant effect was found, indicating that this regional clustering is not relevant regarding the determination of corruption. Surprisingly, initial income is found to have a positive and significant effect (at 10% in the specification with all controls) on the index of corruption (column 7), supporting the thesis that higher initial income is associated with more corruption. This result calls for some comments. One would have expected the opposite effect. Indeed if good quality institutions are considered a “superior good,” higher income should favor them (Lipset 1959; Acemoglu 2008). Yet a possible explanation of this finding may be related to the form of the relationship between corruption and income, which is likely to be nonlinear. Indeed as suggested by Acemoglu, Johnson, and Robinson (2001), an income increase in a set of developing countries is associated with more opportunities for corruption because of the weak quality of their institutions. Then, as income continues to increase, better and stronger institutional mechanisms promoting institutional quality are more likely to be put in place, leading to a reduction in corruption. This suggests an inverse U-shaped relationship between the two variables. As our sample of countries is made up of developing countries, which have, on average, bad institutions, it is likely that the regression results only illustrate the “positive” stage of the relationship. Unsurprisingly, it transpires from the regression results that greater political accountability is associated with less corruption, with a significant marginal effect around 0.3. While some findings suggest that the population may have a strong direct effect on corruption (Tavares 2003), we find the opposite evidence, which is probably explained by the different sampling. We also performed some post–estimations tests to make sure that our results were not driven by a model misspecification or by the fact that the residuals were not normally distributed. For all the main specifications of the model, the Ramsey-Reset test probability is about 0.17, indicating that the model does not suffer from a misspecification. The normality tests also indicate probabilities of about 0.65, confirming a normal distribution of residuals.

(b) Dealing with the potential endogeneity of the uncertainty variable

The previous results could suffer from error measurement bias in the uncertainty variable we used following Aizenman and Marion (1993) and Lensink and Morrissey (2000).

Pagan and Ullah (1988) proposed an instrumental variable (IV) nonparametric estimator, with instruments constructed from the information set. The conditional variance of aid is taken as the unobserved volatility of aid and can be written as:

  • display math(3)

where varprev(.) and Eprev(.) are the variance and the expectation conditional on the previous time period information respectively. From equation (1), we can write that:15

  • display math(4)

Equation (3) can then be rewritten as:

  • display math(5)

To account for the potential endogeneity of the uncertainty variable, we first reestimate aid uncertainty with a nonparametric estimator, which takes advantage of the yearly availability of aid data and the period-based structure we give to our data.16 The estimator was introduced by Schwert and Seguin (1990) and used in Andersen and Bollerslev (1998). The unobserved variability of aid in equation (4) is estimated with:

  • display math(6)

where Atip is aid residuals from the forecasting equation in year t and period p for country i. This estimator has been demonstrated to be consistent for a general conditional variance specification for cases where t-values are high (Andersen and Bollerslev 1998). Contrary to these authors, who estimated daily exchange rate volatility from intraday returns, we estimate aid volatility for each period and each country using yearly data. Ten-year periods give the highest frequency of aid data while keeping a panel structure.

We then rely on the Pagan and Ullah (1988) instrumental variable, which corrects for the large sample bias due to the weak number of subintervals (10 years). The first step of the procedure involves proxying inline image (the residuals from equation (3)) with inline image, since inline image (equation (5)). Our baseline regression equation is then rewritten as:

  • display math(7)

where inline image. The proxy inline image being correlated with ηit and assuming that inline image, Pagan and Ullah (1988) show that:

  • display math(8)

The second step of the instrumental procedure consists in instrumenting inline image with inline image (in equation (5)), which is computed with a set of information correlated with inline image. This procedure has an additional advantage in that inline image should be quite strongly correlated with inline image in spite of the weak number of subintervals (10 years in our case). We checked and confirmed this with our sample data.

Table 3 presents the results of the regression using the instrumented measure of aid unpredictability. Uncertainty still enters significantly with a negative sign. The new coefficient of the uncertainty variable (log value) is higher, at about 1.87 (column 7). Initial income level still significantly affects corruption. Accountability and the Africa dummy enter positively, though the latter effect was unexpected. In keeping with previous findings, ethnolinguistic fractionalization enters with a negative and significant coefficient. The value of the Shea partial R2 (0.98), which records the additional explanatory power of the excluded instruments (taking into account the intercorrelations of the instruments), confirms the strength of our instrumental variable. The reported first-stage coefficient of the instrumental variable strongly further confirms its validity and shows that our results are robust to the treatment of the potential endogeneity of the aid unpredictability variable. Standard post–estimation tests (normality test and Ramsey-Reset specification test) indicate probabilities that are respectively about 0.6 and 0.1 for the main specifications, confirming that our results are not biased by failing to meet these requirements.

Table 3. Aid Unpredictability and Corruption (panel IV regressions, instrumenting for unpredictability)
 Coefficient (std. err.)
(1)(2)(3)(4)(5)(6)(7)
  1. Notes: 1. Beside the coefficient value, the standard errors, which are computed using heteroskedastic-consistent standard deviations, are reported in parentheses.

  2. 2. All the estimations include regional controls, which are not reported for reasons of space.

  3. 3. Data are averaged over two ten-year periods (1984–94 and 1995–2004).

  4. ***, **, and * represent statistical significance at the 1%, 5%, and 10% levels, respectively.

Log(uncertainty)−2.47*** (1.02)−2.19** (1.02)−2.16** (1.04)−2.12** (1.02)−2.22** (1.03)−1.96** (0.932)−1.87** (0.95)
Log(init. income) 0.10 (0.06)0.10 (0.06)0.09 (0.06)0.08 (0.06)−0.03 (0.06)−0.09 (0.09)
British  −0.02 (0.15)−0.009 (0.15)−0.012 (0.15)−0.009 (0.13)0.03 (0.15)
Rents   −0.03 (0.02)−0.04 (0.02)−0.04** (0.02)−0.03 (0.02)
Log (pop.)    −0.04 (0.05)−0.08* (0.04)−0.06 (0.05)
Accountability     0.33*** (0.06)0.36*** (0.06)
Africa      0.35* (0.18)
Eth. frac.      −0.0066*** (0.0029)
Landlock      −0.27 (0.19)
Intercept2.66*** (0.07)1.98*** (0.41)1.98*** (0.44)2.08*** (0.43)2.85*** (1.01)3.22*** (0.91)3.36*** (1.22)
Reg. dummiesYesYesyesYesYesYesYes
Obs.160152150149149149137
R20.040.060.060.080.080.260.32
Instruments quality statistics:
IV first-stage coeff.0.0719*** (0.0006)0.0716*** (0.00067)0.0715*** (0.0007)0.0714** (0.00063)0.0712*** (0.00062)0.0711*** (0.00062)0.07*** (0.0007)
Partial Shea R20.980.980.980.980.980.980.98
4. The importance of the initial institutional conditions

A growing body of empirical research has emphasized the importance of the long-term impact of historic events on institutions as well as their path dependency. The most influential studies examined the importance of the factor endowments and the earlier colonial rule as the historic determinant of institutions, which persisted over time. It is worth noting that most empirical investigations of the determinants of institutions have emphasized the importance of the initial conditions and tried to control for that by including in the vector of control a proxy of the initial institutional quality. We share the same interest in this paper and turn to check whether and how the initial institutional conditions matter for the impact of aid unpredictability on corruption. We test the hypothesis that this impact is more severe in countries with weak initial institutions. As a proxy of those initial institutional conditions, we use the average scores of the constraints on the executive power from Polity IV at the early year of our baseline regressions, i.e., 1984. We indeed assume that the quality and the strength of those constraints proxy well the overall institutional quality, as suggested by some authors (Acemoglu, Johnson, and Robinson 2005). We then split our sample countries into two subsamples, one including the countries having a score of executive constraints lower than the overall sample median value, and the other including the countries with higher scores. The regression results for each sample, summarized in Table 4, provide supportive empirical evidence for our hypothesis; the coefficient of the aid unpredictability variable is only significant for the low initial institutions sample, indicating that the initial quality of the constraints on the executive does matter. Indeed, for that subsample, ceteris paribus, a 1% increase of the log measure of aid unpredictability leads to a 0.32% increase in the corruption index. The Shea partial R2 (about 0.8 in both specifications) still confirms the strength of the instrumental variable, while the Ramsey-Reset and normality test (respective probabilities of 0.15 and 0.95) confirm compliance with the standard regression analysis assumptions. It is worth noting that the main remaining findings are consistent with the previous ones, with political accountability continuing to reduce corruption for both subsamples. These findings seem to be somewhat in line with Dutta, Leeson, and Williamson (2008), who investigated the impact of aid on political institutions and argued that aid has the ability neither to make dictatorships more democratic nor make democracies more dictatorial, but only amplifies the existing political institutions. Our findings also suggest that, regarding corruption and rent-seeking behaviors, aid reinforces the trajectory of institutions by worsening countries' weak institutional performances.

Table 4. Aid Unpredictability and Corruption: The Importance of the Initial Institutions (panel IV regressions)
 Coefficient (std. err.)
Low-Institutions SampleUpper-Institutions Sample
  1. Notes: 1. Beside the coefficient value, the standard errors, which are computed using heteroskedastic-consistent standard deviations, are reported in parentheses.

  2. 2. All the estimations include regional controls, which are not reported for reasons of space.

  3. ***, **, and * represent statistical significance at the 1%, 5%, and 10% levels, respectively.

Log(uncertainty)−0.32*** (0.13)0.05 (0.12)
Log(init. income)−0.70*** (0.30)0.05 (0.24)
British−0.16 (0.38)−0.51* (0.26)
Rents0.01 (0.04)−0.05* (0.03)
Log(pop.)−0.24 (0.19)−0.11 (0.09)
Accountability0.46*** (0.16)0.26*** (0.09)
Africa−0.12 (0.77)−0.75 (0.74)
Eth. frac.−0.01 (0.009)0.001 (0.004)
Landlock−0.39 (0.37)0.29 (0.28)
Intercept6.18 (3.79)5.11*** (2.14)
Regional dummiesYesYes
Obs.4574
R20.460.43
Partial Shea R20.80.8

C. Sensitivity Analysis and Discussion

We conducted several sensitivity analyses. We have already shown that the results are robust to the use of differently structured data (cross-section dataset versus panel dataset, and simple random effects versus two-stage least square). Another important question is about their sensitivity regarding the use of different types of aid. We broke up aid into loans versus grants, bilateral versus multilateral, and project versus program. We also leverage this robustness analysis to test whether a change in the length of the subperiods matters in explaining the findings. Appendix Table A1, which summarizes the robustness tests results are based on four subperiods (1984–89, 1990–94, 1995–99, and 2000–4). Columns 3 and 4 (bilateral versus multilateral aid) show that the effect of aid unpredictability does not seem to vary much when considering this disintegration, even though the coefficient for multilateral aid appears to be higher and more significant (0.2). This result is consistent with the work of Pallage and Robe (2001), who showed that the instability of multilateral aid is greater than it is for bilateral aid (both net receipts and commitments), even though volatility does not necessarily mean unpredictability.

Interestingly, the results reported in the first two columns (1 and 2) show that only the grants part of aid matters in explaining the adverse effects of its unpredictability. A 1% increase in the log measure of grants' unpredictability leads to an increase of the corruption index by 0.17%. Based on data from 37 IMF desk economists, Bulir and Hamann (2003) noted that program aid is more unpredictable than project aid. Unexpectedly, the three last columns of the table show no evidence for program aid, project aid, and financial program aid. None of the related coefficients of aid unpredictability entered significantly. Only political accountability and ethnolinguistic fractionalization kept their previous effect.

In Appendix Table A2, we test the robustness of our findings to the use of a different measure of aid, which is aid measured as a percentage of GNI. Aid unpredictability still enters negatively and significantly with a coefficient of about 0.14. The logged population variable also enters negatively and significantly, indicating that the level of corruption increases with country size, concurring with some of the findings in the literature (Mocan 2004). As expected, being a landlocked country significantly increases the level of corruption. In column 2, we replicate the regression, excluding Guinea-Bissau, Liberia, Somalia, and Mozambique, which were identified as outliers in our base sample using Hadi's 1994, 1992) method. The main previous results stand. Aid unpredictability increases the level of corruption ceteris paribus; population size and democratic accountability respectively enter negatively and positively in the regression and the instruments maintain their good explanatory power as illustrated by the Shea partial R2. Moreover, results of specification and normality tests (respective P-values no less than 0.14 and 0.25 for Appendix Table A1 and 0.11 and 0.10 for Appendix Table A2) rule out a misspecification of the model and a non-normal distribution of residuals.

IV. Concluding Remarks and Policy Implications

  1. Top of page
  2. Abstract
  3. I. Introduction
  4. II. Aid Flow Uncertainty and Rent Extraction
  5. III. Empirical Evidence
  6. IV. Concluding Remarks and Policy Implications
  7. References
  8. Appendix

A number of recent studies have emphasized the need to improve aid predictability, focusing their analyses on the macroeconomic effects of aid unpredictability in recipient countries and particularly in highly aid-dependent countries. This paper addressed the issue of foreign assistance uncertainty from a political economy perspective, by investigating the effect of aid unpredictability on rent-seeking behaviors in aid-recipient countries. We proxied rent-seeking activities by corruption, mainly due to the sparse availability of data and the fact that corruption is one of the main symptoms of rent-seeking activities. Consistent with the literature, statistical analyses in the paper provide evidence of a high unpredictability of aid flows, computed from an econometric forecasting model. Our major empirical findings are threefold: (1) there is a robust statistical relationship between high aid unpredictability and corruption in aid-recipient countries; (2) the effect of aid unpredictability on corruption is more severe in countries with weak initial institutional frameworks, as measured by the constraints on the executive power; and (3) some sensitivity in the findings is found when considering different aid modalities. These findings suggest that donors must keep improving the management and delivery of aid flows, since as well as complicating the fiscal planning and implementation of the development agenda in aid-dependent countries, aid unpredictability may have an adverse effect on institutions through increased corruption. Yet the policy implications must be phrased delicately from this. The findings reveal that the damaging institutional impacts of aid unpredictability should be interpreted as a symptom of weak institutional frameworks, since evidence has been offered that the initial institutional conditions matter. Finally, as aid unpredictability may not be fully controllable by donors, and given that it can potentially create room for corruption in recipient countries, the findings of the paper are a supplementary advocacy for the prioritization of more programmatic forms of aid that are believed to be associated with lower transaction costs.

Footnotes
  1. 1

    In this paper, we use the term “uncertain” interchangeably with “unpredictable.”

  2. 2

    The sources of aid unpredictability are multiple. Aid can be unpredictable due to the fact that aid commitments and disbursements approvals are often made by different actors (e.g., ministries vs. parliaments), creating a gap between what is committed and what is really disbursed. The donors' conditions, which can be process-related or policy/performance-based also contribute to the lack of aid predictability.

  3. 3

    Other works have investigated more broadly the impact of aid on growth through its adverse institutional impacts (see the work of Rajan and Subramanian 2007).

  4. 4

    Interpreted as the probability of a re-election in a democracy.

  5. 5

    Caballero and Yared (2008) reach the same conclusion in the long term.

  6. 6

    See Appendix A for data definition and sources.

  7. 7

    World Development Indicators (http://databank.worldbank.org).

  8. 8

    According to OECD/DAC, “official development assistance” includes grants and loans with a grant element of more than 25%.

  9. 9

    See Bollerslev (1986) and Engle (1982).

  10. 10

    It is also possible to include a quadratic form of the trend and to estimate the model in difference.

  11. 11

    Our data are computed as two ten-year period averages.

  12. 12

    Data are averaged over four periods: 1984–89, 1990–94, 1995–99, and 2000–4.

  13. 13

    The ICRG corruption index was not rescaled; so an increase in the index is associated with improvement.

  14. 14

    Detailed statistics are not shown for reasons of space.

  15. 15

    Figure 1 gives a sense of the zero sample mean of residuals.

  16. 16

    For this purpose, we averaged our data over the 1984–94 and 1995–2004 ten-year periods.

References

  1. Top of page
  2. Abstract
  3. I. Introduction
  4. II. Aid Flow Uncertainty and Rent Extraction
  5. III. Empirical Evidence
  6. IV. Concluding Remarks and Policy Implications
  7. References
  8. Appendix
  • Acemoglu, Daron. 2008. “Oligarchic Versus Democratic Societies.” Journal of the European Economic Association 6, no. 1: 144.
  • Acemoglu, Daron; Simon Johnson; and James Robinson. 2001. “The Colonial Origins of Comparative Development: An Empirical Investigation.” American Economic Review 91, no. 5: 13691401.
  • Acemoglu, Daron; Simon Johnson; and James Robinson. 2005. “The Rise of Europe: Atlantic Trade, Institutional Change and Economic Growth.” American Economic Review 95, no. 2: 546579.
  • Acemoglu, Daron; James Robinson; and Thierry Verdier. 2004. “Kleptocracy and Divide-and-Rule: A Theory of Personal Rule.” Journal of the European Economic Association, Papers and Proceedings 2, nos. 2–3: 162192.
  • Aizenman, Joshua, and Nancy Marion. 1993. “Macroeconomic Uncertainty and Private Investment.” Economics Letters 41, no. 2: 207210.
  • Alesina, Alberto, and Beatrice Weder. 2002. “Do Corrupt Governments Receive Less Foreign Aid?American Economic Review 92, no. 4: 11261137.
  • Andersen, Torben, and Tim Bollerslev. 1998. “Answering the Skeptics: Yes, Standard Volatility Models Do Provide Accurate Forecasts.” International Economic Review 39, no. 4: 885905.
  • Battaglini, Marco, and Stephen Coate. 2008. “A Dynamic Theory of Public Spending, Taxation, and Debt.” American Economic Review 98, no. 1: 201236.
  • Bollerslev, Tim. 1986. “Generalized Autoregressive Conditional Heteroskedasticity.” Journal of Econometrics 31, no. 3: 307327.
  • Bulir, Ales, and Javier Hamann. 2001. “How Volatile and Unpredictable Are Aid Flows, and What Are the Policy Implications?IMF Working Paper no. 01/167. Washington, D.C.: International Monetary Fund.
  • Bulir, Ales, and Javier Hamann. 2003. “Aid Volatility: An Empirical Assessment.” IMF Staff Papers 50, no. 1: 6489.
  • Bulir, Ales, and Timothy Lane. 2002. “Aid and Fiscal Management.” IMF Working Paper no. 02/112. Washington, D.C.: International Monetary Fund.
  • Caballero, Ricardo, and Pierre Yared. 2008. “Future Rent-Seeking and Current Public Savings.” NBER Working Paper no. 14417. Cambridge, Mass.: National Bureau of Economic Research.
  • Celasun, Oya, and Jan Walliser. 2008. “Predictability of Aid: Do Fickle Donors Undermine Aid Effectiveness?Economic Policy 23, no. 55: 545594.
  • Dehn, Jan. 2000. “Commodity Price Uncertainty in Developing Countries.” World Bank Policy Research Working Paper no. 2426. Washington, D.C.: Word Bank.
  • Dutta, Namabita; Peter Leeson; and Claudia Williamson. 2008. “The Amplification Effect: Foreign Aid's Impact on Political Institutions.” http://nyudri.files.wordpress.com/2011/12/driwp821.pdf (accessed April 9, 2013).
  • Engle, Robert F. 1982. “Autoregressive Conditional Heteroscedasticity with Estimates of the Variance of United Kingdom Inflation.” Econometrica 50, no. 4: 9871007.
  • Fielding, David, and Georges Mavrotas. 2005. “The Volatility of Aid.” WIDER Discussion Paper no. 2005/06. Helsinki: UNU-WIDER.
  • Hadi, Ali S. 1992. “Identifying Multiple Outliers in Multivariate Data.” Journal of the Royal Statistical Society. Series B (Methodological) 54, no. 3: 761771.
  • Hadi, Ali S. 1994. “A Modification of a Method for the Detection of Outliers in Multivariate Samples.” Journal of the Royal Statistical Society. Series B (Methodological) 56, no. 2: 393396.
  • Knack, Stephen. 2001. “Aid Dependence and the Quality of Governance: A Cross Country Empirical Test.” Southern Economic Journal 68, no. 2: 310329.
  • Lensink, Robert, and Olivier Morrissey. 2000. “Aid Instability as a Measure of Uncertainty and the Positive Impact of Aid on Growth.” Journal of Development Studies 36, no. 3: 3149.
  • Lipset, Seymour Martin. 1959. “Some Social Requisites of Democracy: Economic Development and Political Legitimacy.” American Political Science Review 53, no. 1: 69105.
  • Mocan, Naci. 2004. “What Determines Corruption? International Evidence from Micro Data.” NBER Working Paper no. 10460. Cambridge, Mass.: National Bureau of Economic Research.
  • Organization for Economic Cooperation and Development (OECD). 2008. “Survey on Monitoring the Paris Declaration; Effective Aid by 2010? What Will It Take?” Paris: OECD.
  • Ouattara, Bazoumana. 2005. “A New Database on Project Aid and Programme Aid Disbursements and a Re-Examination of the Aid-Savings Nexus.” Economics Discussion Paper no. 0501. Manchester: University of Manchester.
  • Pagan, Adrian, and Aman Ullah. 1988. “The Econometric Analysis of Models with Risk Terms.” Journal of Applied Econometrics 3, no. 2: 87105.
  • Pallage, Stéphane, and Michel A. Robe. 2001. “Foreign Aid and the Business Cycle.” Review of International Economics 9, no. 4: 641672.
    Direct Link:
  • Rajan, Raghuram, and Arvind Subramanian. 2007. “Does Aid Affect Governance?American Economic Review Papers and Proceedings 97, no. 2: 322327.
  • Ramey, Garey, and Valerie Ramey. 1995. “Cross-Country Evidence on the Link between Volatility and Growth.” American Economic Review 85, no. 5: 11381151.
  • Robinson, James A.; Ragnar Torvik; and Thierry Verdier. 2006. “Political Foundations of the Resource Curse.” Journal of Development Economics 79, no. 2: 447468.
  • Schwert, G. William, and Paul Seguin. 1990. “Heteroskedasticity in Stock Returns.” Journal of Finance 45, no. 4: 11291155.
  • Svensson, Jakob. 2000. “Foreign Aid and Rent-Seeking.” Journal of International Economics 51, no. 2: 437461.
  • Tavares, Jose. 2003. “Does Foreign Aid Corrupt?Economics Letters 79, no. 1: 99106.
  • Vargas Hill, Ruth. 2005. “Assessing Rhetoric and Reality in the Predictability of Aid.” UNDP Human Development Report Office Occasional paper no. 2005/25. New York: United Nations Development Programme.
  • Ventelou, Bruno. 2001. “Équilibres et stabilité de la corruption dans un modèle de croissance: l'effet de la rémunération des politiciens.” L'actualité Economique 77, no. 3: 339356.

Appendix

  1. Top of page
  2. Abstract
  3. I. Introduction
  4. II. Aid Flow Uncertainty and Rent Extraction
  5. III. Empirical Evidence
  6. IV. Concluding Remarks and Policy Implications
  7. References
  8. Appendix

A Data Definition and Sources

Official development assistance (ODA): Total net disbursements include grants and loans with a grant element of more than 25% (Source: OECD/DAC).

  • Uncertainty: See Section III.B.1 for calculation details (Source: Author's calculation).
  • Aid%GDP: ODA measured as a percentage of GDP (Source: Author's calculation from OECD/DAC aid statistics and World Development Indicators 2008).
  • Loans: Net ODA loans, current US$ millions (Source: OECD/DAC statistics).
  • Grants: ODA grants, current US$ millions (Source: OECD/DAC statistics).
  • Multilateral aid: ODA from multilateral donors, current US$ millions (Source: OECD/DAC statistics).
  • Bilateral aid: ODA from bilateral donors, current US$ millions (Source: OECD/DAC statistics).
  • Project aid: Total net of project ODA, current US$ millions (Source: Ouattara 2005).
  • Program aid: Total net of program ODA, current US$ millions (Source: Ouattara 2005).
  • Financial program aid: Total net of programme ODA minus food aid, current US$ millions (Source: Ouattara 2005).
  • APD%GNI: Aid (% of gross capital formation). Aid includes both ODA and official aid. Ratios are computed using values in US dollars converted at official exchange rates (Source: OECD/DAC statistics and World Bank estimates).
  • Initial income: Value of the GDP divided by midyear population (constant 2000 US dollars) at the beginning of the periods (1984) (Source: Word Development Indicators 2008).
  • Corruption: Indicator of corruption as reported by international consultants. Scaled from 0 to 6, higher values denote less corruption (Source: International Country Risk Guide).
  • Rent: Exports of mineral (%GDP) (Source: World Development Indicators, 2008).
  • Accountability: Measure of how responsive government is to its people, on the basis that the less responsive it is, the more likely the government will fall, peacefully in a democratic society, but possibly violently in a nondemocratic one. Positively scaled from 0 to 6 (Source: International Country Risk Guide).
  • Landlock: Dummy variable taking the value of 1 for landlocked countries and 0 otherwise (Source: Global Development Network Growth Database).
  • Constraints on the executive power: Extent of institutionalized constraints on the decision-making powers of chief executives, whether individuals or collectivities, 1 (worst) – 7 (best) (Source: Polity IV).
  • Ethnolinguistic fractionalization: Probability that two random selected individuals within the country belong to the same religious and ethnic group (Source: Atlas Narodov Mira).
  • Legal origin (British): Origin of country legal system. Dummy variable taking the value of 1 for British legal origin and 0 otherwise (Source: Global Development Network Growth Database).
  • Population: Total population (Source: World Development Indicators, 2008).
  • Africa: Dummy taking the value of 1 for African countries (Source: Author).
  • Regions: Dummies indicating whether the country is part of East Asia and the Pacific, East Europe and Central Asia, the Middle East and North Africa, South Asia, sub-Saharan Africa or Latin America and the Caribbean (Source: Global Development Network Growth Database).

B The Sample Countries

 1.Algeria21.Ethiopia41.Madagascar61.Senegal
 2.Angola22.Gabon42.Malawi62.Sierra Leone
 3.Argentina23.Gambia, The43.Malaysia63.Somalia
 4.Bahrain24.Ghana44.Mali64.Sri Lanka
 5.Bolivia25.Guatemala45.Malta65.Sudan
 6.Botswana26.Guinea46.Mexico66.Suriname
 7.Brazil27.Guinea-Bissau47.Morocco67.Syrian Arab Republic
 8.Burkina Faso28.Guyana48.Mozambique68.Tanzania
 9.Cameroon29.Haiti49.Myanmar69.Thailand
10.Chile30.Honduras50.Nicaragua70.Togo
11.Colombia31.India51.Niger71.Trinidad and Tobago
12.Congo, Dem. Rep.32.Indonesia52.Nigeria72.Tunisia
13.Congo, Rep.33.Iran, Islamic Rep.53.Oman73.Turkey
14.Costa Rica34.Iraq54.Pakistan74.Uganda
15.Cote d'Ivoire35.Jamaica55.Panama75.Uruguay
16.Cuba36.Jordan56.Papua New Guinea76.Venezuela
17.Dominican Republic37.Kenya57.Paraguay77.Vietnam
18.Ecuador38.Korea, Rep.58.Peru78.Yemen, Rep.
19.Egypt, Arab Rep.39.Lebanon59.Philippines79.Zambia
20.El Salvador40.Liberia60.Saudi Arabia80.Zimbabwe
Table A1. Aid Unpredictability and Corruption (panel regressions, by aid types (2SLS); five-year subperiods: 1984–89; 1990–94; 1995–99; 2000–2004)
VariableCoefficient (std. err.)
Loans (1)Grants (2)Bilateral (3)Multilateral (4)Project (5)Prog. (6)Fin. Prog. (7)
  1. Notes: 1. Beside the coefficient value, the standard errors, which are computed using heteroskedastic-consistent standard deviations, are reported in parentheses.

  2. 2. All the estimations include regional controls, which are not reported for reasons of space.

  3. ***, **, and * represent statistical significance at the 1%, 5%, and 10% levels, respectively.

Log(uncertainty)−0.13 (0.09)−0.17** (0.08)−0.18* (0.10)−0.20** (0.09)−0.18 (0.12)−0.02 (0.07)−0.08 (0.08)
Log(init. income)−0.22 (0.18)−0.33* (0.19)−0.28* (0.17)−0.37* (0.20)−0.25 (0.19)−0.09 (0.19)−0.14 (0.18)
British−0.02 (0.20)−0.08 (0.21)−0.08 (0.21)−0.005 (0.20)−0.04 (0.23)−0.06 (0.25)−0.06 (0.25)
Rents−0.004 (0.03)−0.04 (0.03)−0.04 (0.03)−0.03 (0.03)−0.05* (0.03)−0.05* (0.03)−0.05* (0.03)
Log(pop.)−0.21*** (0.08)−0.26*** (0.08)−0.24*** (0.08)−0.28*** (0.10)−0.18* (0.11)−0.00 (0.09)−0.13 (0.09)
Accountability0.15*** (0.05)0.16*** (0.05)0.15*** (0.05)0.16*** (0.05)0.16*** (0.05)0.16*** (0.06)0.16*** (0.06)
Africa−0.08 (0.41)−0.08 (0.40)−0.02 (0.40)0.02 (0.41)−0.20 (0.58)−0.14 (0.58)−0.19 (0.57)
Eth. frac.−0.006* (0.004)−0.007* (0.004)−0.007* (0.004)−0.005 (0.004)−0.008* (0.005)−0.009** (0.004)−0.008* (0.004)
Landlock−0.41 (0.27)−0.42 (0.267)−0.42 (0.27)−0.42 (0.27)−0.43* (0.26)−0.41 (0.32)−0.45 (0.31)
Intercept5.57*** (1.73)6.65*** (1.83)5.84*** (1.70)6.70*** (1.89)5.13*** (1.75)4.22** (1.97)5.04*** (1.72)
Regional dummiesYesYesYesYesYesYesYes
Obs.271271271271235235235
R20.180.190.190.190.200.190.20
Partial Shea R20.60.70.60.60.50.520.5
Table A2. Aid Uncertainty and Corruption (aid types, robustness)
VariableCoefficient (std. err.)
Aid%GNIExcl. Outliers
  1. Notes: 1. Beside the coefficient value, the standard errors, which are computed using heteroskedastic-consistent standard deviations, are reported in parentheses.

  2. 2. All the estimations include regional controls, which are not reported for reasons of space.

  3. ***, **, and * represent statistical significance at the 1%, 5%, and 10% levels, respectively.

Log(uncertainty)−0.14* (0.09)−0.20** (0.10)
Log(init. income)−0.24 (0.16)−0.27 (0.18)
British−0.05 (0.20)−0.006 (0.22)
Rents−0.04 (0.03)−0.04 (0.03)
Log(pop.)−0.22*** (0.09)−0.24*** (0.09)
Accountability0.15*** (0.04)0.15*** (0.05)
Africa−0.005 (0.45)−0.002 (0.41)
Eth. frac.−0.006* (0.004)−0.004 (0.005)
Landlock−0.43* (0.24)−0.43 (0.27)
Intercept7.81*** (2.44)5.44*** (1.73)
Regional dummiesYesYes
Obs.271255
R20.180.1712
Partial Shea R20.70.72