Citation for published version (APA): Hai, L. T. D. (2003). The organization of the liberalized rice market in Vietnam s.n.

Transcription

1 Universiy of Groningen The organizaion of he liberalized rice marke in Vienam Hai, L.T.D. IMPORTNT NOTE: You are advised o consul he ublisher's version (ublisher's PDF) if you wish o cie from i. Please check he documen version below. Documen Version Publisher's PDF, also known as Version of record Publicaion dae: 2003 Link o ublicaion in Universiy of Groningen/UMCG research daabase Ciaion for ublished version (P): Hai, L. T. D. (2003). The organizaion of he liberalized rice marke in Vienam s.n. Coyrigh Oher han for sricly ersonal use, i is no ermied o download or o forward/disribue he e or ar of i wihou he consen of he auhor(s) and/or coyrigh holder(s), unless he work is under an oen conen license (like Creaive Commons). Take-down olicy If you believe ha his documen breaches coyrigh lease conac us roviding deails, and we will remove access o he work immediaely and invesigae your claim. Downloaded from he Universiy of Groningen/UMCG research daabase (Pure): h:// For echnical reasons he number of auhors shown on his cover age is limied o 10 maimum. Download dae:

2 Chaer 6. Rice Marke Performance nalysis 6.1 Inroducion s menioned in Chaer 2, marke erformance refers o economic resuls ha include effeciveness, equiy, roduciviy and rofiabiliy. In his chaer, rice marke erformance will be eamined along several dimensions: (1) Measuring effeciveness of he eising markeing channels. Tesing wheher he eising markeing channels offer roer service ouus or he righ services in relaion o consumer references, (2) nalyzing markeing coss, rice margins and rofiabiliy among he differen rice markeing aciviies in order o measure he degree of rice markeing efficiency; and (3) Through emoral rice analysis o eamine o wha een rice aerns on differen marke-laces cohere (See also Secion and Secion in Chaer 2). Based on he rimary and secondary daa, he average rice a differen marke levels will be esimaed. Then he average rice margins for differen yes of rice raders will be comued. The ercen share of each cos iem for each ye of rice rader will be calculaed. Marke inegraion analysis will be used o measure he degree of rice variaion beween differen surveyed rice markes. Trend and seasonal rice flucuaion are also considered. 6.2 The effeciveness of rice markeing channels The effeciveness of he markeing channel is assessed by eamining how well he markeing channels mee consumers and roducers demand for services in relaion o heir references. In a roerly funcioning marke, markeing channels have o guaranee ha consumers can buy and roducers can sell heir roducs a reasonable rices in a marke lace; hey have o balance suly and demand in each marke segmen a any ime. In his sudy, he effeciveness of rice markeing channels is mainly assessed by differen services ha he channels have offered in he marke in order o maimize consumer saisfacion: assembling, disribuion (echange funcions); sorage, ransor (hysical funcions); grading, financing (faciliaing funcions). Regarding he major asks of he rice marke: assembling and disribuion, he resuls of he markeing channel analysis in Chaer 4 indicaed ha a number of differen markeing channels eis o serve rice farmers (roducers) and final consumers. Boh rivae and governmen channels suly he domesic and foreign markes. Moreover, barriers o enry ha hamer he rice formaion rocess in he marke were no deeced.

3 The Organizaion of he Liberalized Rice Marke in Vienam No major local suly shorages have been revealed by he surveys. By roducing hree rice cros er year, rice roducers can suly he marke more or less coninuously hroughou he year. Rice is always available for final consumers a any ime. However, for eors he sorage aciviy is an imoran issue. Rice millers and eor comanies (he SOEs) usually have o kee in sore a cerain amoun of addy/rice in order o fulfill he requiremens of he rice imorers and o deliver he rice in ime. s he eor marke is sill dominaed by he SOE s, having access o credi from he sae banks in he framework of he floor-rice olicy, no major roblems were deeced during he eriod under sudy. Wih resec o he ransor funcion, mos rice raders in he Mekong River Dela have he advanage ha hey can use heir own boas (chea ransor) or use ransor services from rivae ransor co-oeraives. The rice raders didn signal any ransoraion roblem in boh domesic and eor markes. Grading or form convenience is usually analyzed a he reail marke. his level we can observe he areciaion of final consumers and he degree of roduc suiabiliy. The daa used for analyzing form convenience are based on a SEB under-graduae hesis ha was conduced in 1999 (Vu Ngoc Tuan). In his research, niney rice consumers locaed a Canho Ciy were inerviewed in order o obain heir references wih regard o differen yes of rice. Table 6.1 Major rice characerisics referred by rice consumers in Canho Rice characerisic Number of resonses % of oal 1. Sof 2. Sicky 3. Good smell afer cooking 4. High degree of blooming 5. Easy o cook Toal Source: Vu nh Tuan, Under-graduae hesis 2000, SEB, Canho Universiy Table 6.1 shows ha sof and sicky rice as well as good smell are he mos imoran characerisics for rice consumers. Since Vienam is a roical counry wih ho weaher, rice becomes dry very quickly afer cooking. Therefore, sof and sicky are he mos imoran characerisics of good rice o many consumers. Furhermore, rice wih a good smell afer cooking is referred by medium 138

4 Chaer 6: Rice Marke Performance nalysis and high income consumers. To low income consumers, high degree of blooming of rice afer cooking is also imoran (Tuan, Vu nh 2000). Table 6.2 Major rice varieies markeed by rice reailers in Canho Rice qualiy Percenage of oal sales (%) 1. Mobui Tainguyen IR Jacksmin Nanghom Ohers 1.0 Toal Source: Vu nh Tuan, Under-graduae hesis 2000, SEB, Canho Universiy On he suly side, he figures in Table 6.2 indicae ha rice roducers/farmers and raders in he marke ry o offer suiable yes of qualiy rice o saisfy he consumer references. s menioned in Secion in Chaer 3, Mobui, Tainguyen, and IR64 are he mos oular varieies ha rice roducers/farmers grow and suly in he marke. These kinds of varieies have good characerisics such as sof, sicky, and blooming, ha saisfy he references of rice consumers (See Table 6.1). Nanghom and Jacksmin are very good rice varieies wih a good smell afer cooking. This mees he demand of medium and highincome consumers. By comaring Table 6.1 and Table 6.2 we can conclude ha he channel of rice disribuion is quie effecive in roviding he form and qualiy of rice o saisfy consumer references. However, if we link his resul wih he revious discussion abou he assormen of addy/rice qualiy in Chaer 3 and Chaer 4 we can see ha a he farm gae, wholesale, and reail level, he grading sandard of addy/rice qualiy is no consisen. he farm gae, rice farmers culivae many kinds of rice varieies (See Secion in Chaer 3). he wholesale markes, rice wholesalers, millers, and eorers only classify rice ino wo kinds: C1 and C2 qualiy (See Secion in Chaer 4). I means ha he eising wholesale grading sysem for rice qualiies is somewha biased. This grading sysem reflecs he references of he eor marke. However, reailers sor whie rice in many qualiies o mee consumer references (as discussed above). They assume heir markeing funcions of soring ou and assormen. s long as major rice varieies fulfil he references of consumers no roblems will occur. However, if furher secializaion reduces he suly of hese varieies a higher rice remium may be necessary o encourage farmers o roduce hese varieies. 139

5 The Organizaion of he Liberalized Rice Marke in Vienam In racice, rice reailers in he Mekong River Dela usually sell rice o consumers by kilogram wihou ackaging. 80 Some high qualiy rice such as Nang huong, Nang hom, Jacksmin is sold in 5 or 10 kilogram bags. This form of roduc is very convenien o consumers. 81 Oher rice consumers such as resaurans, schools/universiies, and sae organizaions receive sale services from he rice wholesalers in erms of free delivery of he roduc o heir laces. The rice wholesalers usually conac and negoiae wih wholesale consumers by elehone and deliver a suiable ye of rice wih free ransor service. Some households ha consume large amouns of rice (more han 100 kg er monh) also receive free ransor services from he rice wholesalers and or reailers. Finally, regarding he finance funcion, his sudy found ha in aricular large-scale raders face credi consrains. Mos rivae rice raders borrow money from informal sources such as friends, relaives or rivae moneylenders. In sum, he rice markeing channels in Vienam are organized quie effecively. Rice farmers and raders in he channels ry o offer convenien forms of roduc and good services o final consumers. The effeciveness of he domesic rice markeing channels is quie saisfacory, ece he consrained access o credi is a barrier for many rivae raders for furher develomen of he business. 6.3 Markeing coss and margin analysis The markeing margin refers o he difference beween rices a differen levels in he markeing sysem. The oal markeing margin is he difference beween wha he consumer ays and wha he roducer/farmer receives for his addy or rice, in oher words i is he difference beween reail rice and farm rice. wide margin means usually high rices o consumers and low rices o roducers. The oal markeing margin may be subdivided ino differen comonens: all he coss of markeing services and he rofi margins or ne reurns. The cos of markeing includes all he coss involved in he creaion of lace, ime, and form uiliies. Such coss should be recovered lus a reasonable reurn o invesmen. Coss vary deending uon he services rendered. The markeing coss usually include wages as reurn o labour; ineres as reurn o borrowed caial; ren as reurn o land and buildings; and rofi as reurn o enrereneurshi and risk caial. n analysis of markeing coss would esimae how much eenses are incurred for each markeing aciviy. I would also comare markeing coss incurred by differen acors in he channel of disribuion. 80 Normally, i varies from 5 o 20 kilograms er ransacion. 81 See also Secion in Chaer 5 for furher elanaion. 140

6 Chaer 6: Rice Marke Performance nalysis Paddy/rice markeing margins or rofi margins are comued as follows (See also secion 2.4.3): Profi margin of each ye of rice rader = Toal markeing margin of each ye of rice rader In his sudy, rice markeing coss and margins incurred by differen rice raders in one channel will be esimaed, channel 6 in able 4.16 (Farmers ssemblers Millers Wholesalers Reailers Consumers). This channel is he mos imoran channel in domesic rice markes (See Secion and Table 4.16). _ Markeing cos of each ye of rice rader Markeing coss of differen rice raders In oal 18 rice assemblers, 49 rice wholesalers, 53 rice millers/olishers, and 30 rice reailers were inerviewed. Differen yes of markeing coss relaed o he laes ransacion of rice assemblers, and markeing coss relaed o he las monh of business of rice wholesalers, millers/olishers, and reailers were recorded. This informaion on markeing coss allowed us o calculae he average markeing cos. 82 Table 6.3 shows ha he oeraing coss of rice assemblers are on average 46,500 or 76,000 VND/on of addy/rice deenden on he ye of ransor used. The mos imoran cos iem is ransoraion. I is abou 65 ercen o 68 ercen of he oal eenses. The ransoraion cos by ruck is quie high as comared o ransor by boa. The rice assemblers usually ay he coss for loading and un-loading. These coss do no vary ha much: 4, VND/on. The rice assemblers also ake ino accoun some losses (quaniy and qualiy reducion of addy/rice) and oher service coss during each shimen. 83 In case of ransor by ruck, he service coss are usually higher han hose of ransor by boa. Markeing coss of rice wholesalers are summarized in Table 6.4. On average, he oal markeing coss of rice wholesalers are 115,000 VND er on. Labour coss and maerial coss such as waer, fuel, elecriciy are he highes cos iems (24 25 ercen of he oal coss). Oher imoran cos iems concern ransoraion and some oher coss like ineres on borrowed caial, losses and dereciaion on equimen and he warehouse. The rice wholesalers also have o ay he business a and a license fee. he sar of heir business hey have o ay 82 The average markeing cos of differen rice raders in his secion are comued for he same addy/rice qualiy (25 ercen broken). 83 I includes arking fee, fee for ferry boa, ec. 141

7 The Organizaion of he Liberalized Rice Marke in Vienam around 550,000 o 850,000 VND for he license fee. In line wih he new a olicy, rice wholesalers have o ay 5 ercen VT. In sum, as shown in Table 6.4, medium and large-scale wholesalers, have o ay 18,000 VND er on as heir business aes and license. Table 6.3 verage markeing cos of rice assemblers in he Mekong River Dela (25 ercen broken rice) Markeing cos iems Transor by boa (1) (n=15) Transor by ruck (2) (n=6) Mar. cos (VND/Ton) (5) Percenage of he oal Mar. cos (VND/Ton) Percenage of he oal 1. Transoraion cos (3) 32,000 [0.72] ,000 [0.65] Loading 4,000 [0.33] 3. Un-loading 4,000 [0.37] 4. Ohers (4) 6,500 [0.79] (1) The disance of ransoring by boa is around 70 km, and is mainly used by local assem- Noe: blers 8.6 6,000 [0.42] 8.6 5,000 [0.39] ,000 [0.82] Toal 46, , (2) The disance of ransoring by ruck is around 120 km, and is mainly used by disance assemblers. Boh disances of ransoring by boa and ruck are based on he resul of inerviewing addy and rice assemblers. (3) Transoraion cos iem includes fuel, labour cos, and dereciaion. (4) Oher coss included losses in quaniy and qualiy of rice; accommodaion (foods, drinks, ); arking fees; era aymens for loading and unloading. (5) The Weighed average mehod was used o obain he average markeing cos in his able. The survey resuls allowed us o calculae: - Differen yes of markeing coss of he laes ransacion/shimen incurred by each rice assembler ( i ). - The amoun of rice sold in he laes ransacion/shimen for each rice assembler was used as a weighing coefficien (q i ). q i i - verage markeing cos is comued: M i =. qi (6) The figures in [ ] are coefficiens of variaion. Source: Own survey, 18 rice assemblers in he Mekong River Dela,

8 Chaer 6: Rice Marke Performance nalysis Table 6.4 verage markeing cos of rice wholesalers in he Mekong River Dela (25 ercen broken rice) Markeing cos iems Markeing cos Percen of oal markeing (VND/Ton) (4) coss (%) 1. Labour coss (1) 28, [0.425] 2. Transoraion (2) 20, [0.687] 3. Waer, fuel, elecriciy 27, [0.332] 4. Taes, licenses 18, [0.285] 5. Ohers (3) 20, [0.758] Toal 115, Noe: (1) Labour coss: aymen for ermanen, emorary workers and family labours. This cos was comued based on he average man-days ha boh hired and family workers conribued in 1monh and wage of 1 man-day. (2) This is average ransoraion cos by small ruck/van ha rice wholesalers incur when ransferring heir roducs o rice reailers (a he reail marke, ransor by boa is no convenien). (3) Oher coss included he dereciaion on fied asses, ineres on borrowed caial, losses in quaniy and qualiy of rice. (4) The Weighed average mehod was used o obain he average markeing cos in his able: - Differen yes of markeing coss of he las-monh business incurred by each rice wholesaler ( i ). - The amoun of rice sold in he las-monh business for each rice wholesalers was used as a weighing coefficien (q i ). q i i - verage markeing cos is comued: M i =. qi (5) The figures in [ ] are coefficiens of variaion. Source: Own survey, 49 wholesalers in he Mekong River Dela, Comared o he rice wholesalers, rice reailers incur less markeing coss (63,500 VND). Rice reailers and wholesalers have almos he same cos srucure (See Table 6.5). The mos imoran cos iem is labour (46.5 ercen). Of which, mos labour cos is in erms of sending ime for soring and selling he roduc. Labour coss included aymen for hired workers and imued coss in erms of remuneraion for he labour ime of he rice wholesalers and reailers. round 16 ercen o 22 ercen of oal coss are for elecriciy and oher coss like losses, dereciaion, and aymen of borrowed caial. The res, which is more han 14 ercen is for he license and aes. Wholesalers use 4.2 ercen of heir urnover (sales) o finance markeing coss. For reailers his ercenage is around 2.3 ercen Toal revenues of one on of rice in he domesic marke is 2,750,000 VND on average. The markeing coss as a ercenage of urnover is comued from he oal markeing coss divided by urnover of one on of rice. 143

9 The Organizaion of he Liberalized Rice Marke in Vienam Table 6.5 verage markeing cos of rice reailers in he Mekong River Dela (25 ercen broken rice) Markeing cos iems Markeing cos Percen of oal markeing coss (%) 46.5 (VND/Ton) (4) 1. Labor coss (1) 29,500 [0.328] (5) 2. Fuel, waer and elecriciy (2) 14, [0.457] 3. Taes, licenses 8,900 [0.425] 4. Ohers (3) 10,600 [0.773] Toal 63, Noe: (1) Labour coss: aymen for ermanen, emorary workers and family labours. This cos was comued based on he average man-days ha boh hired and family workers conribued in 1monh and wage of 1 man-day. (2) Fuel was used for he moorbike of he reail sho o deliver rice o final consumers. (3) Oher coss included he dereciaion on fied asses, ineres on borrowed caial, losses in quaniy and qualiy of rice. (4) The Weighed average mehod was used o obain he average markeing cos in his able: - Differen yes of markeing coss of he las-monh business incurred by each rice reailer ( i ). - The amoun of rice sold in he las-monh business for each rice reailer was used as a weighing coefficien (q i ). q i i - verage markeing cos is comued: M i = qi (5) The figures in [ ] are coefficiens of variaion. Source: Own survey, 30 rice reailers in he Mekong River Dela, The figures in Table 6.6 resen he markeing coss of rice millers and olishers. On average, milling one on of addy (usually 25 ercen broken rice) coss abou 53,000VND. In he case of olishing from brown rice o olished rice, oal coss are 68,000 VND/on. The major cos iems are fuel and elecriciy used for oeraing he machines; aymen for emorary and ermanen labourers; maerial coss (mosly in erms of bags for acking finished roducs); dereciaion and mainenance of milling/olishing machines and equimen; aes and license fees. Table 6.6 also indicaes ha in general, he rocessing cos srucure for boh rice millers and olishers are similar. Fuel/elecriciy and aymen for workers are he major cos iems (ranging from 25 ercen o 38 ercen). However, rice olishers have higher labour and maerial coss comared o rice millers. The reason is ha olished rice, afer olishing, has o be acked in a sandard eor ackage. I requires, herefore, more maerial and labour for rocess- 144

10 Chaer 6: Rice Marke Performance nalysis ing. Moreover, he rocess of rice olishing is more sric as comared o rice milling, 85 which means quie a high aymen for ermanen workers. Table 6.6 verage rocessing cos of rice millers in he Mekong River Dela (25 ercen broken rice) Cos iems Cos (VND/Ton) Milling Percen of oal coss (%) Cos (VND/Ton) Polishing Percen of oal coss (%) 1. Labor coss (1) 13, , Elecriciy, fuel 20, , Maerials 3, , Dereciaion of machinery 5, , Preare, mainenance 2, , Taes, licenses 3, , Ohers (2) 4, , Toal 53, , Noe: Toal average rocessing cos is direcly comued from he rimary daa of his survey. However, he secificaion of cos iems is based on informaion rovided by a sub-samle of 14 millers/olishers. (1) Labour coss: aymen for ermanen, emorary workers and family labours (imued cos). (2) Oher coss included he ineres on borrowed caial, managerial coss, elehone, era aymens as bonus for members of rice mills. Source: Own survey from 53 rice millers/olishers in he Mekong River Dela, In our survey of rice millers we found ha some rice millers wih old milling machinery face high rocessing coss. The conversion raio from addy o rice afer milling is lower for old machinery as comared o modern mills (See Secion in Chaer 5). Moreover, old rice mills also bear high fuel coss, elecriciy coss as well as coss of mainenance of milling machinery and equimen. Finally old milling machinery affecs he qualiy of rice afer milling, in aricular he degree of broken rice and he degree of whieness. ccording o he Minisry of griculure and Rural Develomen of Vienam, every year he Mekong River Dela can roduce around million ons of addy. 86 Taking ino accoun demand for home consumion and seeds for he ne cro, he suly of addy o he marke is around 14 million ons er year. 85 In erms of monioring he olishing rocess and adjus he olishing machine in order o obain he eac degree of whieness and broken level of finished roducs. 86 Vienam Economy News, Vol.16, May

11 The Organizaion of he Liberalized Rice Marke in Vienam In he Mekong River Dela we have in oal 7,454 rice mills which can rovide a milling caaciy of over 20 million ons er year. This siuaion leads o surlus caaciy for rice rocessing (he surlus milling caaciy is around 5-6 million ons er year). 87 This imlies ha some milling machines are no used efficienly (esecially in he case of small-scale and old-rice mills). In sum, he above discussion indicaes ha in order o reduce markeing coss a he miller level, we have o look for reasonable soluions o solve he roblem of high rocessing coss of local/old rice mills Markeing margin analysis Table 6.7 gives an overview of he disribuion of he markeing margin among differen acors in he channel. Firs of all we recognize ha rice millers ge he highes gross markeing margin (Value added): 1,008,000 VND/on, while rice assemblers ge he lowes (88,000 VND/on). However, i can be observed ha alhough he rice millers ge he highes markeing margin, hey also incur he highes markeing cos (951,000 VND/on) in erms of aymen for drying, soring, manufacuring, and oher service coss. mong hese rice raders, assemblers and reailers have he lowes markeing coss, 46,500 VND/on and 63,500 VND/on, resecively. The las column of Table 6.7 also indicaes ha among differen rice raders, rice wholesalers obain a relaively large rofi as a ercenage of he cos rice (3.19 ercen). The lowes one is obained by rice reailers (2 ercen). In order o analyze he oal ne revenues of differen rice raders, we calculae he oal rofis er monh based on heir monhly urnover. 88 We conclude ha he resuls are all in he same range. Profi raes are somewha higher han he ineres raes alied by sae and/or non-sae commercial banks (varying from 0.75 ercen o 1.2 ercen er monh). The difference is inerreed as an acceable remium for risk aking by raders. This is anoher indicaor ha he marke is oeraing quie efficienly. 87 This figure is an adjused caaciy aking ino accoun he seasonal use of milling machines (75 ercen of acual caaciy is aken as a maimum) griculure\rice economy. 88 The average urnover of differen rice raders ha were obained from he survey: - ssemblers: 40 ons/monh - Millers: 250 ons/monh - Wholesalers: 60 ons/monh - Reailers: 10 ons/monh 146

12 Table 6.7 Chaer 6: Rice Marke Performance nalysis Markeing coss and gross margins of Farmers, ssemblers, Millers, Wholesalers and Reailers in he Mekong River Dela, Cos Iems I. Farmers 1. Producion coss er on (1) 2. verage selling rice of addy er on (2) II. ssemblers 1. verage buying rice of addy (2) 2. Toal markeing coss (3) 3. Cos rice (3 = 1 2) 4. verage selling rice (2) III. Millers 1. verage buying rice (2) 2. Toal coss Drying (7) Sorage (8) Processing (milling) (4) Losses by conversion raio(32%) (*) 3. Cos rice (3 = 1 2) 4. verage selling rice (2) IV. Wholesalers 1. verage buying rice (2) 2. Toal markeing cos (5) 3. Cos rice (3 = 1 2) 4. verage selling rice (2) V. Reailers 1. verage buying rice (2) 2. Toal markeing cos (6) 3. Cos rice (3 = 1 2) 4. verage selling rice (2) Coss and rices VND/Ton 1,028,000 1,454,000 1,454,000 46,500 1,500,500 1,542,000 1,542, ,000 50,000 30,000 53, ,000 2,493,000 2,550,000 2,550, ,000 2,665,000 2,750,000 2,750,000 63,500 2,813,500 2,870,000 Gross markeing margin(*) (1) Toal mar. cos (2) Profi margins (3) = (1) (2) moun s % of cos rice 88,000 6,500 41, ,008, ,000 57, , ,000 85, ,000 63,500 56, Source: (1) Paddy roducion cos: Survey by ngiang Food Comany. See also endi 6.1a (2) verage buying/selling rice a differen levels were based on he own survey of his sudy, See endi 6.3, 6.4, 6.5, 6.6, and 6.7 (3), (4), (5), and (6) based on he figures from Table 6.3, 6.4, 6.5, and 6.6. (7) See endi 6.1b (8) Sorage cos were colleced from SEB sudens (graduaed hesis ). Noe: (*) Losses 32% (based on he conversion raio 68%) (2,550,000 32% = 818,000) Gross markeing margin (Value added) = verage selling rice verage buying rice 147

13 The Organizaion of he Liberalized Rice Marke in Vienam The figures in Table 6.8 indicae ha rice reailers are in a less rofiable siuaion (lowes rofi er monh: 565,000 VND). However, his rofi level is sill higher han he minimum wage in Vienam (around VND er monh in 2002). In conras, rice millers obain he highes rofi er monh (14,250,000 VND). However, we recall ha in comarison wih oher raders, rice millers inves relaively large amouns of caial (See Secion and Table 5.4). Their rofi as a ercenage of own-caial invesmen is reasonable (1.48 ercen). Rice wholesalers obain quie a large amoun of rofi er monh (5,100,000 VND), bu also his looks reasonable if he rofi is eressed as a ercenage of invesed own caial. Finally, he monhly rofi of rice assemblers is a a moderae level (1,660,000 VND). If we comare rofi raes as a ercenage of gross markeing margins (value added) hen he resul is ha rice assemblers and reailers obain he highes margins (around 47 ercen) and rice millers obain he lowes margins (only 5.7 ercen). We noe ha hese figures are averages. In racice he margins vary. Table 6.8 resens he coefficien of variaion of addy/rice rices among differen raders. These figures reflec he degree of rice volailiy o raders. In general, he rice of addy/rice does no vary much. The highes coefficien of variaion is observed for rice assemblers, and reailers accouning o and resecively. The lowes coefficien of variaion is observed for rice millers (0.058). Those resuls imly ha a he assembling and wholesale markes, rices are more volaile. The las row of Table 6.8 resens he oal margins for he whole channel. Paddy was bough on average, a 1,454,000 VND/on and hrough differen markeing agens he reail rice of whie rice increased o 2,870,000 VND/on. I means ha he gross markeing margin caured by he markeing sysem (1,416,000 VND) concerns almos 49 ercen of he reail rice. Toal markeing coss of he whole channel sysem reached 83.1 ercen and he oal rofi margins was 16.9 ercen. The above resuls indicae ha mos rice raders in his sudy obain a relaively large rofi as a ercenage of value added. However, his rofi is reasonable if i is eressed as a ercenage of invesed caial. This indicaes ha he rice marke is oeraing quie efficienly. 148

14 Table 6.8 Traders 1.ssemblers Chaer 6: Rice Marke Performance nalysis Toal Profi margins of differen rice raders in he Mekong River Dela, (Thousand VND/on) verage buying rice 1,454 [0.148 ] 2. Millers 1,542 [0.058 ] 3. Wholesalers 2,550 [0.068] 4.Reailer 2,750 [0.105] Whole sysem verage selling rice 1,542 [0.058 ] 2,550 [0.068] 2,750 [0.105] 2,870 [0.088] Markeing. margin (Value added) Markeing. cos Profi margin Toal rofi er mon h Mar. cos as % of value added (%) Profi as % of value added (%) Profi as % of own caial (*) (%) , , , , ,416 1, Noe: The figures in arenhesis are coefficiens of variaion ha were comued based on endi 6.3 endi 6.7 (*) The figures in he las column were calculaed from Table 5.4 in Chaer 5 and he average urnover of differen rice raders. Source: Own survey among rice raders in he Mekong River Dela, Vienam, n analysis of addy and rice rice aerns The liberalizaion of he agriculural marke sill is a holy debaed issue in Vienam. The imorance of rice incenives for rice roducion is well known from boh macroeconomic and household analysis. Prices are no only imoran for farmers. They are he main incenives for all he oher aricians in he rice markeing sysem including raders, millers, SOEs, and consumers. The sudy of rice behavior is herefore criical o he undersanding of he erformance of he marke sysem. In his secion, we focus on emoral rice analysis and in aricular on addy and rice rice evoluion (including rends and seasonaliy) afer he reform in Vienam. The daa used in his analysis are monhly rices, rovided by he Informaion Cener of he Minisry of griculure and Rural Develomen in Hanoi. They cover he eriod from January 1989 o December Daa on he Consumer Price Inde (CPI) from 1989 o 2000 are colleced from he General Saisical Office of Vienam. 149

15 The Organizaion of he Liberalized Rice Marke in Vienam Trends in addy and rice rices Growh of addy and rice rices during he eriod is sudied considering wo sub-eriods: he eriod , and he eriod The wo sub-eriods are easily associaed wih (1) afer he olicy shif from cenral lanning o he marke sysem, and (2) he consolidaion of ha olicy afer he imressive success achieved in roducion and eor. The rends in addy and rice rices during he eriod in boh nominal and real erms are shown in Table 6.9 as well as Figure 6.1 and Figure 6.2. While he average annual increase in rice rices in he firs hree years of his eriod ( ) was 58 ercen, he annual increase fell o 6.5 ercen for he remaining eriod ( ). This resul can be elained by he fall in inflaion since ddiionally, he las wo columns of Table 6.9 also indicae ha in 1995 and 1998 he growh rae of addy and rice rices increased suddenly because during hose years here was a shock in demand of rice in he inernaional marke. Table 6.9 Paddy and rice rices in Vienam, Year Farm gae addy rice Reail rice rice Consumer rice inde Deflaed addy rice Deflaed reail rice rice Growh rae of addy rice Growh rae of reail rice rice (VND/kilogram) (1989=1.0) (1989 VND/kilogram) (Percen) , ,404 1, ,427 2, ,445 2, ,523 2, ,939 2, ,822 3, ,774 2, ,198 3, ,872 3, ,708 3, nnual growh rae (%) Source: (1) nnual rices of addy and rice from Virane (colleced by Dearmen of Trade of Canho, 2000) (2) Consumer rice inde (CPI) is comued based on endi Unil 1990, inflaion was very high: beween 1986 and 1987 i reached hyer-inflaion levels wih a eak of 650 ercen in 1987 and hen decreased bu sill was a a raher susained level of over 50 ercen in fer 1991, he macroeconomic olicy was very effecive in bringing he level of inflaion below 30 ercen. In fac, during 1992 and 1993, here was a deflaionary siuaion in Vienam. In he laer years, inflaionary ressures have sared o rise again, bu overall i is a levels around 20 ercen (Golei, 1996). 150

16 Chaer 6: Rice Marke Performance nalysis In racice, he incenives for addy roducion and consumion deend, however, no on nominal rices, bu on he rices of addy and rice relaive o oher goods. I is aaren ha since 1989 boh addy and rice rices have risen more slowly han he inde of consumer rices. While addy rices have increased by an average of ercen in he 12 years eriod , consumer rices have increased by ercen, imlying a decline in he real farm addy rice of 2.97 ercen er year. Similar calculaions show a decline in he real reail rice of rice of 1.92 ercen (Table 6.9). In oher words, he relaive rices of addy and rice have decreased over he eriod. This imlies a benefi for rice consumers and a disincenive for rice farmers. The basic idea is ha he change in farm income deends on he ah of rices, roducion, and cos of roducion. s menioned, real addy rices have declined 2.97 ercen er year bu his decline has been comensaed by an increase in rice yield er hecare Price (VND/kilogram) Reail rice rices Paddy rices Year Figure 6.1 Trends in addy and rice rices in Vienam, Source: Virane (colleced by Dearmen of Trade of Canho, 2000) 90 Rice yield er hecare has grown more han 3 ercen over he same eriod (See Table 3.7 in Chaer 3) 151

17 The Organizaion of he Liberalized Rice Marke in Vienam 1000 Price (VND/kilogram) Reail rice Paddy rice Year Figure 6.2. Real rices of addy and rice in Vienam, The variabiliy of addy and rice rices The variabiliy in addy and rice rices is measured by he yearly coefficien of variaion of monhly rices (Figure 6.3 and Table 6.10). In he firs eriod, from 1989 o 1991, he average rice coefficien of variaion declined from a value of 0.14 o a value of 0.05 for rice and from 0.16 o 0.04 for addy. The macroeconomic sabiliy (lower inflaion) ha reurned in he early 1990 s in Vienam reduced he volailiy of boh addy and rice rices. During he laer ar of he eriod he coefficien of variaion of addy and rice rices seems o be sable (See Figure 6.3). I is jus a very moderae levels of 0.02 o 0.04 for addy and 0.03 o 0.05 for rice (See Table 6.10). This means ha he variabiliy of addy and rice rices is very much relaed o general macroeconomic insabiliy. From 1994 o 1997, he variabiliy of addy and rice rices increased again. The coefficien of variaion of addy rices varies from 0.02 o 0.13 and from 0.03 o 0.12 for rice. Those resuls can be elained by he flucuaion of eor rices in he inernaional marke. 152

18 Chaer 6: Rice Marke Performance nalysis Table 6.10 Coefficien of variaion of monhly farm gae addy and reail rices by year in Vienam, Year Paddy rice Reail rice Noe: The figures in his Table concern nominal rices. Source: Informaion Cener of he Minisry of griculure and Rural Develomen in Hanoi (based on he informaion from Sae Price Commiee) Paddy rice CV Reail rice Year Figure 6.3. Coefficien of variaion of addy & rice rices in Vienam,

19 The Organizaion of he Liberalized Rice Marke in Vienam The las eriod from 1998 o 2000, he volailiy of addy and rice rices has become more sable again. s shown and illusraed in Table 6.10 as well as Figure 6.3, he coefficien of variaion of addy rices is abou 0.10 o 0.11 and in he case of rice rices, i varies from 0.07 o However, addy rices are sill more volaile han reail rices even hough he governmen has imosed a new olicy and inervened in order o hel rice farmers o sell heir addy a a floor rice. 91 Demand for addy, used for rocessing for he eor marke, highly flucuaes due o variabiliy of eor rices, while he domesic consumer rice demand is relaively sable Seasonal rice flucuaions In he revious secion, long-erm movemens of addy and rice rices were sudied. noher source of variabiliy in addy/rice rices is seasonaliy. In his secion, he seasonal rice movemen is looked a. For mos of he analysis in his secion, monhly daa from he eriod are used. 92 Firs, we look a he seasonal inde of addy and rice rices. 93 Then, we discuss he sabiliy of he seasonal movemens for hese roducs. The seasonal inde of addy and rice rices was comued by using he 12 monh-moving average mehod. Deails of he comuaion of he average seasonal inde, as well as he adjused seasonal inde, are resened in endi 6.8, endi 6.9, endi 6.10, and endi Finally, he resuls of he seasonal rice movemens are illusraed in Figure 6.4. s shown in his figure, a he aggregae level, seasonal flucuaions seem o be similar for addy and rice. Over he eriod he range beween seasonal eak and rough is abou 12 ercen for addy. I is abou 9 ercen for rice (Figure 6.4). Seasonaliy of rices is usually a reflecion of seasonaliy of roducion, wih rices showing a endency o decline during he harves season and o increase during he off season. s menioned in Chaer 3 (Figure 3.1), mos rice farmers in he Mekong River Dela culivae 3 rice cros er year: Summer- uumn, uumn-winer, and Winer-Sring. Of which, Summer-uumn and Winer-Sring are considered as major cros. s shown in Figure 6.4, boh addy and rice rices reach he highes level during March-ril (he off season of he year afer harvesing he Winer-Sring rice cro ha usually roduces he highes qualiy of addy/rice) and fall down o heir lowes level from ugus o Seember (he main harves season of he Summer-uumn rice cro wih low qualiy of addy/rice because of he rainy season). 91 See also Secion in Chaer 4 for furher elanaion. 92 By he limiaion on daa source, unforunaely he monhly daa of addy and rice rices are no available for he eriod before Then, his analysis jus based on he monhly daa of addy and rice rices from 1996 o See endi 6.9 for furher elanaion abou he seasonal inde of addy and rice rices. 154

20 Chaer 6: Rice Marke Performance nalysis Percenage Rice Paddy Jan Feb March ril May June July ugus Se Oc Nov Dec Figure 6.4 djused seasonal inde of addy and rice rices in Vienam (Based on he monhly nominal rices from he rend) In general, seasonal movemens of addy/rice rices reflec sorage coss in a comeiive rice marke. fundamenal decision facing mos markeing agens is when heir roducs should be sold. s menioned in he revious secion, he rice ga of addy/rice beween he lean and harves eriod is no so high. I jus varies by 10 ercen. The duraion from he eak o he lean season is only hree o four monhs. Therefore, sorage of addy/rice afer he harves (eak season) and resale a he lean season is no very rofiable. Sorage coss and risks/losses are nearly equal or even greaer han he available rice margin Marke inegraion analysis Saial rice behaviour in regional rice markes is an imoran indicaor of overall marke erformance. Markes ha are no inegraed may convey inaccurae rice informaion disoring he markeing decisions of rice roducers and conribuing o inefficien roduc movemens. Therefore, an imoran ar of marke erformance analysis focuses on regional rice analysis and rice marke inegraion beween differen marke laces. Several mehods for measuring 94 s shown in Chaer 5, sorage cos of 1 on of addy is 160,000 VND (Table 5.19), however he rice margins of 1 kg of addy from he eak o he lean season only varies from VND (Reor from he Minisry of griculure and Rural Develomen of Vienam, Oc 2001). 155

21 The Organizaion of he Liberalized Rice Marke in Vienam rice inegraion have been used beginning wih simle bivariae correlaion coefficiens. This is he simles way o measure he saial rice relaionshis beween wo markes. However, his mehod clearly has some limiaions. 95 One mehod for measuring he degree of rice inegraion akes he abovemenioned criique ino accoun: he co-inegraion rocedure. This economeric echnique rovides more informaion han he correlaion rocedure, as i allows for he idenificaion of boh he inegraion rocess and is direcion beween wo markes. Regional rices move over ime because of various shocks. If in he long run hey ehibi a consan linear relaion, hen we say ha hey are coinegraed. In general, he resence of co-inegraion beween wo series is indicaive of iner-deendence. In oher words, co-inegraion indicaes nonsegmenaion beween he wo series. Co-inegraion analysis is a useful ool o give an answer abou he eisence of a relaion beween wo economic imeseries. To es for co-inegraion, one can aly he Engle Granger as well as he Johansen rocedure. lhough he Engle-Granger rocedure is easily imlemened, i does have several imoran defecs. One of he serious defecs of he Engle-Granger rocedure is ha he mehod has no sysemaic rocedure for he searae esimaion of he mulile co-inegraing vecors. noher defec of he Engle-Granger rocedure is ha i relies on a wo se esimaor (Enders, 1998). Forunaely, several mehods have been develoed ha avoid his roblem. The Johansen (1998) and Sock and Wason (1998) maimum likelihood esimaors circumven he use of wo-se esimaors and can esimae and es for he resence of mulile co-inegraing vecors. Moreover, hese ess allow he researcher o es resriced versions of he co-inegraing vecors and he seed of adjusmen arameers. This sudy mainly alies he Johansen rocedure o es for long run and shor run co-inegraion. To use he co-inegraion rocedure, several ses need o be carried ou on he rice series: firs, he ugmened Dickey Fuller es (DF) and hen he Johansen rocedure The daa and model secificaion When saial rice inegraion is sudied, he analysis should referably be based on day-rices or a leas average week-rices. I is clear ha average monhrices may be very rough esimaes of day-rices. Esecially during eriods of changing marke relaionshis, 96 hese monhly averages may deviae subsanially from marke day-rices. Moreover, average monh-rices are no he righ variables o indicae he shor-run rice inegraion rocess. In Vienam, dayrices are available for only a few marke laces and only for a shor eriod in 95 See also Secion 2.4 in Chaer For eamle a he end of he off season and he beginning of he harves season 156

22 Chaer 6: Rice Marke Performance nalysis 157 he as. Therefore, we use weekly rice series, from 1998 o ugus Esecially, he rice series of 25 ercen broken rice are used for he analysis, since his kind of rice is oular o consumers in he domesic rice markes. Nine major marke laces are chosen o analyze marke inegraion. Of hese, four markes are locaed in he Mekong River Dela (ngiang, Canho, Socrang, and Tiengiang), wo markes are in he middle and cenral high-land area (Lamdong, and Danang), and wo are Ciy markes (HCM Ciy and Hanoi). Finally, he major eor marke a HCM Ciy is also included in his analysis. The weekly rice daa were rovided by he Canho Trade Dearmen, Informaion Cener of Minisry of griculure and Rural Develomen, and he Minisry of Trade of Vienam. Mahemaically, he saring oin for a model for esing marke inegraion based on mulile co-inegraing vecors is he following: ε = ) 1 ( K (6.1) Where: = 1, 2,.. refer o he weeks from 1998 o ugus 2001 n = number of markes included in he analysis = ineger, he volume of which is o be deermined = he (n 1) vecor of variables ( 1, 2,. n ) : [rice of n marke laces] i = an (n n) mari of coefficiens ε = an (n 1) vecor of error erms Wih: 1 = Equaion (6.1) can be u in a more suiable form by subracing 1 from he equaion: ε = ) ( K and so on: ε = ) ( ) ( K.. ε = K K K K ) ( ) ( ) ( and subracing 1 for boh sides I ε = ) ( ) ( K K K

23 The Organizaion of he Liberalized Rice Marke in Vienam Finally we obain: = 1 i= 1 π ε (6.2) i i π 1 Where: I = an (n n) ideniy mari and π and π i is defined by π = I i i= 1 π = i j j= i 1 Here he rank of π or ( 1 I) is equal o he number of indeenden coinegraing vecors. If ( 1 I) consiss of all zeros, so ha rank(π) = 0, clearly, he mari is null (all elemens of π equal zero) and equaion (6.2) is a radiional vecor auo-regression (VR) in firs differences. This imlies ha all he { i } sequences are uni roo rocesses. Since here is no linear combinaion of he { i } rocesses ha is saionary, he variables are no co-inegraed. In such circumsances, here is no error-correcion reresenaion since { i } does no resond o he revious eriod s deviaion from a long-run equilibrium, since ha equilibrium does no eis. If insead rank (π) = n, he vecor rocess is saionary. In inermediae cases, if rank (π) = 1, here is a single co-inegraing vecor and he eression π -1 in equaion (6.2) is he error-correcion facor. For oher cases in which 1 < rank (π) < n, here are mulile co-inegraing vecors. If rank (π) > 0, a vecor error correcion model (VECM) will be used o analyze he shor-run co-inegraion. VECM is a resriced VR designed for use wih non-saionary series ha are known o be co-inegraed. The vecor error correcion has co-inegraion relaions buil ino he secificaion so ha i resrics he long-run behaviour of he endogenous variables o heir coinegraing relaionshis while allowing for shor-run adjusmen dynamics. The co-inegraion erm is known as he error correcion erm since he deviaion from long-run equilibrium is correced gradually hrough a series of arial shor-run adjusmens. One of he mos ineresing asecs of he Johansen rocedure is ha i allows for esing resriced forms of he co-inegraing vecors. In order o es resricions on he co-inegraing vecor, Johansen defines he wo marices α and β, boh of dimension (n r), where r is he rank of π. The roeries of α and β are such ha: π = α β 158

24 Chaer 6: Rice Marke Performance nalysis The mari β is he mari of co-inegraing arameers, and he mari α is he mari of weighs wih which each co-inegraing vecor eners he n equaions of he VR. In a sense, α can be viewed as he mari of he seed of adjusmen arameers. Due o he cross-equaion resricions and he arameers are considered non-linearly in he model, i is no ossible o esimae α and β using OLS. 97 However, wih maimum likelihood esimaion, i is ossible o (i) esimae (6.1) as an error-correcion model; (ii) deermine he rank of π; (iii) use he r mos significan co-inegraing vecors o form β ; and (iv) selec α such ha π = α β I is easy o undersand he rocess in he case of a single co-inegraing vecor. Given ha rank (π) = 1, he row of π are all linear mulilies of each oher. Hence, he equaions in (6.2) have he form: 1 = L π π122 1 L π1 nn 1 ε1 2 = L s2 ( π π 1221 L π 1nn1) ε 2 n = L s π ε n( π L π1 n n 1) n Where he s i are scalars such ha s i π 1j = π ij. nd for noaional simliciy, he marices π i -i have no been wrien ou. Now define α i = (s i π 1j ), so ha each equaion can be wrien as = L α ε or in mari form, i ( 1 1 β2 2 1 L βn n 1) i = i i αβ 1 π ' ε (6.3) Where he single co-inegraing vecor is β = (1, β 2,β 3,.,β n ) and he seed of adjusmen arameers are given by α = (α 1,α 2,.,α n ). Once α and β are deermined, esing various resricions on α and β is sraigh-forward based on he fundamenal oin ha if here are r co-inegraing vecors, only hese r linear combinaions of he variables are saionary. Thus, he es saisics involve comaring he number of co-inegraing vecors under he null and alernaive hyoheses. 97 The Johansen rocedure consiss of he mari of vecors of he squared canonical correlaions beween he residuals of and -1 regressed on lagged values of. The coinegraing vecors are he rows of he normalized eigenvecors (Waler Enders, 1998). 159

25 The Organizaion of he Liberalized Rice Marke in Vienam lying he Johansen mehodology o es for co-inegraion: four ses rocedure Se 1: Pre-ess and lag lengh Firs of all, i is good racice o re-es all variables o assess heir order of inegraion. The order of inegraion will be esed wih an ugmened Dickey Fuller (DF) es on a uni roo. Secondly, we coninue o es for he lag lengh, because he resuls of he co-inegraion es can be quie sensiive o he lag lengh. The mos common rocedure is o esimae a vecor auo-regression (VR). The lags are secified as lags of he firs differeniaed erms. Begin wih he longes lag lengh deemed reasonable and es wheher he lag lengh can be shorened. We can use he kaike informaion crieria (CI) and Schwarz crieria (SC) o selec a suiable lag lengh. Using he DF es, he resuls resened in Table 6.11 indicae ha he rice series for he nine markes under sudy are I(1). For all he rice series he uni roo es on rice levels (hree columns in middle of he able) shows ha he coefficiens are no significanly differen from zero and his imlies ha none of he rice series is saionary. Furhermore, he uni roo es on firs differences (hree las columns in he able) confirms he oosie, which leads us o conclude ha all our rice series are inegraed of order 1 and ha he series of he firs differences are inegraed of order 0. This resul imlies ha inclusion of firs differences as variables in he model, insead of normal rice series, will eliminae he sochasic rend o which he nominal series are eosed. Table 6.11 Uni roo es on rice rice series in differen marke laces No. of Marke lace Uni roo-es on rice levels Uni roo-es on firs differences observaions DF (1) δ -value (2) DF (1) δ -value (2) 1. ngiang 188 DF (2) DF (1) Canho 187 DF (3) DF (2) Danang 188 DF (2) DF (1) HCM Ciy 188 DF (2) DF (1) Hanoi 188 DF (2) DF (1) Lamdong 185 DF (5) DF (4) Socrang 188 DF (2) DF (1) Tiengiang 188 DF (2) DF (1) VN Eor 186 DF (1) DF (0) Noe: (1) In he column DF he number of lags ha was allowed for in he uni roo es is indicaed in brackes. No serial correlaion was deeced (5% significance level). (2) Criical value are given in Maddala 1992: = , 5% level of significance. Source: Weekly addy and rice rice series from 1998 o DF analysis was carried ou in EVIEWS

26 Chaer 6: Rice Marke Performance nalysis Wih resec o he lag lengh, he resul of he VR analysis on firs differences of he rice series shows ha he smalles value of IC and SC is obained wih one lag lengh (See endi 6.12). The resul of he VR analysis is resened in Table 6.12 below. In general, Table 6.12 shows ha he resen rice changes in all marke laces are highly correlaed wih heir own rice change in he las eriod. The resuls in he firs column indicaed ha he rice changes in mos major rice markes in he Mekong River Dela and HCM Ciy are srongly relaed o he eor rice changes. I is clear ha rice raders in Vienam have o follow he rice se in he inernaional marke. The rice series in he domesic marke show ha mos markes wihin he Mekong River Dela and HCM Ciy srongly cohere (HCM, CT, TG). The markes locaed in he cener and in he norh have a weak relaionshi wih he oher markes. This resul may be elained by he ransoraion coss involved. There is a grou of four markes: Canho, Tiengiang, HCM Ciy, and Lamdong ha show a very srong correlaion. Here, a major grou of rice raders is locaed ha handle large amouns of rice in he domesic markes. Table 6.12 VR analysis on rice rice series in differen marke laces VN(-1) HCM(-1) CT(-1) LD(-1) TG(-1) HN(-1) ST(-1) G(-1) DN(-1) VN [-4.327] HCM [2.820] [-2.295] [2.744] CT [1.870] [-2.257] [-3.846] [1.873] LD [2.012] [2.074] [-2.481] [1.792] TG [1.820] [1.813] [2.154] [-3.644] HN [2.067] [-3.068] ST [1.761] [-1.763] G [1.805] [-3.698] DN [2.070] [-2.406] Noe: The resuls of he VR analysis are based on one lag ( he IC and SC are smalles). ll figures in arenhesis [ ] are -values. Source: See Table