econstor Make Your Publications Visible.

Transcription

1 econstor Make Your Publications Visible. A Service of Wirtscaft Centre zbwleibniz-informationszentrum Economics Hefeker, Carsten; Kessing, Sebastian Working Paper Competition for Natural Resources and te Hold-Up Problem CESifo Working Paper, No. 60 Provided in Cooperation wit: Ifo Institute Leibniz Institute for Economic Researc at te University of Munic Suggested Citation: Hefeker, Carsten; Kessing, Sebastian (06) : Competition for Natural Resources and te Hold-Up Problem, CESifo Working Paper, No. 60, Center for Economic Studies and Ifo Institute (CESifo), Munic Tis Version is available at: ttp://dl.andle.net/049/47374 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenscaftlicen Zwecken und zum Privatgebrauc gespeicert und kopiert werden. Sie dürfen die Dokumente nict für öffentlice oder kommerzielle Zwecke vervielfältigen, öffentlic ausstellen, öffentlic zugänglic macen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt aben sollten, gelten abweicend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewärten Nutzungsrecte. Terms of use: Documents in EconStor may be saved and copied for your personal and scolarly purposes. You are not to copy documents for public or commercial purposes, to exibit te documents publicly, to make tem publicly available on te internet, or to distribute or oterwise use te documents in public. If te documents ave been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise furter usage rigts as specified in te indicated licence.

2 Competition for Natural Resources and te Hold-Up Problem Carsten Hefeker Sebastian G. Kessing CESIFO WORKING PAPER NO. 60 CATEGORY : PUBLIC CHOICE OCTOBER 06 An electronic version of te paper may be downloaded from te SSRN website: from te RePEc website: from te CESifo website: Twww.CESifo-group.org/wpT ISSN

3 CESifo Working Paper No. 60 Competition for Natural Resources and te Hold-Up Problem Abstract We study te role of competition for te old-up problem in foreign direct investment in resource-based industries. Te ost country government is not only unable to commit not to expropriate investment ex post, but is also unable to commit to te provision of local resources. In te case of competition for local resources tis dual commitment problem triggers iger investment levels and increases ost country revenues, but urts pro.ts of international investors. JEL-Codes: F0, F30, Q340. Keywords: foreign direct investment, natural resources, old-up problem. Carsten Hefeker University of Siegen Department of Economics Unteres Scloss 3 Germany Siegen carsten.efeker@uni-siegen.de Sebastian G. Kessing University of Siegen Department of Economics Unteres Scloss 3 Germany Siegen kessing@vwl.wiwi.uni-siegen.de 6 September 06 We are grateful for comments from our referees and te associate editor Andrés Carvajal, participants at te European Public Coice Meeting 04 at Cambridge University, te CESifo Public Sector Conference, and te Natural Resource Governance conference in Siegen, in particular Roberto Bonfatti, Urs Steiner Brandt and Marco Sam. We tank Artur Grigoryan for excellent researc assistance. Sebastian G Kessing tanks te Centre for Business Taxation (CBT) at te University of Oxford and te Centre for Competitive Advantage in te Global Economy (CAGE) at te University of Warwick for teir ospitality to carry out tis researc.

4 Introduction Te sustained growt of emerging economies as substantially increased worldwide demand for natural resources. Not surprisingly, countries tat depend on imports of suc resources increasingly discuss te issue of resource security, and increased competition for scarce resources as lead to warnings about "resource wars", a "new scramble for Africa" or a "new great game" for access to Central Asian gas and oil (Cooley 0, Halper 00, Sambaug 03). As a corollary of tis development, international rms in natural resource industries from emerging economies ave been seeking better access to resources and upstream integration in te form of substantially increased foreign direct investment in resource-ric countries. Governments of ost countries ave welcomed te arrival of tese new investors, wereas international companies from traditional and developed source countries ave seen teir pro tability treatened as tese competing investors were moving in. Investing in natural resources is particularly problematic for international investors in te presence of weak institutions in ost countries. Non-renewable resources suc as oil, gas or metals are particularly vulnerable to weak property rigts because tey are usually regionally concentrated, need ig expenditures for searc and exploration, and take considerable time before tey become marketable. Tis exposes investors to te risk of renegotiation or expropriation, i.e. being old-up, and tis can be regarded as one dimension of te resource curse (Battacaryya and Hodler 00, Sacs and Werner 00, Bon and Deacon 000). Recent prominent cases of con icts about revenue saring and outrigt expropriation of foreign investments in resource-based industries include Repsol in Argentina, Rio Tinto in Guinea and First Quantum Minerals in te Democratic Republic of Congo, wic alone are estimated to ave cost some $3bn (Stevens et al. 03). Te World Bank (009) lists 30 countries tat revised oil contracts and taxation between 999 and 00, and Stevens et al. (03) mention 5 cases in wic increases in taxes and royalties were implemented unilaterally. Famous cases of rivalry between foreign rms for access to oil resources are te US-UK con ict between te world wars in te Middle East and Latin America (Venn 986, Yergin 99) or te Cinese- US con ict in Central Asia more recently. Maurer (03), for instance, describes ow Peru exploited te con ict between te US and te UK in te 970s, and Cooley (0) sows ow Kazakstan exploited te con ict between te US and Cina after independence. Host governments expropriating and transfering foreign investments to national oil rms, in turn, as a long istory wit a peak in te 970s and 980s and a renewed increase over more recent years (see Hogan et al. 00, Tomz and Wrigt 00, Bremmer and Jonston 009, Maurer 03). Notorious cases were te terms of contracts and agreements ave been canged or revoked are Venezuela (Manzano and Monaldi 00) or Russia (Gustafson 0). As Engel and Fiscer (00) report, owever, also countries like te US, Canada, Australia and te UK ave "renegotiated" contracts and

5 Te old-up problem in foreign direct investment as been analyzed extensively, starting wit te seminal analysis of Eaton and Gersovitz (983). Tomas and Worall (994), Doyle and van Wijnbergen (994), Svensson (998), Scnitzer (999), Konrad and Lommerud (00), Kessing et al. (007, 008, 009), Stroebel and von Bentem (03) among oters, ave discussed various factors a ecting te interaction between foreign investors and ost country governments and addressed te question ow te old-up problem may be overcome and ow optimal contracts may look like. Tese existing studies of te old-up problem only consider ost governments wic cannot commit temselves not to take te lion s sare of te returns of an investment and/or te invested capital after a foreign investor as irrevocably invested in a ost country. Moreover, in tis traditional formulation te old-up problem is independent of te competition between investors, and is terefore not suitable to analyze te potential e ects of increased competition for natural resources on investment, ost government revenues and pro tability. In tis study, we provide a new insigt tat makes it possible to understand te role of competition between foreign investors. It is based on te observation tat ost countries can not only directly increase taxes on foreign pro ts or expropriate foreign investors, tey can also "renegotiate" contracts, revise concessions and restrict te amount of resources tat a particular investor is allowed to exploit. In a resource-based industry, a key determinant of an investor s returns is te actual access to local resources. Consider te investment in drilling equipment and transportation infrastructure in a given oil eld. Te pro tability of tis investment will not only be determined by te capital investment, te tecnical and managerial expertise deployed in te ost country, and te tax and royalty payments, but will depend crucially on te investor s rigt to exploit tis speci c eld. Wile a ost government may grant suc rigts before te investment is carried out, just as it may promise abstaining from nationalization and con- scatory taxation in te traditional formulation of te old-up problem, suc promises may not be time-consistent. After te investment is sunk, it may be in te ost country s interest to renege on te granted rigts to exploit local resources, and instead grant tese rigts to anoter interested party, suc as anoter foreign investor or a domestic, possibly state-owned, rm. Since capital is sunk, te investor cannot recoup is investment, and altoug te investment is not touced and no expropriation is taking place, it becomes useless. Wat te investor can do, owever, is to witdraw is tecnical expertise, leaving te ost country wit potentially lower output from given capital investments and resources. imposed windfall taxes or increased royalties in te natural resource sector.

6 We study tis additional commitment problem and ow it interacts wit te potential existence of alternative international investors. We demonstrate tat te existence of alternative foreign investors substantially canges te interaction between foreign investors and te ost government. Intuitively, te presence of competing rms provides an alternative for te ost government to put te natural resource to economic use. Given te ost country s inability to commit, tis drags investors into over-investing to secure te provision of local resources. Tis unambiguously bene ts te ost country and urts foreign investors. Contrary to te classic old-up problem of taking too muc ex post, wic typically as negative consequences for te ost country, te inability to commit to provide local resources can bene t ost countries if tere is more tan a single use for local resources. Tus, contrary to te standard old-up approac, our formulation can generate important e ects of competition for natural resources in countries wit weak governance, wic are in line wit te empirical evidence and te cases discussed above. 3 Our study contributes to te vast literature on expropriation of sunk investment already mentioned, but adding te ex-post provision of local resources as an important new dimension of te problem. Te analysis relates to Janeba (000), wo sows tat foreign investors can overcome te old-up problem by investing in excess capacity in an alternative location. Tis treatens te ost government wit a sift in economic activity in case of ex post con scatory taxation. In our analysis, te existence of a competing investor provides an alternative possibility to put local resources to productive use and tereby serves a similar end for te ost government, triggering iger investment. Finally, our study adds an important aspect to te literature on resource nationalism and te scramble for resources. It is often argued tat in particular Cina is trying to secure monopolistic access to resources, mostly in Africa. Te argument put forward for tis is tat by securing a monopoly on suc resources, Cina could improve its termsof-trade (see Bonfatti 009). 4 We add anoter view to tis argument by sowing tat restricted access would increase te pro ts tat a monopolistic investor could make. A restriction of competition from oter investors would tus directly sift more pro ts to te investor, independent from potential terms-of-trade considerations. 3 Note tat setting up a domestic, resource-extracting rm may serve a similar purpose as a second foreign investor. Tis also provides an alternative for te domestic government to put its resources to economic use, and can accordingly also generate incentives for foreign investors to increase investment in order to receive better access to local resources in te absence of commitment. 4 In particular te popular press often claims tat Cina aims at securing and ten exploiting a monopolistic position. It is disputed, owever, weter tis is really te case (Moran 00). 3

7 Foreign investment in natural resources In wat follows we develop a model to illustrate a stylized story of foreign investment, te risk of expropriation, te allocation of resources to be combined wit te capital stock, and tecnical expertise provided by te investor. We tink of tis as a foreign rm tat invests in a resource ric country, for instance in drilling equipment, a mine, or an oil or gas pipeline to exploit tose resources. Tis revenue from tis investment is subject to ex-post expropriation after te investment as been sunk. Expropriation can be full or partial in te sense tat tax arrangements or pro t-saring agreements can be broken. Tis gives rise to te standard old-up problem. Te second old-up problem is tat te ost country s government can renege on te concession and access to te resource by, for instance, not supplying te pipeline or blocking access to te mine or oil or gas eld. We assume tat tere is xed amount of exploitable resources, suc as an oil eld, to wic te ost countries grants access because it is unable to exploit te resource on its own, eiter because it is capital-constrained or lacks te necessary tecnical expertise. Wit a competing investor, owever, te country can renege on its promise to grant access and instead sift te rigt of access to anoter investor wo can provide te necessary additional expertise. Lastly, te investors ave to decide ow muc e ort to provide after capital as been invested and access to resources as been given. We tink of tis as, for instance, sending quali ed personal or additional (non-sunk) tecnical equipment to operate te initially invested equipment. Extraction of te resource depends critically on tis additional e ort by te investor and is impossible witout it. As our bencmark, we rst sow tat, wit commitment, for a given amount of domestic resources it does not matter for total investment, output and tax revenues weter a single, or two competing foreign investors are present in te ost country. We ten contrast te di erential outcome between te single investor case and te case of two competing investors witout commitment to sow ow tis will lead to iger investment, iger output, iger government revenues and lower pro ts for foreign investors. Finally, we can compare total investment and tax revenues wit and witout commitment. Tis allows us to sow tat te positive e ect of competition on investment witout commitment can dominate te negative e ect of te absence of commitment and can result in iger total investment and tax revenues. 4

8 . Te basic framework We start by considering a resource-based industry wic relies on foreign investors, denoted by subscript i, to bring in capital and expertise to process te natural resource in te ost country. For bot scenarios, full commitment and te lack tereof, we contrast two cases. Eiter tere is only a single foreign investor, i = s, or, alternatively, tere are two competing investors i = ; present in te ost country. Total output of investor i is produced according to te constant returns to scale production function f (k i ; m i ; z i ), wit te usual > f < f > 0; q @q@l i ; m i ; z i ; and l = k i ; m i ; z i, and q 6= l, were capital investment k i, te local natural resource m i, and tecnical and business expertise z i are te factors of production. For our results, we make te assumption tat tecnology is of te Cobb-Douglas form, i.e. f (k i ; m i ; z i ) = ki m i z i, 0 < ; ; <, =. For te exposition of te model we use a more general formulation, since te dynamics and te strategic interaction are better visible. Te formal results and teir extensive derivation using te Cobb-Douglas assumption are relegated to te Appendix. Factor costs, r (for capital k i ) and w (for non-resource input z i ) are given, constant and equal for all investors. Output is sold in te world market at an exogenous price of one. Capital investment is sunk once deployed in te ost country and, witout commitment, subject to con scatory taxation ex post. Similar to Janeba (000), we do not explicitly study repeated interaction. However, te investor also provides is tecnical and business expertise z i, and can adjust tis input depending on is tax burden t i and te amount of local resources m i allocated to im. Te possibility to adjust anoter factor of production does not eliminate te old-up problem entirely but alleviates it. We assume tat taxation takes te form of a speci c tax t i per unit of output tat is applied uniformly to all foreign investors. Tis amounts to a situation were te ost country carges an excise on output, suc as, for example, te number of barrels of oil exported. In many real world contracts, te pro rata payments are complemented by a lump sum up-front payment (Hogan et al. 00). Suc a payment could be easily added into our analysis witout canging te results. Te upfront investment may be interpreted narrowly to include only capital in te actual business venture. Alternatively, it may be interpreted more widely to also include oter investments made by foreign investors into te local infrastructure, wic is often te case wit Cinese investments in Africa (Brautigam 009, Economy and Levi 04). Similarly, te local resource provided by te ost government may be interpreted narrowly as access to natural resources, or more widely to additionally include inputs provided 5

9 by te ost government suc as administrative and legal support or publicly provided infrastructure wose use can be restricted. Importantly, te ost government may not be able to commit ex ante on te level of tis input. In addition, only an exogenous level of tese resources m is available. Treating te overall level of m as xed can eiter be justi ed by te fact tat te total amount of te available non-renewable resources is limited by nature. Alternatively, it may be interpreted as te maximum level of te natural resource tat can be processed witin a given time period, were te local government does not inter-temporally optimize te exploitation of its resource. 5 We contrast te commitment case, were te government can commit to tax rates and te allocation of domestic resources ex ante, and te no commitment case, were it cannot. Moreover, we consider te di erences between te case of a single investor and te case of two competing investors in eac of tese cases. Te latter situation is called te competing investor case or te competition case. In tis competition case, we assume trougout tat te two investors take teir decisions non-cooperatively. In te absence of government commitment bot rms use teir up-front investment to a ect te government s decision ow to allocate te domestic resources between investors since teir pro ts depend on te sare of resources tat tey receive. Finally, we assume tat rms are symmetric and focus on te symmetric equilibrium. Te stages of te game di er between te commitment and te no-commitment case. Te timing under commitment is as follows: in stage, te ost government commits to a tax rate/tax rates t i and to te allocation of te natural resource m i to eiter a single investor or ow it sares access to it among competing investors. In stage, te foreign investor/investors decides/decide about capital investment and about te complementary input z i. Witout commitment tere are tree stages in te game. Teir sequence is as follows. In stage te investor/te investors decides/decide ow muc capital k i to invest, and tis capital is sunk. In te case of two competing investors tey decide simultaneously. In stage, te ost government government sets a tax rate/tax rates t i and decides on te allocation of te natural resource m i. Finally, in stage tree, te investor/te investors decides/decide on te level of te complementary input z i, and tax revenues and net-of tax pro ts are ten realized. 5 Tis may be due to te fact tat te local non-benevolent government is facing a ig probability of losing power and aims at maximizing its revenue over a sort time orizon (Robinson et al. 006, Svensson 998). 6

10 . Full commitment As our bencmark, we rst analyze te case in wic te ost government can fully commit on te level of taxation and te amount of resources made available to eac foreign investor. We focus on te di erence between te single investor case and te case of two competing investors, since our main interest is to identify te di erential e ect between tese two cases in te absence of commitment. We summarize our bencmark nding: Proposition In a symmetric equilibrium wit competing investors wic produce using a Cobb-Douglas tecnology, total factor inputs, total output, te tax rate, total tax revenues, and total after tax pro ts are equal to te single investor case, if te government can commit to its policy. Proof. See Appendix. Intuitively, under commitment, te decision problem of eac individual investor is structurally equivalent to tat of a single investor, given a certain level of local resources provided to te investor and given te level of taxation. Te ost government divides te local resource arbitrarily between investors and cooses te same tax rate as wit a single investor to maximize its revenue. Given te constant returns to scale tecnology, te aggregate inputs of all investors equal te total inputs of te single investor. 6 Accordingly, total output is equal in bot cases as well as total tax revenues and total after-tax pro ts. Te bottom line is tat under commitment it makes no di erence weter te ost government deals wit a single or multiple investors. Finally, note tat, under commitment, our restriction on te set of admissible contracts prevents an e cient investment level, since te linear tax distorts te investment decision. If we allowed for more general contracts, i.e. a contract specifying te e cient input levels of k i and z i and a lump-sum payment to te investor tat allowed im to break even, e ciency could be acieved. However, also in tis case, te e cient outcome is independent of te number of investors..3 No commitment: A single investor We consider rst te case of a single investor. Te sequence of events witout commitment is as follows. At stage, te investor decides ow muc capital to invest in te ost 6 Tis result is not restricted to te Cobb-Douglas tecnology but can be derived for any constant returns tecnology. We focus on te Cobb-Douglas case for consistency wit our furter propositions below. 7

11 country. At stage, te ost country government cooses te tax rate t s on revenues and allocates te local resource m. At stage 3, te investor cooses te additional factor input level z s. Solving backwards, at stage 3, te investor cooses is input to wit rst order condition wic implicitly de nes z s = z s(t s; k s; m s). max z s s = ( t s ) f (k s ; m s ; z s ) wz s ; ( t s = w; () At stage, te ost country government decides on t s and m s. We assume tat it maximizes revenue. Tis can be motivated by a non-benevolent government, but it is also fully compatible wit welfare maximization since output is sold in te world market and te foreign investor s pro ts accrue to non-local residents. Since te ost country as no alternative use for te local resource witout te foreign investment, it sets m s = m. Te government s problem wit respect to te optimal ex post taxation ten is to wit rst order condition wic implicitly de nes t s = t s (k s ). max t s R s = t s f (k s ; m; z s(t s ; k s ; m)) ; f (k s ; s ; k s )) t s = s At stage, te single investor cooses k s to maximize wit rst order condition max k s s = ( t s(k s )) f (k s ; m; z s(m; t s (k s ) ; k s )) rk s ; f (k s ; m; ( s s = r: s Te investor equates te direct marginal e ect on net pro ts, te indirect e ect on net pro ts via te corresponding adjustment of te additional input, and te indirect marginal e ect from te resulting adjustment of te tax rate wit te opportunity cost of capital. 8

12 .4 No commitment: Competing investors Consider now te case were te ost country admits two investors i = ; into te country. Te sequence of events is as in te single investor case. At stage 3, bot investors decide on z i wit rst order conditions analogous to (). At stage te government now needs to decide ow to allocate te total amount of resources available m between te two rms, wit sare m i eac, and to coose te tax rates t and t tat are levied on te investors output. Te government s problem is tus max R c = t f (k ; m ; z (t ; m )) t f (k ; m m ; z (t ; m m )), t ;t ;m were we ave substituted te resource constraint m m = m. Te solution to tis problem is caracterized by f (k i ; m i ; z i ) @z i = 0; for i = ; ; and (4) ; (5) were condition (5) illustrates ow te ost government allocates local resources between te two competing investors suc tat te resulting marginal tax revenue is equalized across investors. Since te ost government participates in te rms revenue wit sares t and t, respectively, it is in its interest to maximize total output by allocating te resources e ciently between te rms wenever t = t. In stage bot rms simultaneously decide on teir investment k i, solving max k i i = ( t i (k i )) f (k i ; m i (k i ); z i (k i ; m i (k i ) ; t i (k i ))) rk i ; wit rst order f (:) ( t i @m i = r; (6) for i = ;. For bot investors, adding an additional unit of capital may cange te level of taxation. Moreover, tere are now tree e ects of an additional unit of capital on after tax pro ts as can be seen by te terms in squared brackets on te left and side of (6). First, te direct e ect and te indirect e ect via te adjustment of zi are analogous to te single investor case. Second, tere is now an additional strategic e ect, wic results from te response of te ost government to reallocate te available local resources across 9

13 investors. Te latter consists of a direct e ect and an indirect e ect stemming from te increased leverage of adjusting te complementary input. Te rst order conditions (6) implicitly de ne te investors best responses vis-à-vis eac oter, k (k ) and k (k ). Witout commitment te question arises weter te ost government could also be tempted not only to sift domestic resources between investors or to increase taxes ex post, but to also expropriate one investor and sift tis capital stock to te oter investor. We ave not allowed for tis possibility, but it is straigtforward to see tat tis is not a binding restriction in our model. In fact, te government can never increase its tax revenues from suc an action, since total production and tus tax revenues will not be increased from suc an action under constant returns to scale..5 Comparison We can now study te di erential e ects of competing investors relative to te situation wit a single investor. First, we consider te overall level of investment. Proposition Let tecnology be of te Cobb-Douglas type. Ten, witout commitment, tere exists a unique symmetric equilibrium, in wic we ave k i > k s= for eac individual investor, and total investment is iger tan in te single investor case, i.e. P ki > ks. i=; Proof. See Appendix. As we sow in te Appendix, te comparative statics are straigtforward wit Cobb- Douglas tecnology. In i = 0, wic implies tat te equations (3) and (6), wic determine te investment in bot cases, simplify substantially. i > 0, i.e. increased investment results in a iger sare of domestic resources allocated to te investor by te ost government. Intuitively, since te ost government cannot commit to te provision of resources, bot investors can a ect te allocation of local resources in teir favor troug teir investment. Since te ost country participates in te overall production via ex post taxation, additional investment, wic increases overall output, is a suitable instrument to a ect te distribution of resources in te respective investor s direction. Of course, in te symmetric equilibrium, te strategic e ects of additional investment cancel out, and bot investors receive only alf of te available resources. 7 We turn now to te e ect of competition on tax revenues. 7 Te situation is structurally similar to a non-perfectly discriminating contest, see Konrad (009). Bot investors make sunk investments to receive a iger sare of domestic resources, analogous to exerting iger contest e ort to increase te own probability of obtaining te contested prize. Te only di erence is tat in our case te domestic resource can be continuously divided between te investors. 0

14 Proposition 3 Wit Cobb-Douglas tecnology, tax revenues are iger wit competing investors. Proof. According to Proposition bot rms togeter coose a iger level of investment. Moreover, wit Cobb-Douglas tecnology te ost government cooses te same tax rate as in te single investor case. Given te same tax rate and iger total investment P bot investors coose a level of zi, so tat > zs=. Tus, total output increases zi i=; and government revenues, wic are a xed fraction of tat, must also increase. Tis result sows tat te ost government unambiguously pro ts from increased competition. Wit competing investors, and a production tat relies on domestic inputs controlled by te government, te inability to commit turns into a strategic advantage for te ost government. 8 Next, we consider pro ts. Proposition 4 Given Cobb-Douglas tecnology, total pro ts in te single investor case are iger tan aggregate pro ts in te multiple investor case, i.e. s > P i. Proof. Te single investor could coose an alternative, iger level of investment ~k s = P ki. Tis would result in te same level of taxes as in te competing investors i=; case and te same overall level of additional factor inputs. Accordingly, output and after P tax pro ts would equal i. Te fact tat tis level of investment is not cosen by a i=; single investor reveals tat total pro ts are lower in te competing investors case. Te result sows ow competition for local resources not only induces iger investment, but tat increased investment urts te rms relative to te single investor case. Our Propositions -4 rely crucially on te inability of te government to commit ex ante on te available level of local resources. If te production function did not require m as a factor of production, te old-up problem would only consist of te ost government setting excessively ig taxes after investors ave irrevocably invested. In tis case, tere would be no strategic interaction between investors. Tus, te inability to commit "to give" as speci c implications tat are not present in te traditional old-up situation were te government only cannot commit "not to take". Only by taking tese additional considerations into account it is possible to understand te e ects of competition for te 8 Note tat an increased investment level is typically su cient for tis result, even if te tax rate is not independent as in te Cobb-Douglas case. Te government can always coose te same level and would realize a iger revenue level tan wit a single investor. It will cange te tax level only if tis will increase its revenue. i=;

15 overall level of investment, tax revenues and pro tability in resource based-industries in countries wit weak property rigts protection. Finally, te assumption of Cobb-Douglas tecnology allows us to explicitly solve for total investment wit and witout commitment. Wile te investment level will obviously always be iger under commitment wit a single investor, tis is less evident wit multiple investors. In tis case te second old-up problem wit respect to providing te domestic resources, wic triggers iger investments, may overcompensate te old-up problem wit respect to ex-post taxation. Tis possibility is stated in Proposition 5. Proposition 5 Wit Cobb-Douglas tecnology total investment wit two competing investors, and tax revenues can be iger witout commitment tan wit commitment. Proof. See Appendix. As we sow in te Appendix, tis is more likely to appen wen te domestic resource is more important for production. Intuitively, te strategic over-investment e ects are particularly strong in tis case, and tey can overcompensate te iger level of ex-post con scatory taxation. 3 Conclusion Our main argument is tat te competition for access to natural resources in an uncertain legal environment gives rise to a "second" old-up problem. Not only do international investors su er te standard risk of being taxed beyond initial agreement or being expropriated, but tey also face te risk of denied access to local resources. Tis risk depends on te competition between foreign investors and te scope of te ost government to sift domestic resources to competitors. Competition for natural resources elps ost countries wit weak property rigts protection to induce iger investment in te absence of commitment. Our approac, we would argue, is particularly relevant for te recent surge in natural resources-related foreign investments by investors from emerging countries. After all, rms from traditional "Western" countries ave been competing wit eac oter for access to natural resources in developing countries for many decades (Venn 986, Yergin 99). However, te problem as analyzed above may be less relevant for suc investors. Tese investors may turn for elp to te legal system in teir ome countries and pursue formal or informal action against oter investors wo are bene ting from te re-allocation of local resources by te ost government (Jo é et al. 009). Suc measures are likely to

16 be particularly e ective were international investors ave business interests and tangible property in te ome countries of tose competing rms, facilitating collusion among tem. 9 In te case of competing investors from emerging countries, owever, suc action is muc less e ective, since tese new source countries temselves ave weaker protection of foreign investment and often no unbiased access to te legal system. Moreover, investors from tese countries typically ave less developed business interests and fewer assets in OECD countries, and are tus less vulnerable to defensive legal action. Te scope for suitable actions by an investor from a traditional source country wo is negatively a ected by a decision of a ost country to reallocate access to natural resources or oter essential local inputs to a rival from an emerging country is tus more limited compared to actions against a traditional competitor. Our framework captures te latter situation and is accordingly well suited to explain wy te recent expansion of foreign direct investment in resource-based and resourceextracting industries in countries wit weak governance is perceived to treaten te profitability of investments by traditional investors. It is also in line wit te observation tat te arrival of tis new class of investors as been very welcome to te rulers of resource ric ost countries. In te ligt of tese arguments, te position of multinationals from OECD countries in te competition for natural resources in developing countries may bene t from stronger governance in emerging countries and from iger levels of business activities of te emerging investors in te traditional source countries. 4 Appendix 4. Proof of Proposition Wit Cobb-Douglas tecnology production is given by f(k i ; m i ; z i ) = k i m i z i, wit =, 0 <, ; <, and i = ; ; s. Te subscript s denotes te single investor case, te subscripts i = ; refer to individual investors in te case wit two investors. Under commitment, te government rst cooses te tax rate(s) t i and te division of te local resource among investors m i. Te rm(s) ten coose(s) investment k i and te complementary input z i. Consider rst te case of a single investor. Solving backwards, 9 Also, tere migt be a direct involvement of te ome governments of competing international rms and a political solution for tose con icts. A famous case for tis is te separation of speres of in uence and exclusive access to te Middle East oil elds between te US and te UK up to te 950s (Venn 986). 3

17 te investor maximizes its pro ts wit rst order conditions wic yield k s = max = ( t s ) ks m s zs rk s wz s ; k s = ( t) m z k ( ) r s = ( t) m z () k w = 0; ( t) i m, and z r w s = ( t) i m. At te r w rst stage te government cooses m s = m and maximizes R = t s k m z by its coice of t s wic gives t s =. Substituting, te optimal factor inputs are k s = i i ( ) m, and z i i r w s = ( ) m: r w (A) For notational purpose we ave added te bar over te equilibrium factor inputs wit commitment to distinguis tese variables from te outcomes derived for te non-commitment case below. Wit two investors bot rms maximize i = ( t i ) k i m i z rk i wz i by coosing k i and z i wit rst order conditions so tat k i = ( t i ) m z i k i r = 0; = ( t i ) m z i ki w = 0; i ( t i ) i m r w i and z i = Note tat bot are linear in m i. Te ost government s problem is max t ;t ;m R = t ( t i ) i m r w i, i = ;. " i i # " i i # ( t ) m m ( t i ) m r w r w " i i # " t ( t ) [m m ] [m m ] ( t i ) r w r i i # [m m ] : w Te rst order conditions wit respect to t and t are ( t i ) t i ( t i ) = 0; i = ; ; (A) 4

18 so tat t i = ; i = ; : Moreover, government revenue is linear in m, suc tat any split of te local resource is optimal. Assume tat te government arbitrarily splits te resource, allocating a sare to rm, and a sare = X k i = i=; = to rm. Total investment is ( ) w r ()i m ( ) w r ()i ( ) m ( ) w r ()i m = k s: (A3) Total investment is independent of te number of investors under commitment. Likewise P z i = z s. Tus, depending on te split of m, te investors factor inputs will be split in te same proportion and total factor inputs are equal to te single investor case. Given te constant returns to scale property of te Cobb-Douglas tecnology total output is te same wit a single investor or two competing investors. Finally, given tat te tax rate is te same in bot cases, total tax revenues are also equal. 4. Proof of Proposition Consider rst te single investor case witout commitment. Solving backwards, te single investor s cooses z s at stage 3 to maximize s = ( t s ) ks m s z wz s, wit solution zs = ( t s )ks m s w. At stage te government maximizes R = t s ks m s zs = t s ks m ( t s )ks m w by its coice of t s, were we ave already substituted m s = m as te optimal allocation of te domestic resource. Te solution to tis problem yields t s = : Substituting we ave z s = k s m w. At stage te single investor ten cooses k s to maximize s = ks m s i i ks = A ; were A r () w m. i w k s m s rk s, wit solution For comparison of tis investment level we now solve te competing investor case. At te tird stage bot rms maximize i = ( t i )ki m i z i wz i by coice of z i, wic i yields zi = ( t i )k i m i w ; i = ;. Te government s problem in stage is max R = t k m t ;t ;m z t k m z = t k m ( t k [m m ] ( t )k [m m ] w i i t )k m w : 5

19 Te rst order conditions are k t [( t )] [( t i )] t i [( m k t i)] t [( t )] [m m ] = 0: = 0; i = ; From tese we ave t = t =, as wit a single investor. Te rst order condition wit respect to m simpli es to m = k k k m, and te comparative @k = = k m > 0: (A4) [k k ] Te problem of investor i; i = ; at te rst stage is max k i i = ( t i ) k i m i z i rk i = k i " k i m 4 ki k j k i # k i m w k i k j 3 5 rk i for i; j = ;, and i 6= j. Te corresponding rst order conditions are w Note j m " k i k i k j k i k i k i k i k j # k j (k i k j ) = r: (A5) < 0, wic implies tat te investors reaction functions are downward sloping and te symmetric equilibrium is unique. Using symmetry, we ave w m " ki k i k i k i # ki k i k i (k i ) = r: (A6) i wic can be solved for ki = A : We can now compare te total investment in te competing investor case to te investment in te single investor case. Total investment in te competing investor and te single investor case are ki = A ; and ks = A, (A7) respectively. Since > 0 we ave k i > ks, i.e. total investment in te competing investor case is always iger. 6

20 Figure : Parameter combinations wic ful ll (A8), i.e. for wic total investment witout commitment and competing investors is iger tan in te commitment case (black area) 4.3 Proof of Proposition 5 Proposition establised tat tere is no di erence in total investment, production and tax revenues between te single investor case and te case wit two competing investors. We ere compare tis commitment outcome wit te case of two competing investors witout commitment. Let B mw r. From (A3) and (A7) investment wit and witout commitment (te latter wit two competing investors) equals k i = B ; and k i = k s = B ( ) ; respectively. Total investment witout commitment and competing investors will terefore be iger tan investment under commitment if > ( ) (A8) Since inequality (A8) cannot be solved analytically, we rely on simulations to illustrate te conditions under wic investment will be iger witout commitment and competing investors relative to te commitment case. Tis is illustrated in Figure and sows ow tis typically depends on te importance of te domestic resource as a factor of production. Finally, we sow by example tat tax revenues in te competing investor case witout commitment can be iger tan tax revenues in te commitment case. To tis end, it is useful to express te tax base as a function of te invested capital stock in bot cases. 7

21 Tax revenues wit and witout commitment are, respectively, T = t k i T = tk i m m i m ( t) i z m i = tki i ks w = t ( t) i i ( t i )k i m i w D; and = t( t) D wit D ks i m ks i m i w, and D ki m k i m w i, respectively. If total investment is iger in te absence of commitment, i.e. if k i > k s, wic is te case if inequality (A8) olds, tis implies D > D. wenever (A8), and additionally Accordingly, T > T t( t) > t ( t), or > ( ) ( ) (A9) olds, were we ave substituted t = and t =. Now let = 0:; = 0:3; = 0:5, suc tat tis inequality (A9) is true, and also te inequality (A8) olds. Tus, T > T for tis parameter constellation. References [] Battacaryya, S. and R. Hodler (00) Natural Resources, Democracy and Corruption, European Economic Review 54, [] Bon, H. and R. Deacon (000) Ownersip Risk, Investment, and te Use of Natural Resources, American Economic Review 90, [3] Bonfatti, R. (009) Foreign In uence and te Cinese African Trade in Natural Resources, London Scool of Economics, mimeo. [4] Brautigam, D. (009) Te Dragon s Gift, Oxford: Oxford University Press. [5] Bremmer, I. and R. Jonston (009) Te Rise and Fall of Resource Nationalism, Survival 5 (), [6] Cooley, A. (0) Great Games, Local Rules, Oxford: Oxford University Press. [7] Doyle, C. and S. van Wijnbergen (994) Taxation of Foreign Multinationals: A Sequential Bargaining Approac to Tax Holidays, International Tax and Public Finance, -5. [8] Eaton, J. and M. Gersovitz (983) Country Risk: Economic Aspects, in: R. J. Herring (ed.), Managing international risk, Cambridge: Cambridge University Press, [9] Economy, E. and M. Levi (04) By All Means Necessary: How Cina s Resource Quest is Canging te World, Wasington: Council on Foreign Relations. 8

22 [0] Engel, E. and R. Fiscer (00) Optimal Resource Extraction Contracts under Treat of Expropriation, in: W. Hogan and F. Sturzenegger (eds), Te Natural Resources Trap, Cambridge: MIT-Press, [] Gustafson, T. (0) Weel of Fortune: Te Battle for Oil and Power in Russia, Cambridge: Belknap Press. [] Halper, S. (00) Te Beijing Consensus, New York: Basic Books. [3] Hogan, W., F. Sturzenegger and L. Tai (00) Contracts and Investment in Natural Resources, in: W. Hogan and F. Sturzenegger (eds), Te Natural Resources Trap, Cambridge: MIT-Press, -43. [4] Janeba, E. (000) Tax Competition wen Governments Lack Commitment: Excess Capacity as a Countervailing Treat, American Economic Review 90, [5] Jo é, G. P. Stevens, T. George, J. Lux and C. Searle (009) Expropriation of Oil and Gas Investments: Historical, Legal and Economic Perspectives, Journal of World Energy Law and Business (), 3-3. [6] Kessing, S., K. A. Konrad, C. Kotsogiannis (006) Federalism, Tax Autonomy and te Limits of Cooperation, Journal of Urban Economics, 59(), [7] Kessing, S., K. A. Konrad, C. Kotsogiannis (007) Foreign Direct Investment and te Dark Side of Decentralization, Economic Policy 49, [8] Kessing, S., K. A. Konrad, C. Kotsogiannis (009) Federalism, Weak Institutions and te Competition for Foreign Direct Investment, International Tax and Public Finance 6, [9] Konrad, K. A., K. E. Lommerud (00) Foreign Direct Investment, Intra-Firm Trade and Ownersip Structure, European Economic Review 45, [0] Konrad, K. A. (009) Strategy and Dynamics in Contests, Oxford: Oxford University Press. [] Manzano, O., F. Monaldi (00) Te Political Economy of Oil Contract Renegotiation in Venezuela, in: W. Hogan, F. Sturzenegger (eds), Te Natural Resources Trap, Cambridge: MIT-Press, [] Maurer, N. (03) Te Empire Trap, Princeton: Princeton University Press. [3] Moran, T. (00) Cina s Strategy to Secure Natural Resources, Wasington: Peterson Institute. [4] Robinson, J., R. Torvik and T. Verdier (006) Political Foundations of te Resource Curse, Journal of Development Economics 79, [5] Sacs, J. and A. Werner (00) Te curse of natural resources, European Economic Review 45, [6] Scnitzer, M. (999) Expropriation and Control Rigts: A Dynamic Model of Foreign Direct Investment, International Journal of Industrial Organization 7, [7] Sambaug, D. (03) Cina Goes Global, Oxford: Oxford University Press. [8] Stevens, P., J. Kooroosy, G. Lan and B. Lee (03) Con ict and Coexistence in te Extractive Industries, Catam House Report, London: Catam House. 9

23 [9] Stroebel, J. and A. van Bentem (03) Resource Extraction Contracts under Treat of Expropriation: Teory and Evidence, Review of Economics and Statistics 95, [30] Svensson, J. (998) Investment, Property Rigts and Political Instability: Teory and Evidence, European Economic Review 4, [3] Tomas, J., T. Worrall (994) Foreign Direct Investment and te Risk of Expropriation, Review of Economic Studies 6, -08. [3] Tomz, M. and M. Wrigt (00) Sovereign Teft: Teory and Evidence about Sovereign Default and Expropriation, in: W. Hogan and F. Sturzenegger (eds), Te Natural Resources Trap, Cambridge: MIT-Press, [33] Venn, F. (986) Oil Diplomacy in te Twentiet Century, Houndsmill: Macmillan. [34] World Bank (009) Global Economic Prospects: Commodities at te Crossroads, Wasington: World Bank. [35] Yergin, D. (99) Te Prize, New York: Simon and Scuster. 0