Effects of Tariffication: Tariffs and Quotas under Monopolistic Competition Jan G. Jørgensen∗and Philipp J.H. Schr¨oder† February 2003
Abstract Recent rounds of GATT and later WTO have advocated widespread tariffication, meaning that existing non-tariff barriers be converted into import equivalent tariffs. From an economic point of view, the effects of such tariffication are not entirely clear. The paper presents a trade model with monopolistic competition to examine the welfare effects of tariffication. The ranking of pre- and posttariffication welfare depends crucially on the nature of the initial trade barrier and the tariff tool applied. Tariffication using a specific (an ad valorem) tariff results in the same (a reduced) welfare level compared to an initial sold quota, whereas welfare is increased (the same) compared to an initial shared quota. Key Words: Trade policy, tariffication, specific tariff, ad valorem tariff, quota, tariff rate quotas, VER. JEL: F02, F12, F13
1
Introduction
In the past ten years tariffication – the conversion of non-tariff barriers such as quotas and voluntary export restraints (VERs) into import equivalent tariffs – has been promoted and implemented on a global scale. Yet, the welfare effects of such policies have not been fully understood. The present paper addresses this issue for the case where industries feature monopolistic ∗
Department of Economics, University of Southern Denmark, Denmark. DIW Berlin, Germany. Corresponding author: Philipp J.H. Schr¨oder, DIW Berlin, German Institute for Economic Research, Department of International Economics, K¨onigin-Luise-Straße 5, 14195 Berlin, Germany, Tel.: +49 30 89789-692, Fax:+49 30 89789-108, E-mail:
[email protected]. †
1
competition. In particular the paper finds that, in terms of consumers’ utility, tariffication using a specific tariff is preferable to an ad valorem tariff, even though an ad valorem tariff generates more tariff revenue and all revenues are reallocated to the population. The initial trade regime – in particular, if and how rents accrue (sold versus shared quota) – in combination with the tariffication tool applied determines whether consumer welfare de- or increases under the process of tariffication. Different trade policy instruments have different impacts on the involved countries’ welfare and different visibilities, with a tariff being a clear and straightforward policy rule, while policies such as quotas or VERs are hard to quantify and hence blur the picture of the true level of protectionism (see Anderson (1988)). Because of this, GATT, and now its successor the WTO, have initiated major breakthroughs in tackling protectionism by – among other things – promoting widespread tariffication: ‘Members shall not maintain, resort to, or revert to any measures of the kind which have been required to be converted into ordinary customs duties, . . . ’ (Final Act of the Uruguay Round, 1994)1 The importance and impact of tariffication in the recent GATT (WTO) rounds has been discussed among others by Ingco (1996) and Nguyen et al. (1993) (see also OECD (1996) and WTO (1998, chapter 3)). Further, Lawrence (1989) provides an assessment of the significance of tariffication for US trade policy. The widespread tariffication trend can also be detected in the tariff levels of the major trading nations. Even though tariffs are generally falling, table 1 shows that, for example, for the sector of food, beverages and tobacco, the mid 1990s see a rise in the average tariff level. The reason is tariffication. Table 1: Average external tariff levels1 (percent) USA2 Japan EU15
Food, beverages and tobacco 1988 1993 1996 7.6 8.2 15.9 15.6 17.5 18.9 27.4 27.1 32.5
For all products 1988 1993 1996 4.4 4.7 5.2 4.2 3.6 3.4 8.2 8.4 7.7
Note: 1. Average tariff levels are estimated using production weights based on the destination countries’ composition of value added. 2. 1989 values instead of 1988. Source: OECD (1996), tables 1 and 2; OECD (1999) table 7.1; authors’ calculation. 1
Article 4, Agreement on Agriculture. The footnote to this article explains: ‘These measures include quantitative import restrictions, variable import levies, minimum import prices, discretionary import licensing, non-tariff measures maintained through statetrading enterprises, voluntary export restraints, . . . ’.
2
The issue of tariff-quota equivalence – the underpinning of tariffication – has received considerable attention in economics, see for example Bhagwati (1965). It is well known that bilateral tariff-quota conversions are welfare neutral for perfectly competitive markets if the quota rent accrues to domestic residents. A country may, however, benefit from a tariff if it can extract rents that under the quota regime accrued to foreign residents. For imperfectly competitive markets, however, tariff-quota equivalence becomes a more complex issue, see for example Helpman and Krugman (1989). Despite of the substantial literature on tariff-quota equivalence, or rather the lack thereof, few models exist that explicitly study the case of tariffication, and to the best of our knowledge there is no formal approach that addresses the issue of tariffication for industries that feature monopolistic competition. There are, however, a number of contributions that relate to the present work. Kowalczyk and Skeath (1994) have shown that in a setting where a country faces a foreign monopolist, ad valorem tariffs are welfare superior to specific tariffs. This result is opposed to our finding and driven by the fact that the ad valorem tariff is superior in terms of revenue extraction. On the other hand, in line with our result, Das and Donnenfeld (1987) show that for a country facing a foreign monopolist that has a quality choice, the specific tariff may generate the higher welfare. In a dynamic two-country game of setting optimal tariffs with retaliation, Lockwood and Wong (2000) show that the move from specific tariffs to ad valorem tariffs improves welfare in at least one country. Their model and in particular the mechanism driving it is very different from the present approach; again, what drives their result is the superiority of ad valorem tariffs in terms of revenue generation. As to the effects of tariffication Kaempfer and Marks (1994) present a model where the profitability of an exporting monopolist is affected by tariffication. They show that global efficiency – in the sense of purchasing from the lowest cost producer – may be reduced by the switch from a quota to a tariff. Again their setting is not one of monopolistic competition and increasing returns, rather, and contrary to the present model, producers vary as to their cost efficiency. Yet, all the above contributions (as well as Helpman and Krugman (1989) when addressing non-equivalence of ad valorem tariffs, specific tariffs and quotas) deal with situations of monopoly or oligopoly but not with symmetric situations of monopolistic competition – the topic of the present paper. The contributions most closely related to the present paper are those of Gros (1987) and Collie and Su (1998). Both deal with trade policies in monopolistic competition settings. Gros (1987) builds a two-country single industry model based on Krugman (1980). His central results concern welfare effects of retaliation in tariff wars. However, Gros (1987) focusses primarily 3
on ad valorem tariffs and does not address the welfare issues of tariffication. Collie and Su (1998) use the framework of Romer (1994) and find – in line with the present paper – that a VER (with rents realised by the foreign industry) might be welfare superior to an ad valorem tariff. In addition to being situated in a different framework, the present paper goes beyond the results in Collie and Su (1998) by including the case of specific tariffs as well as distinguishing how the quantitative trade restriction is administered, i.e. whether or not the rents of the initial quantitative trade restriction accrue (sold versus shared quota). To examine tariffication for industries that feature monopolistic competition we use a Krugman (1980) type model. We model two symmetric countries, with two industries, one of which is internationally traded. The paper is novel both in terms of treating tariffication explicitly and in terms of distinguishing the tariffication tools (specific versus ad valorem tariffs) as well as the nature of the initial quantitative import restriction (sold versus shared quotas). All policy tools are applied bilaterally and are compared according to the same exogenously given and binding import restriction.2 The differentiation into sold and shared quotas (e.g the classification of the initial quantitative import restriction regime) that is proposed in this paper is motivated as follows. Under a sold quota well defined property rights exists and rents accrue either to the home or the foreign country, i.e. the government or industry (home or foreign) has a clear property right to the quota and can control export volumes such that the quota rent can actually be realised. This case would occur in a situation of auctioned or sold import quotas or a VER that the foreign country administers via selling export licenses. Under a shared quota, on the other hand, property rights to the quota are not well defined, i.e. no one owns the quota and access is uncontrolled. Hence, the shared quota suffers from the ‘tragedy of the commons’, i.e. firms, by engaging in export activity, capture part of the overall quota from other exporting firms. Such a situation might typically occur under a VER if entry into the export activity cannot be controlled or under a regime of tariff rate quotas with very low within-quota tariffs and prohibitively high outquota tariffs. In the latter case all potential exporters can take a share of total export sales until the quota is filled, after this no further exports take place. With free entry and exit, such a tariff rate quota system will not generate any rents or tariff revenue (e.g. the within-quota tariff is zero). From the formal model the following results are derived. First, it is found 2
It should be noted that this paper does not deal with the emergence or rationale of the initial quantitative import restriction but with its conversion into a tariff.
4
that there is a significant difference in the welfare impact between specific and ad valorem tariffication. Even with complete reallocation of all tariff revenues, a specific tariff generates more consumer utility than an ad valorem tariff. This result is driven by the larger number of variants in the traded sector in the case of a specific tariff. The specific tariff allows for more firms to exist, because industry profitability is higher. Or put differently, as a revenue extractor, the ad valorem tariff is more efficient than the specific tariff. This is what drives the usual result of ad valorem tariffs being welfare superior to specific tariffs in models featuring monopoly or oligopoly market power (see Helpman and Krugman, 1989, chapter 4). However, under monopolistic competition, it turns out that this superior ability of the ad valorem tariff to extract revenue reduces industry profits, therefore, the number of firms is reduced, and this corresponds to lower welfare. Second, we establish that the change in consumers’ welfare under the process of tariffication depends crucially on the initial trade regime. If the import restriction is initially given by a shared quota (sold quota), then under the process of tariffication, welfare will be increased (remain unchanged) in case of a specific tariff and remain unchanged (be reduced) under an ad valorem tariff. In any case free trade dominates any initial quota situation as well as the post-tariffication situations in terms of consumer utility. Section two introduces the formal model. In section three we analyse the impact of the initial sold quota, shared quota and subsequent tariffication on quantities, prices and the number of variants. Section four presents the welfare results of the analysis. Section five concludes the paper.
2
The Model
We develop a setting of Chamberlinian monopolistic competition. The specific starting point for our model are its applications to international trade developed by Krugman (1980 and 1981). Assumptions of the model It is assumed that the world consists of two symmetric countries, each with two industries. In both countries market conditions are characterised by monopolistic competition, increasing returns to scale in production and differentiated goods. Each industry has a large number of potential variants which enter symmetrically into demand. For convenience we assume that the first industry is a non-traded industry and that the other industry is a pure export industry, i.e. the second industry in the home country exports 5
its entire output to the foreign country and vise versa. All our results can be extended to the case of actual intra-industry trade, however, what is decisive for our results to occur is some degree of market segmentation, i.e. the assumption that products within each industry are closer substitutes than products from home and abroad. The model adopts the utility function of Krugman (1981). However, for each industry we apply the functional form utilised in Krugman (1980). As the two countries are completely identical, we only show the specification for one of the countries throughout the analysis. All the individuals in the two countries are assumed to have the same utility function, U = ln
N X
θ
ci + ln
i=1
ˆ N X
cˆθj
(1)
j=1
where 0 < θ < 1 and ci is consumption of the ith variant of the foreign export industry and cˆj is consumption of the jth variant of the non-traded home industry. Due to symmetry between the two countries, the imports of ˆ one country equal the exports of the other country and vise versa. N and N define large numbers of potential variants in each industry. The number of variants actually produced n and n ˆ are assumed to be large, although smaller ˆ than N and N . For the moment we examine the properties of the export industry alone, bearing in mind that the free-trade equilibrium of the export industry is identical to the equilibrium of the non-traded industry. On the supply side we assume that there exists only one factor of production which is labour. All variants will be produced with the same cost function: li = α + βxi
i = 1, . . . , n
(2)
where li is labour used in the production of the ith variant in the traded industry and xi is output of that variant. This specification includes fixed costs α and constant marginal costs β and hence average costs decline at a diminishing rate. This assumption ensures that each variant is produced by only one firm, hence the number of variants equal the number of firms. Also since (2) implies that the labour requirements are identical for every variant, we can restrict our analysis to one variant, since all other variants will behave identically. Hence, in the remainder of the paper the subscript i can be omitted. Labour requirements are converted into nominal costs by multiplying (2) by the wage rate, w. The market clearing condition demands that the output, x, of each variant should be equal to the total consumption of all individuals in the economy 6
of that variant. We will assume full equality between the number of workers, L, and consumers. Hence, x = Lc (3) Equilibrium with free trade Finding equilibrium in this model follows the standard procedure assuming profit maximisation, free entry and exit of firms resulting in the zero-profit condition, labour market clearing at full employment, and goods market clearing (see e.g. Krugman 1980). The equilibrium is characterised by prices p, output per firm x, and the number of firms n. From (1) the demand an θ−1 individual firm faces is given as p = λθcP cθ , where λ is the shadow price on the budget constraint. It is assumed that – since there are a large number of goods – the pricing decision of any one firm has a negligible effect on λ. Then, using the fact that c = Lx , profits π = px − (α + βx)w are maximised by setting the price, βw (4a) p= θ Equating the profit-maximising price with the price implied by zero profits, p0 = (α+βx)w , gives the per firm output quantity, x, under free entry and exit: x x=
θα (1 − θ)β
(4b)
Finally, the number of firms actually producing in equilibrium can be deduced via market clearing conditions using the x just derived. In particular labour market clearing demands that ˆlˆ n + ln = L. More useful in our case is the following: Individuals, by maximizing utility (1), will attempt to use one-half . of their income on imports and one-half on home goods; hence, pxn = wL 2 Using the p and x just derived this condition determines the number of firms in equilibrium. (1 − θ)L n= (4c) 2α Since (4a–4c) characterises the export industry in both countries, it also states the import conditions. In fact, (4a–4c) also states the equilibrium in the non-traded sector under free trade. Thus we have pˆ = p, xˆ = x and n ˆ = n. Defining a restriction on import volume In order to model tariffication we need to define an initial restriction on imports. Noticing that the import volume of a country under free trade is 7
given by χ = nx =
Lθ , 2β
a restriction on imports is defined as: χ¯ = γχ =
γLθ 2β
0≤γ Uτ > Ut Proof is given in appendix A.2. Total consumer utility is larger under a specific tariff than under an ad valorem tariff (given the same trade restriction), yet both tariff regimes are dominated by free trade. The superiority of the specific tariff compared to the ad valorem tariff stems from the fact that under a specific tariff (and in fact also under a sold quota) more profits remain in the traded sector, allowing more firms to exist. Since consumers love variety, this policy generates the higher welfare, even though the total import restriction is the same. Given the results in proposition 1 and 2 one can draw conclusions on the impact of tariffication. Consumers prefer specific tariffication to ad valorem tariffication. However, whether the tariffication process is at all desirable from the consumers’ point of view depends on the nature of the initial import restriction. Corollary 1. (a) If the initial import restriction is given by a shared quota, then tariffication will at least be welfare neutral (using ad valorem tariffs) and at best be welfare improving (using specific tariffs). In particular, Uτ > Uv = Ut (b) If the initial import restriction is given by a sold quota, then tariffication will at best be welfare neutral (using specific tariffs) and at worst be welfare reducing (using ad valorem tariffs). In particular, Uτ = Uq > Ut From corollary 1 it follows that when the initial trade barrier (to be converted into a tariff) is constructed from a mix of sold and shared quantitative restrictions, then specific tariffication will be welfare increasing, while ad valorem tariffication will be welfare reducing.
5
Conclusion
The 1990s have seen widespread tariffication for the members of GATT and later WTO. This process of converting quotas, VERs, import licenses and other trade barriers into their tariff import-equivalents has been endorsed 16
by governments and economists alike. This paper argues that, in a world of monopolistic competition, tariffication may cause a welfare reduction, depending on the method of tariffication (ad valorem versus specific tariffs) and the nature of the initial trade protection regime (sold versus shared quotas). The distinction into sold and shared quotas is driven by the property rights to the quantiative restriction. In particular with clear property rights a quota can be sold and entry be limited such that rents accrue. However, without clear property rights – shared quotas – a quantitative restriction suffers from the ‘tragedy of the commons’ where excessive market entry reduces the sales of each firm such that neither rents nor revenue can be harvested. The model of the paper builds on Krugman (1980 and 1981). In a symmetric, two country, general equilibrium model, the case of bilateral tariffication of a sold quota and a shared quota is addressed. All results are obtained under the assumption of complete redistribution of all tariff revenues and quota rents. The main results of this analysis are that, under the assumption of monopolistic competition, sold and shared quotas and their import-equivalent tariffs can result in different effects in terms of prices, the number of firms/product variants and the output per firm. In particular, in the traded goods sector a sold quota (shared quota) results in higher (the same) prices, less (the same) output per firm, and an even higher (lower) number of firms than under free trade. Enforcing the same amount of total imports via an ad valorem tariff results in the same number of firms as the shared quota (hence, also the same output per firm), however prices are higher than in the shared quota case. Enforcing the same amount of total imports via a specific tariff results in exactly the same equilibrium as under the binding quota. In terms of the effect on consumers’ utility, it is established that utility under specific tariffication – though less than under free trade – is higher than under ad valorem tariffication; a result that is opposed to existing findings in oligopoly or monopoly settings. Furthermore, utility under ad valorem tariffication is identical to utility under a shared quota, whereas utility under specific tariffication is identical to utility under a sold quota. This paper has thus established that when evaluating the welfare impact of tariffication for industries that feature monopolistic competition, it is important to distinguish between both the tariff tool used and the initial trade regime. Nevertheless, despite the findings of this paper, ad valorem tariffs might be preferred to specific tariffs on grounds of transparency, ease of administration, and fairness – issues left aside in the present analysis. Yet what this paper has shown is that an undue reliance on ad valorem tariffication under the rules of WTO might have an opportunity cost in terms of the lost number of product variants and could even be welfare reducing. 17
A A.1
Appendix Impact on the Non-traded Industry
Both labour and spending power (including redistributed funds from quota rents and tariff revenue), stemming from the constrained export industry will move to the non-traded industry. Formally, the labour supply for the ˆ s , can be written as: non-traded industries, L ˆ s = 1 L + ∆Ls L 2
s = q, v, τ, t
(A.1)
The increase in available labour does not influence the equilibrium output and price of the individual firm in the non-traded industry and hence output and price is equal to pˆ and xˆ given implicitly in (4a) and (4b). This reallocation of labour is identical to an increase in market size, and thus only influences the equilibrium number of variants, and hence also total industry output. The equilibrium number of variants, n ˆ s , is found by using the labour market clearing condition for the non-traded industry given by: ˆ s = (α + β xˆ)ˆ L ns
s = q, v, τ, t
(A.2)
Combining (A1) and (A2) and inserting ∆Ls found in (8), (10) and (14), we find the equilibrium number of variants in the non-traded industry under the different trade policies: 2 − θγ n 2−θ n ˆv = n ˆ t = (2 − γ)n n ˆq = n ˆτ =
(A.3) (A.4)
From (A3) and (A4) it immediately follows that n ˆq , n ˆv , n ˆt, n ˆ τ > n and n ˆq = n ˆτ < n ˆv = n ˆt.
A.2 A.2.1
Proofs Proof that Uq , Uv , Uτ , Ut < U
Consumer utility under trade protection, but with full reallocation of tariff revenues and quota rents, is less than utility under free trade. Recall from proposition 1 that U τ = U q and U t = U v . Proof. U t < U . From (17c) it follows that U t = U + ln(2 − γ) + ln γ. Hence, one has to show that: K t = ln(2 − γ) + ln γ < 0 . 18
(A.5)
It follows from (A.5) that limγ→0 K t = −∞ and limγ→1 K t = 0. Since 2−2γ ∂K t = (2−γ)γ > 0, K t is monotone increasing in γ for all 0 < γ < 1, (A.5) is ∂γ fulfilled. Proof. U τ < U . From (17d) it follows that U τ = U +(2−θ) ln 2−θγ +θ ln γ. 2−θ Hence, one has to show that: 2 − θγ τ + θ ln γ < 0 . (A.6) K = (2 − θ) ln 2−θ It follows from (A.6) that limγ→0 K τ = −∞ and limγ→1 K τ = 0. Since 2(1−γ) ∂K τ = θ (2−θγ)γ > 0, K τ is monotone increasing in γ for all 0 < γ < 1, (A.6) ∂γ is fulfilled. A.2.2
Proof that Uτ > Ut
Tariffication with a specific tariff and complete reallocation of all tariff revenues gives higher consumer utility than tariffication with an ad valorem tariff. Proof. U τ > U t . From (17c) and (17d) it follows that: 2 − θγ τ t U > U ⇔ (2 − θ) ln + θ ln γ > ln(2 − γ) + ln γ 2−θ
(A.7)
Define the function: F (z) = (2 − z) ln
2 − zγ 2−z
+ z ln γ
(A.8)
If F (z) is monotone decreasing in z, then b > a implies F (a) > F (b), and hence (A.7) is fulfilled as 1 > θ. From (A.8) it follows that ∂F (2 − z)γ 2(1 − γ) = ln + (A.9) ∂z 2 − zγ 2 − zγ One has to show that for a given z (A.9) is negative for all γ, 0 < γ < 1. It follows from (A.9) that ∂F ∂F = −∞ and lim =0 γ→0 ∂z γ→1 ∂z lim
and since
(A.10)
∂ ∂F 4(1 − γ) ∂z = (A.11) ∂γ γ(2 − zγ)2 > 0 (A.9) is monotone increasing in γ for all z and thus negative in the relevant parameter intervals, such that (A.8) is monotone decreasing and hence U τ > U t holds. 19
References Anderson, J., 1988. The Relative Inefficiency of Quotas. MIT Press. Anderson, J., 1997. Revenue Neutral Trade Reform with Many Households, Quotas and Tariffs. NBER Working Paper, No. 6181. Bhagwati, J., 1965. On the Equivalence of Tariffs and Quotas, in: Baldwin, R. et al. (Eds.) Trade, Growth and the Balance of Payments. Rand McNally, Chicago. Collie, D. R., Su Y. 1998. Trade policy and product variety: When is a VER superior to a tariff? Journal of Development Economics, Vol 55, 249–255. Das, S., Donnenfeld,S., 1987. Trade Policy and its Impact on Quality of Imports: A Welfare Analysis. Journal of International Economics, Vol 23, 77–95. Gros, D., 1987. Protectionism in a Framework with Intra-industry Trade: Tariffs, Quotas, Retaliation, and Welfare Losses. International Monetary Fund Staff Papers, Vol. 34 (1), 86–114. Helpman, E., Krugman, P., 1989. Trade Policy and Market Structure. MIT Press. Ingco, M., 1996. Tariffication in the Uruguay Round: How Much Liberalisation?. World Economy, Vol. 19 (4), 425–46. Kaempfer, W., Stephen, M., 1994. The Possibility of Inefficient Liberalisation through Tariffication. Review of International Economics, Vol.2 (2), 123–130. Kowalczyk, C., Skeath, S.E., 1994. Pareto Ranking Optimal Tariffs under Foreign Monopoly. Economic Letters, Vol. 45, 355-359. Krugman, P., 1980. Scale Economies, Product Differentiation, and the Pattern of Trade. American Economic Review, Vol. 70 (5), 950–959. Krugman, P., 1981. Intraindustry Specialisation and the Gains from Trade. Journal of Political Economy, Vol.89 (5), 959–973. Lawrence, R., 1989. Protection: Is There a Better Way? American Economic Review, Vol. 79 (2), 118–126.
20
Lockwood, B., Wong, K.-Y., 2000. Specific and Ad Valorem Tariffs are Not Equivalent in Trade Wars. Journal of International Economics, Vol. 52, 183–195. Nguyen, T., Perroni, C., Wigle, R., 1993. An Evaluation of the Draft Final Act of the Uruguay Round. The Economic Journal, Vol. 103, 1540–1549. OECD, 1996. Indicators of Tariff and Non-tariff Trade Barriers. Paris. OECD, 1999. OECD Economic Outlook No.65. Paris. Romer, P., 1994. New goods, old theory, and the welfare costs of trade restrictions. Journal of Development Economics, Vol 43, 5–38. WTO, 1998. Annual Report 1998. World Trade Organization, Geneva.
21
