Competition of Imported Wooden Bedroom Furniture in the United States

Transcription

1 Competton of Imported Wooden Bedroom Furnture n the Unted States Yang Wan 1, Changyou Sun 2, and Donald L. Grebner 3 1 Graduate Research Assstant, Department of Forestry, Msssspp State Unversty, Box 9681, Msssspp State, MS ywan@cfr.msstate.edu. (662) Assstant Professor, Department of Forestry, Msssspp State Unversty, Box 9681, Msssspp State, MS csun@cfr.msstate.edu. (662) Assocate Professor, Department of Forestry, Msssspp State Unversty, Box 9681, Msssspp State, MS dgrebner@cfr.msstate.edu. (662)

2 Abstract Competton of Imported Wooden Bedroom Furnture n the Unted States The Unted States has been ncreasng the mport of wooden bedroom furnture from the world to meet ts domestc needs. Tradtonal supplers have been substtuted by developng countres such as Chna and Vetnam over the past decade. To explan ths market structure change, a dynamc Almost Ideal Demand System (AIDS) was employed to analyze ths mport market. Monthly mport data of the top seven countres from 2001 to 2008 were collected from the Internatonal Trade Commsson. The Engle-Granger contegraton test showed that the contegraton relatonshps exsted n ths demand system. Both homogenety and symmetry propertes were not rejected n the dynamc AIDS. The superortes of dynamc AIDS were also revealed n passng several msspecfcaton tests. The U.S. antdumpng nvestgaton on Chna was effectve n the short-run, but the trade dverson occurred from Chna to Vetnam, Indonesa, and Brazl. Fnally, the expendture elastctes suggested that U.S. consumers would spend more money on the wooden bedroom furnture from Vetnam, Chna, and Malaysa. All the own-prce elastctes were sgnfcantly negatve and the cross-prce elastctes ndcated the competton among those supplers. The results from ths study are helpful n understandng the competton among supplers, the consumer behavor n ths market, and the mpact of antdumpng polcy. Keywords: Almost Ideal Demand System (AIDS), antdumpng nvestgaton, Engle-Granger contegraton, wooden bedroom furnture 95

3 Introducton The Unted States has been experencng a rapd growth n the consumpton of furnture and ts furnture ndustry has made sgnfcant contrbuton to the domestc economy over the last decades. Its domestc retal sales steadly grew n recent years, exceedng $100 bllon n 2003 (Gazo and Quesada 2005). Accordng to U.S. Bureau of Census (2006), the total value of shpments of furnture reached $85.6 bllon n 2006, whch was equal to 5.4 % of the manufacturng ndustres GDP. However, an ncreasng share of the rsng furnture demand has been met by the large mport from foregn countres. The mport value of wooden bedroom furnture clmbed from $0.6 bllon n 1996 to $3.8 bllon n Tradtonally, the Unted States mported furnture from Canada, Italy, and Tawan, to name a few. At present, the newly developng countres such as Chna, Vetnam, and Malaysa have substtuted the furnture from tradtonal supplers and begun to domnate the U.S. furnture mport market n recent years. In partcular, Chna has been leadng the wooden bedroom furnture market and accounts for 44% market share over Ths trade phenomenon has led to serous threat to the furnture ndustry and aroused wde concerns about the future of the domestc furnture ndustry. To protect the furnture ndustry, antdumpng nvestgaton (U.S. ITC. 2004) on Chna was conducted durng the perod of 2003 and The fnal antdumpng dutes (0.83% %) have been mposed on the wooden bedroom furnture from Chna, but ts total mport value contnues to rse at present. The objectve of ths study was threefold. Frst, the consumer behavor n the mported wooden bedroom furnture was examned n ths changng market settng. Second, how dfferent suppler countres compete n ths market was analyzed. Thrd, the effectveness of the antdumpng nvestgaton on Chna durng the perod of was evaluated. To complete the above objectves, the dynamc AIDS was employed and varous tests were used to examne theoretcal propertes and model robustness. The results from ths study are helpful n understandng the competton among supplers, the consumer behavor n ths market, and the mplcatons of the current polcy. Market Overvew The U.S. mport of wooden bedroom furnture (wbf) has been steadly growng over the perod of , as shown n Fgure 1. U.S. monthly mport value was about $50 mllon n 1996 and reached ts peak value of $353.8 mllon n August Bed s one of the major products of wooden bedroom furnture as defned by the Internatonal Trade Commsson (ITC). There was an upward trend over the same perod. The rapd growth of mported wbf was largely due to the ncreasng mport of bed from all over the world. In 1996, bed s monthly mport value was about $10 mllon, one ffth of wbf. It rapdly clmbed to $100 mllon per month snce 2004, about one thrd of wbf. Tradtonally speakng, the U.S. manly mported the bed from supplers such as Canada, Indonesa, and Italy. The newly developng Asan countres such as Chna and Vetnam have demonstrated ther potentals to domnate the U.S. market. Chna has been steadly ncreasng ts export to the U.S and become the leadng suppler n the U.S. market snce The mport 96

4 value from Chna was only $5.6 mllon n 1996, but clmbed to $418 mllon n 2008, accountng for 35.4% market share. Although the annual export value of Vetnam was lesser than Chna, t stll was an mportant suppler n the U.S. mport furnture market. Vetnam began to export ts bed to U.S. from 2001 and ht $ mllon n 2008, closely followng Chna. Import Value ($ Mllon) Value - wooden bedroom furnture 2 Value - bed a Oct b Jul c Jan a b c Fgure 1. Monthly mport value of wbf and bed from January 1996 to December 2008 Methodology In ths study, a dynamc AIDS was employed to examne the consumer behavor, evaluate the competton among dfferent suppler countres, and assess the effectveness of antdumpng nvestgaton n the mported wooden bedroom furnture market. The semnal AIDS created by Deaton and Muellbauer (1980) has become one of the most popular demand analyss tools for researchers over the past three decades. The frequent applcaton of AIDS over the other demand models s largely due to ts outstandng advantages. AIDS s hghly consstent wth consumer theory. It can exactly satsfy the axom of choce. AIDS s also a desrable and flexble demand system, whch can take dfferent functonal forms to ncorporate dynamc factors. Varous functon forms have been appled n the emprcal demand studes so far. Anderson and Blundell (1982)frst put forward the dynamc adjustment of consumers expendture. Wth the development of econometrc technques, autoregressve dstrbuted lagged technque allows for several perods of short-run adjustments to long-run equlbrum status n AIDS. Due to the superortes mentoned above, dynamc AIDS has been wdely used n consumer s expendture on food (Balcombe and Davs 1996; Karaganns and Mergos 2002), meat products (Fanell and Mazzocch 2002; Karaganns et al. 2000), olve ol consumpton (Duffy 2003), and alcohol market (Seale et al. 2003). It also has been employed frequently n the feld of toursm demand analyss worldwde (De Mello and Fortuna 2005; L et al. 2004) 97

5 Consderng ths mported bed market, AIDS can be derved from mnmzng the bed expendture functon to attan a specfc utlty level at gven prces. It can be expressed as the functons of the logarthms of prces and total expendture n the followng system: n * (1) ln ln ( / ) s = α + γ p + β x P + δ D + ε j j j = 1 where s s budget share of mported wbf from suppler country ( = 1,2,,7). p j s the prce of bed from country j, x s the total expendture on all of the mported bed. The Stone s Prce Index s the lnear approxmaton to orgnal prce ndex P, whch usually takes the form n * * * of log p = s log p. Then x s deflated by P to get the real total expendture x / P. D ( = 1) s = 1 the antdumpng polcy dummy, and ε s the normal dsturbance term wth zero mean and constant varance. To comply wth the economc theores, the system of equaton [1] s requred to satsfy the propertes of addng-up, homogenety, symmetry, and negatvty. The addng-up property mples that the sum of all budget shares equals to one, whch requres α = 1; β = γ = 0; 0 Homogenety requres γ = 0 and suggests that the proportonal change n the expendture and j j all of the prces has no mpact on the quanttes purchased or the budget allocaton. Symmetry mples that the matrx of the prce dervatves s symmetrc, γ j = γ j. In ths study, tme seres property of all the varables have been nvestgated by Augmented Dckey-Fuller (ADF) unt root test frst. The Engle-Granger (1987)contegraton and error correcton technques are ntegrated nto the AIDS to consder both long-run and short-run consumer behavor. The dynamc AIDS (Feleke and Klmer 2007; Karaganns et al. 2000)can be formulated as follows: n * (2) ψ γ ln β ln( / ) η δ Δ s = Δ s + Δ p + Δ x P + ECT + D + u t, t 1 jt jt t t t t t j = 1 s the frst-dfference operator and ψ measures how the habt persstence nfluences the current n * ECT = a + γ ln p + β ln x / P + δ D s and η measures the consumpton. ECT s calculated as ( ) speed of short-run adjustment. j j j = 1 In the specfcaton of above models, the adequacy of the models should be examned, especally n modelng dynamc AIDS. Frst, the data used n ths study are tme seres data whch are often auto correlated. The Breusch-Godfrey (BG) test s used to test the hypothess of no seral correlaton n those varables. Second, the assumpton of homoscedastcty means that the varance s constant at each observaton pont. A falure of ths assumpton may result n the nvald nferences. The Breusch-Pagan (BP) test s employed to test heteroscedastcty. The assumpton of normally dstrbuted error term s tested by Jarque-Bera LM test. For the test of functonal msspecfcaton, Ramsey s Regresson Specfcaton Error Test (RESET) s adopted. 98

6 In addton, the assumpton of parameter constancy s tested by cumulated sum of squares (CUSUMSQ). Researchers and polcymakers are usually nterested n the expendture elastcty, own-prce elastcty, and cross-prce elastcty. In ths study, all those elastctes were calculated by the estmated parameters from AIDS and the average budget shares (s ) throughout the whole sample perod. The formulas are lsted as follows: (3) β Expendture elastcty: η = 1+ s (4) γ j s M j Marshallan (uncompensated) elastcty: ε = δ ( ) j j + β s s (5) Hcksan (compensated) elastcty: γ H j ε = δ j j + + s j s where δ j = 1f = jand δ j = 0 f j calculated by the above formula.. In addton, both long-run and short-run elastctes can be Data Sources and Varables In order to analyze the consumer behavor and competton n ths mport market, several factors were taken nto consderaton n ths study. Frst, the defnton of mported wooden bedroom furnture was ntroduced. The subject of ths study was manly focused on the specfc furnture under HTS , whch s defned as the wooden furnture of a knd used n the bedroom (bed) by Internatonal Trade Commsson (ITC). Second, the tme perod covered from January 2001 to December One major reason was that ths market structure has undergone a dramatc change durng the past decade. Another mportant reason was that Vetnam began to export ts bed to the Unted States from January 2001 and has experenced tremendous growth n ths market durng the perod of Hence, to analyze the current market structure and the competton among newly developng countres, ths study focused on the perod from January 2001 to December After dentfcaton of the tme perod, major and representatve supplers need to be selected. The aggregate mport value of top seven supplers represented 85% of the total mport durng the perod of 2001 and They were Chna (44.2%), Vetnam (11.7%), Indonesa (7.8%), Malaysa (6.4%), Canada (6.3%), Brazl (4.4%), and Italy (4.2%). All the other countres were aggregated nto the Rest-of-world (ROW) (15.0%). Next, the monthly cost-nsurance-freght (CIF) (dollars) and quantty (pece) data were collected from ITC to construct the varables n AIDS. The total expendture was the total CIF value of bed mported from all over the world. The budget share of each suppler was the percentage of the CIF value of that suppler over the total expendture. The mport prce measured by the unt value of mported bed was calculated as CIF value over quantty. 99

7 Emprcal Results Model estmaton and tests The tme seres propertes of budget share, mport prce, and total expendture were formally examned by ADF unt root test frst. All the results were lsted n Table 1. The null hypothess of unt root n budget share faled to reject at 5% sgnfcance level, whch suggested that all of them had unt roots over All of the mport prces were not statonary except Malaysa and Canada. Fnally, the ADF test suggested that the total expendture also had unt root. However, all the frst-dfference varables were statonary. These results ndcated that the level of data seres were ntegrated n order 1, but the frst-dfference data were zero. The next step was to examne the long-run equlbrum relatonshp n the mported bed market by Engle-Granger contegraton test. The results n Table 2 ndcated that all the equatons were contegrated at the 5% sgnfcance level. Due to the exstence of long-run relatonshp n ths market, the dynamc AIDS can be further establshed to consder the short-run consumer behavor. After the frst-dfference of all the varables, the dynamc AIDS [2] were estmated by SURE agan. Table 1. ADF unt root test of budget share, mport prce, and total expendture Jan to Dec Varable Level ADF Frst-dfferenced ADF T & C Trend None Lag Order None Lag Order Budget share s.cn I(1) I(0) s.vn I(1) I(0) s.id I(1) I(0) s.my I(1) I(0) s.ca I(1) I(0) s.br I(1) I(0) s.it I(1) I(0) s.row I(1) I(0) Import prce ln(p.cn) I(1) I(0) ln(p.vn) I(1) I(0) ln(p.id) I(1) I(0) ln(p.my) I(0) ln(p.ca) I(0) ln(p.br) I(1) I(0) ln(p.it) I(1) I(0) ln(p.row) I(1) I(0) Total expendture ln(totexp) I(1) I(0) 100

8 Table 2. Contegraton test of equatons n AIDS Equaton ADF (T & C) ADF (T) ADF (N) Lag Contegrated? eq. Chna Yes eq. Vetnam Yes eq. Indonesa Yes eq. Malaysa Yes eq. Canada Yes eq. Brazl Yes eq. Italy Yes As mentoned prevously, msspecfcaton tests were essental for the parameter explanaton and further elastcty calculaton. A summary of msspecfcaton tests n dynamc AIDS were presented n Table 3. All the error terms dd not have seral correlaton problems at the 1% sgnfcance level by BG test. The results of BP test ndcated that all the error terms have constant varance. In addton, all the functons except Italy were correctly specfed at the 1% sgnfcance level. The error terms n fve out of seven equatons were normally dstrbuted. The hypothess of stable parameters faled to reject n all the equatons. Therefore, the superortes of the dynamc AIDS were further supported by passng those msspecfcaton tests. Demand elastctes In ths study, the long-run expendture, own-prce, and cross-prce elastctes were calculated by the formula [3], [4], and [5] from the estmated parameters of equlbrum AIDS. The short-run elastctes were calculated by the coeffcents of dynamc AIDS wth the same formula. Both long-run and short-run expendture elastctes and ther p-values were gven n Table 4. All the long-run expendture elastctes were postve and statstcally sgnfcant at 1% sgnfcance level except Canada and Italy. The demand of bed from Vetnam, Malaysa, and Chna were elastc. Among them, Vetnam had the hghest expendture elastcty of 2.737, followed by Malaysa wth and Chna wth These results ndcated that the more expendture spent on the mported bed, the more bed would be mported from Vetnam, Malaysa, and Chna n the long-run. As for the short-run expendture elastctes, all of them except Canada were also postve and statstcally sgnfcant at 1% sgnfcance level. Vetnam and Chna were detected as the major suppler countres n the short-run. Wth regard to the change of quantty n response of own-prce, Marshallan own-prce elastctes were calculated and lsted n Table 5. Both long-run and short-run values were negatve and sgnfcantly at the 1% sgnfcance level, as expected from consumer theores. All the values were elastc and senstve to the change of ts own prce n addton to Chna. In addton, both long-run and short-run own-prce elastctes of Chna were relatvely lower than the other compettors. Ths result mpled that trade polcy n order to ncrease the prce of bed from Chna would have lmted mpacts on ts demand quanttes, whch was consstent the neffectveness of antdumpng nvestgaton on Chna. 101

9 Table 3. Msspecfcaton tests of dynamc AIDS Equaton BG BP RESET J-B CUSUMSQ eq. Chna 0.01 ** eq. Vetnam 0.10 * * 0.00 *** 0.06 * eq. Indonesa 0.06 * ** eq. Malaysa eq. Canada 0.02 ** eq. Brazl 0.07 * ** 0.78 eq. Italy 0.09 * *** 0.00 *** 0.02 ** ***, **, * denotes sgnfcance at 1%, 5%, and 10%, respectvely. Table 4. Long-run and short-run expendture elastcty Country Long-run Short-run Estmates p-value Estmates p-value Chna Vetnam Indonesa Malaysa Canada Brazl Italy ROW Hcksan cross-prce elastctes were calculated under the assumpton of keepng the utlty constant. The short-run cross-prce elastctes were reported n Table 6. If the prce of bed from Chna went up by 1%, the mported quanttes from Vetnam and Malaysa went up by 0.43% and 0.53%, respectvely. The results suggested that the bed from Chna can be substtuted by the mport from Vetnam and Malaysa. In contrast, the 1% ncrease of prce from Vetnam and Malaysa resulted n the 0.11% and 0.08% ncrease of mport from Chna. Moreover, the degrees of substtuton were asymmetrc. Malaysa and Vetnam had hgher cross-prce elastctes than Chna to each of them. Antdumpng nvestgaton effects The purpose of antdumpng polcy on Chna was to curtal the mport from Chna, and furthermore, to protect the furnture ndustry n the Unted States. In ths study, both the longrun and short-run antdumpng effects were detected n the AIDS, lsted n Table 7. In the longrun, after the prelmnary determnaton was mplemented to collect antdumpng dutes n July 2004, the mports share of Chna decreased by 6.3% whle the mport from Vetnam, Indonesa, and Brazl ncreased by 3.9%, 0.9%, and 2.6%, respectvely. Ths result ndcated that the reducton of mport from Chna dverted to the other countres such as Vetnam, Indonesa, and Brazl. In the short-run, the mport from Chna decreased by 13.8%, but the mport value from Vetnam, Indonesa, and Brazl ncreased by 5.4%, 2.2%, and 2.0% at the same tme. There was a sharp drop of Chna s budget share and were sudden jumps of Vetnam, Indonesa, and Brazl. 102

10 Table 5. Long-run and short-run Marshallan own-prce elastcty Country Long-run Short-run Estmates p-value Estmates p-value Chna Vetnam Indonesa Malaysa Canada Brazl Italy ROW Table 6. Short-run Hcksan cross-prce elastcty Impact on Change n prce of demand Chna Vetnam Indonesa Malaysa Canada Brazl Italy ROW Chna (0.001) (0.733) (0.000) (0.402) (0.428) (0.298) (0.092) Vetnam (0.001) (0.078) (0.556) (0.152) (0.763) (0.055) (0.218) Indonesa (0.733) (0.078) (0.386) (0.004) (0.546) (0.905) (0.009) Malaysa (0.000) (0.556) (0.386) (0.063) (0.148) (0.072) (0.092) Canada (0.402) (0.152) (0.004) (0.063) (0.008) (0.862) (0.002) Brazl (0.428) (0.763) (0.546) (0.148) (0.008) (0.204) (0.200) Italy (0.298) (0.055) (0.905) (0.072) (0.862) (0.204) (0.115) ROW (0.092) (0.218) (0.009) (0.092) (0.002) (0.200) (0.115) Table 7. Long-run and short-run antdumpng nvestgaton effects Dummy Chna Vetnam Indonesa Malaysa Canada Brazl Italy Long-run effect Antdumpng ** Short-run effect Antdumpng *** *** * ** Antdumpng : Imposed antdumpng dutes from July

11 Conclusons The Unted States has been ncreasng the mport of wooden bedroom furnture from the world to meet ts domestc needs. Over the last decade, tradtonal supplers have been substtuted by the newly developng countres n Asa such as Chna and Vetnam. The antdumpng nvestgaton aganst Chna was mplemented by ITC n order to protect the furnture ndustry n the U.S. To explan ths market structure change, dynamc AIDS was employed to analyze the consumer behavor, and furthermore, evaluate the effectveness of ths antdumpng nvestgaton. Monthly dsaggregate data of the top seven supplers from 2001 to 2008 were collected from the ITC. The emprcal results n ths study reached several conclusons about the consumer behavor, market competton, and antdumpng effectveness n ths market. Frst, both consumers longrun and short-run choces on the mported wooden bedroom furnture were detected. The expendture elastctes dsclosed that the U.S. consumers wll spend more expendture on the wooden bedroom furnture mported from Vetnam, Malaysa, and Chna n the long-run. Among them, Vetnam has demonstrated ts great potental to be the top one suppler to the U.S. market. As ndcated by the own-prce elastctes, the mported quanttes from most countres were senstve to the change of ts own prce n both short-run and long-run. In partcular, Vetnam had the hghest own-prce elastcty. In partcular, both long-run and short-run ownprce elastctes of Chna were nelastc, whch ndcated the less senstve change n response of prce and further explaned the neffectveness of antdumpng nvestgaton. The degrees of competton and substtuton were revealed by the cross-prce elastctes between those suppler countres. The degree of substtuton among those pars was not symmetrc. The potental of Vetnam to domnate ths market were further proved ts sgnfcant cross-prce elastctes n both long-run and short-run. The cross-prce elastctes were relatvely small, whch were not elastc. The small magntude of those cross-prce elastctes mpled that they were far from perfect substtutes. These results further ndcated that the U.S. consumers do have dfferent preferences on the mported wooden bedroom furnture from all over the world. As a trade protecton nstrument, antdumpng nvestgaton dd not work as t ntended to reduce the mport. The affrmatve determnaton and mposton of antdumpng dutes led to 13.8% drop on the mport value of Chna. However, the trade dverson took place from Chna to Vetnam, Indonesa, and Brazl at the same tme, whch was consstent wth the dscusson by Brenton (2001). Overall, the effectveness of ths antdumpng polcy was not obvous, whch can be further evdenced by the contnuous growth of total mport value. Ths study brngs up several nterestng questons. For example, the competton among domestc and mported bed can be further studed, whch wll descrbe a more accurate pcture of ths market. However, the same defnton of bed and sales data are needed to ncorporate nto ths demand system. Moreover, whether ths market s ntegrated or not s also of great nterest. Further examnng the welfare change after the mplementaton of antdumpng polcy can mprove our understandng of the benefts of domestc retaler, producers, and consumers. 104

12 Lterature Cted Anderson, G., and Blundell, R Estmaton and hypothess testng n dynamc sngular equaton systems. Econometrca 50(6): Balcombe, K.G., and Davs, J.R An applcaton of contegraton theory n the estmaton of the almost deal demand system for food consumpton n Bulgara. Agrcultural Economcs 15(1): Brenton, P Ant-dumpng polces n the EU and trade dverson. European Journal of Poltcal Economy 17(3): De Mello, M.M., and Fortuna, N Testng alternatve dynamc systems for modellng toursm demand. Toursm Economcs(57): Deaton, A., and Muellbauer, J An almost deal demand system. Amercan Economc Revew 70(3): Duffy, M On the estmaton of an advertsng-augmented, contegratng demand system. Economc Modellng 20(1): Engle, R.F., and Granger, C.W.J Contegraton and error correcton: representaton, estmaton, and testng. Econometrca 55(2): Fanell, L., and Mazzocch, M A contegrated VECM demand system for meat n Italy. Appled Economcs 34: Feleke, S.T., and Klmer, R.L Analyss of the demand for mported meat n Swtzerland usng a dynamc specfcaton: mplcatons for the European Unon. Agrbusness 23(4): Gazo, R., and Quesada, H.J A revew of compettve strateges of furnture manufacturers. Forest Products Journal 55(10): Karaganns, G., Katrands, S., and Velentzas, K An error correcton almost deal demand system for meat n Greece. Agrcultural Economcs 22(1): Karaganns, G., and Mergos, G.J Estmatng theoretcally consstent demand systems usng contegraton technques wth applcaton to Greek food data. Economcs Letters 74(1): L, G., Song, H., and Wtt, S.F Modelng toursm demand: a dynamc lnear AIDS approach. Journal of Travel Research 43(2): Seale, J.L., Marchant, M.A., and Basso, A Imports versus domestc producton: a demand system analyss of the U.S. red wne market. Revew of Agrcultural Economcs 25(1): U.S. Bureau of Census Annual survey of manufacturng, statstcs for ndustry groups and ndustres. Washngton, DC: Department of Commerce U.S. ITC Wooden bedroom furnture from Chna. Investgaton No. 731-TA-1058, Publcaton U.S. Internatonal Trade Commsson, Washngton, D.C. 105