Pelle, Nes""kh and semMMir L Internatnal Trade International Economics Department The World Bank February 1991 WPS 597 Rent Sharing in the. Multi-Fibre Arrangement Theory and Evidence from U.S. Apparel Imports from Hong Kong Refik Erzan Kala Krishna and Ling Hui Tan The actual cost of the multi-fibre arrangement (MFA) quotas to exporting developing countries could be considerably higher than conventional estimates that assume that exporters seize all the scarcity rents. For U.S. apparel imports from Hong Kong, the authors' findings point to a 50-50 sharing of the rents between the exporters and the importers. 1hs Policy. Research, and External Affairs Complex distributes PRE Working Papers to disseminate the findings of wosk in progress and to encourage the exchange of ideas among Bank staff and all otheri interested in development issues. These papers carry the naes aof the authors, reflect only their views, and should be used and cited accordingly. The findings. interpretations, and conclusions ae the authors' own. They should not be attributed to the World Bank, its Board of Directors, its managenant, or any of its member coanries. Polle. Mt*a and mWalernal Allaire International Trade WPS 597 This paper- a product of the International Trade Division, International Economics Department -is part of a larger effort in PRE to quantify the effects of protectionism, particularly in the developing countries. Copies are available free from the World Bank, 1818 H Street NW, Washington DC 20433. Please contact Grace Ilogon, room S7-033, extension 33732 (70 pages). Available estimates of tariff equivalents of model. The authors' method essentially wts quotas and welfare calculations on the costs of whether the license price inclusive Hong Kong MFA quotas for developing countries are based price, adjusted for tariffs and transport costs, is on the premise of perfect competition in both equal to the domestic (U.S.) price. A deviation product and license markets. It is also assumed between the two prices indicates rent sharing. that the exporting countries that administer the MFA quotas receive all the scarcity rents. Erzan, Erzan, Krishna, and Tan test the hypothesis Krishna. and Tan argue that, in the presence of with homogenous goods, modify it to take into market power on the buyers' side in the product account compositional differences, and, finally, markets combined with concentration in the consider differentiated goods. They find that license markets, the importing countries might historical data do not conform with the predic- retain part of this rent - that is, share it with the tions of the competitive model. There is strong exporters. Although the impact of imperfect evidence that importers retain a substantial competition on rent appropriation - non- portion of the MFA quota rents. The determi- equivalence of tariffs and quotas, and the size of nants of the price differential are also studied. the rent - has been analyzed in literature, rent sharing has so far been ignored in both analytical Rent sharing substantially affects the and empirical work. estimated magnitude of welfare losses that exporting developing countries suffer because of The paper makes a theoretical case for rent MFA quotas - and, for that matter, because of sharing, and then analyzes U.S. imports of any voluntary export restraint (VER) in other apparel products from Hong Kong. It does not sectors. Not only do these countries have specify a particular model of imperfect competi- reduced export volumes, but, contrary to the tion but investigates whether the data conform to prevailing wisdom, they do not receive all the all the relevant predictions of the competitive scarcity rents the quotas generate. The PRE Working Papcr Set ies disseminates the findings of work under way in the Bank's Policy. Research, and External AffairsComplex. An objpctivc of the series is to get these findings out quickly, even if presentations are less than fully polished. T"he findings, interpretations, and conclusions in thcse papers do not necessarily rpresent official Bank policy. Produced by the PRE Dissemination Center "RENT SHARING" IN TE MULTI-FIBRE ARRANGRMENTO Theory and Evidence from US Apparel Imports from Bong Kong by Refik Ersan, Kala Krishna and Ling Hui Tan* TABLE OF CONTENTS 1. INTRODUCTION 2. THE BASIC MODEL 6 2.1 Perfect Competition 6 2.2 Market Power on the Seller's Side: 6 "Rent Appropriation" 2.3 Market Power on the Buyer's Side: 10 "Rent Sharing" 3. THE DATA 15 4. HONG KONG'S TEXTILE QUOTA SYSTEM 17 5. TESTING FOR RENT SHARING: HOMOGENOUS GOODS 20 6. ALLOWING FOR A COMPOSITION EFFECT 29 7. ALLOWING FOR DIFFERENTIATED PRODUCTS 36 8. DETERMINANTS OF THE US-HONG KONG PRICE DIFFERENTIAL 42 9. CONCLUSIONS 45 END NOTES 48 REFERENCES 51 DATA APPENDIX i 1. Data Source and Product Group Definitions 2. Estimations for Missing Observations xv 2.1 Production data xv 2.2 Import data xv End Notes For Data Appendix xvii *Ref ik Ersan, the World Bank; Kala Krishna, Harvard University and NBR; and Ling Hui Tan, Harvard University. We would like to thank Ronald Chan, Carl Hamilton, Paula Holmes and Yun-Wing Sung for providing us with valuable data; Philip Swagel for helpful comments; and Monique Skrusny and Christopher Holmes for excellent research assistance. 1 1. Introduction The post-World War II era is often hailed as a period of great trade liberalization -- through successive rounds of negotiations conducted under the auspices of GATT, industrial country tariffs have been pared down to their present average level of about 4 per cent on manufactured goods. Unfortunately, this reduction in tariff rates has been accompanied by a proliferation of non-tariff barriers. Among the most important of these non- tariff barriers for developing countries is the MFA, or Multi-Fibre Arrangement, which sanctions a structure of country- and product-specific quotas on apparel and textiles, often the most important area of manufacturing advantage for developing countries. The MFA has been widely studied and much attention has been devoted to its welfare consequences.1 However, this literature has been based on the presumption of perfect competition in all relevant markets and has suppressed dynamic aspects of the issues. In such models, as is well known, tariffs and quotas are equivalent and license prices, when available, equal the implicit specific tariff. This makes it straightforward to calculate the welfare effects of the system and to identify losers and winners from the MFA and any of its proposed reforms. The assumption of competitive markets in the study of the MFA is usually defended on the grounds that there are a large number of producers in the textile and apparel market. In the case of some advanced exporters, notably Hong Kong, further justification is provided by the fact that the quotas are efficiently implemented and are, to a large extent, transferable. For example, Morkre (1984) estimates that US clothing import quotas on Hong Kong in 1980 spawned quota rents of $218 million, or 23 per cent of the total value of 2 clothing imports fror long Kong. The central fasture of his methodology is that "... the price of rights to export textiles from Hong Kong measures the ;ap between import price and unit cost in Hong Kong. The rationale is that textile quotas are openly traded in Hong Kong so that the market price for transfers Is expected to reflect the value of the price-cost difference."2,j Hamilton (1986) also uses Hong Kong quota prices to measure rent income. In addition, he uses specially-compiled sets of clothin- import statistics to calculate f.o.b. unit values for Hong Kong exports to the US, which serve as proxies for tie rent-inclusive export prices. Dividing rent income by the rent-inclusive US value of exports, he arrives at the US export tax equivalent rate of textile and apparel quotas on Hong Kong, which he then converts to the import tariff equivalent (MTE) rate by taking the ratio of c.i.f. and f.o.b. values of clothing imports from Hong Kong. He calculates this MTE rate as 9 per cent in 1981 and 37 per cent in 1982. Trela and Whalley (1988, 1990) employ a general equilibrium model to analyze the effects of bilateral quota restrictions imposed by the US, Canada and the EC on 14 product categories from 34 developing countries (including Hong Kong). Like Morkre, their methodology for obtaining the supply price of quota restricted products involves the use of data on Hong Kong quota prices: they compute the Hong Kong supply price by subtracting the quota price from the US price, then they compute the production costs of quota restricted products in other exporting countries by multiplying the unit cost in Hong Kong with the ratio of the exporting country's relative wage rate in the textile and apparel industry compared to Hong Kong. Using 1986 data, they estimate both global and national welfare costs of the KFA. Their results suggest global gains from the elimination of quotas and tariffs of more than 3 $17 billion ** of which $11 billion will accrue to developing countries -- and gains to the US from the removal of quotas of $3 billion. We do not, in this paper, question the assumption of perfect competition a priori. Rather, we ask if it is possible to test whether all the results of the static competitive model are borne out in the data. Oddly enough, this basic issue has never been addressed in the literature. Note that conceptually, there exist two markets: the market for products and the market for licenses. In the product market, there can be imperfect competition on the side of the buyers, i.e. monopsony or oligopsony, and/or on the side of the sellers, i.e. monopoly or oligopoly. In the license market, we need to consider as well who owns the licenses -- that is, whether the licenses are in the hands of the exporters (who may or may not be the producers themselvei) or the importers or buyers (be they consumers or independent retailers). Again, there can be market power on either or both sides of the license market. Clearly, many different combinations of imperfections can arise in the two markets and it is beyond the scope of this paper to study them all. Our approach is not to point to and model a particular form of imperfection but to see whether the implications of perfect competition in all markets do in fact hold. It is well understood that when market imperfections exist, product prices become endogenous, thus allowing producers to appropriate quota rents by raising their supply price. In this case, a quota has very different effects from a tariff which generates the same level of imports. In particular, the price difference between the quota-restricted and world markets, which we will call the potential rent per license, need not equal the tariff which would induce this level of imports. For this reason, licence 4 prices need not reflect import-equivalent tariffs. (See Krishna (1990a).) More ver, the p *ential rbit per license need not equal the actuAal license price If there is "rent sharing". This is to be distinguished from "rent appropriation". The distinction is crucial. By affecting product prices themselves, producers with market power in effect appropriate rents from quotas. We will call this "rent appropriation" and use "rent sharing" to denote the sharing of potential rents between the license holders and other agents, given the price differential created by the quota. In other words, rent sharing is said to occur when the the license price falls short of the price differential in the quota-restricted and world markets. As we argue in the next section, the theoretical literature in the area has focused on rent appropriation and has ignored rent sharing. Our paper attempts to fill this gap by testing for rent sharing in the MFA. We base our empirical case on Hong Kong because it is often held up as the best-functioning and most competitive exporter of clothing to the US. Licenses are relatively freely traded in Hong Kong compared to other MFA- restricted countries, and the quota implementation process is clearly documented. As a result, it is the least likely to exhibit behavior consistent with market imperfections. Evidence of any such behavior from Hong Kong would therefore cast substantial doubt on the suitability of the static competitive model for analyzing the effects of the MFA. The paper is organized as follows. In the next section, we first outline the competitive model, then we discuss how different kinds of imperfections may affect the results of this model. We argue that rent sharing can occur when there is market power on the side of buyers, which may or may not be combined with imperfections in the license market. As our focus is not 5 theoretical, we us- very simple versions of these models. However, the flavor of che results carries over in more general versions. In Section 3, we briefly discuss the data we use. Details of how th* data were put together can be found in the Data Appendix, which carefully describes our sources and procedures as well as the problems that remain with the data. In Section 4, we give a quick summary of Hong Kong's textile quota system. Section 5 sets up the first model we test. This model deals with testing for rent sharing in the pzesence of quotas when imports and domestically- produced goods are homogeneous. We extend this in Section 6 to incorporate compositional differences in the aggregate goods whilst maintaining the homogeneity assumption of its component parts. Section 7 deals with the implications of product differentiation. Section 8 analyzes the factors influencing rent sharing. Section 9 summarizes our results and makes some concluding remarks. 6 2. The Basic Models 2.1 Perfect ComDetition The base case (which constitutes our null hypothesis in Section 5) is one with competition in all the relevant markets. Both the demand and supply sides of the product market are assumed to be competitive and, in addition, license holders act competitively and are willing and able to sell at the price that clears the license market. This iodel is illustrated diagramatically in Figure 1, which is the standard textbook depiction. In Figure 1, RD represents residual damand from the importing country which we will call the US. It is given by subtracting US supply and supply from sources other than Hong Kong from total demand in the US. RS depicts the residual supply from Hong Kong. This is supply from Hong Kong less demand from all sources other than the US. The intersection of the two gives the world price in the absence of quotas and the level of imports from Hong Kong to the US. If a quota is set allowing only V units to be imported, the home price at which this level of imports is demanded exceeds the world price at which it is supplied. Their difference gives the license price, which can be interpreted as the implicit tariff. That is, if the quota were replaced by a specific tariff at this level, the same amount of imports would be induced. Tariffs and quotas are therefore equivalent. 2.2 Market power on the seller's side: "rent approDriation" As is well known, the above argument breaks down with imperfectly competitive markets. Bhagwati (1965) analyzes three departures from the base model of competition (which he calls case 1). He looks at the effect of monopoly in domestic supply and in the license market, as well as some combinations of these. When there is monopoly only in domestic supply, a quota 7 Price Pesidual supply PS (V) Pesidual lemand ) 'uantity V Figure 1 8 makes demand less elastic for price increases, thereby augmenting monopoly power. As a tariff does not eliminate the foreign supply response, we get non- equivalence between the two. This is his case 2. In his case 3, he considers competitive supply at home and abroad, but monopoly in license holdings. By affecting the utilization of licenses, the monopolist holder of licenses affects their value. The uti..ization rate is chosen to maximize total license value. This makes the effective quota endogenous and creates non-equivalence between tariffs and quotas. In his case 4, he adds monopoly in domestic supply to his case 3. He considers two sub-cases: 4a, where the license holder is not the domestic monopolist; and 4b, where he is the domestic monopolist. Thus, 4a becomes a case where the home market is a duopoly. The solution concept chosen is essentially a Cournot- Nash equilibrium. Here, a quota changes the market structure whereas a tariff does not, so the two are not equivalent. In 4b, the monopolist holds all the licences and so he can even further augment his monopoly power over case 2 by effectively choosing his utilization ratio to maximize the sum of profits and license revenues. As a quota enhances his monopoly power, it is again not equivalent to a tariff. Bhagwati does not address the possibility of foreign market power. If the foreign sellers have market power, no supply curve exists as the supply price is chosen to maximize profits. This makes the world price a choice variable and its determination the result of profit-maximizing decisions of the suppliers. If the sellers have no licenses, thL, will have an incentive to raise their price to obtain the rents from the quota.4 Take, for example, the case where there is a single foreign supplier of the product and markets are segmented. It is clearly optimal for the 9 monopolist to raise his price in response to a quota so as to appropriate the entire quota rent, By closing the gap between the demand price and the supply price, the monoDolist effectively strips the licenses of any value. rhis model with segmented markets is developed diagramatically in Takscs (1987) and is mentioned in Shibata (1968) as well, and mo&c recently in Krugman and Helpman (1989). Krishna (1990b, 1990c) further develops a model in which there is costless arbitrage between the markets so the foreign monopolist cannot practise price discrimination.5 The monopolist's price is an endogenous variable -- by charging a high price, he can appropriate rants and he chooses to do so if this is profitable. Of course, this price depends on the quota level and his allocation of licenses. Even here, as long as the license market is frictionless and competitive, i;- is still the case .hat the value of a license equals the difference in the domestic price and the world price. However, in this mcn'el as well as in those of Bhagwati discussed previously, Tie licen *. price is endogenous and depends on other oarameters such as the allocation of licenses and the product market structure and behavior. To summarize, the existence of markpt imperfections in general can result in the non-equivalence of tariffs and quotas. However, our focus is not on the equivalence issue but on what the licence price reflects. In other words, we are concerned with rent sharing and not rent appropriation. When there is product market power on the seller's side, the license price bec.cass an endogenous variable used by the producer to effect rent appropriation. However, the identity continues to hold that the license price is equal to the difference between the domestic (demand) price and the world (supply) price, so that there is no rent sharing. 10 2.3 Market power on the buyer's side: "rent sharing" If there is monopsony power, that is, if there is a single buyer, then the story is quite different. Assume that the licence market is competitive as is the supply side. In Figure 2, the monopsonist retailer has a marginal revenue curve, MR, which is derived from the market demand for apparel, D, and he faces an upward sloping supply, S. His marginal cost curve, MC, lies to the left of S; this is because he has to pay a higher price for all the inframarginal units in order to purchase an additional unit of apparel. Under free trade, the monopsonist will import VF units of apparel, which is given by the intersection of MC and MR. The lowest price at which this quantity will be supplied is pF, and the monopsonist is willing to pay up to P* Since he is the sole importer, however, he can choose his price and so he will offer the lowest price, pF, and sell the goods in the home market at price P*. Now suppose a quota, V, is imposed on apparel imports. The monopsonist's supply curve then becomes the kinked line, ABE and his marginal cost curve becomes ACE. The lowest price at which the quota amount will be supplied is PS(V), and the monopsonist will not pay more then PD(V), which is the price for which he will sell the imports in the home market. If he pays P, where PS(V) : p pD(V), then the price of a quota license will be P-PS(V), i.e. the difference between the price paid and the supply price charged. Of course, the monopsonist will never choose to pay more than the supply price charged so he will buy the V units at price PS(V) and the license price will be zero. Note that this occurs not because there is no price differential in the home and world markets, but because the monopsonist, as the only importer of the good, can prevent trade from equalizing these prices. The more restrictive the quota &. &, 긷 蝦 . p C& 」 。신― , I : 7「 Xl \ 띠 1 0룔 띠 -－·ㄱr7仁．－■－주근닒－－·■눼 & 섬 n· g F 쎄 O (I) O· O, 12 is, the lower will be the price paid by the monopsonist. In any case, the competitive exporters are paid exactly enough to induce them to sell and they receive no rent.6 It is important for us to emphasize that in this case, unlike all the previous ones discussed, the license price is not given by the deviation of the domestic price from the world price. The license to import only has value if the price offered by the monopsonist exceeds the supply price. Since the monopsonist has sole buying power, the license price is always zero. The difference between the home price and the world price, however, is given by PD(V)_pS(V) in Figure 2, and it is = equal to zero. Thus, monopsony power causes the license price to eiverge from the difference in the supply and demand price. Now suppose there is competitive supply but concentration in license holdings as well as market power on the buyer's side. This seems like a better assumption for the US-Hong Kong apparel trade situation, since the mere existence of active trading in quota licenses in Hong Kong is evidence that the license have value. In this case, there is bilateral monopoly power, and the issue becomes one of sharing the potential license rents. The potential rent from a license equals the difference in the supply price in Hong Kong and the demand price in the US market. If a price between these is the outcome of the bargaining process, then the license price is positive. However, the two prices are not separated by exactly the license price because of rent sharing.7 As a stark illustration of this argument, suppose that all the import licenses are hold by a single exporter (who may or may not be a producer), and that the license price is determined by a Nash bargaining process between the 13 monopsonist and the license holder. The license holder's objective is to maximize his profit wL, where: wL - VL. The monopsonist's objective is to maximize his profit xM, where: XM (PD(V) - L - PS(v)]V pD(.) is the inverse residual demand function and PS(.) is the inverse residual import supply function, so PD(V) is the demand price and PS(V) is the supply price for the quota, V units. The license price is found by maximizing the weighted product of both parties' deviation from their fall-back payoff: II - (VL)P[(PD(V).PS(V)-L)V]lf. For simplicity, we assume both parties receive nothing in the absence of an agreement, so their fall-back payoffs are equal to zero.8 The parameters P and (1-P) represent the bargaining strengths of the license holder and the monopsonist respectively, where 0 : P < 1. The first order condition is: VL-(1-P)tPD(V)-PS(V)-L]f(#(PD(V).PS(V))-L] - 0, which yields the solutions: L - 0, L - pD(v)pS(v) and L - p(PD(V)-PS(V)), of which only the third satisfies the second order condition for a maximum. Therefore the license price which is the outcome of the Nash bargaining setup between the license holder and the monopsonist is given by: L - p(PD(V).pS(V)]. The more powerful the license holder is, the higher is the license price. In the extreme case when P-1, the license holder has all the bargaining power and 14 so he extracts the entire quota rent VL, where the license price L is exactly equal to the difference between the demand and supply prices. At the opposite extreme when P-0, the monopsonist calls all the shots: for each unit, he pays only the supply price P and reaps the rent given by the difference between the demand and supply prices. The license holder gets nothing as the license price is equal to zero. For a value of f between 0 and 1, an intermediate result will obtain and the license price will not reflect the full difference between the demand and supply prices. In Section 5, we look at the relationship between the US price and the Hong Kong price, which includes the license price as well as tariffs and transport costs. In the absence of rent sharing, as argued above, these two should be equal. Moreover, their difference should not depend on factors such as concentration in quota holdings and the quota size and utilization ratio. In this way, we estimate the extent of rent snaring and the factors that seem to be influencing it. In Section 6, we allow for compositional effects to create differences in aggregate prices; in Section 7, we test if prices differ because of product differentiation; and in Section 8, we analyze the determinants of rent sharing. 15 3. The Data The data utilized in this study cover the time period 1981-88 and pertain to three broad areas: domestically-produced apparel, imported apparel from Hong Kong and license holdings for apparel imports from Hong Kong. We did not attempt to obtain data on all categories of apparel. There are severe difficulties in assembling a consistent panel of data due to the different and changing classification systems used in reporting information on imports and domestic production. Therefore, we chose groups of apparel such that these consistency problems were minimized. Our objective was to get as many relatively consistently-defined, disaggregated groups that we could find or develop concordances for between the different classification systems employed. We identified ten such groups. They are: (1) dresses; (2) skirts; (3) playsuits; (4) sweaters; (5) trousers; (6) men's coats; (7) women's coats; (8) woven shirts; (9) knit shirts; and (10) underwear. We obtained data for these groups for the following variables between 1981 and 1988. The variables are defined below. In our notation, the subscript i indexes the apparel group, and t indexes the year. Pit us- Unit value of US production. Pit HK - F.O.B. Hong Kong price. This includes the license price. tit - Ad valorem tariff in the US. Tit - Transport cost per unit from Hong Kong to the US. PitHK - Adjusted Hong Kong price, where: PitHK _ PitHK it) +it Qit us- US sales of US production. QitHK - Imports from Hong Kong. 16 Hit - Numbers equivalent of the Herfindahl index of concentration in licence holding. Vit - Quota level for imports. Uit - Utilization ratio of imports, where: Utt Qit /vit. The sources of these data and details on how they were created can be found in the Data Appendix. 17 4. Hong Kong's Textile Ouota System Hong Kong prides itself on administering an efficient textile quota system. The initial quota allocation is historically based. Past performance, transfers and quota level changes guide the process by which these allocations change in subsequent years. When a product category is newly brought under restraint, the quotas are allocated according to past performance , i.e. each company gets a quota amount corresponding to its share in total shipments of that particular category to the market concerned. Where the manufacturer and the exporter are not the same company, they each share the quota pertaining to a shipment on a 50/50 basis.10 If the level of total shipments exceeds the restraint limit, the allocations are scaled down proportionately. If the quota is larger than total past performance, then the balance remaining is put into a "free quota pool", which is open to any firm registered with the Hong Kong Trade Department which has documentary proof of an overseas order. Quota holders are allowed to transfer a part of their quota to other firms. There are two types of quota transfers: permanent transfers, in which the transferee obtains the use of the qLota for the year in question and, based on its performance against the transferred amount, receives a quota allocation in the following year; and temporary transfers, in which the transferee obtains the use of the quota for the year in question, but the performance against the transferred quantity is attributed to the transferor. In order to allow sufficient time for the transferee to obtain the quota, transfer applications are not normally accepted after the middle of November. Free quotas are not transferable. Under Hong Kong's textile quota system, both the utilization rate and 18 the amount of transfers are important factors in determining a firm's future quota allocation. A firm which uses less than 95 per cent of its quota holding will obtain an allocation in the subsequent year equal to the amou'. it used; a firm which uses 95 per cent or more of its quota holding will be given an allocation equal to 100 per cent of its holding; and a firm which uses 95 per cent or more of its quota holding and does not transfer out any of its quota (on either a temporary or permanent basis) will be awarded an additional amount equivalent to the growth factor for that category provided for in the restraint agreement. In addition, a firm which transfers out 50 per cent or more of its quota holdings on a temporary basis in a year is liable to have its quota allocation reduced in the following year,11 whereas a firm which transfers in 35 per cent or more of its quota holdings on a temporary basis during the year is eligible for a bonus allocation in the following year. Finally, a firm which obtains a free quota and utilizes 95 per cent or more of it qualifies for a quota allocation in the subsequent year; a firm which fails to utilize at least 95 per cent of its free quota may be debarred from future participation in free quota schemes for a period of time. To a certain extent, unused quotas may be transferred between categories (under the "swing provision") and between years (under the "carry-over" and "carry-forward provisions"). As quota entitlements in a subsequent restraint period are based on shipment performance in the preceding period, quotas can only be allocated after this performance has been fully verified against shipping documents. This verification process usually takes two to three months. In order to make a portion of the quotas available during the first few months of the year, 19 therefore, the Trade Department makes preliminary quota allocations to companies. Final quota allocations are normally made in March and they supersede any preliminary allocations. All textile and apparel exports from !ong Kong have to be covered by valid export licenses issued by the Director of Trade. Export licenses are only issued to firms which are able to supply quota to cover the consignment in question. Valid licenses are required to bring the shipment on board. An export license is normally valid for 28 days from the date of issue (or, where applicable, until the end of the year, whichever is earlier). The consignment must be shipped within this period. The final licensing date is the first day of December. All licenses covering shipments applied for against quotas held by a company have to be taken out not later than this date, although shipments may be effected up to the last day of the year. Further details of Hong Kong's textile quota system can be found in the Hong Kong Trade Department publication, Textiles Export Control System. A good description of the system is also contained in Morkre (1979, 1984). 20 5. Testing for Rent Sharing: Homogeneous Goods In this and the following two sections, we develop and implement procedures to test for the existence of rent sharing. In the case of homogeneous goods, if there is no rent sharing, the license price equals the difference in the price of US goods and the price of imports from Hong Kong when the latter is adjusted for tariffs and transportation costs. Our data-set for the ten apparel groups does not contain license prices explicitly. However, the license price is included in the f.o.b. Hong Kong price. Therefore, we can test for rent sharing by looking at whether the f.o.b. price in Hong Kong, adjusted for tariffs and transport costs, equals the US price. Figures 3(i)-(x) are scattergrams of these two sets of prices for each of the ten apparel groups. In each scattergram, the points either lie entirely above the 45 degree line or entirely below it, indicating that the US price and the licensa-inclusive Hong Kong price are not equal. However, when the data are pooled, as in Figure 3(xi), the points appear to lie more or less around the 45 degree line -- this prima facie evidence, the. seems to discount the existence of rent sharing! Yet, while it seems reasonable to assume that Hong Kong producers are competitive, it is not clear that market power does not exist in the market for quota licenses and on the side of the US purchasers. For example, an editorial in the Hong Kong trade journal, Textile Asia, alleges that: "Quota price fluctuations do not in fact reflect normal supply and demand but the course of manipulation by the quota holders;"12 and Goto (1989) claims that: "Although governments of exporting countries under the MFA often allocate export licenses in a manner that helps exporters capture the quota rent, many of these exporters face large importing enterprises that can negotiate prices that capture some of the rent for 21 APPAREL GROUP NO. 1 (DRESSES) APPAREL GROUP NO. 2 (SKIRTS) Price Per Untt ha U Dellarn Prmc Per Unit lo U 3 DUare Nst-19ta I-sIG 26 4 21/ i •i 0 N N o 1 19g - lo I Io N 19 g o10 0 0 cS G IS to 57 Ie 3t 33 26 4 b 10 53 14 UNiTD M MAT UpIT) STATM igur* Vii i- . . . Figure 3(i) Figur. 3(1i) APPAREL GROUP NO. 3 (PIAYSUITS) APPAREL GROUP NO. 4 (SWEATERS) Pnce Per Untin U S Dollars Price Per Unit in U S Dollarn 4 54 0N 0 Pi Pil i 2 8 6 Pi 1 a1 UNrDMT5 UN8D TT 3 w ienws e•s...sua. .wws. Figur. 3(a.ii) Figuare 3<1) 22 APPAREL GROUP NO. 5 (TROUSERS) APPAREL GROUP NO. 8 (MEN'S COATS) Priee Pa Uat in US Dellare Pee Per UMwt to U S Dellers 9 04 1961 * 19M 1 24-// H N 0 O0 K 0 Ias N N S *S3 *I 7 I I I I I I I I I I I I I 6 0 7 8 9 1o II 10 a 14 to o 30 Am 24 UNITED STATES UilD STATD a. gW ne. ". smes Ma" **vm No Figuse 3(v) Figure 3(vi) APPAREL GROUP NO. 7 (WOMEN'S COATS) APPAREL GROUP NO. 8 (WOVEN SHIRTS) Price Per Unit to US Dellere Priee Per Unit in US Doellars 19683-1988los 98-1I- SS to 24 0 0 o ,/ c, o 0 N N7 C a S t*o*- so I4 3b I 8 to UNMCD STATES UNI STU 'Pigure 3(vii) rigure 3(viti) 23 APPAREL GROUP NO. 9 (KNIT SHIRTS) APPAREL GROUP NO. 10 (UNDERWEAR) Price Per Unit in US Dollers Price Per Unit in US Dollare 7 . G /3 G 03 o 0 M N 0 C 11111 0 4 0 05 03 3 I I I 00 / 3 4 6 6 7 00 03 @S 08 10 13 I6 UNITED STATES UNITED STATES mw & . * a . . . m ~ * aA ~ m t M *W * . . ~ w M . * Figure 3(ix) Figure 3(x) APPAREL GROUP NOS. 1-10 (ALL) Price Per Unit in U S Dollars 30- 20 0 H OF 0 N mo 4 G 300 10 , '4'. 0 6 30 16 20 20 30 UNITED STATES 84.. Fu JeMt .emp(t.U Figure 3(xi) 24 themselves..13 As argued in Section 2, the existence of monopsony power can lead to rent sharing. This is separate from the issue of whether or not product markets are perfect on the sellers' side. Imperfect product markets per se do not imply rent sharing as we define it, although they do affect who gains and who loses from a quota. We focus only on rent sharing, which results from buyer power and not on rent appropriation, which results from seller power. In regression (1), we regress the adjusted Hong Kong price on the US price, a constant, the quota utilization ratio, the quota level and the numbers equivalent of the Herfindahl Index, which measures concentration in the license holdings and proxies for market power in the license market. The numbers equivalent of the Herfindahl Index is defined as l/Esi , where si equals the share of license holder i in total licenses. Rigression (1) is therefore of the form: (1) PitHK - a + OPit + rHit + 6Uit + it + it The right-hand-side variables can be considered as exogenous variables. If the US is a large country, PitUS is properly taken as given. As quota license allocations are historically determined, Hit can be also taken as given though it does vary over time with the composition of exports. The quota level, Vit, is exogenously determined. The utilization rate, Uit, should be unity if the quota is binding, and any departure from unity is assumed to reflect exogenous difficulties in attaining full utilization due to frictions in the implementation system. The regression is run on pooled data across the ten apparel groups for the years 1981 through 1988. If there is no rent sharing and the goods are homogeneous, we should expect to observe PitHK _ PitUS. In other words, in 25 regression (1), the constant should be zero and the coefficient on the US price should be unity; furthermore, none of the other variables should be significant. The assumption that US and Hong Kong apparel are perfect substitutes ensures that the license-inclusive Hong Kong price has to equal the domestic price in the US. This means that the Hong Kong supply price (exclusive of the license price) has to vary one for one with the license price. For example, if the US is a large country, a reduction in the quota level will tend to raise the license price; but this will be wholly absorbed by the Hong Kong suppliers, who will have to reduce their supply price so as to remain competitive. Therefore, a quota reduction will make licenses costlier, but reduce the supply price at the same time so that PitHK remains unchanged overall. Similarly, changes in the license utilization rate will affect the license price but not PitHK since the supply price will adjust to maintain the equality between PitHK and PitUS. The concentration of license holdings should not affect PitHK unless there are substantial search costs. The results of regression (1) are reported in Table 1. Note that a is significantly different from zero at the 1 per cent level, and that Hong Kong prices are lower than US prices in general. This suggests that the license price embodied in PitHK falls short of the gap between the domestic price (PitUS) and the world price. Moreover, we are able to conclusively reject the null hypothesis of perfect competition everywhere, i.e. the hypothesis that f-1 and a-r-6-*0- jointly. The hypothesis that P-1 can be rejected at the 1 per cent level. The regression results may be interpreted as follows. We can think of a as the fixed component of rent sharing and I as the marginal component of rent sharing. A $1 increase in the US price, therefore, is associated with a $0.53 increase in the Hong Kong price -- this 26 TABLE 1 RESULTS FOR THE HOMOGENEOUS GOODS MODEL (EQUATION 1) Dependent variable - PitHK Independent Variables Coefficient t Statistic Constant -7.4140 -2.7111a (2. 7374) PitUS 0.5271 9.7766a (0.0539) Hit 0.1255 4.5439a (0.0276) Uit 0.0964 3.7754a (0.0255) Vit -0.4104 x 10-7 -4.1771a (0.9825 x 10-8) R2 - 0.8089, Adjusted R2 - 0.7979 Number of observations - 75 * Standard errors are in brackets beneath the estimates of the parameters. (These standard errors do not differ appreciably from those obtained with the White (1984) correction, therefore we discount the possibility of heteroskedasticity in our sample.) a Significant at the 1 per cent level. Results of hypothesis-testing: F-statistic for joint test of 0-1 and a-r-6-0-0: F(5,72) - 25.0740 -- reject the null hypothesis at the 1 per cent level. t-statistic for test of A-1: t - -8.6726 -- reject the null hypcthesis at the 1 per cent level. 27 may indicate that $0.47 of the price differential or rent is retained in the US.14 In addition, note that ceteris paribus, increasing license market concentration lowers the Hong Kong price. We would expect that greater concentration increases the bargaining power of the license holders who will, in turn, seek to raise the license price. As the license price is included in the Hong Kong price, we would therefore expect greater concentration in license holdings to be associated with a higher Hong Kong price. However, it is also possible that fragmentary quota holdings make it inconvenient for firms to obtain sufficient export licenses when they have large overseas orders. This increased search cost would then be reflected in a higher Hong Kong price. Our results seem to suggest that the second effect outweighs the first; this is not too surprising since the Hong Kong quota system penalizes license hoarding, thereby weakening the first effect. In any case, we reiterate that if there is no rent sharing, the degree of concentration in license holdings should not affect the equality of PitUS and PeHK Yet the observed coefficient on Hit is significant at the 1 per cent level, which directly contradicts this prediction. Although the coefficients on Uit and Vit are small, they are also significant at the 1 per cent level. A higher quota utilization rate, with everything else held constant, raises the Hong Kong price. This could be because a higher utilization rate makes licenses harder to get. This increases the bargaining strength of the license holders and consequently raises the license price and the Hong Kong price of which it is a component. A higher quota reduces the Hong Kong price, all else constant. Again, the reason for this may be due to the fact that a larger quota increases 28 license availability and reduces the power of license holders. This in turn reduces the license price and the Hong Kong price. Of course, in the absence of rent sharing, neither of these variables should be significant -- PitHK should always adjust to exactly match Pit The results of regression (1) therefore seem consistent with the existence of monopsony power in the market for a homogeneous good in the face of imperfections in the license market -- there seems to be a gap between the world price and the domestic price which is not completely closed by the license price. 29 6. Allowing for a Composition Effect In the previous section, we found a price differential between US- produced apparel and imports from Hong Kong. This seems to suggest that some rent sharing does exist. However, there may be an alternative explanation for this price differential, namely, that the Hong Kong product mix is not the same as that of the US. In other words, the null hypothesis described in the beginning of this section may be valid for the component MFA categories but not for the aggregate apparel groups. For example, the prices of cotton dresses, wool dresses and dresses made of manmade fibre may be the same in both the US and Hong Kong, but if the US produces relatively more wool dresses, which are relatively more