[雙語翻譯]貿(mào)易壁壘外文翻譯--貿(mào)易壁壘和貿(mào)易品的相對價格（英文）

1、Trade barriers and the relative price of tradables☆Michael SposiResearch Department, Federal Reserve Bank of Dallas, 2200 N Pearl Street, Dallas, TX 75201, United Statesa b s t r a c t a r t i c l e i n f oArticle histor

2、y:Received 21 May 2013Received in revised form 17 February 2015Accepted 24 February 2015Available online 5 March 2015JEL classification:F1O4Keywords:Relative pricesTradablesNontradablesTrade barriersIn this paper I quant

3、itatively address the role of trade barriers in explaining why prices of services relative totradables are positively correlated with levels of development across countries. I argue that trade barriers playa crucial role

4、 in shaping the cross-country pattern of specialization across many tradable goods. I construct amulti-country, general equilibrium model of trade and derive tractable predictions that show how specializationaffects rela

5、tive prices. I calibrate the model to match the patterns of prices, levels of development, and bilateraltrade across 103 countries. Through counterfactuals I find that removing trade barriers eliminates more thanhalf of

6、the gap in the relative price of services between rich and poor countries with only a minimal systematiceffect on the absolute price of tradables.© 2015 Elsevier B.V. All rights reserved.1. IntroductionA striking fe

7、ature of cross-country price data is that the price ofservices relative to tradable goods correlates positively with levels ofdevelopment.1 It is important to understand the fundamental sourcesof what causes differences

8、in relative prices across countries for atleast two reasons. First, relative prices lie at the heart of understandingreal exchange rates (see for instance Burstein et al. (2005)). Second, rel-ative prices play a crucial

9、role in understanding income differencesacross countries; sectoral productivity differences across countries areoften inferred from differences in relative prices (e.g., Restuccia andUrrutia, 2001; Hsieh and Klenow, 2007

10、).Modern globalization has resulted in tighter integration in goodsmarkets across countries. The world trade-to-GDP ratio increasedfrom 0.14 in 1990 to 0.25 in 2011.2 However, there remain substantialbarriers to internat

11、ional trade. In this paper I explore how trade barriersaffect the distribution of the price of services relative to tradables acrosscountries. I argue that trade barriers affect relative prices through theirimpact on pro

12、ductivity in the tradables sector. My main finding is thatremoving trade barriers eliminates more than half of the gap in therelative price of services between rich and poor countries with only aminimal systematic effect

13、 on the absolute price of tradables.One of the first explanations for why relative prices co-vary positive-ly with development is due to Balassa (1964) and Samuelson (1964).The Balassa–Samuelson hypothesis postulates tha

14、t there are largercross-country productivity differences in tradable goods than innontradable goods.3 There are four main components that drive theBalassa–Samuelson result: 1) large cross-country productivity differ-ence

15、s in tradables, 2) small cross-country productivity differences innontradables, 3) free trade, and 4) perfectly mobile labor within a coun-try across tradables and nontradables. These four components implythat, in each c

16、ountry, the relative price of nontradables is equal to theJournal of International Economics 96 (2015) 398–411☆ This paper is a revised version of the second chapter of my dissertation at theUniversity of Iowa. I thank B

17、. Ravikumar for his continual guidance and RaymondRiezman for his many suggestions. Special thanks also go to Mario Crucini and HakanYilmazkuday for sharing their micro-level estimates of distribution margins. This paper

18、also benefited greatly from the comments made by two anonymous referees, EliasDinopoulos, Elisa Keller, Alex Monge-Naranjo, Michael Plante, Sergio Rebelo, JohnRogers, Ina Simonovska, Gustavo Ventura, Chia-Chi Wang, Jing

19、Zhang, and the audiencesat the Dallas Fed, St. Louis Fed, University of Iowa, Midwest Macroeconomics Meetings,Southern Economics Association Annual Meeting, System Committee for InternationalEconomic Analysis, and the We

20、stern Economics Association Annual Conference. ValerieGrossman provided excellent research assistance. All errors are my own. The views inthis paper are those of the author and do not necessarily reflect the views of the

21、 FederalReserve Bank of Dallas or the Federal Reserve System.E-mail address: michael.sposi@dal.frb.org. 1 Further empirical documentation of relative prices can be found in Kravis and Lipsey(1988), and Heston et al. (199

22、4). More recently, using detailed disaggregate price datafrom the 2005 World Bank's International Comparison Program, Marquez et al. (2012)show that prices of nontradable goods co-vary closely with levels of developm

23、ent whileprices of tradable goods co-vary much less with development. They show that this obser-vation remains robust even after removing goods and services that are “comparison resis-tant” across countries, for reasons

24、that may stem from unobserved quality differences forinstance.2 These data come from the 2012 World Bank's World Development Indicators. In thedatabase, merchandise exports correspond to TX.VAL.MRCH.WL.CD and GDP cor

25、respondsto NY.GDP.MKTP.CD. 3 Herrendorf and Valentinyi (2012) argue that the Balassa–Samuelson effect holds em-pirically in a large cross section of countries.http://dx.doi.org/10.1016/j.jinteco.2015.02.0030022-1996/

26、9; 2015 Elsevier B.V. All rights reserved.Contents lists available at ScienceDirectJournal of International Economicsjournal homepage: www.elsevier.com/locate/jieAt the end of the paper I examine how well the model can e

27、xplainprices as an out of sample prediction. I estimate trade barriers using tar-iffs and transport costs and calibrate the productivity parameters usingdata on bilateral trade and levels of development only. This altern

28、ativecalibration approach does not make use of any price data. The resultsmatch the cross-country variation in the price of tradables in the datareasonably well, but over-predict the gap in the price of services be-tween

29、 rich and poor countries.2. ModelThe model builds on the framework of Eaton and Kortum (2002),Alvarez and Lucas (2007), and Waugh (2010). There are I countriesindexed by i = 1, …, I. There are two primary sectors: tradab

30、le(manufactured) goods, and services—all value added takes place inthese two sectors. The tradables sector is denoted by m and the servicessector is denoted by s. There is a third sector called retail goods, denotedby r,

31、 where no value added takes place, and serves as an intermediarybetween households and producers of tradables. Each country i admitsa representative household that is endowed with a measure Li ofworkers. The representati

32、ve household owns its country's capitalstock denoted by Ki. Both capital and labor are immobile across coun-tries but perfectly mobile across sectors. Returns to capital and laborare spent on consumption of retail go

33、ods and services.Within the tradables sector there is a continuum of individual goodsthat are each potentially tradable across countries. Individual tradablegoods are aggregated into a composite good. Each individual tra

34、dablegood is produced using capital, labor, the composite good, and services.Services are produced using capital, labor, the composite good, andservices as well. Services are combined with the composite good toconstruct

35、retail goods. Neither retail goods nor services are tradable.4I choose world GDP to be the numéraire of the economy,i.e., ∑iwiLi + riKi = 1, where wi and ri denote the rental rates forlabor and capital respectively

36、in country i. Hence, all units are expressedin terms of global value added.2.1. TechnologiesThere is a continuum of individual tradable goods indexed byx ∈ [0, 1], and each individual good is potentially tradable.2.1.1.

37、Composite goodAll individual tradable goods in the continuum are aggregated into acomposite good M according toMi ¼Zqmi x ð Þη?1 η dx? ? η η?1 ;where qmi(x) denotes the quantity of good x purchased, either

38、 domesti-cally or imported, by country i.2.1.2. Individual tradable goodsEach country has access to technologies for producing each individ-ual tradable good as followsmi x ð Þ ¼ zi x ð Þ?θ Kmi x

39、 ð ÞαLmi x ð Þ1?α ? ?νm Smi x ð ÞσmMmi x ð Þ1?σm ? ?1?νm:For each factor used in production, the subscript denotes the sectorthat uses the factor, and the argument in the parenthes

40、es denotes theindex of the good in the continuum. For example, Kmi(x), Lmi(x), Smi(x),and Mmi(x) respectively denote the amounts of capital, labor, services,and composite good that country i uses to produce tradable good

41、 x.The parameter νm ∈ [0, 1] determines the share of value-addedin gross production, σm ∈ [0, 1] determines the share of services inintermediate inputs, while α ∈ [0, 1] determines capital's share invalue-added.As in

42、 Alvarez and Lucas (2007), zi(x) represents country i's cost ofproducing good x, which is modeled as an independent random drawfrom an exponential distribution with parameter λmi N 0 in country i.This implies that zi

43、(x)?θ, country i's level of efficiency in producinggood x, has a Fréchet distribution. The expected value of zi(x)?θ is pro-portional to λmi θ , so average efficiency across the continuum of goodsis proportional

44、 to λmi θ , (fundamental productivity henceforth).5 Ifλmi N λmj, then on average, country i is more efficient than country j.The parameter θ N 0 governs the coefficient of variation in efficiencydraws across the continuu

45、m. A larger θ implies more variation inefficiency levels, and hence, more room for specialization.Since the index of the good x is irrelevant, from now on I followAlvarez and Lucas (2007) and identify each individual goo

46、d x by itsvector of cost draws: z = (z1(x), z2(x), … zI(x)). I denote country i'sdensity for cost draws by ?i(z).2.1.3. ServicesServices are produced using capital and labor in addition to twointermediate inputs: ser

47、vices and the composite good.Si ¼ Asi Kα siL1?α si? ?νs Sσs si M1?σs si? ?1?νs:The term Asi denotes country i's productivity in the services sector.2.1.4. Retail goodsRetail goods are constructed by combining th

48、e composite goodtogether with domestic services.Ri ¼ Sσr ri M1?σr ri :This technology makes explicit the idea that a domestic service mar-gin is applied to tradable goods (the composite good) after trade hastaken pl

49、ace. The parameter σr ∈ [0, 1] corresponds to the distributionmargin. The retail sector “rents” retail services, which are assumed tobe identical to all other types of services, from the service sector. Inturn, both the

50、productivity of services and the value added from theservice sector (i.e., distribution, marketing) are embodied in the retailgoods as intermediate inputs.2.2. PreferencesThe representative household derives utility from

51、 consuming retailgoods and services as follows:Ci ¼ Cρ siC1?ρ ri :The terms Cs and Cr denote aggregate consumption of services andretail goods respectively, and ρ is the share of services in final consump-tion expen

52、ditures. Ci denotes total consumption.4 Throughout the paper I assume that services are not tradable. First, in the data servicesare simply not traded very much. Using import data, export data, and total production dataf

53、rom input–output tables of the Structural Analysis (STAN) database from the OECD, theaverage home trade share across the countries is 0.93 for services, compared to 0.46 formanufacturing. Second, gross production data in

54、 services is available for only 32 countries.The income gap between these 32 countries is small and would thus provide a weak con-text to study price variation across rich and poor countries.5 As in Uy et al. (2013) I re

55、fer to λθ as fundamental productivity. In equilibrium, eachcountry produces only a subset of the tradable goods and imports the rest. Therefore, mea-sured productivity is endogenous because it depends on the set of goods