Fiscal Policy and Civil Conflict

Alvaro Aguirre∗

Central Bank of Chile

Cesar Calderon†

The World Bank

May 18, 2015

[PRELIMINARY DRAFT]

Abstract

We explore empirically the effect of fiscal policy responses to economic shocks on the like-lihood of conflict. Our main finding is that a countercyclical fiscal response to these shockslowers the likelihood of conflict, particularly in Africa and when considering shocks to the priceof mineral commodities as the triggers to macroeconomic cycles. These results are stronger whenconsidering more recent periods and only non-military spending. Although macroeconomic in-come shocks are behind conflict onset in our framework, the effect of the fiscal response seemsto be associated to mechanisms that go beyond macroeconomic stabilization.

∗aaguirre@bcentral.cl. Click here for the latest version of this paper.†ccalderon@worldbank.org

1

1 Introduction

The disastrous effects of civil wars and their prevalence in poor countries have spawned a research

agenda that tries to understand their occurrence and persistence.1 Two different theoretical per-

spectives frame this analysis (Rule, 1989). The first one emphasizes relative deprivation as the

motivation for political action. The second stresses that, regardless of grievances, conflict is deter-

mined by the ability to mobilize resources to create and sustain rebel groups. In this context of

greed versus grievances, as put forth by Collier and Hoeffler (2004), the role of income inequality,

lack of political rights, and ethnic and religious divisions, may be downplayed by geography and

external support.

Perhaps since it relates to both greed and grievances hypotheses, economic growth has been one

of the most studied determinants of civil conflict (See for instance Miguel et al., 2004; Hegre and

Sambanis, 2006; Bruckner and Ciccone, 2010; Dube and Vargas, 2013; Bazzi and Blattman, 2014).

Under the greed hypothesis it influences the opportunity costs of fighting (Chassang and Padro-

i-Miquel, 2009), the gains from state appropriation (Besley and Persson, 2011), and the capacity

of the state to bargain or fight insurgencies (Fearon and Laitin, 2003; Bazzi and Blattman, 2014).

In term of grievances, the larger availability of resources may be used to alleviate the problems of

marginalized groups, reducing the likelihood of conflict, or, alternatively, to increase it if resources

are appropriated by political elites.

But despite the abundant literature on the effect of economic shocks on conflict, the role of

stabilization policies has been neglected. The macroeconomic framework, particularly the conduct

of fiscal policy, may influence both the impact of shocks on economic growth, and the effects of

growth on conflict. The first, although studied in different contexts, is well known. The second has

not been explored. In the case of poor countries fiscal management may downplay the link between

growth and the income of the poor, providing social insurance and guaranteing the provision of

public goods and investment programs.2 Hence different macroeconomic policies could generate an

heterogeneous relationship between economic shocks and conflicts, both across countries and time.

The aim of this paper is to fill this gap, and explore the different strategies on the macroeconomic

front, and specifically fiscal management, to lower the vulnerability of countries to civil conflict.3 We

modify previous empirical work introducing interactions of growth shocks, specifically commodity

1See Sambanis (2002), Collier and Hoeffler (2007), and Blattman and Miguel (2010) for detailed reviews of this

literature.2Calderon and Levy-Yeyati (2009) show that the adverse effects of aggregate volatility on inequality and poverty

are mitigated by public expenditure. Their focus is on the size of the public sector, but ours is on the short-term

response of fiscal spending. This feature differentiates our paper from others who explore the relationship between

fiscal variables and conflict (Savun and Tirone, 2012; Singh et al., 2014). See the literature review for a comparative

analysis of the two approaches.3We focus on fiscal policy because it is more likely to influence the relationship between economic activity and

the risk of conflict. Additional macroeconomic policies, such as monetary and exchange rate policies, have a clearer

role smoothing shocks to output.

2

prices shocks, and indices capturing fiscal responses. Since these and other economic policies

depend on deeper, slow-moving country characteristics, that are also deep determinants of conflict

(i.e. political institutions), we focus on time, instead of cross-country variation, to identify the

effects of policies. We are aware however that, still in this framework, the effect of policies on

conflict is dependent on these features, and hence the improvement of policies is not likely to

change the long-run incidence of conflict in countries that do not improve institutions at the same

time. In our view this last step, although difficult, may be facilitated by temporarily reducing the

incidence of violent conflict through a better macroeconomic management.

We find a significant effect of commodity price shocks on conflict onset. This effect varies across

countries depending on institutional and geographical features, and it is stronger when considering

not only full blown civil conflicts but small armed conflicts as well. Fiscal spending growth does

not seem to affect the likelihood of conflict when not interacted with commodity price shocks. The

novel and main finding is that the sensitivity of violent conflict to economic shocks depends on

the fiscal response to them. Although we consider a large cross-section of countries, these results

are mostly found in the sample of African countries. This is the region with the highest incidence

of conflict, and the one where the growth-conflict nexus seems to be the strongest (Bruckner and

Ciccone, 2010; Bazzi and Blattman, 2014). The interaction between commodity price shocks and

the change in fiscal spending is positive and significant, meaning that a countercyclical response to

these shocks lowers the likelihood of conflict. The facts that more spending increases the likelihood

of conflict when facing a good price shock, and that when it is not interacted with shocks spending

does not affect conflict, reduce the likelihood that our results are capturing reverse causality from

conflict to spending. Results are mostly driven by minerals and persist when we differentiate

between positive and negative shocks. Although weaker, we also find some conditionality to fiscal

responses in the case of large civil wars. In this case minerals are also important, and increases in

spending, both after positive and negative shocks, reduce risks. When exploring possible changes

in the relationship between conflict, economic shocks, and fiscal policy we find that this hasn’t

been stable. The response of fiscal spending has been much more relevant since the 1990s. Finally,

using a smaller sample, we confirm that our results are not driven by changes in military spending.

When considering this type of spending we find a positive relationship with conflict risk, probably

due to reverse causality.

Additional work is needed to understand the mechanisms behind our findings. What we find

when considering the effect of our explanatory variables on growth is that, although the uncondi-

tional effect of commodity shocks is significant, validating growth as the main channel in that case,

the response of fiscal policy to shocks is not. Hence results suggest a role for fiscal policy beyond

its effects on macro stability, which may be related to social insurance or productive investments.

After a brief literature review, section 2 presents the empirical strategy and describes the data.

Section 3 shows the econometric results and section 4 concludes.

3

1.1 Literature Review

There is a large empirical literature on the effects of growth on civil conflict. Hegre and Sam-

banis (2006), after an extensive empirical exploration of the literature on civil war determinants,

conclude that the relationship between growth and conflict is particularly robust. More recently

the literature has focused on finding a causal relationship between growth and civil war trough

the use of instrumental variables and panel regressions. Miguel et al. (2004) use rainfall growth

as an instrument for economic stagnation to explain, successfully, the onset of civil wars in sub-

Saharan Africa. Bruckner and Ciccone (2010) use variations in international commodity prices

as the exogenous determinant. They find that, between 1981 and 2006, a 20% drop in countries

export price indices raised the probability of civil war outbreak by around 2.8 percentage points

in Sub-Saharan Africa. They show that commodity prices affect the likelihood of conflict because

of their effect on economic growth. Bazzi and Blattman (2014) find weaker results for a larger

sample of countries, but they still find a significant effect of commodity price shocks, particularly

on the intensity of fighting. Introducing policies may help to explain the apparent instability in

the relationship between shocks and conflict.4 Dube and Vargas (2013) focus on Colombia and find

differential effects depending on the type of commodities, with those easily-captured commodities

raising violence, and the rest, basically agricultural and labor intensive commodities, lowering it.

Our empirical framework is based on this papers. We use commodity price shocks as an exogenous

determinant of growth to see how stabilization policies may reduce the vulnerability to conflict,

and we explore heterogenous effects depending on the type of commodities.

The role of macroeconomic policies intermediating the link between shocks and conflict has not

been studied before. Singh et al. (2014) explore if public spending has a mitigating effect on conflict.

But they study a different dimension of fiscal policy. They focus on how differences across countries

in terms of the size of public spending to GDP, and its allocation, affect the incidence of conflict in

natural-resource abundant countries. We focus on variations over time of total spending growth,

without considering cross-section variability. Hence our results are associated to the short-run

response of fiscal policy to shocks (i.e. fiscal policy cyclicality), while theirs is associated to state

capacities and conflict (on this see also Fjelde and De Soysa, 2009; Thies, 2010). Moreover they

focus on the link between natural resources and conflict –oil reserves are the main driver of conflict

in their framework–, while we focus on the link between growth and conflict, and use commodity

prices only as an exogenous determinant of growth following the literature just described. Relatedly,

Savun and Tirone (2012) explore the interaction of shocks with foreign aid in generating civil wars,

but they results are influenced by cross-section differences as well, they don’t explore how fiscal

4For instance the relationship between rainfall shocks and civil conflict found by Miguel et al. (2004) for the period

before 2000 in sub-Saharan Africa disappears afterwards (Ciccone, 2011), apparently due to the weakening of it effect

on GDP growth (Miguel and Satyanath, 2011). Similarly the different strength in the findings by Bruckner and

Ciccone (2010) and Bazzi and Blattman (2014) are in part due to different samples of countries and periods.

4

policy responds to shocks, and their focus is on agricultural commodities only.5

2 Empirical Strategy

We extend Bruckner and Ciccone (2010) framework to investigate the effect of fiscal policy on the

impact of commodity prices on conflict onset. Our baseline specification, for county i and year t,

is the following,

coi,t = αi + δt + β0dlnpi,t + β1dlnpi,t−1:2 + β2dlngi,t−1:2 + β3

(dlnpi,t−1:2 × dlngi,t−1:2

)+ εi,t

where αi and δt are fixed and time effects respectively, and the subindex t − 1 : 2 denotes the

two-years average lag of the corresponding variable.

The dependent variable is coi,t, a conflict onset dummy. It takes a value of 1 if there is conflict

in period t but not in period t− 1, and a value of 0 if there is no conflict in period t. Otherwise, if

in both periods t and t− 1 there is conflict, the country-year pair for t is dropped from the sample.

Hence the specification includes all peace years and (the first year of) onset. Neither persistence

nor conflict ending are included. We do this because violent conflict constraints macroeconomic

management, the focus of our study, and because during years of conflict data on fiscal spending

is less reliable.6

The explanatory variables are the percentage change in commodity prices, dlnp, and in public

spending, dlng. For the last variable we use its lagged value to reduce the likelihood of capturing

reverse causality. Additionally we use the two-years average to minimize the effects of measurement

problems. Hence we split the effect of commodity prices in a contemporary effect, and in a two-

years lagged effect. Although we drop from the sample large commodity exporters, separating

contemporary and lagged effects of commodity shocks also helps in minimizing the effect of reverse

causality in our estimations.7 We explore heterogenous effects decomposing the commodity price

index by type of commodity, and by the sign of its change. We also differentiate public spending

between military and non-military spending, although in a smaller sample.

Behind the vulnerability to conflicts there are deep determinants, mainly associated with ge-

ography and political institutions. These, on the other hand, may be important determinants of

macroeconomic policies. Since most of the variation of these deep determinants is observed across

countries, we exploit only time variation in our specification to try to isolate the effect of macro

5Our paper is also related to the literature that studies the effects of fiscal consolidation and social unrest (See

for instance Morrisson et al., 1994; Haggard et al., 1995; Ponticelli and Voth, 2011; Voth, 2013). These papers focus

on fiscal adjustments and not on the response of fiscal policy to shocks, as we do.6Bazzi and Blattman (2014) implement this strategy to avoid the bias when the dynamics of the dependent variable

is ignored. They also include a separate regression for episodes when conflicts end.7African countries dropped from the sample are Ivory Coast (35% of cocoabeans’ World exports) and Senegal

(47% of ground nut oil).

5

policies on conflict onset. In addition to use the change in spending instead of its level as the ex-

planatory variable, which allows us to focus on the stabilization role of fiscal policy, we adjust the

variation in public spending by country means, so the interaction does not capture cross-section dif-

ferences in public spending behavior. Additionally we control for interactions between commodity

price shocks and variables capturing deeper determinants of conflict and policies.

Note that the elasticity of conflict onset to commodity price shocks in our framework is defined

as,∂coi,t

∂dlnpi,t−1:2= β1 + β3dlngi,t−1:2.

We expect from the results in previous papers that β1 < 0. Under a positive value for β3 a

procyclical fiscal policy (i.e. positively correlated with the shock) lowers the drop in the probability

of conflict after a positive price shock, and increases the raise in the probability when facing a

negative price shock. A countercyclical policy has the opposite effects and hence it would be the

type of policy that minimizes the probability of conflict onset under β3 > 0.

2.1 Data

We collect data on civil conflict, commodity prices, and public spending for 105 countries, of which

42 are from Africa, from 1960 to 2013. The list of the countries included is presented in Table 1.8

In the following paragraph we describe the sources and present basic statistics.

Civil Conflict: The source is the UCDP/PRIO Armed Conflict Dataset (Harbom et al., 2008).

UCDP/PRIO defines armed conflict as “a contested incompatibility that concerns government

and/or territory where the use of armed force between two parties, of which at least one is the

government of a state, results in at least 25 battle-related deaths in a year.” In particular internal

armed conflict occurs between the government of a state and one or more internal opposition

group(s) regardless of intervention from other states. This data allows us to distinguish between

small and large conflicts, an exercise we implement below. In particular we distinguish between

civil conflicts, an episode with more than 25 battle-related deaths in a year, and civil wars, where

the threshold raises to 100 deaths.

Table 2 describes the conflict data in our sample of 105 countries. We show the probability of

conflict onset conditional on observing peace, which is our dependent variable, and the uncondi-

tional probability of conflict onset (i.e. considering both peace and conflict years in the denom-

inator). The first are 4.3% and 1.1% for civil conflicts and wars respectively. These are almost

8For most of our specifications we present results for three groups of countries; African countries (42), which

include both North and Sub-Saharan Africa; Developing Countries (89), which include the group of African countries

and the other developing countries listed in Table 1; and World (105), including the group of developing countries

and the group of industrial countries listed in Table 1. We also consider a smaller sample of African countries that

are resource rich. They are marked with an asterisk in Table 1. The sample is obtained after dropping observations

for which data on the variables considered is not available, dropping countries with less than 5 observations, and, as

mentioned above, dropping large exporters (see footnote 7).

6

completely concentrated in developing countries, particularly in Africa where the probabilities rise

to 5.8% and 1.6%, respectively. Among these countries, those that are resource rich are signifi-

cantly more prone to conflict, as these probabilities raise to 8.4% and 2.4%, respectively. When

considering the unconditional probability conflict onset occurs in 2.7% and 0.9% of our sample in

the case of civil conflicts and wars, respectively. Table 2 also shows how prevalent is conflict in

the countries analyzed. While a 12% of the time the average country is in conflict, a 2.5% is in

civil war. These numbers rise to more than 14% and 3.5% in the case of African countries, and to

18% and 5% when taking into account only resource-rich economies. It is clear from these numbers

that civil wars are rare events, and perhaps because of this it is more difficult to find systematic

determinants of them, as our results will show.

Figure 1 shows the temporal pattern of conflict. The left-hand side panel shows conflict preva-

lence, for each type of conflict, and for Africa and the rest of the world separately. In the last

group we can see an inverse U-shaped prevalence of conflict, with a peak around 1990. For Africa,

although weaker, we observe a similar shape but with a different peak, closer to the year 2000. In

the right-hand side panel we show conflict onset, which doesn’t seem to show a pattern as clear as

in the case of prevalence.

Commodity prices: We construct a commodity price index following the methodology proposed

by Collier and Goderis (2012), which is based on Deaton and Miller (1995) and Dehn (2000).

We collect yearly commodity price indices for 50 different commodities, which is as many as data

availability allows. For each country we first construct a weighted average of commodity prices,

considering only commodities in which the country is a net exporter. The weight for each commodity

is its participation on total net exports in 1990 of the commodities considered.9 Additionally we

construct subindexes for minerals, agricultural commodities, and oil.10

Table 3 shows summary statistics of the commodity indexes. Panel A presents statistics for the

aggregate index, and panels B, C, and D for minerals, agricultural, and oil subindexes, respectively.

We denote by wi the fraction of net commodity exports in GDP in the base year, which is 1990.

In our sample of 42 African countries the average fraction of net commodity exports is close to

10%.11 A similar number is observed in other developing countries, but not in developed countries,

which show a smaller fraction. In terms of volatility the standard deviations are similar across

groups, with Africa showing a slightly higher number. In the case of subindexes most countries

9As in Collier and Goderis (2012), prices are deflated by an annual world average of export unit values. We also

followed the adjustments proposed by these authors to treat gaps and missing values. See their paper for details.10The 50 commodities are aluminium, natural gas, phosphatrock, uranium, urea, coal, ironore, nickel, silver,

copper, lead, tin, and zinc, which we classify as minerals; bananas, cotton, oliveoil, pulp, sugar, barley, fish, oranges,

rice, sunfloweroil, butter, groundnutoil, palmkerneloil, rubber, swinemeat, cocoabeans, groundnuts, palmoil, sisal,

tea, coconutoil, hides, pepper, sorghum, tobacco, coffee, jutes, plywood, soybeanoil, wheat, copra, maize, poultry,

soybeans, and wool, classified as agricultural commodities; and gasoline and oil, which we classify as oil.11We don’t show the medians to save space but they are generally lower than the means. In the case of Africa the

median wi is 5.4%.

7

have positive net agricultural exports, while in the case of minerals and oil only two thirds and

one sixth show positive numbers. This last two fractions are lower in Africa, although conditioning

on being positive, the mean and median values are in general higher than in the other groups.

Volatilities are similar across groups and subindexes, although the price of commodities related to

oil seems to be more volatile when considering developing countries in and out of Africa.

Public Spending: The third main variable is the change in public spending. We use data from

the World Bank’s WDI. We obtain a measure of real public spending multiplying GDP in LCU and

the public spending to GDP ratio. This increases the sample size considerably relative to using

public spending in LCU, and the correlation with this last variable is close to 1. The data shows

extreme observations so we drop the highest and lowest 2.5% changes in spending of the entire

sample.12 On average spending grows at a rate of 4.3% per year in our sample, with developed

countries showing a slightly lower rate, around 3.9%. The standard deviation is much higher

in developing countries. In Africa the average standard deviation is 10.8%, while in developing

countries out of Africa it is 8.6%. We also consider differences between military and non-military

public spending in our regressions. The World Bank’s WDI publishes data on military spending as

a fraction of GDP only since 1988, so the sample size is reduced importantly. The average growth

rate of military spending in the 87 countries for which we have data is 2.4%, and the standard

deviation is 13.7%. The highest numbers are observed in Africa, with an average growth of 3.1%,

and a standard deviation of 19%.

Before showing regression results we explore the relationship between our three main variables of

interest. To do this we first identify all the episodes of conflict onset for which there was no conflict

in the previous three years. Then we compute the median value of the changes in commodity prices

and public spending across all the episodes, before and after the onset of a conflict. Figure 2 shows

the results, both for African countries and for the rest of the world. It also presents the median

values for the two variables of interest for all the years and countries in the sample. On the left-

hand side panel civil conflict onset episodes are considered. The clearest feature is the large drop,

from a relatively high value, in commodity prices during the two years before the onset of conflict

in African countries. This is not observed in the rest of the countries, where the drop occurs in

the first year after the onset. We also find a more pronounced change in fiscal spending in Africa,

which growth rate starts rising two years before the onset, from a null change to a rate close to 6%

in the year the conflict begins. In the case of civil wars we observe a similar pattern, but with a

relatively smaller change in commodity prices and a relatively larger fall in fiscal spending growth.

12The fact that we use two-years average also helps with this problem. The standard deviation of the percentage

change in public spending is reduced from 8.8% to 6.6% after averaging it for two years.

8

3 Regression Results

In this section we present the regression results. Before considering public spending we first explore

in detail the relationship between commodity price changes and conflict onset. Table 4 shows the

results. The first 9 columns show the results using civil conflicts as the dependent variable, and the

remaining columns use civil wars. First we restrict the sample to African countries. In column (1)

we include the contemporary change in the commodity index and its two lags. While negative in all

cases, only the first lag is significant. We include the two-lags average in column (2), our preferred

specification for the reasons indicated above, and this is significant at the 10% confidence level.

When we include the average of the three years in column (3) the coefficient becomes not significant.

We repeat the same for developing countries (including Africa), and for the entire sample. In none

of the cases we find a significant effect of commodity prices. Something similar occurs in the case

of civil wars (columns 10-18), although in this case commodity prices are not significant in Africa,

and have a positive, and sometimes significant effect, when adding other developing countries.

As already mentioned we control for interactions of commodity prices with deeper determinants

of conflict and policies. After testing different variables we selected the ones that were more signif-

icant. One is associated to geography, the wheat-sugar ratio, which is a measure of land suitability

for wheat versus sugarcane, used by Easterly (2007) as an instrument for income inequality. The

second is associated to political institutions, the index of executive constraints from the Polity IV

database, for which we use its average value from 1960 to 2013.13 Results are shown in Table 5,

where these interactions are included in our preferred specification with the average of two lags of

commodity prices. Both the wheat-sugar ratio and executive constraints are evaluated at the mean

of each group. The interactions between lagged commodity prices and both of these variables are

significant for all the groups of countries in the case of civil conflicts (left panel). This means that a

higher wheat-sugar ratio is associated with a higher sensitivity of civil conflict onset to commodity

price changes. The opposite is observed in the case of average executive constraints, meaning that

countries with better institutions are less sensitive to commodity price shocks. In the case of civil

wars (right panel), only the interaction with executive constraints is significant, and the sign is

negative. This means that better institutions make the conflict elasticity to price shocks higher.

This may be explained by a possible nonmonotonicity in the relationship between institutions and

conflict (Hegre et al., 2001). These results show that the relationship between commodity prices

and civil conflict is heterogenous across countries. But it is important to notice that since we

exploit only time variation in spending our results are not influenced by these interactions, as these

variables are fixed over time.14 However, when interpreting the results, magnitudes are valid for

13Additionally we test the significance of mountainous terrain, ethnolinguistic fractionalization, polarization, and

the Gini-Greenberg index proposed by Esteban et al. (2012), the share of agricultural land occupied by family farms

in 1858, constructed by Vanhanen (2003), a measure of rain variability, the ICRG index, average urbanization and

population density, the fraction of agricultural land and fertile soil, latitude, and a dummy for tropical climate.14We try time-varying variables as well, but none of them where significant. These were the time-varying component

9

the average country in each group, since we evaluate the interaction with fixed effects at that level.

3.1 Baseline Results

Table 6 shows the results from estimating the baseline specification, for each type of conflict and for

African, developing (including Africa), and all the countries separately. We show results with and

without public spending using our preferred specification, which is separating the contemporary

effect from the average two-years lagged effects.

As already shown in the case of civil conflicts and Africa we find a significant effect of commodity

prices (column 1). In column (2) we include public spending, and although it is not significant by

itself, its interaction with commodity prices is. To have a better idea of the magnitudes involved

we adjust the change in public spending by its mean for African countries, which is 4,4%. Hence,

for a country with that level of spending growth, the coefficient rises from -0.09 to -0.15, and

becomes significant at a 1% level of significance. The interaction effect means that the likelihood

of civil conflict becomes more sensitive to changes in commodity prices as the growth rate of public

spending goes down. To see this we plot in Figure 3 the effect on the probability of civil conflict to

an average 10% decrease in commodity prices for two years, as a function of public spending growth.

The unconditional elasticity, reported in column 1 of Table 6, is very close to the elasticity estimated

by Bruckner and Ciccone (2010) (in green), if we assume in this last case that the average drop of

10% occurs for three years. In black it is shown the unconditional elasticity of civil conflict onset

in the region, which is close to 6% as reported in Table 2. The estimated conditional elasticity (in

red) means that an increase in spending larger than 10%, which is less than one standard deviation

above the mean for African countries, eliminates the effect from the drop in commodity prices.

It also implies that negative growth rates in spending make the response significantly larger than

the unconditional response. For instance, if public spending remains constant, the increase in the

probability of conflict onset is twice as large than in the unconditional case, and one-third of the

unconditional likelihood.

Going back to Table 6 we can see that these effects are only present in the case of Africa. When

we include other developing countries in columns (3) and (4), and developed countries in columns

(5) and (6), we don’t find any significant coefficient. The same happens when we consider civil

wars for every group of countries in columns (7) to (12).

To further illustrate the implications of our baseline estimations we explore actual events of

large drops in the price of commodities in Africa. In particular we compute the change in the

predicted probability of conflict using the coefficients in column 2 of Table 6 when, instead of using

the observed change in spending, we use the sample average to evaluate this predicted probability.

Table 7 shows the results. We select episodes where the commodity index fell more than 10%, and

pick the 10% with the largest increase in the predicted probability, and the 10% with the largest

of executive constraints, democracy (both from Polity), the ICRG index, urbanization, and population density.

10

reduction. In the upper panel are presented the episodes where an expansionary response of fiscal

policy generated the largest reduction in the probability of conflict onset. The table shows the

country and the year of the episode, the change in the commodity price index and in government

spending, and the change in the fitted value of the probability of conflict (dco). Despite the large

drops in commodity prices, in none of these episodes a conflict actually started. Botswana shows

the three largest episodes, with reductions in the predicted likelihood of conflict larger than 7%.

Most of these episodes are concentrated in the 1970s and 1980s. The latest events are Nigeria in

1999 and Mali in 2000, when drops of roughly 15% in the commodity indexes were accompanied by

increases of more than 20% in fiscal spending, implying reductions in the likelihood of conflict of 4%

in both cases. In the lower panel we show the largest increases in the predicted probability. In two

out of these 19 episodes a conflict actually started; Chad in 1976 and Sudan in 1983. According to

our estimates the fiscal contraction, of 2.8% and 16.4%, increased the probability of conflict onset

in 2.5 and 3.6% respectively. The latest event is Botswana 2005, when a 5.2% fall in fiscal spending

increased the fitted probability in 2.4%.

3.2 Heterogeneous Effects

Not all commodities may have the same effect on conflict risk. In Table 8 we explore heterogene-

ity across different commodity types for our sample of African economies, again considering civil

conflicts and wars. For comparisons, in columns (1) and (5) we present the baseline results, which

correspond to those in columns (2) and (8) of Table 6. In columns (2) to (4) we split the commodity

index in agricultural, minerals, and oil subindexes, and show the results using civil conflict onset

as the dependent variable. We include all of the subindexes in a single regression, so the results

in columns (2) to (4) are estimated jointly. Although the signs of the variables of interest are the

same than in column (1), significance is only found in the case of minerals (column 3). Although

agricultural commodities show a similar pattern, it seems that minerals are the ones generating

the non-linear effect of commodity prices on conflict. In the case of civil wars (columns 6-8), the

coefficients are not significant for every subindex. For larger samples, which we don’t show to save

space, results are similar. In developing countries and civil conflicts the interaction of minerals and

spending is significant, although close to one half of the coefficient found for Africa.

Results so far indicate that public spending reduces the sensitivity of conflict onset to commodity

price shocks, particularly in the case of minerals. Note that, although we interpret the results

considering a drop in commodity prices, the same would happen in the case of a positive shock.

Higher spending growth would reduce the fall in the likelihood of conflict in this case. Then, a

countercyclical fiscal policy during booms would be desirable not only to generate resources to

spend in bad times, but also because of a direct effect in the risk of conflict. A lower rate of

spending growth may make the country less vulnerable to other type of shocks or saving resources

may reduce the incentives to fight over them. To see if spending acts in this way both after a fall

11

and rise in prices, we split commodity shocks by their sign. Results are shown in Table 9. In column

(1) we consider the index including all of the commodities. The interaction is positive for positive

as well as negative shocks, although the coefficient is larger, and significant, in the first case. To

better understand the result we plot in the left-hand side panel of Figure 4 the response to a 10%

positive and negative shock, with the corresponding 10% confidence intervals. Although the slope

is small in the case of negative shocks, the prediction on the desirability of a countercyclical fiscal

policy is validated: to reduce the likelihood of a civil conflict not only a country need to spend

more when facing a bad shock, but even more efficiently, it has to spend less when receiving a good

shock.

To see if this depends on the type of commodity we present the results from estimating a

single equation including the three commodity prices subindexes in columns (2) to (4) of Table 9.

Significance is against observed for minerals, but unlike in column (1), the effect is stronger in the

case of a negative shock. In the center panel of Figure 4 we present the corresponding elasticities

and confidence intervals. Hence, although not significant, the difference in terms of magnitudes

obtained in column (1) seems to come from agricultural goods, for which the interaction in the case

of positive shocks is larger.

On the right-hand side panel of Table 9 we show results considering civil wars. Unlike in the

other cases this variable is used, when differentiating between negative and positive shocks we do

find significant coefficients. And again these correspond to minerals. In the case of positive shocks

the sign is the opposite than in the case of civil conflicts. As shown in the right-hand side panel of

Figure 4, higher spending lowers the likelihood of conflict, in a statistically significant magnitude,

when facing both good and bad shocks. It is worth to notice that the coefficient on public spending,

not shown in Table 9 to save space, is not significant. Note also that the change in sign is obtained

for agricultural goods as well when moving from civil conflicts to wars (columns 2 and 6). Finally,

when we include other developing countries results are very similar, and even the interaction with

positive shocks in column (6) becomes significant at the 5% significance level. Although we cannot

test the mechanism behind this change with respect to smaller conflicts with our simple framework,

it support the idea that political elites need to signal that they won’t steal the resources (see e.g.

Alesina et al., 2008, for a related mechanism).

3.3 Resource-Rich Countries

Another dimension in which heterogeneity may play a role is the dependency of the country in

natural resource rents. To explore this possibility we split the sample of African countries into

those where rents from natural resources are above and below 10% of GDP according to the WDI.15

15We use natural resource rents excluding forests. We include Botswana, Liberia, Namibia, Sierra Leone, and

Niger in the group of resource-rich countries despite having rents below the threshold, because of their possession of

diamonds (the first three), and uranium (Niger).

12

These are the countries marked with an asterisk in Table 1. Note that we still drop big exporters of

a particular commodity, to reduce the likelihood of measuring causality from conflict to commodity

prices. Results for civil conflicts are presented in Table ??. For comparisons, the first three columns

show the results for the whole group of African countries. The three columns in the middle present

the results when only the 19 resource-rich countries are included, and the last three columns do

that for the rest of the African countries in the sample. First we can see that the unconditional

effect is larger in resource-rich countries (columns 4 and 7). The interaction, introduced in columns

5 and 8, is also greater in the first group. It is not significant though, probably because of the

smaller size of the sample. When differentiating between positive and negative shocks we observe

significant differences between the two groups. While for resource-rich countries the interaction is

positive and significant in the case of positive shocks only (column 6), it is positive and significant

only in the case of negative shocks for countries that are not rich in natural resources. Therefore

the results imply that resource-rich countries reduce the probability of civil conflict after reducing

spending facing positive shocks, while non-resource-rich countries do the same increasing spending

when facing negative shocks.16

3.4 Temporal Changes

As shown in Figure 1 the pattern of conflict incidence in Africa since 1960 differs from the one

observed in the rest of the developing world. To see if the nature of the relationship between

conflict, commodity prices, and fiscal policy has changed over time as well we estimate the baseline

regressions for Africa, those in columns (2) and (8) of Table 6, in windows of 20 years, from 1960

to 2013. The resulting series for the interaction between commodity prices and fiscal spending, and

its 10% confidence interval, are presented in Figure 5. On the left-hand side panel we can see that

the interaction is more or less stable until de mid eighties, when the point estimate goes from 1.15,

a similar level than the one estimated for the entire sample, to close to 4 in the samples starting

around 1990. In the case of civil wars, shown in the right-hand side panel, we find an increase in

the coefficient as well, although sooner (around 1985), and the coefficient remains not significant

as in the entire sample.

To explore more in detail the change in the interaction for civil conflicts we present in Table 11

the results from estimating our baseline specifications, including differences by signs of the shocks,

for the years before and after 1990. The first thing to notice is that the unconditional effect of

commodity prices (columns 1 and 4) drops to one half and becomes not significant in the second

period. Then, as expected from the rolling regressions just analyzed, the interaction term with

16We don’t present results for civil wars to save space. Results in this case show that the unconditional effect

of commodity prices becomes significant in both groups. The sign differs though. Commodity price shocks have a

positive effect on the probability of civil war onset in the case of resource-rich countries, but a negative effect, in line

with the result for civil conflicts, in resource-poor countries.

13

public spending has the opposite variation. It rises from 0.85 to 3.05 (columns 2 and 5). This

means that the lower sensitivity to commodity prices is due to the fiscal response to shocks. To

better understand these results in columns (3) and (6) we differentiate the interaction term by the

sign of the shock. We can see that before 1990 a positive spending response to a drop in commodity

prices could reduce their effect on conflict, while a response to an increase in prices didn’t change

the response of conflict probability. After 1990 the pattern is the opposite. A negative spending

response to a positive shock make even less likely the onset of a conflict, while in the case of

a negative price shock spending does not have a significant effect moderating the increase in the

probability. Splitting the commodity index by type of commodity increases importantly the number

of regressors, and since the sample size is smaller, results are very unprecise. Because of this we

don’t show the results, which seem to indicate that most of the changes come from the effects of

agricultural commodities.

3.5 Military Spending

Results may be influenced by military spending. If that were the case then the mechanisms behind

our results would not be related to the stabilization or insurance roles of fiscal policy. Moreover it

would be more likely that the results were capturing reverse causality. To explore this possibility

we differentiate between military and non-military spending. However, as already mentioned, data

on military spending is available for a shorter period of time. The WDI publishes data on this

category only from 1988 onwards, which roughly coincides with the first year of the second sub-

sample considered in the previous sub-section. Regression results for the three groups of countries,

considering separately military and non-military spending, are presented in Table ??. For each

group the first column present the baseline results for comparisons, using total spending as before,

but adjusting the sample to countries and periods where military spending is available. Column

(1) shows that the interaction is still positive in the case of Africa, and similar to the coefficient

estimated for the second sub-sample in Table 11. However, due to the smaller sample, the coefficient

is not significant at the 10% confidence level. In column (2) we split public spending into military

(M) and non-military (NM) spending. The coefficient on non-military spending is positive, large,

and highly significant. Military spending on the other hand is negative and not significant. This

would imply that it is non-military spending which is behind the results obtained so far in the

paper. In the left-hand side panel of Figure 6 we show the corresponding elasticity and confidence

interval for the case of a drop in commodity prices.

In columns (3) and (4) we explore heterogeneous effects with respect to the sign of the shock.

Again for comparisons, column (3) shows the results with total spending. Although results are still

similar to those for the second sub-sample in Table 11, the coefficients are still not significant. But

in column (4), where we distinguish again by the type of spending, results are significant for each

type of spending, and for each sign of the shock. In the case of non-military spending results are

14

in line with what we have found in the previous specifications; a countercyclical fiscal policy lowers

the likelihood of conflict when facing a commodity shock, irrespectively of its sign. Results are even

stronger than the ones found in the second sub-sample using total spending (column 6 of Table 11).

The conditional elasticities and their respective confidence intervals are shown in the center panel

of Figure 6. Results for military spending differ. As shown in the right-hand side panel of Figure 6,

lower increases in this type of spending after commodity shocks are always associated with a lower

probability of conflict onset. Hence, it is very likely in this case causality is running from this last

variable to spending changes.

We repeat the estimations for the group of developing countries (columns 5 to 8) and for the

entire sample (columns 9 to 12). As before results become weaker. The exception is the effect of a

response of non-military spending to a positive price shock, which is large and significant in both

groups. The pattern is similar to what we found for Africa when considering the second sub-sample

(column 6 of Table 11). It implies that the response of conflict probability to negative price shocks

is unconditional on spending. However, in the case of a positive price shock, a countercyclical

response reduce the probability of conflict risk in a statistically significant magnitude.

3.6 Growth Effects

Bruckner and Ciccone (2010) show that the significant effects of commodity price shocks on conflict

onset are due to the effects on growth. To explore the same question but regarding the interaction

between these shocks and public spending we run the baseline regressions but using GDP growth

as the dependent variable. For comparisons we present again the results when civil conflict onset

is the dependent variable in columns (1) to (4). In column (5) we can see that the commodity

price index do have a significant effect on GDP growth, both contemporarily and with lags. This is

consistent with the results presented by Bruckner and Ciccone (2010). But note that in the case of

the interaction with public spending, unlike when explaining civil conflict in column (1), the effect

is not significant. When decomposing the index (columns 6 to 8) we can see that only in the case

of agricultural commodities the effect of commodity prices is significant, and more importantly for

our results, even in the case of minerals, the interactions are not significant. Hence it is likely

that the response of fiscal spending to commodity shocks influence the likelihood of conflict trough

channels that are not necessarily related to economic growth. The effect of fiscal policy on the

relationship between growth and conflict, and not between commodity shocks and growth, seems

to be the relevant dimension behind our results.

4 Conclusions

The results confirm the negative effect of commodity shocks on the likelihood of conflict onset

in Africa (for the average country), and economic growth as the main channel behind this. More

15

importantly for our analysis is that the effect is conditional on public spending: drops in commodity

prices that coincides with an increase in public spending, and hikes that coincides with low spending,

reduce the risk of conflict. And this doesn’t seem to be driven by military spending or the effect of

fiscal policy on growth.

16

References

Alesina, A., F. R. Campante, and G. Tabellini (2008): “Why Is Fiscal Policy Often Pro-

cyclical?,” Journal of the European Economic Association, 6(5), 1006–1036.

Bazzi, S., and C. Blattman (2014): “Economic shocks and conflict: the Evidence from com-

modity prices,” American Economic Journal: Macroeconomics, Forthcoming.

Besley, T. J., and T. Persson (2011): “The Logic of Political Violence,” Quarterly Journal of

Economics, 78(1), 1–34.

Blattman, C., and E. Miguel (2010): “Civil War,” Journal of Economic Literature, 48(1),

3–57.

Bruckner, M., and A. Ciccone (2010): “International Commodity Prices, Growth and the

Outbreak of Civil War in Sub-Saharan Africa*,” The Economic Journal, 120(544), 519–534.

Calderon, C., and E. Levy-Yeyati (2009): “Zooming in: From Aggregate Volatility to Income

Distribution,” World Bank Policy Research Working Paper Series 4895.

Chassang, S., and G. Padro-i-Miquel (2009): “Economic shocks and civil war,” Quarterly

Journal of Political Science, 4(3), 211–228.

Ciccone, A. (2011): “Economic shocks and civil conflict: A comment,” American Economic

Journal: Applied Economics, 3(4), 215–227.

Collier, P., and B. Goderis (2012): “Commodity prices and growth: An empirical investiga-

tion,” European Economic Review, 56(6), 1241–1260.

Collier, P., and A. Hoeffler (2004): “Greed and Grievance in Civil War,” Oxford Economic

Papers, 56(4), 563–595.

(2007): “Civil War,” Handbook of Defense Economics, 2, 711–739.

Deaton, A., and R. I. Miller (1995): International Commodity Prices, Macroeconomic Per-

formance, And Politics In Sub-Saharan Africa. International Finance Section, Department of

Economics, Princeton University.

Dehn, J. (2000): “Commodity Price Uncertainty In Developing Countries,” World Bank Policy

Research Working Paper N.2426.

Dube, O., and J. F. Vargas (2013): “Commodity price shocks and civil conflict: Evidence from

Colombia,” The Review of Economic Studies, 80(4), 1384–1421.

17

Easterly, W. (2007): “Inequality does cause underdevelopment: Insights from a new instrument,”

Journal of Development Economics, 84(2), 755 – 776.

Esteban, J., L. Mayoral, and D. Ray (2012): “Ethnicity and Conflict: An Empirical Study,”

American Economic Review, 102(4), 1310–42.

Fearon, J. D., and D. D. Laitin (2003): “Ethnicity, Insurgency, and Civil War,” American

Political Science Review, 97(1), 75–90.

Fjelde, H., and I. De Soysa (2009): “Coercion, Co-optation, or Cooperation? State Capacity

and the Risk of Civil War, 1961-2004,” Conflict Management and Peace Science, 26(1), 5–25.

Haggard, S., J.-D. Lafay, and C. Morrisson (1995): The political feasibility of adjustment in

developing countries. Development Centre of the Organisation for Economic Co-operation and

Development.

Harbom, L., E. Melander, and P. Wallensteen (2008): “Dyadic Dimensions of Armed

Conflict, 1946-2007,” Journal of Peace Research, 45(5), 697–710.

Hegre, H., T. Ellingsen, S. Gates, and N. P. Gleditsch (2001): “Toward A Democratic

Civil Peace? Democracy, Political Change, And Civil War, 1816–1992,” American Political

Science Association, 95(01), 33–48.

Hegre, H., and N. Sambanis (2006): “Sensitivity Analysis of Empirical Results on Civil War

Onset,” Journal of Conflict Resolution, 50(4), 508–535.

Miguel, E., and S. Satyanath (2011): “Re-examining economic shocks and civil conflict,”

American Economic Journal: Applied Economics, 3(4), 228–232.

Miguel, E., S. Satyanath, and E. Sergenti (2004): “Economic Shocks and Civil Conflicts:

An Instrumental Variable Approach,” Journal of Political Economy, 112(4), 725–753.

Morrisson, C., J.-D. Lafay, and S. Dessus (1994): Adjustment programmes and politico-

economic interactions in developing countries: lessons from an empirical analysis of Africa in

the 1980s. Royal African Society.

Ponticelli, J., and J. Voth (2011): “Austerity and anarchy: Budget cuts and social unrest in

Europe, 1919-2008,” .

Rule, J. B. (1989): Theories of civil violence. Univ of California Press.

Sambanis, N. (2002): “A review of recent advances and future directions in the quantitative

literature on civil war,” Defence and Peace Economics, 13(3), 215–243.

18

Savun, B., and D. C. Tirone (2012): “Exogenous Shocks, Foreign Aid, and Civil War.,” Inter-

national Organization, 66(3).

Singh, R. J., C. Bodea, and M. Higashijima (2014): “Oil and civil conflict: can public spending

have a mitigation effect?,” World Bank Policy Research Working Paper, (7100).

Thies, C. G. (2010): “Of rulers, rebels, and revenue: State capacity, civil war onset, and primary

commodities,” Journal of Peace Research, 47(3), 321–332.

Vanhanen, T. (2003): Democratization: A Comparative Analysis of 170 Countries, Routledge

Research in Comparative Politics, 7. Taylor & Francis Group.

Voth, H.-J. (2013): “Tightening Tensions: Fiscal Policy and Civil Unrest in Eleven South Amer-

ican Countries, 1937-1995,” in Fiscal Policy and Macroeconomic Performance, ed. by L. F.

Cespedes, and J. Gali, pp. 59–92. Central bank of Chile.

19

Africa Developing World

1 * Zambia 43 Yemen, Rep. 90 Norway

2 Somalia 44 Nicaragua 91 Finland

3 * Gabon 45 Paraguay 92 New Zealand

4 Swaziland 46 Sri Lanka 93 Switzerland

5 * Namibia 47 Iran, Islamic Rep. 94 Greece

6 * Liberia 48 Papua New Guinea 95 Japan

7 Kenya 49 Saudi Arabia 96 Korea, Rep.

8 * Angola 50 Guyana 97 United Kingdom

9 Djibouti 51 Mexico 98 Denmark

10 * Mauritania 52 Syrian Arab Republic 99 Ireland

11 * Congo, Rep. 53 Mongolia 100 Portugal

12 Burkina Faso 54 Jamaica 101 Sweden

13 * Equatorial Guinea 55 Honduras 102 France

14 * Nigeria 56 Peru 103 Austria

15 * Sierra Leone 57 Uruguay 104 Italy

16 Guinea-Bissau 58 Guatemala 105 Belgium

17 Malawi 59 Jordan

18 Rwanda 60 Bolivia

19 Madagascar 61 Dominican Republic

20 Morocco 62 Vietnam

21 Lesotho 63 Tunisia

22 * Guinea 64 Cambodia

23 Mali 65 Kuwait

24 * Niger 66 India

25 Central African Republic 67 Hungary

26 * Libya 68 Costa Rica

27 * Egypt, Arab Rep. 69 Suriname

28 * Chad 70 Fiji

29 * Sudan 71 Turkey

30 Togo 72 Panama

31 South Africa 73 Pakistan

32 Cameroon 74 Lao PDR

33 Mozambique 75 Ecuador

34 Ghana 76 Thailand

35 Burundi 77 Nepal

36 * Botswana 78 Lebanon

37 * Congo, Dem. Rep. 79 Czech Republic

38 Gambia, The 80 Bulgaria

39 Benin 81 Oman

40 * Algeria 82 Albania

41 Tanzania 83 Qatar

42 Zimbabwe 84 Poland

85 Romania

86 Haiti

87 El Salvador

88 Venezuela, RB

89 United Arab Emirates

Notes: In the tables showing regression results Developing includes African and OtherDeveloping countries, and World includes African, Other Developing, and Industrial coun-tries. Asterisks denote African countries classified as resource-rich.

Table 1: Countries in the Sample

20

Conflict Onset Conflict Prevalence

Conditional on peace Unconditional

Countries Civil Conflict Civil War Civil Conflict Civil War Civil Conflict Civil War

Africa 42 5.79 1.58 3.59 1.30 14.27 3.47

Africa, Resource Rich 19 8.39 2.40 4.27 1.84 17.78 4.77

Other Developing 47 4.16 0.99 2.67 0.85 13.24 2.50

Developed 16 0.43 0.00 0.25 0.00 2.70 0.00

Total 105 4.25 1.07 2.67 0.90 12.05 2.51

Table 2: Civil Conflict: Summary Statistics

Africa Other Developing Developed Total

A. All commodities

wi > 0 42 47 16 105

Mean wi 9.9 9.8 2.0 8.6

Mean sd(∆Pi,t) 17.4 16.8 16.3 17.0

B. Agricultural

wi > 0 42 47 15 104

Mean wi 4.0 4.3 1.0 3.7

Mean sd(∆Pi,t) 16.4 15.8 16.2 16.1

C. Minerals

wi > 0 21 33 16 70

Mean wi 3.0 1.2 0.4 1.6

Mean sd(∆Pi,t) 16.0 18.4 19.0 17.8

D. Oil

wi > 0 9 15 6 30

Mean wi 20.7 14.6 1.8 13.9

Mean sd(∆Pi,t) 20.9 21.3 14.5 19.8

Notes: wi is net exports of the commodities considered in each index as a fraction of GDP in 1990. Itsmean a median values, and the corresponding statistics for the volatility of the index, are computedacross all the countries in the group for which its value is positive.

Table 3: Commodity Prices: Summary Statistics

21

Civ

ilC

on

flic

tC

ivil

Wa

rs

Afr

ica

Dev

elo

pin

gW

orl

dA

fric

aD

evel

op

ing

Wo

rld

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)

(11)

(12)

(13)

(14)

(15)

(16)

(17)

(18)

dlnPt

−0.0

0−

0.0

10.0

30.0

30.0

20.0

20.0

0−

0.0

0−

0.0

0−

0.0

0−

0.0

0−

0.0

0

0.03

0.03

0.02

0.02

0.02

0.02

0.02

0.02

0.01

0.01

0.01

0.01

dlnPt−

1−

0.0

7∗

−0.0

0−

0.0

10.0

00.0

10.0

1

0.04

0.03

0.02

0.02

0.01

0.01

dlnPt−

2−

0.0

1−

0.0

1−

0.0

10.0

30.0

2∗

0.0

2∗

0.04

0.02

0.02

0.02

0.01

0.01

dlnPt−

1:2

−0.0

8∗

−0.0

1−

0.0

20.0

40.0

30.0

3∗

0.05

0.03

0.03

0.03

0.02

0.02

dlnPt:3

−0.0

90.0

10.0

00.0

30.0

30.0

2

0.07

0.04

0.04

0.04

0.02

0.02

R2

0.1

40.1

40.1

40.1

10.1

10.1

10.1

10.1

10.1

10.1

20.1

20.1

20.0

80.0

80.0

80.0

80.0

80.0

8

Obs

erva

tio

ns

1260

1260

1260

3127

3127

3127

3887

3887

3887

1434

1434

1434

3516

3516

3516

4298

4298

4298

Note

s:T

he

dep

end

ent

vari

ab

leis

aco

nfl

ict

on

set

du

mm

y.T

he

sub

ind

exes

t−

1:2

an

dt

:3

ind

icate

the

aver

age

of

the

firs

ttw

o-l

ags,

an

dth

ese

an

dth

ecu

rren

tvalu

e,of

the

expla

nato

ryvari

ab

lere

spec

tivel

y.D

evel

op

ing

incl

ud

esA

fric

an

an

dO

ther

Dev

elop

ing

cou

ntr

ies,

an

dW

orl

din

clu

des

Afr

ican

,O

ther

Dev

elop

ing,

an

dIn

du

stri

al

cou

ntr

ies.

Fix

edan

dti

me

effec

tsare

incl

ud

edin

all

the

regre

ssio

ns.

Sta

nd

ard

erro

rsare

clu

ster

edby

cou

ntr

y.

Tab

le4:

Com

mod

ity

Pri

ces

and

Civ

ilC

onfl

ict

22

Civil Conflict Civil Wars

Africa Developing World Africa Developing World

(1) (2) (3) (4) (5) (6)

dlnPt −0.01 0.03 0.03 −0.00 −0.01 −0.00

0.04 0.03 0.02 0.02 0.01 0.01

dlnPt−1:2 −0.09∗ −0.01 −0.02 0.03 0.03 0.03∗

0.05 0.04 0.03 0.03 0.02 0.02

dlnPt × −0.05 0.05 0.03 0.07 0.01 0.02

wheat-sugar 0.12 0.11 0.08 0.06 0.04 0.03

dlnPt × −0.04 −0.06 −0.07 −0.11∗ −0.02 −0.01

executive constraints 0.16 0.09 0.07 0.06 0.03 0.02

dlnPt−1:2 × −0.50∗∗ −0.40∗∗ −0.34∗∗∗ −0.03 −0.04 0.00

wheat-sugar 0.22 0.17 0.13 0.10 0.07 0.04

dlnPt−1:2 × 0.32∗∗ 0.19 0.16∗∗ −0.27∗∗ −0.11∗∗ −0.07∗

executive constraints 0.16 0.12 0.08 0.10 0.05 0.04

R2 0.14 0.11 0.11 0.12 0.09 0.08

Observations 1225 2591 3352 1400 2980 3763

Notes: The dependent variable is a conflict onset dummy. The subindex t − 1 : 2 indicates the average of the first two lagsof the explanatory variable. Developing includes African and Other Developing countries, and World includes African, OtherDeveloping, and Industrial countries. Executive constraints and the wheat-sugar ratio are adjusted by the mean of each group.Fixed and time effects are included in all the regressions. Standard errors are clustered by country.

Table 5: Commodity Prices and Civil Conflict, Interactions with Country-Fixed Variables

Civil Conflict Civil Wars

Africa Developing World Africa Developing World

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12)

dlnPt −0.01 −0.01 0.03 0.03 0.03 0.03 −0.00 −0.00 −0.01 −0.01 −0.00 −0.00

0.04 0.04 0.03 0.02 0.02 0.02 0.02 0.02 0.01 0.01 0.01 0.01

dlnPt−1:2 −0.09∗ −0.15∗∗∗ −0.01 −0.02 −0.02 −0.03 0.03 0.04 0.03 0.04 0.03∗ 0.04∗

0.05 0.06 0.04 0.04 0.03 0.04 0.03 0.04 0.02 0.03 0.02 0.02

dlnGt−1:2 −0.04 0.04 0.02 −0.03 −0.01 −0.01

0.08 0.07 0.06 0.06 0.04 0.04

dlnPt−1:2 × dlnGt−1:2 1.15∗∗ 0.24 0.23 −0.20 −0.21 −0.22

0.48 0.36 0.33 0.33 0.20 0.19

R2 0.14 0.14 0.11 0.11 0.11 0.11 0.12 0.12 0.09 0.09 0.08 0.08

Observations 1225 1225 2591 2591 3352 3352 1400 1400 2980 2980 3763 3763

Notes: The dependent variable is a conflict onset dummy. The subindex t − 1 : 2 indicates the average of the first two lagsof the explanatory variable. Fixed and time effects, and interactions of commodity prices with institutions (average executiveconstraints from polity) and the wheat-sugar ratio, are included as additional regressors (both variables are adjusted by eachgroup’s mean). The change in fiscal expenditure is adjusted by country means, and normalized by the mean for Africancountries. Developing includes African and Other Developing countries, and World includes African, Other Developing, andIndustrial countries. Standard errors are clustered by country.

Table 6: Commodity Prices, Fiscal Spending, and Civil Conflict

23

Country Year dlnPt−1:2 dlnGt−1:2 dco

Botswana 1975 -42.2 25.9 -11.5

Botswana 1976 -38.9 19.7 -7.6

Botswana 1982 -35.5 20.4 -7.3

Swaziland 1977 -39.2 18.0 -6.8

Lesotho 1976 -38.9 18.1 -6.8

Congo, Dem. Rep. 1989 -21.8 26.5 -6.5

Gabon 1987 -37.1 18.0 -6.4

Niger 1976 -31.4 15.9 -4.7

Cameroon 1987 -26.2 17.7 -4.6

Cameroon 1988 -20.4 19.5 -4.2

Mali 2000 -15.6 22.7 -4.1

Ghana 1987 -10.3 29.2 -4.0

Kenya 1973 -13.9 23.6 -3.9

Rwanda 1976 -15.9 20.9 -3.8

Mauritania 1973 -14.3 22.4 -3.8

Nigeria 1999 -14.8 21.4 -3.7

Gambia, The 1988 -14.6 21.4 -3.6

Lesotho 1981 -10.9 24.9 -3.5

Burundi 1988 -19.0 17.2 -3.4

Botswana 1969 -22.5 -1.9 1.8

Mali 1983 -11.0 -6.7 1.8

Togo 1983 -10.4 -7.8 2.0

Niger 1982 -20.7 -2.9 2.0

Rwanda 1980 -18.5 -4.1 2.1

Lesotho 1965 -18.3 -4.6 2.2

Burkina Faso 2000 -15.1 -6.2 2.3

Egypt, Arab Rep. 1988 -14.1 -6.8 2.3

Botswana 2005 -18.4 -5.2 2.4

Benin 1976 -27.1 -2.8 2.5

Congo, Dem. Rep. 1980 -19.4 -5.2 2.5

Chad 1976 -27.4 -2.8 2.5

Namibia 1990 -22.0 -4.6 2.6

Congo, Dem. Rep. 1973 -14.9 -9.2 2.9

Guinea 1993 -16.2 -9.5 3.1

Nigeria 1987 -37.0 -3.0 3.4

Sudan 1983 -11.4 -16.4 3.6

Zambia 2000 -14.3 -15.9 4.2

Liberia 1976 -47.7 -6.2 6.1

Table 7: Fiscal Spending and Conflict Onset: Mayor Events

24

Civil Conflict Civil Wars

Equation 1: Equation 1:

BaselineEquation 2: by type

BaselineEquation 2: by type

Agro Minerals Oil Agro Minerals Oil

(1) (2) (3) (4) (5) (6) (7) (8)

dlnPt −0.01 −0.01 −0.03 0.03 −0.00 −0.01 0.02 −0.02

0.04 0.03 0.07 0.05 0.02 0.02 0.03 0.05

dlnPt−1:2 −0.15∗∗∗ −0.08 −0.15 −0.18 0.04 0.01 0.01 −0.04

0.06 0.06 0.10 0.15 0.04 0.05 0.07 0.04

dlnPt−1:2 × dlnGt−1:2 1.15∗∗ 0.76 1.28∗∗ 0.12 −0.20 −0.25 0.04 −0.08

0.48 0.58 0.57 1.34 0.33 0.40 0.29 0.19

R2 0.14 0.15 0.12 0.12

Observations 1225 1225 1400 1399

Notes: The dependent variable is a conflict onset dummy. The subindex t − 1 : 2 indicates the average of the first two lagsof the explanatory variable. Fixed and time effects, the change in public spending, and interactions of commodity prices withinstitutions (average executive constraints from polity) and the wheat-sugar ratio, are included as additional regressors (the lastvariables are adjusted by their African mean). The change in fiscal expenditure is adjusted by country means, and normalizedby the mean for African countries. Standard errors are clustered by country.

Table 8: Commodity Prices and Civil Conflict in Africa, by Type of Commodity

Civil Conflict Civil Wars

Equation 1: Equation 1:

BaselineEquation 2: by type

BaselineEquation 2: by type

Agro Minerals Oil Agro Minerals Oil

(1) (2) (3) (4) (5) (6) (7) (8)

dlnPt −0.01 −0.00 −0.02 0.03 −0.00 −0.01 0.03 −0.02

0.04 0.03 0.07 0.05 0.02 0.02 0.03 0.05

dlnPt−1:2 −0.15∗∗∗ −0.08 −0.15 −0.18 0.03 0.01 0.02 −0.05

0.06 0.06 0.10 0.14 0.04 0.05 0.06 0.04

dlnPt−1:2 > 0 × dlnGt−1:2 1.70∗ 1.00 0.88 0.37 −0.98 −0.79 −1.23∗∗ 0.06

0.92 1.02 0.59 2.75 0.65 0.71 0.52 0.47

dlnPt−1:2 < 0 × dlnGt−1:2 0.66 0.53 1.65∗ −0.11 0.53 0.23 1.12∗ −0.22

0.57 0.59 0.90 0.47 0.40 0.42 0.65 0.47

R2 0.15 0.15 0.13 0.13

Observations 1225 1225 1400 1399

Notes: The dependent variable is a conflict onset dummy. The subindex t − 1 : 2 indicates the average of the first two lagsof the explanatory variable. Fixed and time effects, the change in public spending, and interactions of commodity prices withinstitutions (average executive constraints from polity) and the wheat-sugar ratio, are included as additional regressors (the lastvariables are adjusted by their African mean). The change in fiscal expenditure is adjusted by country means, and normalizedby the mean for African countries. Standard errors are clustered by country.

Table 9: Commodity Prices and Civil Conflict in Africa, by Type of Commodity and Sign of Shock

25

All Africa Resource-Rich Rest

(1) (2) (3) (4) (5) (6) (7) (8) (9)

dlnPt −0.01 −0.01 −0.01 −0.03 −0.03 −0.04 0.01 0.02 0.02

0.04 0.04 0.04 0.08 0.08 0.08 0.06 0.06 0.06

dlnPt−1:2 −0.09∗ −0.15∗∗∗ −0.15∗∗∗ −0.15∗ −0.21∗∗ −0.22∗∗ −0.11 −0.15 −0.16

0.05 0.06 0.06 0.08 0.10 0.10 0.13 0.13 0.13

dlnPt−1:2 × dlnGt−1:2 1.15∗∗ 1.38 0.72

0.48 0.98 0.45

dlnPt−1:2 > 0 × dlnGt−1:2 1.70∗ 3.89∗∗ 0.47

0.92 1.98 0.77

dlnPt−1:2 < 0 × dlnGt−1:2 0.66 −0.56 1.00∗

0.57 1.10 0.60

R2 0.14 0.14 0.15 0.28 0.28 0.29 0.11 0.11 0.11

Observations 1225 1225 1225 486 486 486 727 727 727

Notes: The dependent variable is the civil conflict onset dummy. The subindex t − 1 : 2 indicates the average of the first twolags of the explanatory variable. Fixed and time effects, the change in public spending, and interactions of commodity priceswith institutions (average executive constraints from polity) and the wheat-sugar ratio, are included as additional regressors(the last variables are adjusted by each group’s mean). The change in fiscal expenditure is adjusted by country means, andnormalized by the mean for African countries. See Table 1 for African countries classified as resource rich. Standard errors areclustered by country.

Table 10: Commodity Prices and Civil Conflict in Africa, Resource-Rich Countries

1960-1989 1990-2013

(1) (2) (3) (4) (5) (6)

dlnPt −0.05 −0.05 −0.05 0.04 0.05 0.04

0.05 0.05 0.05 0.06 0.06 0.06

dlnPt−1:2 −0.13∗ −0.19∗∗ −0.20∗∗ −0.07 −0.20∗∗ −0.19∗∗

0.07 0.08 0.08 0.09 0.10 0.10

dlnPt−1:2 × dlnGt−1:2 0.90∗∗ 3.18∗∗

0.44 1.60

dlnPt−1:2 > 0 × dlnGt−1:2 0.03 4.55∗∗

0.88 2.24

dlnPt−1:2 < 0 × dlnGt−1:2 1.60∗∗ 0.72

0.81 2.79

R2 0.21 0.21 0.21 0.23 0.24 0.24

Observations 622 622 622 603 603 603

Notes: The dependent variable is the civil conflict onset dummy. The subindex t − 1 : 2 indicates the average of the first twolags of the explanatory variable. Fixed and time effects, the change in public spending, and interactions of commodity priceswith institutions (average executive constraints from polity) and the wheat-sugar ratio, are included as additional regressors(the last variables are adjusted by their African mean). The change in fiscal expenditure is adjusted by country means in eachsubsample, and normalized by the mean for African countries. Standard errors are clustered by country.

Table 11: Commodity Prices and Civil Conflict in Africa, Subsamples

26

Africa Developing World

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12)

dlnPt −0.02 −0.04 −0.03 −0.02 0.01 0.00 0.01 0.00 0.01 0.00 0.01 0.01

0.09 0.08 0.09 0.08 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.02

dlnPt−1:2 −0.30∗∗ −0.25∗ −0.29∗ −0.20 −0.06 −0.10 −0.05 −0.06 −0.06 −0.10 −0.06 −0.06

0.15 0.14 0.15 0.13 0.07 0.08 0.07 0.07 0.06 0.06 0.06 0.06

dlnGt−1:2 −0.23 −0.34 0.03 −0.05 0.02 −0.05

0.18 0.23 0.12 0.16 0.11 0.14

dlnNMt−1:2 −0.27∗ −0.18 −0.02 −0.10 −0.03 −0.10

0.15 0.15 0.09 0.08 0.08 0.08

dlnMt−1:2 0.06 −0.24∗∗ 0.07 −0.07 0.07 −0.07

0.08 0.11 0.06 0.07 0.05 0.07

dlnPt−1:2 × dlnGt−1:2 2.78 0.64 0.56

1.72 0.78 0.73

dlnPt−1:2 × dlnNMt−1:2 2.66∗∗∗ 1.70∗∗ 1.66∗∗

0.33 0.72 0.74

dlnPt−1:2 × dlnMt−1:2 −0.94 −0.10 −0.14

1.23 0.69 0.66

dlnPt−1:2 > 0 × dlnGt−1:2 3.96 1.41 1.32

2.62 1.30 1.19

dlnPt−1:2 < 0 × dlnGt−1:2 0.85 −0.52 −0.54

2.83 1.16 0.99

dlnPt−1:2 > 0 × dlnNMt−1:2 1.97∗∗∗ 2.06∗∗∗ 2.01∗∗∗

0.33 0.58 0.60

dlnPt−1:2 < 0 × dlnNMt−1:2 5.06∗∗ −0.08 −0.02

2.31 1.37 1.25

dlnPt−1:2 > 0 × dlnMt−1:2 2.65∗∗ 1.30∗∗ 1.23∗∗

1.22 0.63 0.60

dlnPt−1:2 < 0 × dlnMt−1:2 −5.17∗∗ −2.08 −2.03

2.51 1.83 1.77

R2 0.25 0.30 0.25 0.33 0.24 0.26 0.24 0.26 0.23 0.25 0.23 0.26

Observations 447 447 447 447 1058 1058 1058 1058 1421 1421 1421 1421

Notes: The dependent variable is a conflict onset dummy. The subindex t− 1 : 2 indicates the average of the first two lags ofthe explanatory variable. dlnNM and dlnM are the change in non-military and military public spending, respectively. Fixedand time effects, and interactions of commodity prices with institutions (average executive constraints from polity) and thewheat-sugar ratio, are included as additional regressors (both variables are adjusted by each group’s mean). The change infiscal expenditure is adjusted by country means, and normalized by the mean for African countries. Developing includes Africanand Other Developing countries, and World includes African, Other Developing, and Industrial countries. Standard errors areclustered by country.

Table 12: Commodity Prices and Civil Conflict, Military and Non-military Spending

27

Civil Conflict Output Growth

Equation 1: Equation 1:

BaselineEquation 2: by type

BaselineEquation 2: by type

Agro Minerals Oil Agro Minerals Oil

(1) (2) (3) (4) (5) (6) (7) (8)

dlnPt −0.01 −0.01 −0.03 0.03 0.03∗∗ 0.01 0.03∗∗ 0.05∗

0.04 0.03 0.07 0.05 0.01 0.02 0.02 0.02

dlnPt−1:2 −0.15∗∗∗ −0.08 −0.15 −0.18 0.07∗∗∗ 0.08∗∗∗ 0.02 0.06

0.06 0.06 0.10 0.15 0.02 0.03 0.02 0.04

dlnPt−1:2 × dlnGt−1:2 1.15∗∗ 0.76 1.28∗∗ 0.12 0.11 −0.05 −0.17 0.31

0.48 0.58 0.57 1.34 0.20 0.20 0.26 0.49

R2 0.14 0.15 0.17 0.18

Observations 1225 1225 1224 1224

Notes: The dependent variable in columns 1-4 is the civil conflict onset dummy. The subindex t − 1 : 2 indicates the averageof the first two lags of the explanatory variable. Fixed and time effects, the change in public spending, and interactionsof commodity prices with institutions (average executive constraints from polity) and the wheat-sugar ratio, are included asadditional regressors (the last variables are adjusted by their African mean). The change in fiscal expenditure is adjusted bycountry means, and normalized by the mean for African countries. Standard errors are clustered by country.

Table 13: Commodity Prices, Civil Conflict and GDP growth in Africa

28

Notes:

Figure 1: Civil Conflict, 1960-2013

Notes:

Figure 2: Commodity Prices and Fiscal Spending around Civil Conflict Episodes

29

0.0%

1.0%

2.0%

3.0%

4.0%

5.0%

6.0%

-10% -8% -6% -4% -2% 0% 2% 4% 6% 8% 10% 12% 14%

dlnG

Bruckner & Ciccone

Unconditional elasticity

Conditional elasticity

Unconditional Likelihood

Notes: response of the probability of civil conflict to an average 10% decrease in commodity prices for two years.dlnG is the two-years average change in public spending. The standard deviation for this variable in the sample ofAfrican countries is 6.6%, and its average 4.4%. Elasticities are evaluated at the African mean value of executiveconstraints and the wheat-sugar ratio. Dotted lines are 90% confidence intervals. For Bruckner and Ciccone (2010),whose estimations are for the period 1961-2006, three-years average changes in P and g are assumed to obtain theelasticity.

Figure 3: The Impact of a Commodity Price Shock in Africa

-6.0%

-4.0%

-2.0%

0.0%

2.0%

4.0%

6.0%

-10% -8% -6% -4% -2% 0% 2% 4% 6% 8% 10% 12% 14%

dlnG

Negative price shock

Positive price shock

-6.0%

-4.0%

-2.0%

0.0%

2.0%

4.0%

6.0%

-10% -8% -6% -4% -2% 0% 2% 4% 6% 8% 10% 12% 14%

dlnG

Negative price shock

Positive price shock

-3.0%

-2.0%

-1.0%

0.0%

1.0%

2.0%

3.0%

-10% -8% -6% -4% -2% 0% 2% 4% 6% 8% 10% 12% 14%

dlnG

Negative price shock

Positive price shock

Notes: response of the probability of civil conflict (left and center) and civil war (right) to an average 10% changein commodity prices (left) and mineral prices (center and right), for two years. dlnG is the two-years averagechange in public spending. The standard deviation for this variable in the sample of African countries is 6.6%, andits average 4.4%. Elasticities are evaluated at the African mean value of executive constraints and the wheat-sugarratio. Dotted lines are 90% confidence intervals.

Figure 4: The Impact of a Commodity Price Shock in Africa, by Sign of Shock

-1

0

1

2

3

4

5

6

7

8

1960-1980 1965-1985 1970-1990 1975-1995 1980-2000 1985-2005 1990-2010 -2

-2

-1

-1

0

1

1

2

2

3

3

1960-1980 1965-1985 1970-1990 1975-1995 1980-2000 1985-2005 1990-2010

Notes: interaction term between the growth in commodity price index and public spending estimated by rollingregressions for 20-years windows from 1960 to 2013. The dependent variables are civil conflict (left) and civil war(right) onset. Spending is adjusted by the country mean in each subsample. Dotted lines are 90% confidenceintervals.

Figure 5: Interaction between Commodity Price Shocks and Public Spending in Africa, Rolling

Estimations

30

-3.0%

-1.0%

1.0%

3.0%

5.0%

7.0%

-10% -8% -6% -4% -2% 0% 2% 4% 6% 8% 10% 12% 14%

dlnNM -9.0%

-7.0%

-5.0%

-3.0%

-1.0%

1.0%

3.0%

5.0%

7.0%

9.0%

-10% -8% -6% -4% -2% 0% 2% 4% 6% 8% 10% 12% 14%

dlnNM

Negative price shock

Positive price shock

-8.0%

-6.0%

-4.0%

-2.0%

0.0%

2.0%

4.0%

6.0%

8.0%

-10% -8% -6% -4% -2% 0% 2% 4% 6% 8% 10% 12% 14%

dlnM

Negative price shock

Positive price shock

Notes: response of the probability of civil conflict to an average 10% change in commodity prices, for two years.dlnNM and dlnM are the two-years average change in non-military and military public spending, respectively.The standard deviations for these variables in the sample of African countries are 8.3 and 13%, and their averages5 and 2.6%, respectively. Elasticities are evaluated at the African mean value of executive constraints and thewheat-sugar ratio. Dotted lines are 90% confidence intervals.

Figure 6: The Impact of a Commodity Price Shock in Africa, Military and Non-Military Spending

31