The Pennsylvania State University. The Graduate School. Department of Agricultural Economics and Rural Sociology

Transcription

1 The Pennsylvana State Unversty The Graduate School Department of Agrcultural Economcs and Rural Socology MEASURING THE COMPETITIVENESS OF CHINA S VEGETABLE EXPORTS IN THE US MARKET A Thess n Agrcultural, Envronmental & Regonal Economcs by Na Pu 2009 Na Pu Submtted n Partal Fulfllment of the Requrements for the Degree of Master of Scence August 2009

2 The thess of Na Pu was revewed and approved* by the followng: Davd Blandford Professor of Agrcultural and Envronmental Economcs Thess Advsor Davd G.Abler Professor of Agrcultural, Envronmental and Regonal Economcs and Demography Alessandro Bonanno Assstant Professor of Agrcultural Economcs Stephen M. Smth Professor of Agrcultural and Regonal Economcs Head of the Department of Agrcultural Economcs and Rural Socology *Sgnatures are on fle n the Graduate School

3 ABSTRACT Wth Chna s fast growth n exports of hortcultural products, t s very mportant to study Chna s compettve strength. Snce NAFTA s the domnant trade partner for hortcultural products n the US market and t represents a free trade envronment, I use NAFTA as the benchmark. Based on the theory of Armngton Elastctes, an elastcty of substtuton model s used to compare Chna s compettveness wth that of the rest of the world. Three man conclusons can be obtaned from emprcal results on monthly data. Frst, mports of vegetables from Chna are less responsve to cross prce changes than the rest of the world. Second, Chna s substtuton elastcty s sgnfcantly hgher for exports of processed products than fresh products. The fnal concluson s that mports from Chna are more prce elastc over a longer perod. But there s trade off for processed products.

4 TABLE OF CONTENTS v LIST OF FIGURES... v LIST OF TABLES... v ACKNOWLEDGEMENTS... v Chapter 1 Introducton... 1 Chapter 2 Chna s Agrcultural Exports n the US Market Changes Snce Chna Entered the WTO n Recent Developments n Trade n Vegetables between Chna and US Chapter 3 A Revew of the Lterature on Analyzng Compettveness Constant Market Share Method Armngton Elastctes (Foregn products and domestc products) Analyzng the Impact of Non Tarff Barrers on Imports Implcatons for the Approach Used n Ths Study Chapter 4 Methodology Introducton Model Selecton Sub Group Analyss Data Chapter 5 Emprcal Results Model Selecton Import Demand Functon Comparson of Results for Chna and Its Compettors Import Demand Functon Comparson of Regresson Result for Product Sub Groups Annual Data Chapter 6 The Role of Trade Barrers: A Case Study Chapter 7 Concluson Appendx Table: Chna s major fresh produce exports by regons Bblography... 76

5 v LIST OF FIGURES Fgure 1.1: Major fresh products market value percentages from 2002 to Fgure 1.2: Major fresh products market value rato of Chna to NAFTA from 2002 to Fgure 1.3: Major processed products market value percentages from 2002 to Fgure 1.4: Major processed products market value percentages from 2002 to Fgure 2.1.1: Tme seres plots of major fresh products mported from Chna by the US from 2002 to 2008, Mushrooms, Fresh Fgure 2.1.2: Tme seres plots of major fresh products mported from Chna by the US from 2002 to 2008, Arrowroot, Salep Fgure 2.2.1: Tme seres plots of major processed vegetable products mported from Chna by the US from 2002 to 2008, Legumes, Frozen Fgure 2.2.2: Tme seres plots of major processed vegetable products mported from Chna by the US from 2002 to 2008, Vegetables, Frozen Fgure 2.2.3: Tme seres plots of major processed vegetable products mported from Chna by the US from 2002 to 2008, Vegetables Nes dred Fgure Tme seres plots of major processed vegetable products mported from Chna by the US from 2002 to 2008, Urd, Mung, Black dred Fgure 2.2.5: Tme seres plots of major processed vegetable products mported from Chna by the US from 2002 to 2008, Kdney Beans, Dred Fgure 5.1: Pooled data for the log of the rato of mports and prces for Chna and NAFTA Fgure 5.2: The rato of mports between Chna and NAFTA for January Fgure 5.3: The rato of mports between Chna and NAFTA for July... 51

6 Fgure 5.4: The rato of mports between Chna and NAFTA for December v Fgure 6.1: AD actons ntated aganst Chna relatve to world total Fgure 6.2: Imports of fresh garlc from Chna

7 LIST OF TABLES v Table 1.1: Value of Chna s major exported fresh products and processed products n the US market. Unts: mllon$... 4 Table 2.1: Top 10 of Chna s vegetable exports to the US ranked by quantty, ,Unts: mllon$, mllon metrc tons Table 2.1: Contnued, Top 10 of Chna s vegetable exports to the US ranked by quantty, , Unts: mllon$, mllon metrc tons Table 2.2: Top 10 of Chna s vegetable exports to the US ranked by value, ,Unts: mllon$, mllon metrc tons Table 2.2: Contnued, Top 10 of Chna s vegetable exports to the US ranked by value, , Unts: mllon$, mllon metrc tons Table 4.1: Varable names and defntons Table 4.2: Varables for group category analyss Table 4.3: Subgroups of products determned by correlaton n mport data Table 4.4: Sub groups of products determned by the degree of processng Table 5.1: Regresson results for model selecton Chna versus NAFTA Table 5.2: Varables for the demand functon and group analyss Table 5.3: Results of estmatng the mport demand functon Table 5.4: Regresson results for the elastcty of substtuton model( Model 2) Table 5.5: Import demand prce coeffcents for Chna, NAFTA and ROW Table 5.6: Regresson results of sub groups accordng to correlaton Table 5.7: Regresson results for sub groups accordng to degree of processng Table 5.8: Comparson of results from monthly data and annual data Table 5.9: Comparson of results of sub groups: fresh products vs processed products Table 6.1: Chna s share of U.S. mports of targeted hortcultural products... 70

8 v ACKNOWLEDGEMENTS I would lke to express my deep grattude to my thess advsor and char of my commttee Dr. Davd Blandford for the support and gudance. Thank you very much for all your patence and efforts spent on gudng my thess. The subject of my thess s a pretty new area of agrcultural economcs and I feel frustrated n the begnnng snce I cannot fnd lots of lterature for gudance. Dr. Blandford has such a good personalty and enthusasm that he always nspres me to develop new deas and make them come true. I really enjoy every meetng wth hm! Great grattude goes to my commttee member Dr. Davd Abler. Quck thnker, short words but ht the nal on the head, Dr. Abler always help me to deal wth my dffcultes n my thess. Dr. Blandford and Dr. Abler also guded me n my personal lfe. I really apprecate your advce! Fnally I want to express my thanks to Dr. Bonanno for hs nspred comments and support. Wth all your help, I am very grateful that fnally I can put a small step on the vrgn sol.

9 Chapter 1 Introducton Trade n fruts and vegetables s becomng more and more mportant n Chna s exports of agrcultural products. Durng , Chna ranked eghth n world exports of fruts and vegetables (ncludng pulses and tree nuts) and exports reached more than sx tmes the level of ts mports. 1 Chna s exports of fruts and vegetables rose from $2 bllon n 1992 to $3.7 bllon n Growth stalled n the md- and late-1990s as total volume ncreased, whle average prces declned due to ncreasng competton n major markets. Chna s export volume of fruts and vegetables s about 1 percent of total domestc output. Processed products (ncludng juce) account for more than two-thrds of Chna s frut and vegetable export value. Leadng products nclude vegetable mxes (not frozen), frozen vegetables, and mushrooms. Fresh vegetables such as mushrooms, garlc, onons, and radshes account for 14 percent of total hortcultural exports. The balance s n pulses (8 percent), fresh frut (6 percent), and nuts (4 percent). Hgh-qualty seed mports, advances n transportaton technology, mprovements n domestc seed producton and the avalablty of better agrcultural chemcals and equpment are movng Chna along the road to becomng a major producer of exportqualty vegetables. Chna s already the world s largest producer and one of the world s top exporters of fruts and vegetables. Accordng to the Food and Agrculture 1 Data and background n ths part comes from Huang (2004).

10 Organzaton of the Unted Natons (FAO), Chna produces nearly half of the world s vegetables and 16 percent of the world s frut, ncludng 36 percent of the world s apples. 2 For years, Japan has been Chna s man market for vegetable and frut exports, whch are domnated by processed products (canned, frozen, and dehydrated) and, to a much lesser degree, fresh vegetables. Japan s the major market for processed products, absorbng nearly half of Chna s exports. Chna has now clamed the poston of No. 1 suppler of fresh vegetables to Japan. The Unted States s Chna s bggest compettor n Japan s market. The Unted States s also the second largest market for Chna, accountng for about 6 percent of Chna s total exports for vegetable and fruts. As shown n Appendx D, the US accounted for 7% of Chna s total exports of fresh garlc and 4% of Chna s total exports of fresh mushrooms from 2002 to Table 1.1 summarzes the man export markets and trends n Chna s exports of major hortcultural products from 1992 to US mports of caulflower and broccol have grown steadly. Exports of carrots, beets, radshes, onons and leeks have been growng snce Especally snce 1997, mports of those vegetable types have accelerated. Onons and leeks have grown the most, from 10 tons n 1992 to almost 500 tons n Table 1-1 s taken from Carter and Gunnng-Trant (2006).

11 3 Meanwhle, Chna's own market for vegetable mports s growng. Economc reform and market lberalzaton, removal of prce and foregn exchange controls and the movement towards the removal of mport lcensng have made Chna a more appealng market for agrcultural exports from other countres. Chna s growng presence n world markets for vegetables and frut has led to trade dsputes wth a number of mportng countres. Japan mposed quotas on mports of leeks and fresh shtake mushrooms from Chna n The US mposed ant-dumpng (AD) dutes on mports of Chnese garlc n 1994 n response to a massve mport surge n AD cases ntated aganst Chna by the US are dscussed n Chapter 6.

12 Table 1.1: Value of Chna s major exported fresh products and processed products n the US market. Unts: mllon$ Year 4 Products Hs6code Fresh Total fresh vegetables GARLIC, FRESH MUSHROOMS, FRESH ARROWROOT, SALEP Processed products Total processed products LEGUMES, FROZEN VEGETABLES, FROZEN DRIED MUSHROOMS AND TRUFFLES VEGETABLES NES DRIED URD,MUNG,BLACK DRIED KIDNEY BEANS, DRIED Rato Fresh/processed Data source: USDA, Foregn Agrcultural Servce, accessed March Avalable at And Comtrade data from WITS software of World Bank for data on total processed and total fresh data 2-dgt HS6codes (07) for vegetables. The NAFTA countres (Canada and Mexco) are major supplers of agrcultural products to the US market. Fgures 1.1 to 1.4 show the market share and rato of mports from Chna relatve to NAFTA for fresh and processed products. Major fresh and processed

13 5 products all have a smlar pattern, wth Chna s market share ncreasng at the begnnng of the perod and then decreasng n recent years. The rato of Chna s mports of many major fresh vegetables to NAFTA s ncreased at a very low rate durng ths perod, although salted arrowroot decreased sharply. The entre rato of Chna s major processed vegetables ncreased wth urd mung black dred and dred mushrooms rsng very quckly. But urd mung black dred started to decrease n Fgure 1.1: Major fresh products market value percentages from 2002 to Data source: USDA and Comtrade 3 3 Data source for Fgure1.1 to 1.4 s USDA, Foregn Agrcultural Servce. Avalable at And Comtrade data from WITS software of World Bank for the data on total processed, total fresh data by selectng the 2-dgt Hs codes 07 for vegetables.

14 6 Fgure 1.2: Major fresh products market value rato of Chna to NAFTA from 2002 to Data source: USDA and Comtrade Fgure 1.3: Major processed products market value percentages from 2002 to Data source: USDA and Comtrade

15 7 Fgure 1.4: Major processed products market value percentages from 2002 to Data source: USDA and Comtrade Ams and Structure of Ths Thess A clear dstncton can be made between trade n fruts and n vegetables. Chna s a large net exporter of fresh and processed vegetables. In contrast, the country mports more frut than t exports. In ths thess, I focus on vegetables only to examne export compettveness. It s assumed that all WTO member countres sellng to the US market face the same trade envronment wth respect to tarff levels and nontarff measures, except for members of NAFTA, whch s a free trade agreement among the US, Canada and Mexco. Rather than drectly comparng Chna wth all the other tradng partners n the US market, I use mports from NAFTA as a benchmark and compare Chna wth the rest of the world s compettveness level (on average), whch s a new angle to examne Chna s export compettveness n US market. I compare the prce substtuton elastcty of Chna wth NAFTA and the rest of the world wth NAFTA to acheve ths goal.

16 Further I try to examne whether there s a sgnfcant dfference n the substtuton elastcty for some sub groups such as processed products versus fresh products. 8 The thess s organzed as follows. In chapter 2, the structure and trends n US mports of vegetables from Chna are dscussed. Chapter 3 s the lterature revew about research on Armngton elastctes and substtuton elastctes. The methodology used to analyze compettveness n ths study s dscussed n Chapter 4 and the followng Chapter 5 analyzes the emprcal results. Fnally n Chapter 6 antdumpng measures are examned as a case study about remanng trade barrers faced by Chna n exportng hortcultural products to the Unted States.

17 Chapter 2 Chna s Agrcultural Exports n the US Market 2.1 Changes Snce Chna Entered the WTO n 2001 A major development n Chna s tradng relatonshps took place wth the country s membershp of the World Trade Organzaton (WTO) n Chna s total trade scale has ncreased very rapdly at a rate of over 20 percent per year, whle the less compettve and hstorcally protected agrcultural sector has been greatly affected. Chna became a net mporter of agrcultural goods n 2004 wth the US beng the major source of mports. But there has been lmted research on Chna s agrcultural trade and more specfcally on the trade between Chna and the US. Chna agreed to lower ts tarffs on agrcultural products from 31.5 percent to 14.5 percent overall by January As stated n the agrcultural agreement on enterng the WTO, Chna manly has three oblgatons: (1) Tarff reducton - Chna agreed to lower ts tarffs on agrcultural products from 31.5 percent to 14.5 percent overall by January 2004 ; (2) tarffcaton - the frst step was to replace restrctons such as mport quotas wth tarffs that represented about the same level of protecton; and (3) a reducton n domestc support and export subsdes for agrcultural products - ths nvolved domestc subsdes and other programs, ncludng those that rase or guarantee farmgate prces and farmers ncomes, export subsdes and other methods used to make exports artfcally

18 compettve. The subsdes provded by the Chnese Government for agrculture every year are not to exceed 8.5 percent of ts agrcultural output n that year There have been three major changes n Chna s agrcultural exports snce t joned the WTO. Frst, there have been changes n the types of products beng exported. Imports of land-ntensve crops such as soybeans, other vegetable ols, and cotton have ncreased whle exports of labor-ntensve goods such as garden plants and seafood have expanded. Second, there are growng regonal dfferences. The east coast s the man area nvolved n agrculture trade, wth a 94 percent share of total agrcultural mports and 75 percent share of total agrcultural exports. But trade at the borders of the mddle and west of Chna s growng rapdly. Thrd, trade markets are becomng more concentrated. For agrcultural exports, Asa s Chna s bggest market wth a share of 60 percent and the EU has taken about a 20 percent share. North and South Amerca are Chna s major mport partners wth more than 50 percent of agrcultural mports. Japan, the US, and South Korea are Chna s major export destnatons whle the US, Argentna, and Australa are Chna s major mport sources. Table 2.1 and 2.2 lst the top 10 traded vegetable products from Chna to the US durng the perod from 2002 to These data are derved from the Unted Natons Comtrade Database. 4 WTO The agrcultural agreement and Chow (2001).

19 11 The tables reveal only lmted changes n the rankngs of the most mportant products. Dred vegetables, fresh garlc, frozen vegetables, dred mushrooms and dred beans are usually ncluded n the lst. Most of the top products are processed products. But for those products appearng n the lst n each year, there s a clear ncreasng trend n both quantty and value. Also there have been some trade swtches from dred to fresh for some products, most notably mushrooms. Both dred and fresh mushrooms were n the top 10 US vegetable mports from Chna durng 2002 to But n 2006 and 2007 the quantty of dred mushrooms declned sharply and t fell out of the rankng. More fresh mushrooms are now exported drectly rather than beng processed for export.

20 Table 2.1: Top 10 of Chna s vegetable exports to the US ranked by quantty, , Unts: mllon$, mllon metrc tons. hs6code product value quantty hs6code product value quantty Dred vegetables, nes Dred vegetables, nes Garlc, fresh or chlled Garlc, fresh or chlled Legumnous vegetables, Legumnous shelled or unshelled, fresh vegetables, shelled or unshelled, fresh Roots and tubers wth hgh starch content, frees Dred beans, shelled Dred kdney beans, ncl. whte pea beans, shell Vegetables, frozen, nes Mushrooms, fresh or chlled Vegetables, frozen, nes Roots and tubers wth hgh starch content, fresh Dred beans, shelled Mxtures of vegetables, frozen Dred kdney beans, ncl. whte pea beans, shell Mushrooms, fresh or chlled Mxtures of vegetables, frozen Dred adzuk beans, shelled Dred onons Dred vegetables, nes Dred vegetables, nes Garlc, fresh or chlled Garlc, fresh or chlled Legumnous vegetables, Legumnous shelled or unshelled, 7 vegetables, shelled fresh or unshelled, fresh Vegetables, frozen, nes Vegetables, frozen, nes Roots and tubers wth Roots and tubers hgh starch content, fres wth hgh starch content, fresh Dred beans, shelled Dred beans, shelled Mushrooms, fresh or chlled Dred kdney beans, ncl. whte pea beans, shelled Mushrooms, fresh or chlled Dred onons Mxtures of vegetables, frozen Dred kdney beans, ncl. whte pea beans, shell Spnach, frozen Dred mushrooms and truffles

21 13 Table 2.1: Contnued, Top 10 of Chna s vegetable exports to the US ranked by quantty, , Unts: mllon$, mllon metrc tons hs6code product value quantty hs6code product value quantty Dred vegetables, nes Dred vegetables, nes Garlc, fresh or chlled Garlc, fresh or chlled Vegetables, frozen, nes Vegetables, frozen, nes Legumnous vegetables, shelled or unshelled, Legumnous vegetables, shelled or fresh unshelled, fresh Dred beans, shelled Dred kdney beans, ncl. whte pea Dred kdney beans, ncl. whte pea beans, shell Spnach, frozen Mushrooms, fresh or chlled Mxtures of vegetables, frozen beans, shell Roots and tubers wth hgh starch content, fresh Dred beans, shelled Spnach, frozen Dred onons Onons and shallots, fresh or chlled Truffles, fresh or chlled Data source: USDA, Foregn Agrcultural Servce, accessed March Avalable at

22 Table 2.2: Top 10 of Chna s vegetable exports to the US ranked by value, , Unts: mllon$, mllon metrc tons. hs6code product value quantty hs6code product value quantty Dred vegetables, Dred vegetables, nes nes Legumnous Garlc, fresh or vegetables, shelled chlled or unshelled, fresh Garlc, fresh or chlled Vegetables, frozen, nes Dred beans, shelled Roots and tubers wth hgh starch content, frees Dred kdney beans, ncl. whte pea beans, shell Dred mushrooms and truffles Mushrooms, fresh or chlled Mxtures of vegetables, frozen Dred vegetables, nes Garlc, fresh or chlled Legumnous vegetables, shelled or unshelled, fresh Vegetables, frozen, nes Dred mushrooms and truffles Mushrooms, fresh or chlled Roots and tubers wth hgh starch content, frees Dred beans, shelled Mxtures of vegetables, frozen Legumnous vegetables, shelled or unshelled, fresh Vegetables, frozen, nes Dred mushrooms and truffles Roots and tubers wth hgh starch content, frees Dred beans, shelled Mushrooms, fresh or chlled Dred onons Mxtures of vegetables, frozen Dred vegetables, nes Garlc, fresh or chlled Legumnous vegetables, shelled or unshelled, fresh Vegetables, frozen, nes Dred mushrooms and truffles Roots and tubers wth hgh starch content, frees Dred onons Mushrooms, fresh or chlled Dred beans, shelled Dred onons Dred kdney beans, ncl. whte pea beans, shell

23 15 Table 2.2: Contnued, Top 10 of Chna s vegetable exports to the US ranked by value, , Unts: mllon$, mllon metrc tons. hs6code product value quantty hs6code product Dred vegetables, nes Garlc, fresh or chlled Legumnous vegetables, shelled or unshelled, fresh Vegetables, frozen, nes Dred mushrooms and truffles Mushrooms, fresh or chlled Dred beans, shelled Spnach, frozen Dred kdney beans, ncl. whte pea beans, shell value, mllon quantty, thousand Dred vegetables, nes Garlc, fresh or chlled Vegetables, frozen, nes Legumnous vegetables, shelled or unshelled, fresh Dred mushrooms and truffles Truffles, fresh or chlled Roots and tubers wth hgh starch content, fresh Dred beans, shelled Dred kdney beans, ncl. whte pea beans, shell Dred onons Spnach, frozen Data source: USDA, Foregn Agrcultural Servce, accessed March Avalable at Recent Developments n Trade n Vegetables between Chna and US As a member of the WTO, Chna has trade benefts. But Canada and Mexco are members of NAFTA, a free trade agreement and they have even more benefts n the US market, facng low or zero tarffs. Frut and vegetable trade among the NAFTA countres has ncreased substantally snce the Agreement s mplementaton n Mexcan frut

24 16 and vegetable exports to the Unted States approached $5.6 bllon n 2008, and more than trpled durng the NAFTA perod. Accordng to a report by the USDA that summarzes the achevements of NAFTA, by 2008 lots of products became duty free such as hams (fresh or chlled, not processed) and soybeans 5. Therefore, t wll be meanngful to use NAFTA as the benchmark for trade facng few restrctons. The prncpal dataset used to analyze compettveness n ths thess s monthly mport data from Foregn Agrcultural Servce of USDA for the tme perod from January 2002 to December of As ndcated earler, ths spans the perod snce Chna joned the WTO. The products examned are vegetables wth hs6codes startng wth 07. Usng the data from USDA, I made tme seres plots for each product from 2002 to For most products, the tme seres plot s smlar to the followng example for Edble Brasscas. The data n ths fgure are for the rato of the mport quantty from Chna to the quantty mported from NAFTA. We can see from the followng plots (Fgure 2.1 and Fgure 2.2) that there are seasonal effects wthn years but no clear trend. But we can notce that the data appear epsodc n that there are sudden ncreases from tme to tme (e.g., the mddle of 2007), after whch mports can fall to very low levels. 5 Zahnser and Crago(2009).

25 17 Fgure rqchn g24=70951 MUSHROOMS, FRESH 2002m1 2004m1 2006m1 2008m1 tme Fgure 2.1.1: Tme seres plots of major fresh products mported from Chna by the US from 2002 to 2008, Mushrooms, Fresh 6 Note: The data source for Fgure2.1.1 to s monthly data from the USDA, Foregn Agrcultural Servce, avalable at Monthly data.

26 g61=71490 ARROWROOT, SALEP 2002m1 2004m1 2006m1 2008m1 tme rqchn1 rqchn2 Fgure 2.1.2: Tme seres plots of major fresh products mported from Chna by the US from 2002 to 2008, Arrowroot, Salep. g33=71029 LEGUMES, FROZEN rqchn m1 2004m1 2006m1 2008m1 tme Fgure 2.2.1: Tme seres plots of major processed vegetable products mported from Chna by the US from 2002 to 2008, Legumes, Frozen

27 19 g36=71080 VEGETABLES, FROZEN rqchn m1 2004m1 2006m1 2008m1 tme Fgure 2.2.2: Tme seres plots of major processed vegetable products mported from Chna by the US from 2002 to 2008, Vegetables, Frozen rqchn g49=71290 VEGETABLES NES DRIED 2002m1 2004m1 2006m1 2008m1 tme Fgure 2.2.3: Tme seres plots of major processed vegetable products mported from Chna by the US from 2002 to 2008, Vegetables Nes dred

28 20 g52=71331 URD,MUNG,BLACK DRIED rqchn m1 2004m1 2006m1 2008m1 tme Fgure Tme seres plots of major processed vegetable products mported from Chna by the US from 2002 to 2008, Urd, Mung, Black dred g54=71333 KIDNEY BEANS, DRIED rqchn m1 2004m1 2006m1 2008m1 tme Fgure 2.2.5: Tme seres plots of major processed vegetable products mported from Chna by the US from 2002 to 2008, Kdney Beans, Dred

29 21 There has been a surge n U.S. agrcultural mports from Chna snce 2002, largely explaned by mports of hortcultural products. The U.S. hortcultural sector can be categorzed nto three prncpal areas: (1) vegetables; (2) nursery/greenhouse; and (3) fruts, nuts, and berres. Although the growth of U.S. mports of vegetable products from Chna has been sgnfcant, Chna s not yet a man exporter to the Unted States but rather a strong compettor n thrd markets. Chna s the fourth largest suppler of vegetables to the U.S. market, but only accounts for 4 percent of U.S. mports, behnd Mexco, Canada, and Span. Chna s mports of hortcultural products ncreased greatly from 1990 to In 1990, they were only one thrd of the total mports of agrcultural products, but they now account for one half of all mports of agrcultural commodtes whle lvestock, meat products, sugar and tropcal products, and grans and olseeds account for the other half. 7 Carter and Gunnng-Trant (2006), pp1230, Fgure2, U.S. food mports from Chna by commodty groupng (constant 2004 U.S. dollars).

30 Chapter 3 A Revew of the Lterature on Analyzng Compettveness As global trade n fruts and vegetables has expanded at a hgher rate than trade n other agrcultural commodtes, many countres have tred to take part n the growng market and so has Chna. It s crucal to be compettve on the world market to reap the potental gans from ncreased demand. Compettveness can be assessed by examnng how well a country uses resources. Effcent labor and captal allocaton s a key to comparatve advantage. In a paper by Yurcan and Iskl (2006), about the nternatonal compettveness of Turksh agrculture, they computed Domestc Resource Cost Ratos (DRCs) to measure the compettveness of Turksh hortcultural products. The DRC compares the socal opportunty costs of domestc producton to value added generated at nternatonal prces. Consstent wth statstcs on trade flows, the authors fnd that Turkey has a comparatve advantage n tomatoes, melons, watermelons and tangernes snce ther DRC s smaller than 1. Ths means that the value added from exports exceeds the costs of the domestc resources to produce these, so that scarce domestc resources are beng used effcently n producng these products. Whle these calculatons show that t s to Turkey s domestc advantage to devote resources to these products, they do not reveal anythng about Turkey s compettve poston relatve to other exporters. But n ths thess, I focus on Chna s

31 compettve poston n the US market rather than the overall effcency of resource allocaton n Chna s agrcultural sector. 23 The most mportant factor n global competton s prce, but mport prces do not only depend on producton costs but also on trade polces n mportng countres. Import demand elastctes whch show how demand wll change as the mport prce changes are crucal for an exporter. Armngton elastctes and the compettveness term of the constant market share model are well-known methodologes for examnng the prce correspondence among tradng partners. Elastctes are more meanngful and valuable under the assumpton that mported products and domestc products are not the same. 3.1 Constant Market Share Method Snce ts ntroducton by Tyszynsk (1951), Constant Market Share (CMS) analyss has been a popular tool for analyzng a country s export growth or compettveness. Nevertheless, ts theoretcal foundatons (and polcy relevance) have been questoned (Houston (1967); Rchardson (1971a, b)). Merkes and Meer T. (1988) provde a theoretcal foundaton by relatng CMS analyss to a two-degree homothetc demand functon. The CMS terms are explaned n terms of dfferent behavoral components, whch provde a bass for polcy makng. CMS analyss separates export changes for a country nto four parts: (1) the world term; (2) the market term; (3) the commodty term; and (4) the compettveness term. The frst three represent the growth of exports whch would result f the country had mantaned constant market shares n each

32 24 market, and are jontly referred to as the structural term. The compettveness term captures the effect of changng market shares. Leamer and Stern (1970) and Rchardson (1971b) present CMS analyss by showng that the compettveness term can be explaned by prce changes. The man assumpton for CMS s that prce changes are not demand, but supply determned. Merkes and Meer T. confrm ths assumpton through ther emprcal results. There are two man drawbacks to the CMS method. Frst, the structural term of CMS has not been formally nterpreted, n contrast to methods such as the Armngton Model (dscussed below) whch drectly addresses the mpact of prces on market share. Second, CMS analyss s less relevant for analyzng long term export changes. Its applcablty declnes f we consder longer tme perods. Ths concluson s supported by consderng the polcy relevance of CMS analyss. The extent to whch export structure can be polcy adjusted to antcpated export growth s lmted by the extent to whch ths export growth can be forecast. Besdes, there are many factors to be consdered n export growth and compettveness such as factor avalablty, technology, market structure, demand patterns, and government polces n the focus country, ts customers, and ts compettors. Snce ths thess s focused on compettveness n a sngle mportng country, whch requres nformaton and data on the demand sde, the CMS model does not ft well.

33 Armngton Elastctes (Foregn products and domestc products) The Armngton model explans mports from country A relatve to those from country B as a functon of relatve mport prces. The model dfferentates commodtes by ther country of orgn (natonal product dfferentaton), and assumes mports to be mperfect substtutes for each other. Armngton elastctes are very mportant n trade polcy analyss. Usng economc models to evaluate changes n trade polcy generally requres the converson of polcy changes nto prce effects. Armngton elastctes are a key part of models that reflect ths. They can also show the degree of substtuton between mported and domestc goods due to changes n the relatve prce of those goods. Gallaway et al. (2003) summarze the roles and mportance of Armngton elastctes n the trade lterature. Frst, the magntude of the trade-substtuton elastcty s mportant n the debate regardng the border effect. Second, t s a key varable n testng Grossman and Helpman s Protecton for Sale model 8. Thrd, t plays a key role n appled modelng ncludng partal and general equlbrum models that are often used to assess the ex ante mpacts of polcy changes such as taxes and tarffs. For example, mult-country computable general equlbrum (CGE) models used to analyze tarff and 8 Grossman and Helpman (1994).

34 26 trade polcy changes typcally ncorporate the Armngton structure. The degree to whch a polcy change wll affect a country s trade balance, level of ncome and employment depends on the magntude of the elastctes used n the model. Armngton models have some mportant general propertes. Frst, there s no comparatve advantage and hence no gans from trade due to product specalzaton; second, the number of products s fxed and hence there are no gans from trade due to ncreased product varety; thrd, large terms of trade effects tend to offset other gans from trade. The most crucal assumpton that underles the Armngton model s that the utlty functon s weakly separable n the types of goods, gvng Armngton dfferentaton. Further research has mproved the Armngton model by dstngushng between horzontal and vertcal dfferentaton. Wth the development of CGE models by ncorporatng ntermedate nputs, the Armngton model can be extended by dfferentatng the demand for ntermedate nputs by natonalty (e.g., Yue and Beghn 2008). There s a substantal lterature contanng estmates of Armngton elastctes. The fve man studes avalable for US mports are Stern et al (1976), Shells et al (1986), Renert and Roland-Holst (1992), Shells and Rener (1993) and Gallaway, McDanel et al (2000). Shells and Renet (1993) dsaggregated U.S. mports nto those from the NAFTA members and those from the rest of the world (ROW). But they found estmated elastctes to be relatvely nsenstve to estmaton procedures. McDanel (2002) summarzed the fndngs from emprcal research on Armngton elastctes. Frst, long-

35 27 run estmates are hgher than short-run estmates. Second, more dsaggregated analyses fnd hgher elastctes. Thrd, reduced-form tme seres analyses generally fnd lower elastctes relatve to cross sectonal studes that nclude a consderaton of supply condtons. 3.3 Analyzng the Impact of Non Tarff Barrers on Imports Non Tarff Barrers (NTBs) refer to the wde and heterogeneous range of polcy nterventons other than tarffs that affect and dstort trade of goods, servces, and factors of producton. However, because of uncertanty and mperfect nformaton, t s very challengng to quantfy the sze and effects of NTBs. The most successful methodologes for estmatng NTBs for goods have nvolved some sort of prce comparson to nfer the tarff equvalent of a non-tarff barrer, and must be derved from ndustry-specfc nformaton. There are an ncreasng number of studes on measurng NTB effects. NTBs are lkely to reduce the elastcty of mport demand. Ths suggests that t mght be worthwhle to perform a tme seres estmaton of mport demand elastctes, n the aggregate or for ndvdual commodtes/sectors, usng data from several countres and nteractng country dummy varables n order to determne how these elastctes may dffer across countres. Ths should provde an ndcaton of the extent to whch the prce responsveness of mport demand s reduced by NTBs n some countres more than others. Such a measure could be qute general n ts coverage, snce t would nclude even cultural barrers and other nsttutonal factors that may restrct trade.

36 28 We suppose that the hgher the elastcty of substtuton relatve to mports from a partcular country or country group (e.g., NAFTA), the more compettve that country s exports. What factors wll affect compettveness? Apart from producton costs on the supply sde, tarff or non tarff barrers n the mportng country have a drect mpact on compettveness. Leamer s (1990) comparatve advantage approach s wdely appled for nferrng the effects of NTBs. Ths conssts of predctng mports usng factor endowments. In the presence of NTBs, actual mports are dfferent from predcted mports. Ths dfference represents the mpact of NTBs on trade flows. Ths quantty mpact s then converted nto an ad-valorem tarff equvalent (AVE) usng mport demand elastctes. Leamer compared two models: the comparatve advantage model and a dummy varable model and determned that the former has greater explanatory power. The frst model allows for both nontarff and tarff barrers to deter trade wth effects that vary by mporter and commodty. It also allows trade to be dverted from other mporters f they have relatvely hgh barrers. Leamer analyzed both the effects on mporters and on exporters. He concluded that the model whch accounts for dfferences n comparatve advantage n terms of measurable characterstcs of commodtes and countres does a farly good job of explanng the structure of trade and provdes statstcally sgnfcant and economcally large estmates of the effects of barrers. 9 He fnally concluded that Latn Amercan 9 Leamer (1990), pp362.

37 exports to a group of 14 mporters have been reduced by trade barrers by a total of 34 percent, varyng from 5 percent for Mexco to 75 percent for Argentna. 29 A second method whch also provdes AVEs of NTBs has been developed by Kee et al (2006). Usng Leamer s (1990) comparatve advantage approach, the authors estmate the quantty mpact of two broad types of NTBs (core NTBs and agrcultural domestc support) on mports at the HS sx-dgt tarff lne. They do ths by computng a Trade Restrctveness Index (TRI)), usng nformaton on tarffs, AVEs of NTBs and elastctes of mport demand at the tarff lne level. Kee et al (2006) frst calculate elastctes of mport demand at the tarff lne level, and then followng Anderson and Neary (1992, 1994) and Feenstra (1995), they use ther estmates to construct theoretcally sound TRIs and GDP losses assocated wth exstng tarff structures. The countres analyzed are determned to be 30 percent more restrctve than ther smple or mport-weghted average tarffs would suggest. Thus, dstorton s nontrval. The authors estmate that GDP losses from trade restrctons are largest n the Unted States, Chna, Inda, Mexco and Germany. 3.4 Implcatons for the Approach Used n Ths Study Groupng by measurable characterstcs of commodtes and countres, as suggested by Leamer, s a promsng way to examne compettveness gven specfc trade barrers. In ths study, correlatons between mport patterns and the degree of processng are used to determne group categores.

38 30 An attempt s made to analyze a country s compettveness compared to competng countres n one specfc mportng country (the Unted States). Most of the current lterature analyzes compettveness n global markets or for multple mportng countres. Ths study focuses on compettveness n a sngle market at a more dsaggregated level to deal wth the heterogenety of products. Frst, exporters to the Unted States are dvded nto three groups, NAFTA (whch s the free trade zone), Chna and the rest of the world. Although Shells and Renert (1993) dsaggregated U.S. mports nto those from the NAFTA members and those from the rest of the world (ROW), and found the estmates of Armngton elastctes to be relatvely nsenstve to alternatve estmaton procedures, I examne whether ths s so for more dsaggregated commodtes, n ths case vegetables. Rather than comparng Chna s products wth domestc US products, I use NAFTA as a compettve benchmark snce the focus s on the relatve compettveness of exporters to the Unted States. In choosng ntra-ndustry groups, I use the correlaton of mports from Chna wth those of competng exporters and the degree of processng as crtera for establshng product categores. Fnally, both monthly and annual data are used.

39 Chapter 4 Methodology 4.1 Introducton In ths study I want to examne whether prce s a domnant factor n the pattern of US mports of vegetable products from Chna. The estmaton of an mport demand functon s typcally used for ths purpose. However, based on the assumpton that all WTO members face the same tarff level when they enter the US market, except the countres from the free trade agreement group NAFTA, I also use the quantty and prce of NAFTA mports as a benchmark for deal trade regulatons (most favorable condtons for access to the US market) and examne the relatonshp between relatve mports and relatve prces. To perform the analyss I dvde compettors n the US market nto three groups: Chna; the countres n NAFTA (Canada and Mexco); and the rest of the world. I use Chn to denote mports from Chna, NAFTA to denote mports from NAFTA, and ROW to denote those from the rest of the world. I examne the rato of the quantty of mports from Chna to NAFTA and ROW to NAFTA as a functon of the rato of ther relatve prces, computed from average mport values. Snce there s substantal varaton n the mports of many products by month, I nclude dummy varables for each month n an attempt to capture ths effect. January s chosen as the benchmark month and so there are eleven dummes n total. The varaton among months can be caused by the natural

40 seasonal effects of products and also changes created by trade restrctons at specfc tmes. 32 There are two man steps n the econometrc analyss n ths study. Frst I examne how much the prce rato can explan the rato of mports. I then compare substtuton elastctes across two groups: the elastcty of Chna s mports wth respect to those from NAFTA; and that of the ROW wth respect to NAFTA. Followng emprcal analyss for the full dataset, I examne subsets of products to see f there are sgnfcant dfferences among these. Several characterstcs have been consdered on the bass of the data to determne these groupngs such as correlaton, trend, seasonal and epsodc patterns, as well as the degree of processng of the products concerned. Correlaton and degree of processng are used as the prmary crtera for establshng categores snce these seem to carry the most useful nformaton for establshng dfferences. To sum up, there are three steps for the emprcal analyss. The frst step s to select an approprate model from several dfferent models that have been constructed and compared. The second s to compare the substtuton elastctes for Chna and the Rest of the World. The thrd step s to apply the model to dfferent product categores to examne whether dfferences exst across these. Varables used n constructng the models are defned n Tables 4.1.

41 33 Table 4.1: Varable names and defntons hs6code Hs6code of each product Tme monthly tme data Cp monthly CPI data Pop monthly Pop data rqchn1 rpchn1 logrqchn1 logrpchn1 rpchn2 Qchn Pchn Qnafta Pnafta Prow Qrow Rqrow Rprow Logrqrow Logrprow Imonth_ the quantty rato of mports from Chna and NAFTA the prce rato of mports from Chna and NAFTA log(rqchn1) log(rqpchn1) the prce rato of mport products from Chna and ROW the mported quantty of products from Chna the prce of products mported from Chna, we calculate ths varable from vchn/qchn the mported quantty of products from NAFTA ( ncludes Canada and Mexco) the mported prce of products from NAFTA, we calculate ths varable from vnafta/qnafta the prce of the products mported from rest of the world, we calculate t from vrow/qrow the mported quantty of products from Rest of the world the quantty rato of mports from ROW and NAFTA the prce rato of mports from ROW and NAFTA log(rqrow) log(rprow) dummy varable for each month (February to December)

42 4.2 Model Selecton 34 The essental purpose of the econometrc model s to examne the substtuton elastctes between mports from Chna and the Rest of the World wth NAFTA, whch s used as the benchmark for free trade. The rato of the quantty of mports s used as the dependent varable and the rato of prces (average mport values) s the basc ndependent varable. The followng models are specfed: A lnear model Model 1 and non lnear model Model 2 are compared frst. The prce data n my thess s calculated from the value and quantty data. So t s closely correlated wth quantty data. The smallest value for rpchn and rqchn are zeros, and when rqchn equals zero, rqchn wll also be zero. So I suppress the constants n the two models and obtan two new models Model 3 and Model 4. To account for monthly varaton, monthly dummes have been added to the four models to generate four new models. Therefore, I have constructed eght models n total. Model 1: A lnear model has the quantty rato as the dependent varable and the prce rato as the ndependent varable: rqchn = b + β * rpchn + ε, = 1,2. (1) Model 2: Uses the logarthm of the quantty and prce ratos. Monthly effects refer to the sze of systematc monthly varaton and the logarthm of dependent and ndependent varables can reduce the scale of such varaton. Snce the logarthm does not exst for zero, ths requres treatment of zero observatons: log rqchn = b + β * log rpchn + ε, = 1,2. (2)

43 Model 3: Model 1 wthout constant: 35 rqchn = β * rpchn + ε, = 1,2. (3) Model 4: Model 2 wthout the constant: log rqchn = β * log rpchn + ε, = 1,2. (4) Model 5: Add dummy varables to Model 1 to account for systematc monthly varaton: j 12 = + * + = rqchn b β rpchn γ j * j. month + ε, = 1,2; j = 2,12. (5) = 2 j Model 6: Add dummes to Model 3. Consderng that the smallest value for the quantty and prce rato s zero, when the prce rato or quantty rato equals zero the model wthout constant could be an mproved specfcaton: j 12 = * + = rqchn β rpchn γ j * j. month + ε, = 1,2; j = 2,12. (6) j= 2 Model 7: Add monthly dummy varables to Model 2: j 12 log = + * log + = rqchn b β rpchn γ j * j. month + ε, = 1,2; j = 2,12. (7) j= 2 Model 8: Add monthly dummes to Model 4: j 12 log = * log + = rqchn β rpchn γ j * j. month + ε, = 1,2; j = 2,12. (8) = 2 j

44 4.3 Sub Group Analyss 36 Category 1: Correlaton After selectng the approprate model, that model s used to examne whether there are sgnfcant dfferences between groups of products. Ths addresses the heterogenety of products. The crtera used for determnng categores are gven below. Table 4.2: Varables for group category analyss pro1 pro2 cornafta1 cornafta2 corrow2 corrow3 cornafta corrow processng dummy varable for degree of processng 1 = fresh products dummy varable for degree of processng 2 = processed products dummy varable for negatve correlaton between mports from Chna and NAFTA dummy varable for postve correlaton between mports from Chna and NAFTA dummy varable for negatve correlaton between mports from Chna and ROW dummy varable for postve correlaton between mports from Chna and ROW the correlaton between mport quantty from Chna and NAFTA the correlaton between mport quantty from Chna and ROW The degree of processng, wth fresh products as 1 and processed products as 2. The processed products nclude frozen, dred and preserved products. Ths table shows the varables used for the sub group analyss. Cornafta s a dummy varable relatng to the correlaton between mports from Chna and those from

45 37 NAFTA. If Cornafta equals 1, then the correlaton s negatve, otherwse t s postve. The same logc apples to Corrow. There can be mssng data n runnng the correlaton test for a specfc product. The frst sub group s selected usng the condton that the sum of Cornafta1 and Corrow1 equals 2, whch means that both of the dummes are 1 and correspondngly both the correlaton of mports from Chna wth NAFTA and Chna wth the rest of the world are negatve. Snce most of the correlaton coeffcents are not statstcally sgnfcant, there s no sgnfcant dfference wthn groups. Group1: Cornafta1+Corrow1=2, N+N Group2 Cornafta2+Corrow1=2, P+N Group3 Cornafta1+Corrow2=2, N+P Group4 Cornafta2+Corrow2=2, P+P N+N: denotes that correlaton between the rato of mports of Chna and NAFTA and Chna and rest of the world are both negatve. P+N: correlaton of Chna and NAFTA s postve whle the correlaton of Chna and rest of the world s negatve. N+P: correlaton of Chna and NAFTA s negatve whle the correlaton of Chna and ROW s Postve. P+P: Both of the correlatons are postve. Table 4-3 shows the products ncluded under each of the four sub groups.