The effect of war on a Country’s GDP

(1)

1

University of Amsterdam

The Effect of War on a Country’s GDP

---

Bachelor Thesis Economics and Finance

---

Abstract

As wars have not stopped the effect wars have on the economy remains of relevance. Providing a

conclusive answer would enable a better understanding of the impact future wars might have. Therefore the effect wars have on the economy are researched in this thesis. From the current empirical

macroeconomic literature and the empirical analysis done in this thesis it has been found that the effect of war on GDP are not unambiguous. However for international war it has been found that at the start of the war significantly negative effects are present. Near the end of war the negative effects are reduced and on the short term after war signs of growth and thus positive effects are observed. For civil war negative effects occur at the start of war, but at the end and on the short term there are some countries who experience positive effects, while others are still negatively affected. Additionally a panel data analysis provides insight in the positive effects of trade and investment on GDP on the short term post-war.

Student Mohammed Abdulaziz (10443673)

Field Macroeconomics

Supervisor Mr. R.E.F. Van Maurik

(2)

2

Verklaring Eigen Werk

Hierbij verklaar ik, Mohammed Abdulaziz, dat ik deze scriptie zelf geschreven heb en dat ik de volledige verantwoordelijkheid op me neem voor de inhoud ervan. Ik bevestig dat de tekst en het werk in deze scriptie gepresenteerd wordt origineel is en dat ik geen gebruik heb gemaakt van andere bronnen dan

(3)

3

1. Introduction

Since the start of the Arab Spring in 2011 a vast amount of civil wars have emerged. In the fifty years before that more than 140 wars have taken place, where the majority of these wars have been fought in African and Middle-Eastern countries. At the time of writing there are at least several ongoing wars.

Research in this field has primarily focused on channels of the economy affected by wars. However, what remains unclear is the impact on the economy of a country in which a war takes place. Knowing more about this relationship between wars and income is also important from a policy perspective. In order to address the effect of wars on a country’s economy the GDP forecasts will be compared to the actual GDPs in war time and on the short term after war (has taken place).

According to empirical macroeconomic literature the effect of conflict on the economy is

ambiguous. Some researchers claim that wars can raise productivity and might boost the economy in the short run, others argue that physical and human capital are destroyed by international and civil wars. Marwah and Klein state that increasing military expenditures can lower growth and productivity in Latin America (2005). Despite an increase in globalization, data on wars suggest that there has not been a decrease in the quantity of wars that have been fought during the course of the 20th_century.

The purpose of this thesis is to provide an analysis of the relationship between international and civil wars and income, measured by GDP, of the country where wars have been fought. Therefore the following question is central to this thesis: what is the effect of war on a country’s GDP? To measure when wars have taken place and the amount of years in war data from the Correlates of War (COW) project is used. Wars in the period 1970-2000 will be taken into account. The corresponding GDP data will be extracted from the Penn World Table and the World Data Bank. With use of time series forecasting methods the historical data can provide a forecasted GDP for each country in the time of war. This forecast will be an estimation of what the GDP would have looked like if there was no war, which can then be compared with the actual GDP.

This thesis first gives an overview of the literature, in which the empirical literature and the relationship between conflict and income are discussed. The next section (3) describes the data and war variables obtained from the COW project and other databases. In the following section (4) the method, time series forecasting: exponential smoothing with a trend, is described. Thereafter, in section 5, the results are shown and discussed, which includes a graphical analysis of select countries. Finally the conclusion of this thesis and possible directions for further research are given in section 6.

(5)

5

2. Literature Review

Economic literature generally recognizes that political stability and economic activity are interconnected. The existing literature regarding instability can be grouped into three categories: government instability (for instance coups or government turnovers), social instability (for instance ethnic tensions or

corruption) and political instability (for instance riots or wars). According to Londregan and Poole (1990) government instability has a weak negative effect on the economy. They argue that coups are negatively related to the level of income and the rate of economic growth. Easterly, Kremer, Pritchett and Summers (1993) and Levine and Renelt (1992) come to the same conclusion. Social instability is found to have a negative relationship with economic growth. Mauro (1995) finds that corruption is significantly more present in times of social instability. Which then lowers investment and thereby lowering economic growth. Knack and Keefer (1995) conclude the same but they claim that in the case of social instability property rights are infringed and thus investment and growth would become lower. Furthermore, political violence in Alesina and Perotti (1996) and Barro (1991) is shown to be negatively related to investment and growth on a large sample of countries.

For measuring political instability empirical studies have focused on either years at war or events like revolutions or on an institutional quality index. This thesis, however, occupies a time series analysis and only takes the years at war into account. Barro (1991) argues that property rights and private investment are related and because property rights would be distorted, due to political instability, investment and growth would be negatively affected. This result has also been found by Alesina, Ozler, Roubini and Swagel (1996) and Easterly and Levine (1997). However, Barro and Lee (1993) and Easterly and Rebelo (1993) do not find a statistically significant negative effect of political instability on the economy.

According to Alesina et. al. (1996) countries with periods of a high tendency of government collapse experience times where growth is significantly lower than it otherwise would be. Other studies, which adopt a similar definition of political instability, have found that instability leads to higher inflation (Cukierman, Edwars and Tabellini, 1992) and (Aisen and Veiga, 2006). These papers however don’t take the joint endogeneity between economy and the form of government into account (Alesina et. al., 1996). Which was addressed by Londregan and Poole (1990) and they have found different results. They don’t find evidence of reduced growth as a result of political instability, but political instability is confined to the occurrence of coups this study.

(6)

6

Sevestianova (2009) has found that both in Latin America and in Africa the impact of war is negative and significant, which is also found for civil war in general. It is argued that on average war reduces growth, however international war does not always reduce growth while civil war has been found to do as such. Furthermore it is found that war has a stronger effect on the economy on the short term but over time economies tend to recover. The results of the impact of war on one and two year intervals have been found to be negative and statistically significant while the five year interval was not significant (Sevastianova, 2009).

2.1 The Link between War and Income

According to the literature there are several channels through which wars might affect the economy. Economic performance tends to be impeded by decreases in the stock of human capital where war drafts and battle deaths mean less human capital. Another linkage is the reduction in investments, destruction of infrastructure and disruption of daily market activities which all mean a decrease in the stock of physical capital. Wars also, especially civil wars, negatively affect foreign direct investments as investors would rather invest in countries which are politically more stable (Murdoch and Sandler, 2002). Furthermore, social and private capital can be destroyed as a result of battles. Finally, wars might

enhance negative spillovers to neighboring countries (Sevastianova, 2009).

If the economy is believed to be affected by war through the human capital channel, this would be due to the formation of human capital for war purposes, as well as draft and deaths in combat. An increase in the stock of human capital would lead to higher shares of investment in human and physical capital and eventually to better economic performance (North, 1990). However this is not the case in the prevalence of war and therefore human capital would be misallocated and reduced and has a negative effect on the economy (Sevastianova, 2009).

If, on the other hand, war is linked to the economy through the physical capital channel, the deterioration of savings and therefore investments lowers economic performance and growth. North (1990) argues that in times of political instability when there is low security of property rights on physical capital, opportunities and incentives to invest and innovate are reduced. Also, the delay in providing licenses and permits may slow down the process of innovation in technology to be embodied in the production processes (North, 1990). Furthermore Sevastianova (2009) claims that ambiguity regarding political stability and government policies would lead to a tax on investment which then would lower growth.

(7)

7

linked to physical capital is based upon the effects of uncertainty caused by instability. Uncertainty causes suboptimal productive economic decision making on investment production or labor supply (Alesina et. al., 1996). Often, a high tendency towards political instability is associated with uncertainties which might lead for risk averse economic agents to hesitate to take initiatives. As a result these agents may transition from investing domestically to investing abroad, thereby exiting the economy. Foreign investors, too, prefer a stable political environment with less uncertainty about property rights or new policies (Goodrich, 1992). Furthermore Alesina and Tabellini (1990) argue that governments who are uncertain about their future purposefully partake in suboptimal policies to worsen the situation for their successor. This would then lead to economic inefficiencies.

A similar relation, between political instability and growth, is also described by Grossman’s (1991) analysis of revolutions. Countries with relatively weak rulers have a higher probability of

revolutions, where citizens are found to rather engage in revolutionary activities than productive market activities. While a strong ruler encourages market activities and discourages revolutionary activities.

Considering the effect of political instability on investments Campos and Nugent (2003) claim that there is a causal relationship. In contrast to other studies they have found that this relationship is positive, an increase in political instability would cause an increase in investments. They state three possible reasons for this relationship. The first is that political instability causes uncertainty and

therefore delays investment. Secondly they argue that political instability destroys capital stock and as a result the demand for replacement capital would increase. Thirdly it is believed that the change in government would be beneficial to the economy in the long run. They conclude that even though political instability is related to lower investment, it will lead to larger investments in the future (Campos and Nugent, 2003).

There are a couple other relationships between wars and the economy which are proposed by the literature. For instance, the political channel, in which Persson and Svensson (1989) argue that unstable governments would sooner partake in rent-seeking activities and thus misallocate public funds, adopt taxation policies which are not optimal and have a higher debt to GDP ratio. Murphy, Shleifer and Vishny (1991) further emphasize the negative effects on economic performance associated with rent-seeking activities. They state that a relatively weak government in fear of losing their position may be susceptible to the pressure of lobbyists or advocacy groups which could affect policy decision. Another argument is described by Dollar and Svensson (2000) which states that income growth in the long run might be slowed down due to the possibility of a slower political and economic reform process. Whereby the economy is less capable to adequately adapt to change.

(8)

8

Another claim is that a large industrial defense complex, which could be encouraged by war, stimulates investments and economic growth. But this result would mainly hold in countries who produce military goods (Diehl, 1983). On the contrary, Seiglie (1991) states that rich countries, high in GDP, use their wealth for larger defense expenditures to become a militarily more powerful. This, however, could eventually lead those countries to engage in wars. Others argue that in the case of countries who are rich in natural resources there can be a decrease in growth if rebel groups take control over poorly controlled governments and exploit natural resources (Collier, Elliot, Hegre, Hoeffler, Reynal –Querol and Sambanis, 2003).

In the case of civil wars Murdoch and Sandler (2002) argue that it is not likely that the economic consequences are solely attributed to national problems, such as a decrease of the stock in human and physical capital. Rather, Murdoch and Sandler argue, ‘’there is apt to be negative spillovers to

neighboring nations from disruptions to trade, heightened risk perceptions by would-be investors, severance of input supply lines, collateral damage from nearby battles, and resources spent to assist refugees.‘’ (2002). A further aspect of negative externalities is that the borders become less accessible. Another possibility, which is argued in Collier and Sambanis (2002), is that wars might cause neighboring countries to increase their military expenditure, as a result labor will be diverted from other productive purposes. Additionally, according to Collier and Sambanis (2002) the greater the impact of a civil war the greater these effects will be. Murdoch and Sandler distinguish the effects of civil war on the short run and the long run. They have found that most effects of civil war tend only to have an impact on the short run. However if a civil war is of a greater magnitude those effects are found to have a deeper long run impact (2002).

In recent empirical literature the civil wars in Africa have been studied extensively. Caruso (2010) shows that due to civil wars in Sub-Saharan Africa the size of the manufacturing sector decreases. Furthermore it is argued that ethnically more diverse societies have a higher probability on the incidence of civil war. This higher level of potential conflict may affect the economy through social and political channels. For example, government expenditure may be favored towards certain ethnic groups and trade may be limited between individuals of the same ethnic group (Montalvo and Reynal-Querol, 2005).

(9)

9

3. Data

For the measure of war data this thesis utilizes multiple data sets. First of all the Correlates of War project (COW), which is continuously held up to date by the Pennsylvania State University. This project keeps track of all wars between and within nations, with the first war recorded in 1816. They define war as a militarized conflict of at least 1000 deaths in battle a year. Further description of details, such as participants, duration and fatalities are also provided. This thesis, however, uses civil and international war prevalence in the time frame 1970 – 2000. Second, the Penn World Table (PWT) 8.2 is utilized. This data set provides economic times series data for 189 countries for the years 1950 – 2010. The data use a common set of prices, denominated in 2005 U.S dollar which makes real quantity comparisons possible. More specifically, this data set is used to gather information on GDP and investment of countries in war. Finally the World Bank’s Development Indicators (WDI) is utilized to obtain data on trade, government consumption and foreign direct investment (FDI). Hereafter follows a description of the variables used in the panel data regression.

The dependent variable used is GDP and the independent variables are investment, trade, government consumption and FDI.

The GDP is used as an indicator for income, even though it has been subject to criticism because it does not capture economic well-being, factors such as income inequality are not accounted for. However there is no other standard account which.

First of all investment is taken into account. As is stated in the literature review above, empirical studies have found that investment is lowered due to war. And therefore GDP can decrease. In empirical literature it is argued that trade is disrupted in war time. Therefore it could be expected that after war trade might increase proportionally to achieve pre-war standards. Additionally government

consumption, FDI and the lagged term of GDP are used as independent variables.

4. Methodology

To analyze the effect of war on a country’s GDP, this thesis will conduct two analyses. First a general look at the effect of war on GDP. This is done by a panel data regression on 15 countries which have fought a war between 1970-2000. First a regression on the five years before war, which is then compared with the regression results of the short term of two years after war. The variables that are taken under consideration are investment, trade, government consumption and FDI. If coefficients change we might be able to conclude whether war has had an effect on a particular variable. If the effect of war on GDP is not significant it is expected that the coefficients, or their sign, positive or negative, stay the same.

(10)

10

Furthermore a time series analysis will be done whereby the forecasted GDP is compared with the actual GDP. This comparison will be made on three different time periods. Those are at the start of the war, at the end of the war and finally a comparison after a war has taken place, which will be on the short term of two years. This way the effect of war on GDP can be analyzed over time, thus giving a more conclusive answer to the research question.

4.1 Panel Data Analysis

Because countries differ from one another OLS regression is not suitable. OLS regression would estimate a single constant which would hold for all countries, however this would most likely give less accurate results. Therefore fixed effects are considered. The following equation is then used for estimation:

𝐺𝐷𝑃𝑖𝑡 = 𝛼𝑖𝑡+ 𝛽1∗ 𝐿𝑎𝑔𝐺𝐷𝑃𝑖𝑡 + 𝛽2∗ 𝐼𝑛𝑣𝑒𝑠𝑡𝑚𝑒𝑛𝑡𝑖𝑡+ 𝛽3∗ 𝑇𝑟𝑎𝑑𝑒𝑖𝑡+ 𝛽4∗

𝐺𝑜𝑣𝑒𝑟𝑛𝑚𝑒𝑛𝑡𝐶𝑜𝑛𝑠𝑢𝑚𝑝𝑡𝑖𝑜𝑛𝑖𝑡+ 𝛽5∗ 𝐹𝐷𝐼𝑖𝑡+ 𝜀𝑖𝑡 (1)

Where i = country and t = year.

This regression will then be performed twice. First on the period of five years before war and second on the period of two years after a war took place.

4.2 Time Series Analysis

The time series techniques in general can be grouped in two categories, on the one hand open model time series and on the other hand fixed model time series. Open model time series (OMTS) analyzes the data and then tries to build a unique model to project the forecasts. Fixed model time series (FMTS), however, are based on predetermined assumptions and formulas concerning the existence of certain patterns in the data. The disadvantage of OMTS over FMTS is that it usually requires at least 48 observations to be able to create a model projecting forecasts. On the contrary FMTS are less complex and require little data without being less accurate (Mentzer, 2004). All time series techniques are endogenous techniques. Which means that they look at the underlying patterns of the actual historical data and identify the patterns over time. When these patterns are identified they can be projected into the future, thereby creating a forecast (Mentzer, 2004). Because FMTS techniques are relatively simple and require little data usage they can quickly adjust to changes in the level of GDP.

(11)

11

smoothing is actually the exponentially weighted moving average which essentially develops a moving average but the most recent periods of GDP will be weighted more heavily than the older observations. Those older observations will decrease at an exponential rate. Because economies and GDPs move in waves it’s required to take this movement into account in the projection of the forecasts. This trend component can be added into exponential smoothing, in which it adjust the level after each consecutive period. This gives the following formulas with which the forecasts can be computed:

𝐿𝑡 = 𝛼 𝐺𝐷𝑃𝑡+ (1 − 𝛼) ∙ (𝐿𝑡−1+ 𝑇𝑡−1) (2) 0 < 𝛼 < 1

𝑇𝑡 = 𝛽 (𝐿𝑡− 𝐿𝑡−1) + (1 − 𝛽) ∙ 𝑇𝑡−1 (3) 0 < 𝛽 < 1

With: L = Level T = Trend

This formula requires two estimates of level to exponentially smooth them. The first one is the level of GDP for this period. However, the trend is observed in the change of level from period to period and not in the single observed values of GDP. The second one then is the estimate of level from the previous period and additionally the estimate of what the change in level should have been from last period to this period, which is the trend. With these two estimates level can be exponentially smoothed with α. The two estimates for the measurement of trend, which are also shown in formula (2), are the change of level from last period to this period and the estimate of last period’s trend. These measures of the trend are exponentially smoothed with β. Both α and β are chosen to minimize the sum of squares of the forecast error, which is the difference between the actual GDP and the forecasted GDP squared and summed.

When level and trend have been estimated a forecast can be made for as far in the future as is wanted. This can be done by adding the trend times as many periods as one would want to forecast to the level. The following formula displays this:

(12)

12

𝐹𝑡+𝑚= 𝐿𝑡+ (𝑇𝑡∙ 𝑚) (3)

With:

m = the number of periods to be forecasted into the future.

Furthermore Mentzer (2004) argues that because forecasts do not have a slope they are biased

downwards. More specifically, exponential smoothing forecasts are biased downwards when an upwards trend occurs and biased upwards with a downward trend.

This thesis expects the effects of war on GDP to be in the following way. For international war it is thought to have a significant negative direct effect on a country’s GDP, which is at the start of a war. At the end of a war the effect is expected to still being significantly negative. However it is expected that the economy is going to grow after a war has ended, thus a positive effect. For civil war this thesis hypothesizes a significant negative effect at the start of a war, at the end of war and even on the short term after a war has taken place. It is expected that civil war destroys human and physical capital more heavily which causes the negative effect to last longer.

5. Results

In this section the results follow from the panel data analysis and the time series analysis, at the end of the section a closer look in the form of a graphical analysis is provided.

5.1 Results Panel Data Analysis

The results are presented in table 1, where the coefficients of the independent variables are shown with the standard errors in parentheses. Additionally the R-squared, F-statistic and the number of

observations are presented. The pre-war regression resulted in no significant variables. However the post-war regression resulted in significant variables in investment and trade, which both are positively related to GDP. For example, a one percent increase in investment results in a 0.773% increase in GDP. The difference in significance in the variables investment and trade might indicate that positive effects on GDP occur on the short term post war. Earlier, pre-war, those variables were not significant which shows that there might be a significant increase in activity of both variables. Which is in line with earlier empirical literature, more specifically argued by Campos and Nugent (2003). They state that due to the destruction of capital the demand for replacement capital increases.

(13)

13

PANEL DATA REGRESSION RESULTS

Independent Variable Pre-War Post-War

Lagged GDP -.03410 (.1090) .01196 (.0414) Investment -.03759 (.0785) .77388* (.1742) Trade -.00415 (.0907) .20171* (.0661) Government Consumption -.13011 (.0971) -.03893 (.0688) FDI -.05529 (.0681) -.12165 (.0556) Constant 64.8629* (23.430) 12.9198* (3.051) R2_(within) _0.0414 _0.8493 F Number of Observations 0.47 74 10.14 29 *p < .05 Table 1.

5.2 Results Time Series Analysis

After having conducted the empirical analysis and having computed all of the GDP forecasts we can evaluate and analyze the effects of war. The first analysis is regarding international war and the results are shown in table 2. In the table the difference between actual and forecasted GDP is shown in percentages. Additionally the t-statistic is presented which determines whether the difference, that is actual minus forecast, is statistically significant. Furthermore whether this difference is positive, negative or insignificant. A positive difference, for example, means that actual GDP has done better than what was forecasted. A full table including the actual GDP and the forecasted GDP is to be found in the appendix.

(14)

14

ANALYSIS OF INTERNATIONAL WAR EFFECTS ON GDP.

Difference t-statistic Significant

Start 2,437 % 2,801 Positively

Israel End -0,368 % -0,450 Insignificant

Short term -4,482 % -5,651 Negatively

Start -3,919 % -4,733 Negatively

Egypt End -1,355 % -1,677 Insignificant

Short term 1,489 % 1,938 Positively

Ethiopia End 4,634 % 7,476 Positively

Eritrea End -7,586 % -3,945 Negatively

Start 3,211 % 3,369 Positively

China End 3,211 % 3,369 Positively

Start 1,273 % 0,765 Insignificant

Chad End -2,556 % -1,500 Insignificant

Kuwait End -86,882 % -10,449 Negatively

Short term 4,698 % 0,851 Insignificant

Iran End -9,323 % -2,139 Negatively

Start -6,826 % -1,000 Insignificant

Iraq End 3,054 % 0,444 Insignificant

(15)

15

Difference t-statistic Significant

Bangladesh End -16,392 % -7,686 Negatively

Lebanon End 5,403 % 0,485 Insignificant

Cyprus End -33,257 % -14,513 Negatively

Table 2.

The main findings for the effects of international war on a country’s GDP are as follows. Overall the effect at the start of the war is significantly negative at most countries, which is around seventy per cent of the countries. Then the effect of war at the end of war, this does not have an unambiguous effect, however the negative effects occur less than at the start and there are more countries, than at the start, who have insignificant effects, suggesting a possible transition towards a growing economy. On the short term after the wars have been fought it is found that most countries, sixty per cent, witness positive effects. This means that their actual GDP significantly outperforms the forecasted GDP.

ANALYSIS OF CIVIL WAR EFFECTS ON GDP

Difference t-statistic Significant

Congo End -3,295 % -3,956 Negatively

Angola End 9,874 % 8,777 Positively

Liberia End 74,319 % 25,276 Positively

El Salvador End 5,143 % 4,717 Positively

(16)

16

Difference t-statistic Significant

Rwanda End -128,760 % -94,672 Negatively

Sri Lanka End -2,210 % -10,037 Negatively

Burundi End 6,563 % 10,622 Positively

Start 1,786 % 1,321 Insignificant

Nigeria End -9,001 % -5,783 Negatively

Table 3.

This thesis finds that at the start of the civil wars most countries, around sixty per cent, endure significantly negative effects to their GDPs. At the end of the wars it has been found that the negative effects are still present but some countries experience positive effects to their GDP. The effect on the short term seems to be ambiguous with some countries being affected positively and others negatively or even having insignificant effects. The results for civil war effects on GDP are summarized in table 3, a full table can be found in the appendix.

Overall the results regarding international war coincide with the hypothesized argument made above, however the effects of civil war on GDP are ambiguous and therefore do not coincide with what is earlier hypothesized.

Hereafter follows a graphical analysis of a set of countries in international war and other countries in civil war.

5.3 Graphical Analysis of International and Civil War

Looking at the graphs below, it is not possible to state an unambiguous relationship between war and income. Some countries experience a decline in income before the start of war and a period of growth when the war has ended, Kuwait and Cyprus for example. While others have declining income during war, which is most present in civil war cases. The results vary because the effect of war depends on the severity of the war, the type of war, civil or international, and on location and the type of country. The graphs illustrated in figure 1 and 2 show the relationship between international and civil war and the

(17)

17

GDP of a selected country. The time in war is illustrated by a shaded area. The values on the vertical axis are expressed in 2005 U.S dollar and in millions.

First the examination of the international war graphs, shown in figure 1. The first graph

illustrates the Yom Kippur War fought in 1973 in which Egypt was involved. The war was fought between Israel and Egypt with Syria. At the onset of war Egypt’s GDP was growing, however during the war GDP shows signs of stagnation while after the war the GDP was growing again. The living standard in Egypt did not decline at any time pre-, post- or during war.

In the case of Kuwait, where the Gulf war was fought in 1990-1991, the standard of living did decline. Kuwait was experiencing an upward trend which started in the early 1980s, which was disrupted by the war. However after the war ended GDP returned to the pre-war level.

International conflict in the case of Cyprus, where the Turk Cypriot War took place in 1974, seems to show the same pattern as in Kuwait. At the start of the war GDP was decreasing and even after war it was still in decline. However eventually pre-war standard of living was regained and even

surpassed.

Next, the examination of the civil war graphs, shown in figure 2. The first graph demonstrates the civil war in Angola which started in 1975 and was fought against UNITA, backed by the U.S in the Cold War. The war lasted until 2002 and during war GDP was often reduced. However when the war ended GDP significantly increased and growth was imminent.

In Liberia civil war took place from 1989 to 1997. In this period GDP more than halved and even though the economy grew after the war, pre-war standards of living were not recovered.

The civil war in Rwanda which took place in 1990-1994 initially was not as damaging to the economy as it eventually turned out to be. The war ended with the Rwandan genocide killing around 1.1 million people and significantly harming the economy, even after the end of the war Rwanda failed to recover to pre-war standard of living.

(18)

18 Figure 1. 0 20000 40000 60000 80000 100000 120000 140000 1965 1970 1975 1980 1985 1990 1995

Kuwait

War GDP 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 1945 1950 1955 1960 1965 1970 1975 1980 1985

Cyprus

War GDP 0 10000 20000 30000 40000 50000 60000 70000 80000 1945 1950 1955 1960 1965 1970 1975 1980

Egypt

War GDP

(19)

19 Figure 2. 0 10000 20000 30000 40000 50000 60000 70000 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010

Angola

War GDP 0 500 1000 1500 2000 2500 3000 3500 1965 1970 1975 1980 1985 1990 1995 2000 2005

Liberia

War GDP 0,0000 1000,0000 2000,0000 3000,0000 4000,0000 5000,0000 6000,0000 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000

Rwanda

War GDP

(20)

20

6. Conclusion

This thesis analyzes the effect of war on GDP by doing a panel data and a time series analysis on war incidence during 1970-2000. Exponential smoothing with a trend is used to compute forecasts for the GDP of countries who have experienced war. These forecasts are compared to the actual GDPs.

The main findings are that overall the effect of war on GDP is not unambiguous, especially in the case of civil wars. International war is found to have significantly negative effects at the start of the war, however near the end of the war and after the war has been fought the economy experiences

significantly positive effects. For civil war it is found that at the start of war countries are affected negatively by war. However at the end of war, and on the short term after war, there can be positive effects while other countries experience negative effects. It is also found that war can positively attribute to investment and trade on the short term after war has been fought. Both were significant in the panel data analysis.

Furthermore this thesis provides a visual examination of the graphs of GDPs of select countries in times of war. These have shown that each war varies, due to the severity and type of war, time period and location. Therefore further research is required to find a more conclusive answer. The focus thereby could be on whether specific regions have other effects. Another possibility is to take the standard of living into account with a proxy such as the Human Development Index or average life expectancy.

(21)

21

References

Alesina, A., Özler, S., Roubini, N., & Swagel, P. (1996). Political instability and economic growth. Journal of

Economic growth, 1(2), 189-211.

Alesina, A., & Perotti, R. (1996). Income distribution, political instability, and investment. European

economic review, 40(6), 1203-1228.

Alesina, A., & Tabellini, G. (1989). External debt, capital flight and political risk.Journal of International

Economics, 27(3), 199-220.

Aisen, A., & Veiga, F. J. (2006). Does political instability lead to higher inflation? A panel data analysis. Journal of Money, Credit, and Banking, 38(5), 1379-1389.

Barro, R., (1991), Economic Growth in a Cross Section of Countries, Quarterly Journal of Economics, vol.106, no. 2, pp. 407-444.

Barro, R. J., & Lee, J. W. (1993). Losers and winners in economic growth (No. w4341). National Bureau of Economic Research.

Campos, N. F., & Nugent, J. B. (2003). Aggregate investment and political instability: an econometric investigation. Economica, 70(279), 533-549.

Caruso, R. (2010). BUTTER, GUNS AND ICE‐CREAM THEORY AND EVIDENCE FROM SUB‐SAHARAN AFRICA. Defence and Peace Economics,21(3), 269-283.

Collier, P., Elliott, V. L., Hegre, H., Hoeffler, A., Reynal-Querol, M., & Sambanis, N. (2003). Breaking the Conflict Trap: Civil War and Development Policy (Washington, DC: World Bank).

Collier, P., & Sambanis, N. (2002). Understanding civil war: a new agenda.Journal of Conflict Resolution, 3-12.

Cukierman, A., Edwards, S., & Tabellini, G. (1989). Seigniorage and political instability (No. w3199). National Bureau of Economic Research.

Diehl, P. F. (1983). Arms races and escalation: a closer look. Journal of Peace Research, 20(3), 205-212.

Dollar, D., & Svensson, J. (2000). What explains the success or failure of structural adjustment programmes?. The Economic Journal, 110(466), 894-917.

(22)

22

Easterly, W., Kremer, M., Pritchett, L., & Summers, L. H. (1993). Good policy or good luck?. Journal of

Monetary Economics, 32(3), 459-483.

Easterly, W., & Levine, R. (1997). Africa's growth tragedy: policies and ethnic divisions. The Quarterly

Journal of Economics, 1203-1250.

Easterly, W., & Rebelo, S. (1993). Fiscal policy and economic growth. Journal of monetary

economics, 32(3), 417-458.

Garcia-Montalvo, J., & Reynal-Querol, M. (2004). Ethnic polarization, potential conflict, and civil wars. Potential Conflict, and Civil Wars (July 2004).

Goodrich, S. (1992). Political Instability as a Determinant of US Foreign Direct Investments. Harvard

University Senior Thesis.

Grossman, H. I. (1991). A general equilibrium model of insurrections. The American Economic Review, 912-921.

Knack, S., Keefer, P., (1995), Institutions and Economic Performance: Cross Country Tests Using Alternative Institutional Measures, Economics and Politics, vol. 7, no. 3, pp. 207-227.

Levine, R., & Renelt, D. (1992). A sensitivity analysis of cross-country growth regressions. The American

economic review, 942-963.

Londregan, J. B., & Poole, K. T. (1990). Poverty, the coup trap, and the seizure of executive power. World

politics, 42(02), 151-183.

Marwah, K., & Klein, L. R. (2005). Lost productivity and defense burden of the Southern Cone of Latin America: A page from history. Peace economics, peace science and public policy, 11(1). Mauro, P. (1995). Corruption and growth. The quarterly journal of economics, 681-712.

Min, S., & Mentzer, J. T. (2004). Developing and measuring supply chain management concepts. Journal

of business logistics, 25(1), 63-99.

Murdoch, J. C., & Sandler, T. (2002). Economic growth, civil wars, and spatial spillovers. Journal of conflict

resolution, 46(1), 91-110.

Murphy, K. M., Shleifer, A., & Vishny, R. W. (1990). The allocation of talent: Implications for growth (No. w3530). National Bureau of Economic Research.

(23)

23

North, D. C. (1990). Institutions, institutional change and economic performance. Cambridge university press.

Persson, T., & Svensson, L. E. (1989). Why a stubborn conservative would run a deficit: Policy with time-inconsistent preferences. The Quarterly Journal of Economics, 325-345.

Seiglie, C. (1991). Determinants of military expenditures (Doctoral dissertation, University of Chicago, Department of Economics).

Sevastianova, D. (2009). Impact of war on country per capita GDP: A descriptive analysis. Peace

(24)

24

Appendix A – Full version of Table 2.

Actual GDP Forecast GDP Difference t-statistic Significant

Start 44571,8320 43485,5167 2,437 % 2,801 Positively

Israel End 47411,6836 47586,1649 -0,368 % -0,450 Insignificant

Short term 48900,8750 51092,5899 -4,482 % -5,651 Negatively

Start 56151,5391 58351,9811 -3,919 % -4,733 Negatively

Egypt End 57547,1836 58327,0052 -1,355 % -1,677 Insignificant

Short term 60475,1992 59574,4389 1,489 % 1,938 Positively

Ethiopia End 36897,5977 35187,7811 4,634 % 7,476 Positively

Eritrea End 967,9427125 1041,370116 -7,586 % -3,945 Negatively

Start 668569,6875 647101,8457 3,211 % 3,369 Positively

China End 668569,6875 647101,8457 3,211 % 3,369 Positively

Start 5952,5859 5876,8107 1,273 % 0,765 Insignificant

Chad End 5810,3896 5958,9296 -2,556 % -1,500 Insignificant

Kuwait End 42872,2695 80120,5130 -86,882 % -10,449 Negatively

Short term 64603,5352 61568,2227 4,698 % 0,851 Insignificant

Iran End 368934,4688 403329,6698 -9,323 % -2,139 Negatively

Start 78170,6641 83506,9462 -6,826 % -1,000 Insignificant

Iraq End 77509,1328 75142,3246 3,054 % 0,444 Insignificant

Short term 70633,7734 81677,6127 -15,635 % -2,070 Negatively

(25)

25

Actual GDP Forecast GDP Difference t-statistic Significant

Libanon End 20668,2656 19551,6485 5,403 % 0,485 Insignificant

Cyprus End 2691,6147 3586,7737 -33,257 % -14,513 Negatively

(26)

26

Appendix B – Full version of Table 3.

Actual GDP Forecast GDP Difference t-statistic Significant

Start 8717,5605 8742,4743 -0,286 -0,339 Insignificant

Congo End 8809,8652 9100,1534 -3,295 -3,956 Negatively

Short term 9477,0830 8969,4688 5,356 6,917 Positively

Start 27741,4043 29800,7470 -7,423 -4,152 Negatively

Angola End 44081,3906 39728,6511 9,874 8,777 Positively

Short term 45536,0781 44796,5905 1,624 1,491 Insignificant

Liberia End 903,7239 232,0821 74,319 25,276 Positively

Short term 1171,8108 793,9049 32,250 14,222 Insignificant

El Salvador End 1491,8771 1415,1489 5,143 4,717 Positively

Rwanda End 2652,3865 6067,6103 -128,760 -94,672 Negatively

Short term 3541,5359 6173,6679 -74,322 -72,964 Negatively

Start 33504,1133 33600,3668 -0,287 -0,391 Insignificant

Sri Lanka End 111690,9063 114159,5259 -2,210 -10,037 Negatively

Burundi End 4012,6975 3749,3623 6,563 10,622 Positively

Start 146137,7188 139682,6853 1,786 1,321 Insignificant

Nigeria End 126952,9375 145354,8186 -9,001 -5,783 Negatively

Short term 125615,8906 130381,1619 -14,259 -9,065 Negatively

The effect of war on a Country’s GDP

University of Amsterdam