On the Channel and Type of International Disaster Aid

Produced by the Research Support Team

Abstract

The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development
issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the
names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those
of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and
its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent.

Policy Research Working Paper 4953

Research suggests that a donor country’s decision to

provide post-disaster assistance is not only driven by

the severity of a disaster and the resulting humanitarian

needs in the recipient country, but also by strategic

considerations. The authors argue that the identification

of the determinants of the size of disaster assistance is a

first step in the analysis of the donor’s behavior. Since all

aid is not motivated by the same reasons, the evaluation

of the donor country’s behavior requires a second step

accounting for the type and the channel of aid provided.

Using data on international disaster assistance between

This paper—a product of the Global Facility for Disaster Reduction and Recovery Unit, Sustainable Development Network

Vice Presidency—is part of a larger effort in the department to disseminate the emerging findings of the forth coming

joint World Bank-UN Assessment of the Economics of Disaster Risk Reduction. The team leader for the Assessment—

iApurva Sanghi—ican be contacted at “asanghi@worldbank.org.” We would like to thank Jesus Crespo-Cuaresma, Howard

Kunreuther, Harald Oberhofer, Michael Pfaffermayer, S. Ramachandran, Apurva Sanghi, Mark Schelker, Reimund Schwarze,

Martina Tonizzo and Hannelore Weck-Hannemann as well as participants of a workshop in Zurich 2008 and a seminar

at the World Bank in 2009 for their useful comments.

2000 and 2007, the analysis examines both the donor

countries' decision on the channel (bilateral versus

multilateral) and the type of disaster relief (cash versus

in-kind). The empirical results suggest that international

disaster relief is not as much driven by the needs of the

recipient country, but also by strategic interests (for

example, oil or trade relationships) of the donor country.

Bilateral and cash transfers are used as a vehicle to signal

strategic interests, while multilateral and in-kind transfers

are chosen to control for misuse in badly governed

recipient countries.

On the Channel and Type of International Disaster Aid

Paul A. Raschky* Manijeh Schwindt*

Keywords: Foreign aid, natural disasters, bilateral vs. multilateral, type of aid

JEL classification: O17, O19, Q54

* Institute of Public Finance, University of Innsbruck, Universitaetsstrasse 15, A-6020 Innsbruck, Email:

paul.raschky@uibk.ac.at

and

manijeh.schwindt@uibk.ac.at

asanghi@worldbank.org

This paper –a product of the Global

Facility for Disaster Reduction and Recovery Unit, Sustainable Development Network Vice
Presidency- is part of a larger effort in the department to disseminate the emerging findings of the
forth coming joint World Bank-UN Assessment of the Economics of Disaster Risk Reduction.
The team leader for the Assessment – Apurva Sanghi – can be contacted at
“

”. We would like to thank Jesus Crespo-Cuaresma, Howard Kunreuther,

Harald Oberhofer, Michael Pfaffermayer, S. Ramachandran, Apurva Sanghi, Mark Schelker,
Reimund Schwarze, Martina Tonizzo and Hannelore Weck-Hannemann as well as participants of
a workshop in Zurich 2008 and a seminar at the World Bank in 2009 for their useful comments.

1 Introduction

Research suggests that by 2015 the number of people affected by natural disasters could rise by
more than 50 percent to an average of more than 375 million each year (Bailey 2009). Such an
enormous increase in the number of people facing risks of natural disasters will inevitably
increase the need for humanitarian assistance. Budget constraints call for an improvement of
disaster assistance strategies and a more efficient application of scarce resources. Therefore, a
better understanding of the allocation of disaster relief, including the quantity and quality, is
necessary. Why did the United States assist in 43% of flood relief cases in Kenya with cash
transfers and only in 20% of the flood relief cases in Bolivia (despite  both recipient countries
having  comparable numbers  of fatalities from floods between 2000 and 2007)? Why did
Mozambique receive 59% of United States' flood relief inflows bilaterally and the remaining via
a multilateral agency, while India received only 14% of its United States' flood assistance
bilaterally?

In order to shed more light on these  questions, we combine aspects of two broad strands in the
literature on foreign assistance. First, we refer to the literature on the type of aid, with discussions
on the effectiveness of cash and in-kind transfers, respectively (e.g. Currie & Gahvari 2008). The
argument of paternalism brought forward in this literature, can be seen as an explanation for the
tendency to provide in-kind or restricted aid transfers, in particular to recipient nations with weak
institutions. Recent contributions analyzing moral hazard behavior of the recipient country
include Amegashie, Ouattara & Strobl (2007) and Svensson (2000).

Second, our paper relates to the vast amount of literature investigating the allocation of foreign
aid. The seminal paper by Dudley  & Montmarquette (1976),  which argues that the supply of
foreign aid will be explained by the donor countrys' demand for foreign aid impact,  paved the
way for a discussion of donors' motives behind contributing aid. The majority of these papers
showed that foreign aid allocation is determined by the donor countries' strategic and political
interests rather than by the recipient countries' need. For example, Kuziemko  & Werker (2006)
find that more foreign aid is paid to countries which rotate on a seat in the U.N. security council,
Dudley  & Montmarquette (1976)  show that foreign aid is induced by the level of exports from
donor to recipient countries and the results of Alesina & Dollar (2000) suggest that colonial ties
result in greater aid allocation. More specifically and of great relevance for this paper, Fink  &
Redaelli (2009)  find that not only foreign aid but also international emergency assistance is
determined by political and strategic interests. They show that more emergency aid is paid to
countries which are located closer to the donor country, export oil or were former colonies of the
donor country. In addition, the results of Eisensee & Strömberg (2007) suggest that U.S. disaster
relief is driven by the level of media coverage of disasters. By contrast Olsen, Carstensen &
Hoyen (2003) find a rather limited role of news coverage. They use case-studies in order to show
that the size of humanitarian assistance is determined by the donor country's interest in stability
and security in the affected country as well as by the local presence of international organizations
with lobbying power. The former argument is consistent with what Bermeo (2007) calls strategic
development or strategic stabilization. According to her  argument and in contrast  to the above
mentioned literature,  all foreign aid is used for strategic interests. However, development or
stabilization of the recipient country can be part of the strategic goal as well. In this case, donor
interests and recipient needs coincide.

The analysis so far had a focus on the determinants of the size of aid and emergency assistance,
respectively, including the implicit assumption that all donors give the same type of aid and use
the same channels or that all aid is motivated by the same reasons. We argue that the analysis of
the amount of aid is a necessary first step but not sufficient to derive implications about donor
countries’ behavior. In order to get a more comprehensive picture on the motivation of donor
countries’ incentives to provide humanitarian aid and to design more efficient mechanisms of
international (disaster) assistance, the decision on both the type and the channel of aid need to be
considered. What criteria does a donor country use in order to decide whether to assist by cash
transfers or in-kind transfers? Why do countries pay bilateral aid to one country and multilateral
aid to another? Maintaining  the terminology of Bermeo  (2007)  and distinguishing between
strategic stabilization goals and non-stabilization goals, we would expect that donor countries try
to maximize aid effectiveness by their choice of the type and channel of aid, if strategic
stabilization is the main objective.  Whereas we would  expect them to choose the type of aid,
which the recipient country values the highest, when their primary interests are non-stabilization
goals.

Strategic stabilization requires that transfers reach their desired recipients. However, in
comparison to cash transfers, restricted transfers, e.g. in-kind transfers, might be better suited for
the reduction of moral hazard behavior and efficient targeting (e.g. Amegashie  et al. 2007

Besley & Coate 1991, Gahvari & Mattos 2007, Svensson 2000). Note, that the term cash transfer
does not include conditional cash transfers. Literature on conditional cash transfers suggests that
depending on the design this form of assistance can be very effective. For example, Doocy,
Gabriel, Collins, Robinson & Stevenson (2006) describe the implementation of cash for work
programs after the Tsunami in Aceh.

Not only the type of aid, but also the channel of aid might

be relevant for the ability of transfers to achieve stabilization. Opposed to bilateral assistance,
multilateral agencies might have better information about the risks in aid receiving countries and
hence account for them in their allocation decisions (Weck-Hannemann & Schneider 1991).
Moreover, since donor countries lack commitment power, Svensson (2000) argues that the
delegation of aid to agencies which are less risk averse and have plausible commitment
techniques could provide incentives in the receipient country to generate own effort. On the other
hand, non-stabilization goals are more probable to achieve, if the aid receiving government
values the assistance highly. It is reasonable to assume that governments value cash payments
higher than in-kind payments because they can use it in accordance with their own preferences.
Moreover, donor countries transferring money directly to recipient governments might be more
successful in building up political ties, since bilateral transfers are more visible for recipients than
countries which act anonymously via a multilateral agency. For that reason, strategic stabilization
goals should increase the probability of multilateral and in-kind transfers whereas non-
stabilization goals should induce bilateral and cash transfers.

The remainder of the paper is organized as follows: In section 2 we present the theory and derive

In fact, Amegashie et al. (2007) investigate how donor countries' choice of the composition of cash and in-kind

transfers adjusts to changes of the moral hazard behavior of recipient countries. While multilateral donors reward
(penalize) decreases (increases) in moral hazard behavior by reducing (rising) the proportion of in-kind relative to
cash-transfers, bilateral donors do not react to changes in moral hazard behavior.

For a more detailed analysis of aid programs using conditional cash transfers see Fiszbein, Schady, Ferreira,

Kelleher, Olinto & Skofias (2009).

Needless to say that the decision on the type and channel of post-disaster assistance is determined by other

characteristics (e.g. disaster type) as well. Of course we control for these variables in the empirical part of this paper.

the signs for the partial derivatives of the probability of the different types of aid. Research
design, data and econometric strategy are presented in section 3. Finally, section 4 concludes.

2 Theory

In order to derive the factors which explain the choice of the type and channel of disaster aid
made by donor countries, we use a technical framework applied by Huber  & Nowotny (2008).
They analyze an individuals decision to commute and to migrate in another region opposed to the
possibility to stay in the home region.

First, we allow each donor country to choose between bilateral disaster assistance, multilateral
disaster assistance and no disaster assistance. The donor country's decision to assist or not and its
choice between bilateral and multilateral disaster aid are based on two instruments,  which
according to Bermeo  (2007),  both contribute to foreign policy purposes. First, because of
strategic reasons,  the donor country might be interested in the stabilization of the recipient
country in the aftermath of a disaster. Strategic stabilization could, for example, be motivated by
the intention to mitigate negative spill-over effects on the donor country's economy or to prevent
large flows of disaster-refugees. Hence, under these circumstances donor interests and recipient
needs coincide. Second, the donor country might have other strategic interests not related to the
strategic stabilization goal, which the donor country wants to promote, e.g. political ties to oil
exporting countries.

To be more precise, consider a donor country with GDP

Y which decides a) whether to

contribute an amount of

to a country which was hit by a natural disaster and b) whether to

transfer this amount directly to the recipient government (bilateral (B)) or via a multilateral
agency (M). Empirical results suggest that it is reasonable to assume that disaster assistance
contributes to a higher level of stabilization in the recipient country

S ,

∀

and hence a

higher utility level of the donating country if e.g. good governance indicators in the recipient
country are high (Burnside & Dollar 2000). Therefore, the stabilization effect is determined by
the ability of multilateral and bilateral disaster aid to circumvent the adverse effects of bad
governance, respectively. Due to a more profound knowledge of the specifics of the recipient
countries resulting from the presence in affected countries, we assume that the control
mechanism using multilateral disaster assistance is better compared to the control in the case of
bilateral disaster assistance. Moreover, Svensson (2000) suggests that the delegation of aid to
international agencies with fully developed commitment technologies contributes to a higher
effecieny level. For this reason, stabilization is easier to enforce with multilateral assistance for a
given level of good governance. Moreover, for higher levels of good governance, strict control
mechanisms are less of a constraint. In order to gain utility from better access to strategic
interests

I ,

∀

in the recipient country, it is important for the donor country to ensure

that the source of disaster assistance is visible for the recipient country. We assume that bilateral
disaster assistance is more visible than multilateral disaster assistance and thus secures strategic
interest in a better way. Assuming an additive utility function, the utility

U of a donor country

choosing bilateral disaster assistance is determined by its own governmental purchases

− ,

by the stabilization effect in the recipient country

S and by the strategic advantages

I induced

by bilateral disaster aid. We formulate the donor country's utility function for bilateral emergency
assistance as

−

(1)

Equivalent to equation (1) the donor country's utility in the case of multilateral disaster assistance
is

−

(2)

The terms

ε and

ε are random utility components for donating bilateral and multilateral,

respectively. Apart from bilateral and multilateral disaster assistance, the potential donor country
can decide not to assist. In this case the donor country's utility is

(3)

We use equations (1), (2) and (3) in order to calculate the utility differentials between bilateral
disaster assistance

)

(

, multilateral disaster assistance

)

(

and no disaster assistance

)

(

−

(4)

−

(5)

Equations (4) and (5) state that higher levels of stabilization

S achieved with bilateral and

multilateral disaster aid as well as higher strategic gains, increase the utility of disaster assistance
in comparison to not assisting. However, assistance bears costs in the form of a reduction in a
donor country's government purchases.

(

) (

)

ε −

−

(6)

Equation (6) shows that the stabilization differential

(

−

increases the utility gain from

multilateral disaster assistance, whereas the strategy differential

(

−

increases the

utility gain from bilateral disaster assistance.

In order to simplify (4), (5) and (6) we define the direct utility gains by

−

Ω

(7)

−

Ω

(8)

(

) (

)

−

Ω

(9)

and rewrite (4), (5) and (6) as

Ω

−

(10)

with

−

Ω

−

(11)

with

−

(

)

ξ −

Ω

−

(12)

Equations (10) - (12) state that countries will decide not to assist, if

Ω

−

and

Ω

−

However, for

Ω

−

and

Ω

−

there will be bilateral or multilateral disaster

assistance, depending on the relation of

Ω and

Ω . Countries will choose bilateral disaster

assistance for

Ω

−

and

Ω

−

)

(

, whereas they will assist via a multilateral

agency if

Ω

−

and

Ω

−

)

(

We can now determine the probability that a donor country chooses bilateral assistance

)

(

P ,

multilateral assistance

)

(

and no assistance

)

(

)

(

)

−

Ω

−

Ω

;

(13)

(

)

(

)

−

Ω

−

Ω

;

(14)

(

)

−

Ω

−

Ω

;

(15)

Using comparative statics we can derive the derivatives of probabilities of assisting by bilateral
or multilateral transfers and not assisting (see table 1).

Table 1: The choice between bilateral, multilateral and no disaster assistance subject to

selected variables

Now that we have determined the partial derivatives of the probabilities to pay bilateral disaster
assistance, multilateral disaster assistance and no assistance, we vary the options of the donor
country, by distinguishing between cash (C) and in-kind (IK) transfers in the case of bilateral
assistance. In accordance to (1) - (3), we formulate the donor country's utility depending on the
type of aid, as:

−

(16)

−

(17)

−

(18)

(19)

Note that in-kind transfers are assumed to cause some additional transportation costs

)

, which

will be higher for large distances between donor and recipient country. Moreover, we would
expect that in-kind transfers can better control for the adverse effects of bad governance
compared to cash transfers, whereas cash transfers might be valued higher than in-kind transfers
from the aid receiving country and hence be more suited for achieving strategic goals. Using the
same approach as above, we determine the following signs of the partial derivatives of the
probabilities to choose cash, in-kind, multilateral or not any assistance (see table 2).

Table 2: The choice between cash, in-kind, multilateral and no disaster assistance subject

to selected variables

3 Empirical analysis

3.1 Research design and data

We are interested in the decision of potential donor countries (i.e. every country that has not been
directly affected by a disaster) to provide post-disaster assistance and the channel and type the
actual donors choose. In order to examine the effect of humanitarian needs of a recipient country
and strategic interests of a donor country, we construct a basic dyadic data set for each major
natural disaster (that is included in the EM-DAT data set) in a given country between 2000 and
2007. For any given disaster in a country,  all remaining countries are considered as potential
donor nations. Including only those cases where one potential donor actually provided aid in our
regression  would truncate the data. All potential donors that did not provide post-disaster
assistance are coded zero and this information is used in the first stage selection estimates. The
combination of 228 disasters, where information on both the channel and type of disaster aid is
available,  and 187 potential donor nations,  results in a basic dataset of 42,636 observations.
However, this number is reduced to between 20,077 and 25,836 (depending on the specification)
due to missing data. Of the aforementioned observations, between 2,603 and 3,123 (depending on
the specification) observations are actual aid contribution (dependent variable=1).

The dependent variables are dummies that switch to 1 if a donor has contributed post-disaster
assistance (in the selection equation), switch to 1 if the if the contribution was bilateral (in the
channel equation) or switch to 1 if the type of contribution was cash (in the type equation). In
figures 1 and 2 we show the breakdown of total aid in bilateral vs. multilateral and cash vs. in-
kind by major recipients. Pakistan received 602 contributions. Among these 602 contributions,
463 (77%) were made via a multilateral agency and 139 (23%) were made bilaterally. Peru on the
other end of the list, received 99 contributions where 54 (55%) where made multilaterally and 45
(45%) bilaterally. The differences between recipient countries in the type of disaster aid received
is even bigger. Out of 100 contributions to Haiti only 1 (1%) was bilateral cash, while
Mozambique received 53 (16%) bilateral aid contributions out of 335 disaster assistance flows.
The main explanatory variables include indicators for humanitarian needs by the recipient and
variables accounting for strategic interests of the donor. The former group includes the number of
fatalities in a disaster (in thousand) and the level of gross domestic product per capita (GDP p.c.).
The latter group contains  information on the donor's trade volume with the recipient and the
percentage of fuel exports of total merchandise exports by the recipient. These two variables are
widely used as empirical proxies for strategic interests in the existing aid allocation literature
(e.g. Fink & Redaelli 2009,  Berthelemy  2004). In addition we include dummies for the type of
natural disaster.

In choosing other relevant covariates, we follow the existing empirical literature (e.g. Fink  &
Redaelli 2009,  Alesina  & Dollar 2000): Size of the recipient country (population), distance
between donor and recipient, oneness for trade, colonial history between the donor and the
recipient as well as an updated version of Gartzke's affinity index that is constructed using voting
patterns in the United Nations General Assembly. The index ranges between 1 (recipient and
donor always voted the same way) and -1 (recipient and donor never voted the same way). In
addition we include two dummy variables that account for common religious beliefs and common
language in the donor and recipient countries. The donor nation’s level of development measured
via the GDP p.c. is also included. Table 15 gives an overview of the data sources used.

3.2 Econometric strategy

Based on the theoretical concept in section 2, our goal is to identify the driving factors of the
likelihood of choosing 1) a certain channel for disaster aid (bilateral vs. multilateral) and 2) a
certain type of disaster aid (cash vs. in-kind). The decision on both the channel and the type
however, is conditional on the decision to provide post-disaster assistance at all. The resulting
selection problem can be formulated as follows:

0),

(

)

(

)

(

(20)

where

y denotes the selection variable and equals 1 if aid is given, zero otherwise,

y denotes

the second stage channel (1 if aid is bilateral, zero if multilateral) or type (1 if type of aid was
cash, zero if in-kind) and

x is a vector of covariates. In our case, the sample selection model

consists of two stages: The first stage defines the cases where actual post-disaster aid is given.
The selection variable is a latent variable

y and equals 1 if aid is given. The second stage is the

outcome stage and is estimated in two separate specifications. In the first specification it
describes the cases where bilateral aid was given, while in the second specification it describes
the cases when cash was contributed rather than in-kind. In either of the two specifications we
denote this second stage latent variable as

*
2

y . We derive the following system:

(21)

*
2

(22)

x and

u are the explanatory variables and the error terms for the first and the second stage. The

correlation between the two equations,

ρ , indicates if there is actual sample selection. The

traditional Heckman model (Heckman 1979) requires that the second stage outcome equation is
estimated using OLS. Dubin & Rivers (1989) developed an extended selection model where
second stage is estimated using a probit model, which is applied in this paper. The latent
variables

y are related to the observed variables

in the following way:











*
2

and

(23)

The application of a sample selection model requires unique information in the explanatory
variables

x and

x to separately identify the parameters in the selection and the outcome stage.

To deal with this issue we use the donor nation's population as additional selection variable in the
first stage selection.

3.3 Results

The results are structured by the strategic variables of interest, oil and trade. We start presenting
estimates of the second stage estimates on the decision on the channel of disaster  aid including
the fraction of fuel exports of total merchandise export if a contribution has been made (Table 5).
The dependent variable is a dummy that equals  to 1 if the contribution was bilateral and 0 if it
was multilateral. Fatalities do not appear to have an impact on the choice of the aid channel, but
the number of people affected has a significant positive impact. More distant countries are also
more likely to receive aid via a multilateral agency. Countries with a higher fraction of fuel
exports and better governance indicators are more likely to receive bilateral aid. These results
allow for interpretations: First, the bilateral channel is preferred over the multilateral channel
because it is more attributable to the donor and thus supports strategic interests in a better way.
Second, donors are more likely to give aid via a multilateral agency because these agencies might
have a better competence  in controlling the sound use of the aid in recipient nation's with low
good governance indicators.

The interaction term of oil and good governance indicators have a negative sign and are only
significant at the 5%-level for corruption control (column (5)). The signalling effect of bilateral
aid might by decreasing with better regulatory quality or corruption control.

Table 6 presents the results of the first stage selection equation. As expected, countries that have
suffered more fatalities and where more people where affected from a disaster and that are poorer
are more likely to receive international disaster assistance. In line with the standard ODA
literature, donors appear to favor more open recipient countries. Interestingly, smaller nations are
less likely to receive disaster aid. This result stays in contrast to the small-country bias found in
ODA decisions. The likelihood of receiving aid increases in the number of fatalities and if donor
and recipient country have a common language. In accordance  with the findings of Fink  &
Redaelli (2009) and opposed to the existing literature on ODA decision, affinity does not increase
the likelihood of receiving aid from a donor. Interestingly, if the donor and the recipient share the
same religious beliefs, the likelihood of receiving aid significantly decreases. One possible
explanation for this result is that the data on religion only controls for potential frictions between
large religious groups and not within the large religious groups (e.g. Sunnite and Shiite).

In the specification presented in column (2) we have included oil (percentage of fuel exports from
total merchandise exports) as an additional regressor. A higher fraction of fuel exports increases
the likelihood of receiving disaster aid. These results  remain  robust after the inclusion of
regulatory quality (column (3)) and corruption control (column (5)). Including the oil variable
with the good governance indicators (columns (4) and (6)) reveals that the likelihood of receiving
aid decreases for a recipient nation that has large fuel exports and good regulatory quality and
increases if the recipient has large fuel exports and good corruption control.

We now turn to the estimates of the specifications using trade volume as the second indicator for
strategic interests. The results of the aid selection equations in table 8  are pretty similar to the
first stage results for the oil estimates in table 6. The effects of strategic interests and good
governance as well as the interaction of these two variables reveal a similar trend for the trade
estimates. Donors are more likely to donate via a multilateral agency if there has been a major
disaster causing a large number of fatalities and if the affected country is poor. In contrast to the
oil estimates, there are differences in the second stage estimates on the channel of aid. The

interaction between trade volume and governance indicators is positive and highly significant.

The next step consisted of changing the specification of the second stage estimates and
examining the decision on the type of aid. The dependent variable is the probability of giving
bilateral cash. We only present the results of the second stage for both the oil and trade estimates.
Interestingly, fuel exports do not appear to have a significant effect on the likelihood of receiving
cash (table 9), while trade volume has a significantly positive effect throughout all specifications
(table 10). Humanitarian aspects appear to have adverse effects on the likelihood of receiving
cash. Donors prefer to donate cash to richer nations and after smaller disasters with fewer people
killed. Regarding the strategic variables we find similar patterns as in the choice of the aid
channel.

The analysis so far has assumed that the variables that explain the choice on the composition of
aid do not differ between countries. The empirical literature, however, suggests that donors'
decision on ODA (e.g. Alesina & Dollar 2000, Kuziemko & Werker 2006) and disaster aid (e.g.
Fink & Redaelli 2009) are not the same across donor nations. For expositional convenience, we
limit our analysis to a comparison between OECD and non-OECD countries. This robustness test
basically splits the sample in OECD and non-OECD donor subsamples and repeats the estimates
in tables 5 - 10 for each subsample, respectively.

We only present the coefficients of the second stage estimates for the key variables. The results of the first stage

estimates as well as the full estimation table including all other covariates are available from the authors.

Unfortunately, the the estimates did not

converge using a heckman probit estimator. We therefore applied a simple conditional probit
model, estimating the likelihoods of receiving bilateral aid or bilateral cash conditional that aid
has been given. Table 11  presents the coefficients of the second conditional probit models for
both OECD and non-OECD donors decisions on the channel of disaster aid. A first glance reveals
that the pooled estimates in tables 5 - 10 are mainly driven by the OECD sample. OECD donors
are more likely to provide bilateral disaster aid if the recipient export share of fuel is large and if
the recipient nation has good governance indicators. There is a smaller probability of giving
bilateral aid if the disaster caused a lot of fatalities and if the recipient has a larger per capita
income. In contrast, fuel exports appear to have a negative and not significant impact on the
decision on the channel of disaster aid for non-OECD countries. One possible explanation is that
the majority of donors in this subsample are countries from the Middle East and North Africa
(MENA) which are already endowed with large oil resources. Good governance indicators also
influence the decision in both subsamples, where the coefficients for the non-OECD contries is
actually larger. Table 12  reports the result for the second strategic variable, trade. Again the
results of the OECD donor group are comparable to the results of the main sample. The estimates
for non-OECD again reveal that strategic interests do not play a significant role in their decision.
We find some indication that the trade relationship between the donor and the recipient has a
positive and significant influence on the decision to provide bilateral aid. Tables 13 and 14 show
the results for the probit estimates on the type of aid for each subsample and each strategic
variable, respectively. Again, the decision on the type of aid is driven by strategic interests in the
case of OECD countries, while strategic  interests have no significant influence on non-OECD
countries decision. However, the only explanatory variables that have a significant influence in
the non-OECD subsample are the disaster type dummies and the common language dummy.

4 Conclusion

Research of the past 30 years suggests that foreign aid is allocated due to strategic and political
interests rather than due to recipient needs. The aim of this paper was to show that strategic
concerns not only dominate the choice of the size of aid, but also the decision on the type and the
channel of aid. Our theoretical results suggest that donor countries assist by multilateral or in-
kind transfers rather than by bilateral or cash transfers, if their main interest is the stabilization of
the recipient country, whereas they choose bilateral or cash transfers if they account for non-
stabilization goals. The empirical application shows that a) recipient countries with good
governance are rewarded by bilateral or cash transfers, b) countries with lower levels of GDP p.c.
or higher death tolls are more probable to receive multilateral or in-kind transfers and c) the
probability of getting bilateral or cash assistance increases if the donor country has strategic
interests (e.g. oil or trade) in the receiving country. Moreover, we find that good governance is
less of a constraint for bilateral or cash transfers if the donor country has strategic interests in the
same country. We show that donor countries deliberately choose the type and channel of aid in
conformity with their goals, however, strategic interests seem to dominate stabilization goals.
Interestingly, the reasons to provide bilateral aid as well as cash aid differ between OECD
countries and non-OECD countries. OECD countries are more likely to give bilateral aid and
cash if the recipient has oil resources, is a trading partner and has sound institutions. In contrast,
non-OECD countries’ decisions on the channel and type are not influenced by strategic interests.
These countries are more likely to provide bilateral aid and cash if the recipient has suffered a
large number of casualties in a disaster.

References

Alesina, A. F. & Dollar, D. (2000), Who Gives Foreign Aid to Whom and Why?, Journal of

Economic Growth 5(1), 33-63.

Amegashie, A. J., Ouattara, B. & Strobl, E. (2007), Moral Hazard and the Composition of

Transfers: Theory with an Application on Foreign Aid, CESifoWorking Paper No. 1996,
CESifo.

Bailey, R. (2009), The Right to Survive in a Changing Climate, Oxfam Background Paper,

Oxfam International.

Bermeo, S. B. (2007), Utility Maximization and Strategic Development - A Model of Foreign

Aid Allocation, mimeo, Department of Politics, Princeton University.

Berthelemy, Jean-Claude & Tichit, A. (2004), Bilateral Donors’ Aid Allocation Decision - a

three-dimensional Panel Analysis, International Review of Economics & Finance 13(3),
253-274.

Besley, T. & Coate, S. (1991), Public Provision of Private Goods and the Redistribution of

Income, American Economic Review 81(4), 979-984.

Burnside, C. & Dollar, D. (2000), Aid, Policies, and Growth, American Economic Review,

90(4), 847-868.

Correlates of War 2 Project (2008), Colonial/Dependency Contiguity Data, 1816-2002.,

http://correlatesofwar.org Version 3.0.

Currie, J. & Gahvari, F. (2008), Transfers in Cash and In-Kind: Theory Meets the Data’, Journal

of Economic Literature 46(2), 333–383.

Doocy, S., Gabriel, M., Collins, S., Robinson, C. & Stevenson, P. (2006), Implementing Cash for

Work Programmes in Post-Tsunami Aceh: Experiences and Lessons Learned, Disasters
30(3), 277–296.

Dubin, J. A. & Rivers, D. (1989), Selection Bias in Linear Regression, Logit and Probit Models,

Journal of Economic Growth 5(1), 33–63.

Dudley, L. & Montmarquette, C. (1976), A Model of the Supply of Bilateral Foreign Aid’,

American Economic Review 66(1), 132–142.

Eisensee, T. & Strömberg, D. (2007), New Floods, New Droughts, and U.S. Disaster Relief,

Quarterly Journal of Economics 122(2), 693–728.

Fink, G. & Redaelli, S. (2009), Determinants of International Emergency Aid: Humanitarian

Need Only?, World Bank Policy Research Working Paper No. 4839, The World Bank.

Fiszbein, A., Schady, N., Ferreira, F. H., Grosh, M., Kelleher, N., Olinto, P. & Skofias, E. (2009),

Conditional Cash Transfers, Reducing Present and Future Poverty, The World Bank.

Gahvari, F. & Mattos, E. C. (2007), Conditional Cash Transfers, Public Provision of Private

Goods and Income Redistribution, American Economic Review 97(1), 491–502.

Garcia, S., Harou, P., Montagné, C. & Stenger, A. (2009), Models for Sample Selection Bias in

Contingent Valuation: Application to Forest Biodiversity, Journal of Forest Economics
15, 59–78.

Heckman, J. J. (1979), Sample Selection Bias as a Specification Error. Econometrica, 47(1),

153-161.

Huber, P. & Nowotny, K. (2008), Moving Across Borders: Who Is Willing to Migrate or to

Commute?, WIFO Working Paper No. 322, WIFO.

Kaufmann, D., Kraay, A. & Mastruzzi, M. (2008), Governance Matters VII: Aggregate and

Individual Governance Indicators 1996-2007, World Bank Policy Research Paper No.
4654, The World Bank.

Kuziemko, I. & Werker, E. (2006), How Much Is a Seat on the Security Council Worth? Foreign

Aid and Bribery at the United Nations, Journal of Political Economy 114(5), 905–930.

Marshall, M. G. & Jaggers, K. (2005), Polity IV Project. Political Regime Characteristics and

Transition, 1800-2004, Version 2004.

Olsen, G. R., Carstensen, N. & Hoyen, K. (2003), What Determines the Level of Emergency

Assistance? Media Coverage, Donor Interests and the Aid Business, Disasters 27(2),
109–126.

Svensson, Jakob (2000), When is Foreign Aid Policy Credible? Aid Dependence and

Conditionality, Journal of Development Economics 61(1), 61-84.

Tunali, I. (1986), A General Structure for Models of Double Selection and an Application to a

Joint Migration/Earnings Process with Remigration, in E. G. Ehrenberg, ed., Research in
Labour Economics, Vol. 8, Part B, JAI Press, pp. 235–283.

United Nations Office for the Coordination of Humanitarian Affairs (OCHA) (2009), Financial
Tracking Service (FTS) – The Global Humanitarian Aid Database,
http://ocha.unog.ch/fts/pageloader.aspx

Voeten, E. & Merdzanovic, A. (2009), United Nations General Assembly Voting Data,

hdl:1902.1/12379 unf:3:hpf6qokddzzvxf9m66yltg= =, Georgetown University.

Weck-Hannemann, H. & Schneider, F. (1991), Determinants of Foreign Aid Under Alternative

Institutional Arrangements, in R. Vaubel & T. D. Willett, eds, The Political Economy of
International Organizations, Westview Press, pp. 245–266.

Table 3: List of recipient countries' total fatalities and number of disasters

Recipient

Fatalities

Disasters

Recipient

Fatalities

Disasters

Afghanistan

669

Lao, PDR

Albania

Madagascar

602

Algeria

971

Malawi

567

Argentina

Malaysia

Armenia

n.a.

Maldives

102

Azerbaijan

Mali

Bahamas, The

Mauritania

Bangladesh

2,309

Mexico

Belize

Micronesia, Fed. Sts.

Bolivia

271

Moldova

Botswana

Mongolia

Brazil

Morocco

708

Bulgaria

Mozambique

908

Cambodia

403

Myanmar

307

Central African Rep.

Namibia

Chile

Nepal

657

China

1,185

Nicaragua

Colombia

109

Niger

Comoros

Oman

Costa Rica

Pakistan

74,137

Cuba

Panama

Czech Republic

Papua New Guinea

n.a.

Djibouti

Peru

815

Dominica

Philippines

3,070

Dominican Republic

830

Poland

Ecuador

Portugal

El Salvador

863

Romania

Ethiopia

498

Russian Federation

101

Fiji

Senegal

Georgia

Seychelles

Ghana

Solomon Islands

Grenada

Somalia

350

Guatemala

n.a.

Sri Lanka

35,634

Guinea

n.a.

St. Lucia

Guyana

Sudan

Haiti

2,857

Suriname

Honduras

Tajikistan

Hungary

Thailand

8,449

India

38,730

Togo

Indonesia

172,214

Tonga

n.a.

Iran

28,110

Turkey

219

Jamaica

Uganda

Japan

Ukraine

Kenya

173

Uruguay

Korea, DPR

934

Vanuatu

Korea, Republic of

210

Venezuela

Kyrgyzstan

Vietnam

844

Zimbabwe

Table 4: List of donor countries' total contributions and number of donations

Donor

Total contribution Events

Donor

Total contribution

Events

(in USD)

Afghanistan

500,000

Iceland

473,627

Algeria

2,489,199

India

23,630,944

Andorra

58,386

Indonesia

n.a.

Angola

n.a.

Iran

347,380

Argentina

n.a.

Ireland

40,573,378

131

Armenia

n.a.

Israel

2,357,000

Australia

54,936,086

115

Italy

76,690,358

121

Austria

11,436,846

Japan

445,981,017

195

Azerbaijan

622,000

Jordan

n.a.

Bahrain

n.a.

Kazakhstan

n.a.

Bangladesh

100,000

Kenya

75,000

Belarus

113,018

Korea, DPR

130,000

Belgium

44,886,419

Korea, Republic of

1,576,709

Bolivia

n.a.

Kuwait

3,366,013

Botswana

482,000

Kyrgyzstan

27,093,596

Brazil

200,000

Lao, PDR

75,000

Bulgaria

103,717

Latvia

446,726

Burundi

20,000

Lebanon

n.a.

Canada

108,799,910

204

Lesotho

110,000

Chile

30,000

Libya

1,500,000

China

14,009,631

Liechtenstein

305,278

Colombia

100,000

Lithuania

252,631

Costa Rica

n.a.

Luxembourg

12,165,218

Croatia

n.a.

Malawi

100,000

Cuba

129,965

Malaysia

5,138,948

Cyprus

756,462

Malta

10,854,817

Czech Republic

5,498,495

Mauritania

200,336

Denmark

60,283,135

146

Mauritius

80,000

Dominican Republic

196,370

Mexico

4,127,922

Ecuador

13,237

Moldova

455,307

Egypt

300,000

Monaco

640,081

El Salvador

n.a.

Morocco

496,980

Eritrea

n.a.

Namibia

800,000

Estonia

577,084

Nepal

235,391

Fiji

9,700

Netherlands

101,964,604

139

Finland

25,055,726

New Zealand

15,536,259

France

48,601,080

118

Nicaragua

n.a.

Gabon

200,000

Nigeria

1,150,000

Germany

174,339,341

371

Norway

117,858,752

223

Ghana

100,000

Oman

100,000

Greece

27,047,570

Pakistan

157,560

Guatemala

n.a.

Palau

51,772

Guyana

20,000

Panama

n.a.

Honduras

n.a.

Peru

111,130

Hungary

1,005,267

Poland

6,966,713

Table 5: Aid & bilateral OIL – 2

stage

)

(

bilateral

(1)

(2)

(3)

(4)

(5)

Ln(Fatalities)

0.008

0.026

0.024

0.019

(0.016)

(0.017)

(0.018)

(0.017)

(0.016)

Ln(Affected)

0.032***

0.024***

0.023***

0.010

0.014**

(0.005)

(0.006)

(0.007)

Ln(GDP p.c.)

-0.053

-0.119**

-0.125**

-0.138***

-0.122**

(0.042)

(0.048)

(0.060)

(0.054)

(0.055)

Ln(Population)

-0.103***

-0.100***

-0.096***

-0.099***

(0.026)

(0.027)

(0.028)

Trade (% of GDP)

-0.001

-0.002

-0.001

-0.000

(0.001)

Distance

-0.070***

-0.082***

-0.079***

(0.009)

(0.010)

Affinity index

0.679**

0.470

0.467

0.546*

0.563*

(0.296)

(0.317)

(0.318)

(0.319)

Common religion

-0.028

0.003

0.002

-0.045

-0.048

(0.073)

(0.084)

(0.085)

(0.084)

Common language

0.137

-0.067

-0.019

-0.023

(0.088)

(0.124)

(0.122)

(0.129)

Former colony

-0.091

-0.166*

-0.151

-0.177*

(0.090)

(0.099)

(0.100)

Ln(GDP p.c.)

donor

-0.188***

-0.149***

-0.152***

-0.162***

-0.161***

(0.052)

(0.051)

(0.050)

Fuel exports (% of

0.002**

0.009***

0.008***

0.006***

-0.003

merchandise exports)

(0.001)

(0.003)

(0.001)

(0.004)

Regulatory quality

0.504***

0.522***

(0.091)

(0.122)

Fuel exports

-0.001

Regulatory quality

(0.004)

Corruption control

0.387***

0.477***

(0.104)

(0.110)

Fuel exports

-0.014**

Corruption control

(0.006)

Disaster dummies

Yes

0.717***

0.793***

0.788***

0.773***

0.776***

(0.120)

(0.123)

(0.121)

(0.119)

Constant

1.308

1.731**

1.813*

2.197**

2.077**

(0.800)

(0.843)

(0.943)

(0.909)

(0.896)

Loglikelihood

-7258.304

-6054.057

-6053.534

-6056.025

6032.289

3158

2632

Notes: Probit estimates. Coefficients reported; robust standard errors in parentheses. Dependent variable is bilateral,
a dummy that switches to 1 if the aid flow was bilateral. ***, **, * indicate significance at the 1, 5 and 10%-level,
respectively.

Table 6: Aid & Bilateral OIL- 1

stage

)

(

aid

(1)

(2)

(3)

(4)

(5)

Ln(Fatalities)

0.158***

0.154***

0.155***

0.153***

0.149***

(0.008)

(0.009)

Ln(Affected)

0.064***

0.056***

0.058***

0.059***

0.055***

(0.003)

(0.004)

Ln(GDP p.c.)

-0.094***

-0.141***

-0.130***

-0.084***

-0.094***

(0.017)

(0.023)

(0.025)

(0.024)

Ln(Population)

-0.304***

-0.283***

-0.284***

-0.287***

-0.289***

(0.013)

(0.015)

Trade (% of GDP)

-0.004***

-0.003***

-0.004***

(0.001)

Distance

-0.070***

-0.072***

-0.073***

-0.072***

(0.004)

Affinity index

-0.627***

-0.601***

-0.599***

-0.588***

(0.093)

(0.102)

Common religion

0.115***

0.081**

0.079*

0.074*

(0.036)

(0.041)

Common language

0.278***

0.252***

0.249***

0.264***

(0.049)

(0.059)

(0.060)

(0.059)

(0.060)

Former colony

-0.009

-0.023

-0.020

-0.012

(0.051)

(0.053)

(0.054)

Ln(GDP p.c.)

donor

0.607***

0.602***

0.601***

0.602***

(0.016)

(0.017)

Pop. (in mio.)

donor

0.001***

(0.000)

Fuel exports (% of

0.002***

0.004***

0.006***

0.003***

0.012***

merchandise exports)

(0.001)

(0.002)

(0.001)

(0.002)

Regulatory quality

0.072

0.044

(0.049)

(0.051)

Fuel exports

0.002

Regulatory quality

(0.002)

Corruption control

-0.134***

-0.213***

(0.041)

(0.045)

Fuel exports

0.012***

Corruption control

(0.002)

Disaster dummies

Yes

Constant

-1.694***

-1.589***

-1.688***

-2.042***

-1.910***

(0.350)

(0.390)

(0.398)

(0.400)

(0.407)

26811

22246

Notes: Probit estimates. Coefficients reported; robust standard errors in parentheses. Dependent variable is aid, a
dummy that switches to 1 if the donor has made a contribution. ***, **, * indicate significance at the 1, 5 and 10%-
level, respectively.

Table 7: Aid & bilateral TRADE – 2

stage

)

(

bilateral

(1)

(2)

(3)

(4)

(5)

Ln(Fatalities)

0.017

0.014

0.013

0.018

(0.014)

(0.015)

Ln(Affected)

0.035***

0.028***

0.025***

0.022***

(0.004)

(0.005)

(0.006)

Ln(GDP p.c.)

-0.147***

-0.133***

-0.148***

-0.149***

-0.154***

(0.040)

(0.044)

(0.045)

(0.048)

(0.049)

Ln(Population)

-0.254***

-0.244***

-0.257***

-0.240***

-0.253***

(0.029)

(0.031)

(0.032)

Trade (% of GDP)

-0.002**

-0.002*

-0.002**

-0.002*

(0.001)

Distance

-0.055***

-0.064***

-0.062***

-0.063***

-0.062***

(0.008)

(0.009)

(0.010)

Affinity index

0.469*

0.387

0.356

0.388

0.360

(0.271)

(0.296)

(0.298)

Common religion

-0.007

0.003

-0.001

0.001

0.008

(0.064)

(0.075)

(0.076)

(0.075)

(0.076)

Common language

0.165**

0.049

0.036

0.066

0.015

(0.079)

(0.105)

(0.104)

(0.105)

Former colony

-0.222***

-0.257***

-0.271***

-0.258***

-0.236**

(0.084)

(0.093)

(0.092)

(0.093)

(0.094)

Ln(GDP p.c.)

donor

-0.159***

-0.150***

-0.160***

-0.156***

-0.173***

(0.048)

(0.051)

Trade with donor

0.124***

0.119***

0.142***

0.120***

0.188***

(0.014)

(0.015)

(0.017)

(0.015)

(0.023)

Regulatory quality

0.027

-0.184*

(0.069)

(0.106)

Trade with donor

0.055***

Regulatory quality

(0.021)

Corruption control

0.102

-0.398**

(0.088)

(0.158)

Trade with donor

0.117***

Corruption control

(0.028)

Disaster dummies

Yes

0.973***

0.970***

0.950***

0.964***

0.918***

(0.130)

(0.132)

(0.135)

(0.140)

Constant

3.840***

3.735***

4.071***

3.957***

4.052***

(0.750)

(0.803)

(0.811)

(0.843)

(0.852)

Log likelihood

-7546.738

-6398.397

-6394.050

-6386.357

-6356.830

3284

2757

Table 8: Aid & Bilateral TRADE – 1

stage

)

(

aid

(1)

(2)

(3)

(4)

(5)

Ln(Fatalities)

0.135***

0.128*** 0.128*** 0.128***

0.136***

(0.008)

Ln(Affected)

0.062***

0.055*** 0.055*** 0.060***

0.057***

(0.003)

(0.004)

Ln(GDP p.c.)

-0.197***

-0.205*** -0.205*** -0.149***

-0.161***

(0.017)

(0.021)

(0.022)

Ln(Population)

-0.399***

-0.372*** -0.372*** -0.374***

-0.384***

(0.015)

(0.016)

Trade (% of GDP)

-0.005***

-0.004*** -0.004*** -0.004***

-0.004***

(0.001)

Distance

-0.050***

-0.053*** -0.052*** -0.052***

-0.053***

(0.004)

Affinity index

-0.661***

-0.647*** -0.649*** -0.645***

-0.698***

(0.092)

(0.101)

(0.102)

Common religion

0.130***

0.100*** 0.100*** 0.103***

0.112***

(0.034)

(0.038)

(0.039)

Common language

0.216***

0.212*** 0.211*** 0.216***

0.190***

(0.048)

(0.057)

(0.056)

(0.057)

Former colony

-0.227***

-0.226*** -0.225*** -0.226***

-0.208***

(0.051)

(0.054)

Ln(GDP p.c.)

donor

0.514***

0.516*** 0.516*** 0.516***

0.515***

(0.017)

(0.018)

Pop. (in mio.)

donor

0.001***

0.001*** 0.001*** 0.001***

0.001***

(0.000)

Trade with donor

0.136***

0.126*** 0.127*** 0.127***

0.176***

(0.008)

(0.009)

(0.010)

(0.009)

(0.011)

Regulatory quality

0.018

0.001

(0.035)

(0.049)

Trade with donor

0.005

Regulatory quality

(0.010)

Corruption control

-0.191***

-0.515***

(0.036)

(0.052)

Trade with donor

0.085***

Corruption control

(0.010)

Disaster dummies

Yes

Constant

1.209***

0.937** 0.935** 0.397

0.445

(0.376)

(0.422)

(0.418)

(0.421)

26585

22259

Table 9: Aid & bilateral OIL -CASH - 2

stage

)

(

cash

(1)

(2)

(3)

(4)

(5)

Ln(Fatalities)

-0.060**

-0.042

-0.055*

-0.044

-0.043*

(0.023)

(0.027)

(0.028)

(0.027)

(0.026)

Ln(Affected)

0.014

0.008

0.001

-0.001

0.001

(0.009)

(0.010)

(0.011)

(0.010)

Ln(GDP p.c.)

-0.094*

-0.152**

-0.220***

-0.189***

-0.184**

(0.053)

(0.064)

(0.079)

(0.072)

Ln(Population)

0.045

0.028

0.027

0.030

0.031

(0.041)

(0.046)

(0.045)

(0.046)

Trade (% of GDP)

0.003**

0.002

0.003

0.003*

(0.001)

(0.002)

Distance

0.011

-0.017

-0.014

-0.017

-0.016

(0.013)

(0.015)

Affinity index

-0.129

-0.215

-0.243

-0.191

-0.183

(0.302)

(0.337)

(0.334)

(0.333)

Common religion

0.199**

0.175

0.170

0.149

0.147

(0.092)

(0.109)

(0.108)

Common language

-0.153

-0.289*

-0.277

-0.272

-0.276

(0.120)

(0.168)

(0.169)

(0.170)

Former colony

-0.369***

-0.553***

-0.557***

-0.551***

-0.567***

(0.129)

(0.166)

(0.168)

(0.164)

Ln(GDP p.c.)

donor

-0.187***

-0.154**

-0.164**

-0.159**

(0.060)

(0.068)

(0.066)

Fuel exports (% of

-0.001

0.002

-0.004

0.001

-0.003

merchandise exports)

(0.002)

(0.004)

(0.002)

(0.005)

Regulatory quality

0.228*

0.390**

(0.119)

(0.157)

Fuel exports

-0.008

Regulatory quality

(0.005)

Corruption control

0.302**

0.341**

(0.140)

(0.146)

Fuel exports

-0.006

Corruption control

(0.007)

Disaster dummies

Yes

-0.102

-0.044

-0.057

-0.052

(0.141)

(0.158)

(0.154)

(0.153)

Constant

-0.170

0.448

1.257

0.994

0.901

(0.955)

(1.028)

(1.171)

(1.106)

(1.098)

Log likelihood

-6631.837

-5529.674

-5527.802

-5526.460

-5511.599

3158

2632

Notes: Probit estimates. Coefficients reported; robust standard errors in parentheses. Dependent variable is cash, a
dummy that switches to 1 if the bilateral aid contribution was cash. ***, **, * indicate significance at the 1, 5 and
10%-level, respectively.

Table 10: Aid & bilateral TRADE -CASH – 2

stage

)

(

cash

(1)

(2)

(3)

(4)

(5)

Ln(Fatalities)

-0.035

-0.032

-0.033

-0.039

(0.026)

(0.027)

(0.028)

(0.027)

(0.029)

Ln(Affected)

0.019**

0.008

0.007

0.003

-0.002

(0.010)

(0.011)

Ln(GDP p.c.)

-0.171***

-0.149**

-0.152**

-0.189***

-0.170**

(0.060)

(0.068)

(0.069)

(0.073)

(0.074)

Ln(Population)

-0.082

-0.051

-0.048

-0.030

(0.066)

(0.069)

(0.071)

(0.070)

(0.075)

Trade (% of GDP)

0.002

(0.001)

(0.002)

Distance

0.018

-0.008

-0.005

-0.007

-0.001

(0.012)

(0.014)

(0.015)

Affinity index

-0.304

-0.307

-0.324

-0.325

-0.321

(0.294)

(0.322)

(0.323)

(0.322)

(0.320)

Common religion

0.249***

0.206**

0.203**

0.201**

(0.086)

(0.101)

Common language

-0.134

-0.239

-0.253*

-0.224

-0.282*

(0.116)

(0.154)

(0.152)

(0.154)

(0.150)

Former colony

-0.431***

-0.543***

-0.549***

-0.544***

-0.516***

(0.125)

(0.153)

(0.154)

(0.156)

Ln(GDP p.c.)

donor

-0.173**

-0.182**

-0.177**

-0.216***

(0.070)

(0.071)

(0.073)

(0.071)

Trade with donor

0.078***

0.049

0.057*

0.049

0.091**

(0.029)

(0.031)

(0.032)

(0.031)

(0.041)

Regulatory quality

0.009

-0.100

(0.097)

(0.147)

Trade with donor

0.027

Regulatory quality

(0.029)

Corruption control

0.170

-0.266

(0.132)

(0.227)

Trade with donor

0.094**

Corruption control

(0.038)

Disaster dummies

Yes

0.052

-0.004

-0.026

-0.009

-0.134

(0.197)

(0.198)

(0.206)

(0.199)

(0.210)

Constant

1.650*

1.522

1.623

1.978*

1.920*

(0.980)

(1.053)

(1.058)

(1.106)

(1.100)

Log likelihood

-6908.274

-5846.983

-5846.317

-5784.745

-5784.063

3284

2757

Table 11: Aid & bilateral OIL – 2

stage - Robustness Test OECD vs. NON-OECD

)

(

bilateral

(1)

(2)

(3)

(4)

OECD

Ln(Fatalities)

-0.060***

-0.057***

-0.066***

-0.062***

(0.006)

(0.019)

(0.017)

Ln(Affected)

0.006

0.008

-0.012

-0.008

(0.007)

(0.008)

Ln(GDP p.c.)

-0.067

-0.053

-0.128*

-0.113

(0.064)

(0.078)

(0.069)

(0.070)

Fuel exports (% of

0.009***

0.011***

0.007***

-0.004

merchandise exports)

(0.002)

(0.004)

(0.002)

(0.005)

Regulatory quality

0.540***

0.504***

(0.119)

(0.157)

Fuel exports

0.002

Regulatory quality

(0.005)

Corruption control

0.550***

0.652***

(0.134)

(0.140)

Fuel exports

-0.017**

Corruption control

(0.004)

NON-OECD

Ln(Fatalities)

-0.088**

-0.100**

-0.107**

-0.102**

(0.046)

(0.051)

(0.046)

(0.044)

Ln(Affected)

0.033*

0.022

-0.002

0.001

(0.017)

(0.024)

(0.018)

(0.019)

Ln(GDP p.c.)

-0.064

-0.157

-0.337

-0.290

(0.161)

(0.089)

(0.208)

(0.197)

Fuel exports (%

-0.000

-0.007

-0.003

-0.026**

of merchandise exports)

(0.004)

(0.010)

(0.003)

(0.011)

Regulatory quality

0.631**

0.886*

(0.305)

(0.512)

Fuel exports

-0.009

Regulatory quality

(0.014)

Corruption control

0.901**

1.424***

(0.372)

(0.396)

Fuel exports

-0.035**

Corruption control

(0.017)

Notes: Probit estimates. Coefficients reported; robust standard errors in parentheses. Dependent
variable is bilateral, a dummy that switches to 1 if the aid flow was bilateral. ***, **, * indicate
significance at the 1, 5 and 10%-level, respectively.

Table 12: Aid & bilateral TRADE – 2

stage - Robustness Test OECD vs. NON-OECD

)

(

bilateral

(1)

(2)

(3)

(4)

OECD

Ln(Fatalities)

-0.072***

-0.070***

-0.074***

-0.070***

(0.016)

(0.017)

Ln(Affected)

0.006

0.007

0.001

-0.000

(0.006)

(0.007)

Ln(GDP p.c.)

-0.058

-0.070

-0.111*

-0.108*

(0.061)

(0.065)

Trade with donor

0.106***

0.135***

0.107***

0.153***

(0.021)

(0.024)

(0.021)

(0.030)

Regulatory quality

-0.057

-0.332**

(0.086)

(0.134)

Trade with donor

0.069**

Regulatory quality

(0.028)

Corruption control

0.234**

-0.148

(0.117)

(0.216)

Trade with donor

0.081**

Corruption control

(0.038)

NON-OECD

Ln(Fatalities)

-0.077**

-0.074*

-0.088**

-0.080*

(0.041)

(0.046)

(0.044)

Ln(Affected)

0.022

0.025

-0.013

-0.022

(0.015)

(0.016)

Ln(GDP p.c.)

-0.057

-0.062

-0.354*

-0.335*

(0.140)

(0.145)

(0.190)

(0.192)

Trade with donor

-0.022

0.027

-0.030

-0.113

(0.036)

(0.051)

(0.035)

(0.075)

Regulatory quality

0.432**

0.205

(0.194)

(0.278)

Trade with donor

0.082

Regulatory quality

(0.066)

Corruption control

1.053***

0.294

(0.355)

(0.464)

Trade with donor

0.196**

Corruption control

(0.088)

Notes: Probit estimates. Coefficients reported; robust standard errors in parentheses. Dependent
variable is bilateral, a dummy that switches to 1 if the aid flow was bilateral. ***, **, * indicate
significance at the 1, 5 and 10%-level, respectively.

Table 13: Aid & bilateral OIL - CASH- Robustness Test OECD vs. NON-OECD

)

(

cash

(1)

(2)

(3)

(4)

OECD

Ln(Fatalities)

-0.079***

-0.091***

-0.079***

(0.019)

(0.020)

(0.019)

Ln(Affected)

0.017**

0.009

0.012

(0.008)

(0.009)

Ln(GDP p.c.)

-0.170***

-0.244***

-0.181***

-0.172**

(0.063)

(0.077)

(0.067)

(0.068)

Fuel exports (% of

0.003

-0.004

0.001

-0.008*

merchandise exports)

(0.002)

(0.004)

(0.002)

(0.005)

Regulatory quality

0.299***

0.475***

(0.112)

(0.145)

Fuel exports

-0.010*

Regulatory quality

(0.005)

Corruption control

0.264**

0.347***

(0.124)

(0.128)

Fuel exports

-0.014**

Corruption control

(0.007)

NON-OECD

Ln(Fatalities)

-0.006

-0.023

-0.008

-0.005

(0.048)

(0.051)

(0.049)

(0.051)

Ln(Affected)

-0.000

-0.017

-0.013

-0.019

(0.017)

(0.026)

(0.023)

(0.024)

Ln(GDP p.c.)

0.161

0.012

-0.020

-0.008

(0.152)

(0.208)

(0.199)

(0.196)

Fuel exports (%

-0.010*

-0.019*

-0.007

0.005

of merchandise exports)

(0.005)

(0.011)

(0.005)

(0.025)

Regulatory quality

-0.125

0.173

(0.305)

(0.408)

Fuel exports

-0.013

Regulatory quality

(0.013)

Corruption control

0.480

0.360

(0.426)

(0.437)

Fuel exports

0.021

Corruption control

(0.017)

Notes: Probit estimates. Coefficients reported; robust standard errors in parentheses. Dependent
variable is cash, a dummy that switches to 1 if the bilateral aid contribution was cash.
***, **, * indicate significance at the 1, 5 and 10%-level, respectively.

Table 14: Aid & bilateral TRADE-CASH - Robustness Test OECD vs. NON-OECD

)

(

cash

(1)

(2)

(3)

(4)

OECD

Ln(Fatalities)

-0.065***

-0.064***

-0.060***

(0.018)

Ln(Affected)

0.016**

0.017**

0.013

0.013*

(0.008)

Ln(GDP p.c.)

-0.167***

-0.170***

-0.187***

-0.180***

(0.062)

(0.066)

Trade with donor

0.063***

0.068***

0.064***

0.105***

(0.023)

(0.024)

(0.023)

(0.027)

Regulatory quality

0.082

0.009

(0.098)

(0.172)

Trade with donor

0.015

Regulatory quality

(0.030)

Corruption control

0.143

-0.285

(0.115)

(0.215)

Trade with donor

0.077**

Corruption control

(0.032)

NON-OECD

Ln(Fatalities)

0.010

0.019

0.000

0.022

(0.042)

(0.043)

(0.044)

(0.048)

Ln(Affected)

-0.021

-0.019

-0.033

-0.047**

(0.019)

(0.020)

(0.022)

(0.024)

Ln(GDP p.c.)

0.129

0.081

-0.052

-0.065

(0.145)

(0.150)

(0.181)

(0.180)

Trade with donor

-0.050

0.010

-0.090

-0.080

(0.043)

(0.056)

(0.065)

(0.064)

Regulatory quality

-0.210

-0.554**

(0.231)

(0.270)

Trade with donor

0.130*

Regulatory quality

(0.073)

Corruption control

0.457

-0.161

(0.372)

(0.504)

Trade with donor

0.196**

Corruption control

(0.092)

Notes: Probit estimates. Coefficients reported; robust standard errors in parentheses. Dependent
variable is cash, a dummy that switches to 1 if the bilateral aid contribution was cash.
***, **, * indicate significance at the 1, 5 and 10%-level, respectively.