جنگ و بهره برداری منابع طبیعی

We build a theoretical framework that allows for endogenous conflict behaviour (i.e., fighting efforts) and for endogenous natural resource exploitation (i.e., speed, ownership, and investments). While depletion is spread in a balanced Hotelling fashion during peace, the presence of conflict creates incentives for rapacious extraction, as this lowers the stakes of future contest. This voracious extraction depresses total oil revenue, especially if world oil demand is relatively elastic and the government’s weapon advantage is weak. Some of these political distortions can be overcome by bribing rebels or by government investment in weapons. The shadow of conflict can also make less efficient nationalized oil extraction more attractive than private extraction, as insecure property rights create a holdup problem for the private firm and lead to a lower license fee. Furthermore, the government fights less intensely than the rebels under private exploitation, which leads to more government turnover. Without credible commitment to future fighting efforts, private oil depletion is only lucrative if the government’s non-oil office rents are large and weaponry powerful, which guarantees the government a stronger grip on office and makes the holdup problem less severe.

Natural resource abundance and exploitation fuel political unrest if different factions in society try to get control of the resource rents.1 Natural resource wealth thus increases the potential gains of controlling political power. However, conflict also affects natural resource extraction. Political instability pushes governments to rapacious resource depletion, but also reduces the incentive of especially private oil companies to explore and extract. Conflict thus influences economic decisions with respect to the method and intensity of natural resource extraction. So far the two directions of causality have been analyzed largely independently. While the literature on conflict and rent seeking has mostly treated natural resource extraction as exogenous and has focused on explaining appropriation efforts and outcomes (e.g., Torvik, 2002, Collier and Hoeffler, 2004, Fearon, 2005, Mehlum et al., 2006, Besley and Persson, 2011 and Rohner, 2011), the literature on resource economics has investigated the impact of insecure property rights on extraction rates without taking into account the effects of natural resources on fighting decisions (e.g., Tornell and Lane, 1999; van der Ploeg, 2010). Our objective is to investigate the two-way interaction between natural resource extraction and conflict. We therefore construct what to the best of our knowledge is the first unified framework with both endogenous fighting decisions and an endogenous choice of method and intensity of extraction. This enables us to derive predictions on the relative and absolute levels of fighting effort, the odds of victory, the method of and investment level in resource extraction, the speed of depletion, the size of license fees, as well as the total amounts and payoffs from extraction received by the government, rebel factions and international extraction companies. Our analysis generates a rich set of interesting and empirically relevant predictions. The primary distortion in our setting is that the government and the rebels cannot credibly commit to a cooperative outcome with both groups renouncing fighting and committing to transfers. To address this distortion, second-best policies become optimal. First, the speed of resource extraction is affected. While in a first best world without political instability it would always be optimal to extract resources in a balanced way that spreads out depletion over all periods, in a world of anarchy it is optimal to engage in rapacious over-extraction in order to reduce future rents and the incentives for rebel fighting, especially if the elasticity of world demand for resources is high and the government’s weapon arsenal is small. Also the ownership of resource production is distorted. In a world without threat from rebellion it is optimal to sell exploitation rights of natural resources to the technically most efficient company, which typically is a big private multinational. In contrast, if rebels fight for taking over the state and in case of victory renege on the past government’s oil contracts, this creates a hold-up problem for private investors. The bad property rights protection depresses their initial drilling investment, and also the licence fees that private companies are willing to pay are lower, especially if the government has bad weaponry and the spoils from office are small. This in turns makes private extraction relatively less attractive for the national government, even if on purely economic grounds private depletion by an international oil company would be the first best. In fact, if the government can commit in advance to future fighting efforts, it can reduce the political instability’s negative impact on private extraction. But if the government is bound to its exploitation contracts and treats received licence fees as sunk, it always has incentives to fight less. This time inconsistency problem of the government implies that if extraction is delegated to a private company, the government’s perceived stakes of keeping power are smaller than the stakes for the rebels, as the latter aim not only to appropriate rents from office, but also to expropriate the private exploitation company. This asymmetry of stakes implies that rebels fight harder than the government under private extraction, while the stakes and fighting efforts are symmetric under nationalized depletion. Hence, we predict government turnover to be higher under private than under nationalized resource depletion. These results might contribute to our understanding of the trend towards nationalized oil companies and the demise of the big multinational oil enterprises (the ‘Seven Sisters’) despite strong empirical evidence that private international oil companies are typically much more efficient than national oil firms. 2 Besides characterizing the optimal fighting efforts and resulting winning probabilities for the various extraction speed and ownership options, and deriving what extraction speed and method is selected in equilibrium, our paper analyses more complex oil contracts, endogenizes oil exploration and shows that social transfer policies that “bribe” the rebels to work rather than fight can reduce conflict. As discussed in detail below, our results can address various empirical puzzles. Our paper builds on the economic literature on contests and conflict pioneered by Hirshleifer (1991a), who has established that conflict efforts are increasing in the stakes of contest and the decisiveness of conflict technology. Further, poorer or less productive groups tend to fight harder and hence can have a higher winning probability than richer or more productive groups (e.g., Hirshleifer, 1991a, Hirshleifer, 1991b and Skaperdas, 1992). Favourite (respectively, underdog) groups with a probability of winning above (respectively, below) 1/2 will select a higher (respectively, lower) effort level if they can pre-commit and move first as Stackelberg leader compared to a simultaneous Nash Equilibrium without commitment (Dixit, 1987). Whether governments can pre-commit or not and the timing of decisions thus has a decisive influence on outcomes. With endogenous choice of the timing of moves, the underdog will always move first in equilibrium, so that there will always be under-commitment of efforts with respect to the benchmark without commitment (Baik and Shogren, 1992).3 An excellent survey of the literature on contests is provided by Konrad (2009). There are few theoretical papers linking natural resource exploitation and civil war. Most focus on how larger natural resource stocks increase the incentives for rent seeking and appropriation by boosting the “prize” to be appropriated (Torvik, 2002, Grossman and Mendoza, 2003, Olsson and Fors, 2004, Maxwell and Reuveny, 2005 and Hodler, 2006).4 The geographic location of resource deposits also affects conflict. For example, civil wars are more likely if resources are relatively abundant in the homelands of ethnic minorities, especially if these groups are geographically concentrated (Morelli and Rohner, 2011). Also, two countries with a shared border engage more often in inter-state war if one of them has its oil deposits close to the border and the other has no oil or deposits located far away from the border (Caselli et al., 2011). However, none of these studies endogenizes resource extraction. Natural resource exploitation is not modeled and resources are simply treated as exogenous lump-sum rents that can be appropriated. Extraction method, ownership and speed of extraction are all abstracted from. Our main objective is to endogenize natural resource depletion in a conflict framework, and to model explicitly the extraction speed, ownership, contracts and investment of the exploration companies. Another strand of the theoretical literature analyzes how over-extraction of natural resources can result from uncertainty about property rights (Hotte, 2001 and Hotte, 2005) or future political outcomes (Robinson et al., 2006), but does not distinguish private versus nationalized extraction methods and does not analyze civil conflict.5

We have presented a framework that makes both conflict behaviour and oil extraction endogenous. With a small number of exogenous parameters related to extraction technology and spoils of office, we derive a multitude of predictions on conflict variables such as the equilibrium fighting efforts, variables related to oil depletion like extraction method, extraction speed, exploration investment and licence fees, and political outcomes like regime durability. The looming shadow of conflict results in imperfect property rights protection and policies have to take account of this primary distortion. With a nationalized oil industry, extraction will be rapacious if government and rebels do not cooperate. This depresses the present value of oil revenue, especially if world oil demand is more elastic and the government’s weapon arsenal is small. As far as political turmoil is concerned, a more elastic demand and a bigger weapon arsenal of the government deter rebels, which reduces fighting. But there is also a countervailing effect of heavier weaponry, as it induces less aggressive extraction, which raises the stakes and increases fighting. The stake effect dominates the deterrence effect if resource demand is highly elastic. We have also shown that it is optimal for the government to bribe rebels to encourage them to work and put their arms down and that this wage subsidy is higher if oil and office rents are bigger. As a result, both rebels and government field a smaller army, but the rebels’ army declines relatively more and thus the government’s grip on office becomes stronger. Hence, oil depletion is less rapacious. While private extraction would be most efficient on purely economic grounds, the holdup problem on the oil company’s investment implies that private extraction is only attractive if the threat from rebels is not too large (i.e., with large office rents and a strong army), as fragile governments can only gain modest incentive compatible license fees, and political instability reduces private investment levels. In fact, with privatized extraction and the government unable to commit to future fighting levels, the government fights less intensively than the rebels and its chance of staying in office are less than under nationalized oil extraction. It thus follows that government turnover is higher under private extraction. The government’s grip on power also affects oil exploration: the oil exploitation company invests more, discovers more oil reserves, and pays a higher license fee if the government has superior weapons, office rents are large (which both reduce government turnover), and also if oil demand is more elastic. As discussed earlier, one of the reasons why property rights of private extraction investments are poorly protected is the fact that the government treats received licence fees as sunk and has incentives to select “too low” fighting efforts in the future. There are several ways in which the government can try to overcome this time inconsistency problem. First, “back-loaded” oil contracts that stipulate the payment of a large part of the licence fee ex post keep the government’s fighting incentives high. Holding future oil revenue constant, paying less of the license fee upfront indeed implies that the government fields a bigger army and the rebels a smaller one whilst the government improves its chances of staying in power. However, this effect is partly offset by a countervailing force, as paying less upfront implies that the oil company invests more and future oil revenue is higher which stimulates conflict and reduces chances of staying in power. This latter effect is strong if oil demand is less elastic. Second, if feasible, the government’s commitment to future weapon investments before oil licenses are sold can help to overcome the time inconsistency problem. Let us sketch in some detail what would happen in our model if we allowed for endogenous weaponry investment. If the government cannot pre-commit to invest in weaponry and to fighting efforts before licenses are auctioned, the license fee is a bygone at the moment when decisions are made about fighting efforts and weapon investments, and hence the stake is B and not B+L. It follows that the optimal investment in weapons increases in oil plus office rents, but decreases in the ratio of oil to office rents. The government invests less in weaponry if it cannot commit itself to investing in weaponry and fighting efforts, and thus to safeguarding property rights on natural resources. Not being able to commit to investment in weaponry and fighting efforts means that oil rights are less protected and thus that the government obtains a lower license fee than if it can commit. The incumbent faces a time inconsistency problem: it wants to convince the oil company that it will invest a lot in weapons to stave off rebellion and make oil investments attractive, but once the license fee has been received it has an incentive to renege. Without commitment, investment in weaponry is too low and both the government and the oil company are worse off as a result. In contrast, if the government is able to commit to future weaponry investments, it offers more secure oil rights and thus extracts a bigger fee from the oil company. This way the government also increases its grip on office. 26 Our analysis highlights the importance of explaining conflict and resource extraction simultaneously in empirical work; existing estimates of the effect of resources on conflict may suffer from endogeneity bias. Some of our results with respect to the determinants of fighting or with respect to factors favouring under-investment can provide mechanisms to explain existing empirical results, as discussed above. In contrast, other propositions require new empirical efforts. For example, the expected effects of domestic political institutions and military capacity on the ownership structure and depletion speed of extraction companies, as well as on the share of rents captured by the government through licence fees, are still largely empirical terra incognita. Similarly, our prediction of a “conditional resource curse”, making the conflict inducing impact of natural resources depend on extraction method and speed, as well as on military asymmetry, needs further empirical investigation. Moreover, our findings on the impact of speed and method of extraction on government turnover and army size suggest new topics for empirical testing. Our theoretical analysis can also be extended in the following directions. First, if the government has cash constraints and cannot use future resource revenue as collateral, private oil companies have an incentive to finance an army to fight off rebel coups. These constraints bias the mode of exploration towards more rapacious nationalized or privatized extraction, as this yields more funds upfront to bribe rebels, field a bigger army and have more advanced weaponry. However, looting and “booty futures” can also finance a stronger rebel army which can set off or prolong conflict (Ross, 2004). Second, oil companies may not have an interest to disclose private information about in-situ reserves, necessary investment outlays, or costs of extraction to the government, in which case the design of incentive-compatible contracts under moral hazard with risk sharing is called for (e.g., Bolton and Dewatripont, 2005 and Radon, 2007). Third, with capital-intensive extraction companies (say, oil or gas) a higher price of resources pushes up the relative return on capital and pushes down the wage, which boosts conflict (Dal Bó and Dal Bó, 2011). With less rapacious depletion, fighting intensities increase more, as not only the wage falls, but also the revenues from natural resources and thus the stakes rise. On the other hand, a higher price of labour-intensive resources (e.g., coffee, rice or bananas) pushes up wages and reduces the return on capital, so fighting becomes less intense and work more attractive, which could make balanced resource extraction relatively more attractive. Fourth, we find that fast depletion of natural resources occurs due to insecure control over future ownership of the resource and the consequent threat of expropriation, but too slow depletion may occur due to the holdup problem and the resulting lack of investment in oil exploration (cf., Bohn and Deacon, 2000). Although some work has been done with an exogenous expected chance of expropriation and an exogenous confiscation risk (Strand, 2010), more work is needed to make both these effects endogenous. Of course, vertical integration or nationalization can solve the holdup problem. Finally, aggressive extraction resulting from the threat of expropriation may be less severe with rival oil companies that face a future threat of expropriation. Oil companies are then induced to conserve oil in order to have less competition in the future if rivals get expropriated (Laurent-Luchetti and Santuguni, 2012).