Transcription

Tariff Overhang and Aid:Theory and EmpiricsOliver Lorza and Susanna ThedebNovember 25, 2020AbstractIn this paper, we consider aid payments as a possible explanation for tariff overhangs. We set up a theoretical model in which rich countries use development aidto pay for tariff concessions. Developing countries, in turn, may anticipate sucha policy in the negotiations for tariff bindings. Setting the bound tariff rate at arelatively high level can serve as a mechanism to incentivize rich countries to carryon with aid payments in the subsequent “aid for trade” game. We empirically examine this hypothesis using detailed data on WTO members’ bound and appliedtariff rates under the Uruguay agreement. The data sample contains almost allaid recipients participating in the Uruguay round negotiations. Our results providestrong support for the model’s prediction that larger tariff overhangs are implemented by countries that receive more aid.Keywords: Foreign aid, tariff binding, tariff overhangJEL classification: F13, O19.aRWTH Aachen University, School of Business and Economics, 52056 Aachen, Germany, [email protected] of Malta, Institute for European Studies, 2080 Msida, Malta, Örebro University Schoolof Business, Sweden, [email protected] Corresponding author.

1IntroductionMost world trade takes place under WTO (and former GATT) regulation: 75 percentof all countries are members of the organization and a further 11 percent are underobservation to acquire membership. Almost all tariffs are thereby subject to regulationset in multilateral trade negotiations. The high participation in multilateral agreementsreflects the success of the multilateral trade negotiation system in liberalizing world trade.Thus, it may be surprising that many countries, especially in the developing world, havecommitted to tariff ceilings, so-called “tariff bounds”, which exceed applied rates on mostproducts so that much of world trade takes place under nonbinding tariff constraints (see,e.g., WTO, 2009).This observation of a positive “tariff overhang” is difficult to reconcile with standardtheoretical hypotheses on incentives for governments in trade negotiations. If specialinterest pressure determines protection, tariff outcomes of trade negotiations reflect thestakes of domestic and foreign industries (Grossman and Helpman, 1995). If tariffs arisefrom a motivation of manipulating the terms-of-trade to reap unilateral trade gains, tradenegotiations can result in an internalization of these terms-of-trade externalities (Bagwelland Staiger 1999). In both cases, there seems to be no reason to negotiate bound tariffsthat exceed applied tariff rates in noncooperative equilibrium.The negotiation of tariff ceilings instead of actual tariff rates can be explained, however, if additional factors are taken into account. If interest group pressure influencesdomestic tariff policies and if capital is inter-sectorally mobile, governments can opt fortariff bindings to counteract investment distortions (Maggi and Rodrı́guez-Clare 2007).Other reasons for negotiating tariff ceilings are contracting costs under uncertainty (Hornet al., 2010) or a trade-off between commitment in negotiations with trade partners andflexibility to respond to future changes in political pressure by special interests underthe agreement (Bagwell and Staiger, 2005). The tariff overhang is lower under higherimporter market power because of stronger enforcement to reduce the tariff bound in negotiations and a starker terms-of-trade improvement of applied tariff protection (Beshkaret al., 2015).In this paper, we examine an explanation for tariff bounds that hinges on the fact thatdeveloping countries are overrepresented in the use of tariff overhangs and regularly adoptlarger tariff overhangs than other countries. According to our hypothesis, developmentaid may serve as an instrument to influence tariffs of developing countries. Rich countriesmay buy access to poorer countries’ markets by promising aid payments in return for tariff concessions. Developing countries, in turn, anticipate such a policy when negotiatingthe multilateral trade agreement. Setting the bound rate at a high level then serves as1

a mechanism to incentivize rich countries to carry on with aid payments in the subsequent “aid for trade” game. We empirically examine this hypothesis using detailed dataon WTO members’ bound and applied tariff rates under the Uruguay agreement. Ourcountry sample includes 95 percent of the aid recipients that were WTO members at thetime. The results provide strong support for the predictions that countries with higheraid receipts adopt larger tariff overhangs because they negotiate higher tariff bounds anduse lower applied tariffs.The rest of this paper is structured as follows: In the next section, a description ofrelated studies is provided to place our contribution in perspective. A theoretical modelthat illustrates main mechanisms at work is presented in section 3. A detailed overviewof the tariff data is given in section 4 and the empirical investigations are presented insection 5. The last section concludes.2BackgroundMarket power is a central feature of multilateral trade negotiations. A classic argumentfor tariff protection is that the government of a country with market power can restrictimports to improve its terms-of-trade. As this argument also holds for the foreign trading partners as well, countries are likely to set tariffs at inefficiently high levels, incurringlosses in noncooperative equilibrium. Governments acting under social welfare motivesmay therefore forge international trade agreements to enforce mutual tariff reductions.Two pillars of the multilateral trade negotiation system counteract terms-of-trade manipulation: the nondiscrimination principle and the reciprocity principle (Bagwell andStaiger, 1999). Empirical evidence shows that elasticities of foreign export supply affecttariff rates that are not subject to WTO regulation (Broda et al., 2008) so that largertariff cuts are incurred upon WTO membership if importer market power is stronger(Bagwell and Staiger, 2011). However, protectionism to raise the terms-of-trade maynot be completely eliminated by the multilateral trade negotiation system due to variousreasons: exceptions from nondiscrimination in its regulatory framework (Bagwell andStaiger, 1999), free riding of exporters with low stakes on the most-favored-nation tariff(Ludema and Mayda, 2013),1 and nonbinding MFN tariffs (Beshkar et al., 2015; Nicitaet al., 2018).That industry-specific interests may influence tariff outcomes is well established inthe political-economy literature.2 Governments may take into account interests of the1Recent evidence by Ludema et al. (2019) indicates that the latter effect has been counteracted bythe formation of preferential trade agreements, pointing to a building-block effect of preferential tradeliberalization.2See, e.g., Hillman (1982) or Grossman and Helpman (1994). For an application of the Grossman-2

import competing industry in addition to implications for the aggregate welfare of theirconstituency. The resulting tariff rate is higher at a less elastic import demand due tolimited deadweight loss, and it is higher at a larger ratio of domestic output to importsbecause of larger gains at stake of industry-specific interests and lower welfare costs(Grossman and Helpman, 1994).3 In large countries, the tariff rate is also higher at aless elastic foreign export supply due to the additional incentive to manipulate terms-oftrade (Grossman and Helpman, 1995). While the terms-of-trade effect can be neutralizedby effective negotiation, the uneven political influence of interest groups is certain tofilter into the agreement. A higher negotiated tariff rate results if industry interests inthe importing country exert stronger political power over their government compared tocorresponding interests in the exporting country.Governments can opt for negotiating tariff ceilings (weak bindings) instead of actualtariff rates (strong bindings) to ensure that they have the discretion to respond shouldshocks appear that affect policymaking constraints.4 Weak bindings enable governmentsto levy lower tariffs and to incur higher national welfare in the absence of such events.Contracting costs can be important in explaining the negotiation focus on weak tariff bindings under uncertainty about future conditions (Horn et al., 2010). Bagwell andStaiger (2005) and Amador and Bagwell (2013) model the trade-off between allowing governments to react to the political influence of special interests and restricting the abilityto manipulate the terms-of-trade. They show how a tariff ceiling may arise endogenouslyfrom this trade-off. A tariff ceiling also preserves the political influence of industry interests and reduces the net returns from influencing the negotiation, which counteractsinefficiencies arising from a distorted allocation of capital (Maggi and Rodrı́guez-Clare,2007).Tariff overhang reflects flexibility in policymaking, which is utilized because the future political influence of import-competing producers is uncertain (Bagwell and Staiger,2005). Policymakers face political uncertainty as the demand from the import-competingsector varies over time (due to changing production conditions within and between industries in this sector). Governments in countries where this variability is larger arethereby expected to implement larger tariff overhangs. The tariff overhang decreaseswith importer market power as the terms-of-trade externality of protection stimulatesnegotiation partners to exert more effort to reduce the bound rate and leads to a higherapplied tariff rate under the agreement (Beshkar et al., 2015). Additionally, the gainsHelpman model to a developing country context, see Mitra et al. (2002).3More recently, the empirical relevance of this hypothesis has been placed under scrutiny by Imai etal. (2009; 2013), who show that testing the model using quantile regressions overturns its support anduncovers a positive link between protection and import penetration.4In a related setting, Busch and Pelc (2014) compare the use of tariff bindings to that of trade remediesin the WTO.3

from negotiating binding tariff reductions compared to contracting costs are increasingin importer market power (Nicita et al., 2018).The relationship between foreign aid and tariffs of recipient countries has been analyzed before. To our knowledge, however, none of the existing literature deals with itsimplications for tariff bindings. In Lahiri and Raimondos-Møller (1997), aid increasesdemand for goods exported by the donor. Due to this effect, the donor country allocatesmore aid to a recipient country that has a low tariff rate. Lahiri et al. (2002) extend thisanalysis allowing the donor country to commit to aid payments that are contingent onsubsequently set tariff rates by the recipients. Nanivazo and Lahiri (2015) analyze theimplications of conditional aid that is given as a prize depending on the tariff policy ofrecipient countries.3The modelOur model characterizes in stylized form international trade agreements between developed and developing countries, incorporating voluntary foreign aid payments. We consider two countries, one in the North N and one in the South S. The country in N exportsa final good q to the country in S on which the government in S may set an import tariffwith tariff rate τ 0. The political objectives of the government in N and of its tradingpartner in S with regard to the tariff are given in reduced form by V N (τ ) and V S (τ ).These objective functions can be interpreted as representing aggregate welfare but mayalso incorporate political economy elements as outlined in the preceding discussion of theliterature. Both objective functions are twice differentiable. We furthermore assume thatVN monotonically declines in the tariff rate τ , i.e., VτN (τ ) 0, whereas VS first increasesand then declines in τ such that there exists a strictly positive optimum tariff rate fromthe view of the importing country. Let this optimum tariff be denoted by τ̂ 0, i.e.,VτS (τ̂ ) 0. Due to the negative spillover of the tariff to the exporting country, the tariffrate that maximizes the joint payoff of both countries would be accordingly lower thanτ̂ . With τ denoting this jointly optimal tariff rate, we have τ τ̂ .For the role of development aid, we consider a setting in which aid is not merely alump-sum transfer from rich to poor countries but has positive allocative effects in theSouth. More precisely, we assume a development project that requires aid paymentsof a from country N and yields a benefit of b βa to country S, with β 1. Withaid, the per period payoff for the North is W N V N (τ ) a and that for the South isW S V S (τ ) βa.In a static non-cooperative setting without international agreements, country S wouldset the tariff at the optimum rate τ̂ , whereas country N would pay no development aid.4

By negotiating a trade agreement alone, both countries could reduce the tariff to thejoint optimum rate τ , but there still would be no tariff overhang or aid payment. Theoutcome may change if the aid and tariff game is repeated. With an infinite time-horizon,a cooperative solution can be supported by a subgame perfect equilibrium (see, e.g., Dixit1987, or Bagwell and Staiger, 1990). In our model, tariff concessions by the South canbe incentivized by aid payments from the North and vice versa.5 To characterize suchself-enforcing agreements, we consider the following setting: In an initial period t 0,both countries negotiate a weak tariff binding τ b τ̂ and from period t 1 on, countryS can set the cooperative tariff rate τ c or the tariff bound τ b while country N decideswhether to pay aid a or not. We consider trigger strategies in which country S sets thetariff at the cooperative level τt τ c and country N pays aid at a, but each countryreverts to noncooperative policies τt τ b and at 0 for the remainder of the game if therespective other country has deviated from its cooperative policy.6Given these trigger strategies, we can determine the necessary conditions for enforcement of the cooperative outcome. With δ as discount factor (0 δ 1), country S Pdoes not deviate from the trigger strategy as long as V S (τ b ) βa δ t 1 V S (τ b ) t 2 Pδt 1Sc[V (τ ) βa]. Rearranging yieldst 1Constraint S:V S (τ b ) V S (τ c ) δβa .(1) PPSimilarly, country N does not deviate if V N (τ c ) δ t 1 V N (τ b ) δ t 1 [V N (τ c ) a] ,t 2t 1orConstraint N:δ[V N (τ c ) V N (τ b )] a .(2)Both constraints would be satisfied as equalities for a 0 and τ b τ c . If there were noaid payments, the tariff binding could be lowered to the cooperative tariff rate. Tariffconcessions can already be part of the negotiated tariffs such that there would be noneed for a tariff overhang for their enforcement. North and South would then set tariffsdirectly at the joint optimal rates, and there would be no tariff overhang. The situation isdifferent for aid payments that are not part of the multilateral agreement. With a strictlypositive aid level, constraint N requires a tariff overhang (τ c τ b ) to ensure incentivecompatibility for country N. Similarly, according to constraint S, incentive compatibility5Another possible argument for linking trade agreements with foreign aid is analyzed by Maoz et al.(2011), who consider an endogenous growth model in which aid improves the international allocation ofcapital.6For certain parameter constellations, a self-enforcing agreement may also exist without a tariff binding. In such a setting, the non-cooperative tariff in the trigger strategy would be given by τt τ̂ . Aswill become clear below, a tariff binding below τ̂ may improve upon this possible outcome.5

for country S requires aid payments if there is a tariff overhang.Define τ Sc τ Sc (τ b ) as the critical tariff binding at which constraint (1) for the Southis satisfied as an equality. As the l.h.s. of (1) is decreasing in τ c for a given bound rateτ b , τ Sc determines a lower limit for τ c . It specifies the minimum cooperative tariff thatcan be set without inducing country S to choose the bound tariff rate instead. Similarly,we can define a critical tariff binding for the North τ Nc τ Nc (τ b ), which determines anupper limit, i.e., the maximum cooperative tariff that is incentive compatible for countryN. At a higher cooperative tariff rate, country N would choose to pay no aid given thatcountry S plays the trigger strategy. Since τ Sc determines a lower limit and τ Nc specifiesan upper limit for the cooperative rate, the two incentive compatibility constraints canonly be satisfied jointly for τ Sc τ Nc . Both critical values of τ c are increasing in thecooperative tariff rate τ b :dτ N cVτN (τ b ) 0 dτ bVτN (τ N c )anddτ ScVτS (τ b ) 0. dτ bVτS (τ Sc )(3)Figure 1 illustrates a scenario in which both constraints are satisfied and shows howthe critical values of τ c depend on the bound rate.7 For τ c τ , governments have anincentive to jointly raise the cooperative rate. Thus, τ c is chosen as high as possiblesuch that (2) becomes binding and τ c τ Nc (red segment of the τ Nc line). For τ c τ ,governments set τ c as low as possible such that constraint (1) binds and τ c τ Sc (redsegment of the τ Sc line).τcτNcτScτ*τ*τbFigure 1: Incentive Constraints7In figure 1, τ Nc τ Sc is assumed for all relevant values of τ b . Another possible outcome would bethat both lines intersect such that both constraints are satisfied only for a certain range of τ b . Moredetail is provided in appendix a.6

An increase in the aid payment shifts both constraints to the right yielding1dτ cδβdτ cc 0forτ τand S c 0 for τ c τ .NcdaδVτ (τ )daVτ (τ )(4)To summarize, the applied tariff rate that satisfies incentive compatibility increases inthe bound rate while for a given tariff binding it declines in the level of aid. In themodel developed so far, governments can try to induce the joint optimal tariff as appliedtariff rate, i.e. τ c τ , by setting the bound rate accordingly. If both countries stickto the trigger strategies, the bound rate is never actually chosen and therefore does notinduce any payoff costs for both governments. This outcome in which the bound ratenever materializes and countries can expect with certainty that they will be able to implythe cooperative solution, however, appears to be somewhat unrealistic. Instead, we maythink of several real-world situations in which self-enforcing contracts do not come intobeing. For example, it could be that (i) political conflicts might prevent aid paymentsbetween two countries, that (ii) no appropriate aid projects can be found, or (iii) thatsome governments may be short sighted or regarded as unreliable, etc. To account forthis possibility in a tractable and straightforward manner, we assume in the followingthat a self-enforcing mechanism for trade liberalization and aid can be established onlywith a certain ex ante probability ρ. Otherwise, the country in the South sets the tariffat the bound rate τ b and receives no aid. In addition, we allow side-payments betweenN and S in the tariff negotiations. With these assumptions, governments set the boundand cooperative tariff rates such that the expected aggregate per period payoff of bothcountries is maximized, which is defined as EW ρ[VN (τ c ) VS (τ c ) (β 1)a] [1 ρ][VN (τ b ) VS (τ b )].In this setting, an interior optimum can only exist if τ c τ . Otherwise both countriescould increase EW by negotiating a lower tariff binding that also induces a lower τ c andthereby raise their aggregate payoff.8 Only the incentive compatibility constraint forthe North (2) is binding in such an interior optimum. The optimum tariff binding andthe resulting cooperative tariff rate can be found by maximizing EW over τ b given thatτ c τ Nc (τ b ) according to constraint (2) in equality form. From the first-order conditionsfor an interior solution of this problem, we obtain the following expression:[ρ 1]A(τ b ) ρA(τ c ), with A(τ ) 8VτN (τ ) VτS (τ ).VτN (τ )(5)In our model, governments prefer to keep tariff bindings low as these bindings may determine theactual tariff rates. In a framework that incorporates risk about tariff rates, low bindings may also improvewelfare due to stimulating market entry and trade (Sala et al., 2010; Handley, 2014).7

It is assumed that Aτ (τ ) 0, which is sufficient for the second order condition to besatisfied (as shown in appendix a). Conditions (5) and (2) determine both tariff rates,implying τ c τ τ b .9 This outcome results from the fact that with probability πcountries do not cooperate and set the predetermined bound rate.While equation (4) has shown the effects of a change in the aid level ex post, i.e., afterthe bound rate has been set, equations (5) and (2) can be used to determine the ex anteinfluence of aid on the bound rate and the resulting applied rate. Totally differentiatingthese equations yieldsρAτ (τ c )dτ b 0.daρδAτ (τ c )VτN (τb ) (1 ρ)δAτ (τ b )VτN (τc )(6)Ex ante, the influence of a change in the aid level is smaller than the ex-post effect setout in expression (4) but still negative:dτ c(1 ρ)Aτ (τ b ) 0.daρδAτ (τ c )VτN (τb ) (1 ρ)δAτ (τ b )VτN (τc )(7)To obtain further comparative static results, we specify a simple partial equilibriumsetting with linear demand and supply curves as in Bagwell and Staiger (2005) or Beshkarand Bond (2017). The model accounts for importer market power as well as politicaleconomy motives as reasons for positive import tariffs. Demand in S for the import goodq is given by dS n(1 pS ), producers in S supply the good according to q npS /2,producers in country N supply the good according to qN pN , and demand in N is givenby dN 1 pN . The term n accounts for the relative market size (and market power) ofS. The tariff inserts a wedge between prices in S and in N, i.e., pS pN τ . Equilibriumprices and imported quantities of S, mS dS qS , are given bypS 2 2n 3nτn(1 6τ )2 2n 4τ, pN , and mS .4 3n4 3n4 3n(8)Imports are positive as long as τ 1/6. Producer and consumer surplus in the Northdepend on the tariff rate according toπN 9(2 2n 3nτ )2(2 n 3nτ )2dandπ .N2 (4 3n)22 (4 3n)2(9)As constraint (2) requires τ b τ c , equation (5) can only be satisfied if VτN (τ b ) VτS (τ b ) 0 and VτS (τ c ) 0.VτN (τ c )8

Producer and consumer surplus as well as tariff revenues in the South are given byn (2 n 4τ )2τ n (1 6τ )n (1 n 2τ )2 d, πS , and TS πS .224 3n(4 3n)2 (4 3n)(10)The government in N maximizes the sum of producer and consumer surplus. The government in S considers producer surplus, consumer surplus and tariff revenues in its objectivefunction. We assume that the government in S places a relatively higher weight on thesurplus of its import-competing producers due to political considerations. That is, theobjective function of the government in the South with regard to this particular tradingdrelationship is VS λπS πSd TS , with λ 1, while that in the North is VN πN πN.An increase in the tariff rate has the following effects on government objectives in countryN:3n2 (6τ 1).(11)VτN (τ ) (4 3n)2According to (11), VτN is negative for all non-prohibitive τ . The marginal influence ofthe tariff rate on V S (τ ) given byVτS (τ ) 4(λ 1)n(1 n) 3n2 4nτ [2(λ 1) 3(2 3n)].(4 3n)2(12)We assume that λ is not too large, such that the term in squared brackets in the numeratorof (12) is negative.10 With this assumption, V S (τ ) is first increasing and then decreasingin τ and has an interior maximum atτ̂ 4(λ 1)(1 n) 3n,12(2 3n) 8(λ 1)(13)the optimum tariff rate from the view of an individual country in S. From the view ofboth countries together, V S (τ ) V N (τ ) has its maximum at:11τ 2(λ 1)(1 n).12 9n 4(λ 1)(14)While for λ 1 free trade (τ 0) would be jointly optimal, the tariff rate τ is positivefor λ 1. The joint optimal tariff rate increases in the political weight of the importerindustry λ and in the market size of the importer country n. Given the functionalspecifications of our example, the incentive compatibility constraints for country S and10This term is negative if 2λ 8 9n. The requirement is satisfied for all positive n if λ 4.For an interior optimum, the joint optimum tariff has to be lower than the prohibitive tariff. With(14), it can be shown that this requires λ 7/4.119

country N can be written in equality form asn(τ b τ c )[4(λ 1)(1 n) 3n 4(λ 4 6n)(τ b τ c )] δβa(4 3n)2(1’)3δn2 (τ b τ c )[1 3(τ b τ c )] a,(4 3n)2(2’)andwhile the term A(τ ) in the first order condition (5) becomesA(τ ) 4(λ 1)(1 n)(τ τ ).3nτ (1 6τ )(5’)Differentiating equation (5’) reveals that the requirement Aτ 0 is satisfied in thismodel specification. Figure 2 shows how the bound tariff rate in an interior optimumis determined. From the objective function WE, we can derive iso-payoff curves, whichare ellipses around the unconstrained optimum at which τ b τ c τ . In an interiorequilibrium, governments choose a point on the reaction curve τNc at which the aggregatepayoff is maximized. The tangency point between an iso-payoff curve and the τNc -curvedetermines the optimal bound rate and the applied tariff rate.τcτNcτScτ*τcτ* τbτbFigure 2: Tariff OutcomeFor a comparative static analysis of the tariff outcome, we begin with the situationex post. In this case, the incentive constraint (2’) for country N determines the applied10

rate. Totally differentiating (2’) reveals the following effects:dτ c(4 3n)21 6τ bdτ c 0, 0,dτ b1 6τ cda3δn2 (1 6τ c ) 8(τ b τ c ) 1 3(τ b τ c )dτ c 0.dnn (4 3n) (1 6τ c )and(15)An increase in the bound rate raises the applied rate, but as dτ c /dτ b 1, the tariffoverhang increases. An increase in the aid level makes it more costly for the Northto stick to the cooperative outcome. As a result, incentive constraint N shifts to theright, and the applied rate has to be lowered accordingly. In contrast, a larger marketsize of country S makes cooperation more attractive for country N such that its incentiveconstraint shifts to the left and the applied rate increases. Finally, for a given bound rate,the political weight of importers λ in country S does not affect the incentive constraintfor country N and therefore has no impact on the applied rate.Ex ante, governments can adjust the bound rate in response to a change in exogenousvariables. If the aid level increases, governments raise the bound rate to limit the declinein the applied rate ex post and arrive at a new tangency point with the shifted incentiveconstraint for N, i.e., dτ b /da 0.12 An increase in the political weight has consequencesfor the joint optimum tariff and the iso-payoff curves. Governments prefer a higher jointtariff τ such that the tangency point with the incentive constraint shifts upwards and tothe right, resulting in a higher bound rate, dτ b /dλ 0. An increase in the market powermay raise or reduce the bound rate.4Tariff dataTo empirically examine our model, the investigation is performed using Uruguay agreement tariff data for WTO aid recipients. Specifically, we use 2005-2013 tariff data as theUruguay agreement was implemented in its entirety in the year 2005 and was succeededby the Doha agreement in 2014.13 The tariff data is from the most comprehensive anddetailed data set available, the UNCTAD TRAINS data base. Our data includes allcountries participating in the Uruguay negotiations and most of the countries that wereWTO members by 2013.14 In this section, we start out describing product tariff data forWTO members under the Uruguay agreement and then turn to present product tariff12See appendix b for a derivation of these results.The implementation of the agreement started on 1st January 1995 and ended for sensitive agriculturaland textile products on 1st January 2005.14Data is lacking for countries that became WTO members in 2012 or 2013 (Laos, Montenegro, Russia,Samoa, Tajikistan and Vanuatu).1311

characteristics for aid recipients in this category and our country sample in particular.Product tariff data is reported in the common nomenclature used in negotiations, the6-digit level of the Harmonized System classification.15The product tariff overhang equals the gap, in percentage points, between the importer’s bound (BND) and most-favored-nation (MFN) product tariff. This gap can bepositive, equal to zero in case of binding MFN rates, or negative. Almost all tariff implementations fall into the first two categories in adherence to WTO regulation depictingthat a country can only temporarily deviate from negotiated tariff rates (using ‘safeguard measures’ to avoid serious damage otherwise caused by an import surge). BindingMFN product tariffs (zero overhang) are reported for 20 percent of the observations and76 percent of MFN product tariffs are set below their bounds. The share of bindingtariffs accounts for 48 percent in developed (OECD) countries compared to only 15 percent in developing (non-OECD) countries, which is consistent with stylized evidence thatdeveloping countries are overrepresented in the utilization of tariff overhangs (see, e.g.,WTO, 2009). There is a strong correlation between the tariff overhang and the boundtariff (ρ 0.88) and a weak correlation between the tariff overhang and the MFN tariff(ρ 0.03), implying that the use of bounds gives governments considerable flexibility.Countries often do not use any of the leeway given by set bounds; zero MFN tariffsare adopted for almost 1 out of 5 products. In contrast, tariff bounds are

Tari Overhang and Aid: Theory and Empirics Oliver Lorza and Susanna Thedeb November 25, 2020 Abstract In this paper, we consider aid payments as a possible explanation for tari over-hangs. We set up a theoretical model in which rich countries use development aid to pay for tari concessions. Developing countries, in turn, may anticipate such