Inflation and Relative Price Variability in India- A Decomposition Analysis

Transcription

1 Workng Paper 3/011 nflaon and elave Prce Varably n nda- A Decomposon Analyss Saraj asool S aja Sehu Dura M amachandran DEPATMENT OF ECONOMCS PONDCHEY UNVESTY May 011 sysem8 [Type he company name] 1/11/011

2 Workng Paper No. 3 (011) Deparmen of Economcs nflaon and elave Prce Varably n nda- A Decomposon Analyss Saraj asool 1 esearch Scholar, Deparmen of Economcs, Pondcherry Unversy, Puducherry S aja Sehu Dura Asssan Professor, Deparmen of Economcs, Pondcherry Unversy, Puducherry M amachandran Professor, Deparmen of Economcs, Pondcherry Unversy, Puducherry Absrac Ths sudy ams a decomposng relave prce varably due o real facors, nflaon and her neracon, usng he mehodology proposed by Clemens-Nguyen (1981, 198) for nda. The emprcal evdence from hghly dsaggregaed commody-wse monhly daa for he perod from Aprl-1994 o November-009 suggess ha he real facors such as echncal change, real ncome growh ec. accoun for 55 percen of oal relave prce varably and he res s arbued o nflaon. Also dsrbuon of cross seconal varance shows ha almos 60 percen of s conrbued by 30 commodes ndcang ha varably n he relave prces s drven by only few commodes. The more crucal nference ha emerges from he emprcal analyss s ha he hreshold nflaon rae ha mnmzes he relave prce varably seems o have been 4.5 percen, whch s conssen wh he offcal hreshold rae mananed by he eserve Bank of nda. Keywords: Key words: nflaon, elave prce varably, Decomposon analyss, Mnmumvarance nflaon rae JEL Classfcaon: E31; E3 1 Correspondng auhor emal: sarajrasool@gmal.com 1

3 1. nroducon Prce sably s he predomnan objecve of moneary polcy of cenral banks all over he world. Moneary polcy wheher expressed n erms of neres raes or growh of moneary aggregaes has been ncreasngly geared oward he achevemen of low and sable nflaon. Prce sably means ha keepng he nflaon rae a a level ha s no hnderng he households and busnesses n every day decson makng. Hgh nflaon can hamper growh because of number of facors: lowers economc effcency, redsrbuon of wealh, averson o long erm conracs and excessve resources are devoed o hedgng nflaon rsks (Barrow, 1996; Fsher, 1983). n developng economes, an addonal cos of hgh nflaon emanaes from s adverse effecs on he low and fxed ncome group. n general, he dea s ha busnesses and households are hough o perform poorly when nflaon s hgh and unpredcable. has been found ha nflaon proceeds n a very ragged manner wh some prces rsng by more han average an ohers by less (Parks, 1978). Ths fndng s exremely mporan because hese relave prce dsorons may cause cosly resource reallocaons, whch would have o be ncluded n any evaluaon of he cos of nflaon. Bascally, hgher nflaon effecs real secor of he economy hrough s effec on nflaon uncerany and relave prce varably (PV). Ths follows from he fac ha hgher nflaon uncerany resuls n shorer lengh of conracs han he opmal lengh or nduces rsk prema for long-erm arrangemens, rases coss for hedgng agans nflaon rsks, leads o unancpaed redsrbuon of wealh and also ndexng becomes more prevalen. All of whch reduce economc effcency relave o a world of more sable prces. More mporan, due o ncreasng relave prce varably agens ably o dsngush beween changes n relave and absolue prces decreases. The role of he marke-prce sysem n coordnang economc acvy and ransmng nformaon become mpeded, hus reducng economc effcency furher. Thus, boh, nflaon uncerany and PV reduce he effcency of marke prces as a coordnaor of economc acvy and affec negavely nvesmen decsons (Caballero, 1991). The resulng prce srucure s dsored from nal cos and preference fundamenals and may nduce resource msallocaon and welfare loss (Fsher, 1981). Oher han nflaon real facors such as demand shf varables, resources, weaher and echnology can also cause changes n relave prces of commodes. However, relave prce adjusmens assocaed wh changes n demand shf varables, resources, weaher and echnology are essenal for he effcen allocaon of resources. The exsence of a posve correlaon beween nflaon and relave prce varably was observed by Mlls (197) and Graham (1930) much earler, receved lle aenon unl he md-1970s when Vnng and Elwerowsk (1976) re-examned hs relaonshp. neresngly, he publcaon of her work concded wh he afermah of he hgh and exremely volale nflaon of he early 1970s, whch many economss blamed on so called supply shock. These supply shocks were ypcally assocaed wh ncreases n he relave prces of such as energy and food. The debaes and dscussons of ha perod brough o he fore he mporance of relave prce changes. One sream of he leraure ha aemps o explan he relaonshp beween nflaon and relave prce varably focuses on he sources of relave prce changes. For example, usng a mul-secoral supply-anddemand model ha fs he mspercepons model framework of Lucas (1973) and Parks

4 (1978) show ha changes n real ncome and unancpaed nflaon explan a large par of relave prce varance. Ths model has been exensvely ulzed for he sudy of he relaon beween nflaon, nflaon uncerany, and relave prce varably (e.g. Cukerman and Wach- el, 1979, Hercowz, 1981). Hercowz (1981) exends he mspercepons models of Lucas (1973) and Barro (1976) o show ha unancpaed money growh s responsble for he rse n relave prce dsperson, whch s nflaonary. Fscher (1981,198) descrbes he asymmerc prce response hypohess, assocaed prevously wh Schulze (1959) and Tobn (197), whereby dfferen prces respond dfferenly o dsurbances, so ha any shock o he prce sysem s spread asymmercally across s componens.. These sudes have focused manly on he underlyng facors of relave prce varably wh arbung o unancpaed nflaon, real ncome changes and unancpaed money growh. s clear from he exsng leraure ha varably n relave prces orgnaes from eher from real facors or from nflaonary facors (moneary) or boh, whch may vared from me o me and across commodes. However he ssue of relave mporance of he real and nflaonary facors s no very clear. s of neres o know he proporon of varably n relave prces arbued o real and o nflaonary facors for analysng he relave mporance. wll be also neresng o analyse he dynamc shares of hese sources (real and nflaon) whch may vary from me o me and also across commodes. n oher words wll be very useful o polcy makers o undersand wheher change n a relave prce of a parcular commody s due o real facors such as change n real ncome, echnology, supply shock (whch s essenal for he effcen allocaon) or nflaonary facors such as changes n nflaon, unancpaed nflaon, money supply (whch s very dsoronary) or boh. n hs backdrop hs sudy ams a analysng he relave shares of real and nflaonary facors n causng varably n relave prces of commodes. Ths analyss wll gve an dea abou he varably n relave prces arbued fully o real facors ha s naural rae of relave prce varably as called by Clemens and Nguyen (1981). Also hs sudy gves an esmae of mnmum varance nflaon rae ha s nflaon rae whch s assocaed wh mnmum varably of relave prces. We have used a mehodology proposed by Clemens and Nguyen(1981) ha suded he relaon beween nflaon and varably of relave prce changes. Usng he hghly dsaggregaed commody wse daa on Whole sale prce ndex (WP) for he sample perod from Aprl-1994 o November-009 he sudy esmaes he proporons of relaveprce varably ha can be arbued o real facors and nflaon. Also, he sudy denfes hose commodes whch conrbue mos o oal relave prce varably and analyses her nflaon and real componen n her respecve shares. Ths s gves an dea abou he commodes whose relave prces are dsored manly by nflaonary facors. Nex he sudy ams a esmang mnmum-varance nflaon rae and naural rae of relave prce varably for nda. Ths sudy dffers from Clemens Nguyen (198) n he manner ha we have used hghly dsaggregaed monhly WP daa across 419 commodes whch enables he decomposon a very mcro level. n addon o ha, he sudy asses he me varyng shares of dfferen commodes n oal relave prce and analyses he decomposon of he each commody-share ndvdually n each monh. Ths provdes mcroscopc vew of me varyng componens of nflaon and real facors n he composon of oal relave prce varably for each 3

5 monh and for each commody. Also, snce he daa are monhly and sample sze s farly large, we have run a Ba-Perron (1998) mulvarae srucural break es whch found hree breakpons and dvded sample no four subsamples. A comparson s made of poson for low nflaonary perod wh ha of hgh nflaonary perod, hs gves an assessmen of he composon and dsrbuon of relave prce varably n varous nflaonary regmes. The fndngs of he sudy sugges ha he 60 percen of he oal relave prce varably s arbued o op 30 commodes and res o all oher commodes. Ths shows ha hgh flucuaons n he prces of only few commodes s responsble for major par of he observed oal varably of he relave prces. Also, he decomposon analyss carred shows ha, n oal relave prce varably 45 percen s nflaon componen and 55 percen s real componen for he perod of sudy. Ths proporon of nflaon and real componen seems o vary wh nflaon rae across he subsamples. The share of nflaonary componen seems o ncrease n hgh nflaonary perod and decreases n low nflaonary suaon as s expeced. The more crucal nference ha emerges from he emprcal analyss s ha he hreshold nflaon rae ha mnmzes he relave prce varably seems o have been 4.5 percen, whch s broadly conssen wh he offcal hreshold rae mananed by he eserve Bank of nda. Also, he naural rae of relave prce varably s calculaed a 0.06 percen for he sudy perod. n he followng secon of he paper we dscuss daa and mehodology and n secon 3 emprcal esmaes are gven, subsequenly secon 4 draws concluson of he sudy.. Daa and he Economerc Mehodology a model ha suded he relaon beween nflaon and varably of relave-prce changes. The model enables he denfcaon of hose commodes whose relave prces are dsored by nflaon (or by real facors). also provdes measures of he naural rae of relave prce varably whch s assocaed wh zero nflaon and he mnmum-varance nflaon rae ha s he nflaon rae whch mnmzes he varably of relave prces. Subsequenly, hey nroduced an neresng mehodology for decomposng relave prce varably no he componens due o he real facors, nflaon and neracon (see, Clemens and Nguyen 198). They appled hs decomposon procedure o Ausralan daa for he perod across 7 commody groups and fnd ha nflaon accouns for 4 percen of relave prce varably durng he perod. To compue he aggregae nflaon rae and he momens of he dsrbuon of relave prce changes, we use he Wholesale Prce ndex (WP) and s sub-ndces a hghly dsaggregaed level. The old measure of WP comprses 435 commodes under hree dfferen caegores. The complee me seres derved from a unform defnon of he sub-ndces of each commodes coverng he enre sample perod were avalable on 419 commodes ou of oal 435 commodes. Monhly daa on hese 419 commodes for he perod from Aprl 1993 o November 009 s used for he analyss. The choce of he sample perod s dcaed by he avalably of conssen and unform me seres on prce ndces of hese commodes over he perod. The daa on WP and her respecve weghs s colleced from he webse of Offce of he Economc Advsor, Mnsry of Commerce and ndusry, Governmen of nda. Clemens and Nguyen (1981, 198) proposed 4

6 .1. nflaon and varably of relave prce changes The sample momens of he dsrbuon of relave prce changes s compued as follows. The realzaon of he prce n h commody (=1,..,419) observed n perod s denoed by p and ln denoes he log change operaor. The rae of change n h commody prce ndex from -1 o me perod wll be denoed by and s calculaed as gven n equaon.1. Ths measures he annual rae of nflaon for each monh. ln( p / p 1) (.1) The aggregae nflaon rae n perod can hen be compued as he weghed average of he changes of he sub-ndces, as follows n 1 (.) where n denoes he number of commodes n he sample and ω represens he wegh of h commody n he WP baske. Usng aggregae nflaon rae, he hgher order cenral momens of he dsrbuon of relave prce changes are easly compued. The relave prce varably ha s cross-seconal varance of hs dsrbuon denoed as gven by: n 1 s (.3) Ths s a measure of he degree o whch prce changes are dsproporonae, or a measure of changes n relave prces. Thus f all prces move equproporonally (.e, f all relave prces are consan) herefore, π = and =0. Clemens and Nguyen (1981, 198) pon of deparure s he followng relaonshp beween he relave prce change of a commody and he rae of nflaon. ( ) (.4) Where and parameers and are commody specfc s a random dsurbance erm wh he usual classcal properes. follows from he (.4) ha ( / ) s he auonomous rend n he relave prce of commody, reflecng real changes n he economy and ( / ) s he elascy of he relave prce of wh respec o nflaon rae. Also, f >0 (< 0), rses more (less) han proporonaely wh he general nflaon rae, so ha relave prce of commody rses (falls). Varance of relave prces n equaon (.3) can be equvalenly expressed as n 1 ( / ) ( / ) (.5) Fnally he componen due o he neracon of real changes and nflaon s.. Decomposon of varance of relave prces The share of commody n s ( / ) ( / ) / (.6) The s are non-negave and have a un sum. Ths wll gve he dsrbuon of relave prce varably across he commodes n me perod. The effecs of real changes on he relave prce of commody are gven by he coeffcen n equaon (.4), hus he real componen of s gven by ( / ) / (.7) whch s non-negave. The correspondng nflaon componen s ( / ) / (.8) whch s also non-negave. The equaon s esmaed for all 419 commodes ncluded n he sample usng Ordnary Leas Squares esmaon. 5

7 ( / ) / (.9) whch may be eher negave or posve. follows ha (.10) Summng, and over, respecvely gves he share of real componen, nflaon componen and neracon componen, n. Tha s ( / ) / (.11) gves he real componen. Smlarly he share of nflaon componen s ( / ) / and he share of neracon s ( / ) / follows ha, (.1) (.13) 1 (.14) Summng, ; and over, ha s ; (.15) gves he componens of share of commody n oal relave prce varance. And he oal share of commody n oal relave prce varably s gven by (.16) To calculae oal, and ha s real, nflaon and neracon componen for he overall sample perod,, and has been summed respecvely, over. Tha s ; ; (.17).3. Mnmum-varance nflaon rae and naural rae of relave prce varably By usng he equaon (.5) we can derve he naural rae of relave prce varably, n 1 ( / ) ( / ) ( ) (.18) 0 ( / ) (.19) ha s relave prce varably correspondng o zero nflaon. Snce hs value s deermned enrely by real facors such as ncome, ases, echnology ec. hence s called as naural rae of relave prce varably. By dfferenang equaon (.5) w r and pung / 0 gves mnmumvarance nflaon rae *,.e * 3. The Emprcal esuls (.0) Ths secon provdes emprcal resuls and dscussons under dfferen headngs Commody shares and Cross seconal decomposon of relave prce varance The share of a commody ( ) n he oal varance of relave prce changes s gven n column 4 of he Table1 3. neresngly op 30 commodes almos conrbue 60 percen of he oal relave prce varably for he sample perod of he sudy, he shares of only hese domnan commodes has been presened 4. The relave prce of Texursed Yarn alone accouns for 6 percen of oal varably n relave prce. Nex o hs s hgh speed desel ol wh 5 percen share. The enres n columns 3 o 4 decompose ( ) no real ( ) and nflaon ( ) ( / ) / ( / ) componen for each commody. The resuls shows ha 6 percen share of Texursed Yarn n oal relave prce varance s whole made up of real facors and no vsbly of nflaon 3 The resuls are shown for only op 30 commodes, full nformaon can be oban from auhor on reques. 4Monh wse decomposon and commody share s no presened here so as o save he space. However hese fgures gves some dea abou he monh wse decomposon. 6

8 componen. Smlarly for Hgh speed desel ol 93 percen of s share s real componen and 7 percen s nflaon 5. n Table hese domnan commodes are arranged n erms of her percenage share of real and nflaon componen n her respecve conrbuons o oal varance of relave prces. For he commodes lke Lquefed peroleum gas, ron ore, Texursed Yarn, Hgh speed desel ol, Kerosene, T.V. ses B &W and he caegory Capsule oher han vamn & anbocs, he major componen of her share n relave prce varance s of real facors. evden ha all he percen share ) of ron ore comes from real facors such ( as supply shocks, real ncome growh ec. As gven n Table he varably n he relave prces of hese op few commodes commodes s due o real facors. From boom, Table shows commodes whose major componen n her respecve shares s of nflaon facors. For he commodes lke, Napha, Furnace ol, Compuer &Compuer Based Sysem and Capsule oher han vamn & anbocs he major share (almos 100 percen for few) s of nflaon componen n her respecve conrbuons o he oal relave prce varance. Thus varably n relave prces of hese commodes s manly due o nflaonary facors. s evden from esmaes ha here are only few commodes among hese op 30 whose relave prces are affeced manly due o he nflaon facors. eal componen s domnan n mos of commodes. 5fgures gven are afer adjusng for neracon componen. The approach used o allocae no real and nflaon componen, s o arbue he fracon /( ) of neracon o real componen and /( ) o nflaon. Table 1: Commody share and s decomposon* Commody ) ( ) ( ) Texursed Yarn Hgh speed desel ol Polyser Yarn aw Coon TV ses Color Compuer & Compuer Based Sysem Polyser saple fbre Capsule oher han 8 9 vamn & anbocs ron ore Lqufed Peroleum 96 4 Gas Table execp 94 6 vamn & penclln Coon Yarn-'Hanks Acd all Knds Coon Yarn-Cones Napha Ammonum sulphae N conen Synhec Yarn Fsh-nland T.V. ses B&W Deoled cake Elecrcy for 1 98 Agrculure Poulry chcken Lme sone Purfed Terephalc Acd Pcure Tubes (Color) Furnace ol ecfed spr ( Sugar Mlk Kerosene 1 98 TOTAL *all fgures gven are n percenage 7

9 Table : Hgh nflaon and real componen commodes* COMMODTY ) ( ) ( ) T.V. ses B&W ron ore Texursed Yarn Elecrcy for Agrculure 1 98 Kerosene 1 98 Lqufed peroleum gas 96 4 Table execp vamn & 94 6 penclln Hgh speed desel ol Pcure Tubes (Color) TV ses Color Polyser Yarn Lme sone Synhec Yarn Polyser saple fbre Fsh-nland Mlk Poulry chcken Purfed Terephalc Acd Coon Yarn-Cones Coon Yarn-'Hanks Acd all Knds aw Coon Sugar ecfed spr Deoled cake Ammonum sulphae and conen Compuer & Compuer Based Sysem Furnace ol Capsule oher han vamn 8 9 & anbocs Napha TOTAL *all fgures gven are n percenage Thus ndcaes ha he major componen of relave prce varably s bul of real facors and relavely less vsbly of nflaon. ( 3.3. Decomposon of oal relave prce varably across varous perods Below Table 4 (column 1, and 3) gves he componens of oal relave prce varably arbued o real facors ( ), nflaon ( ) and neracon ( ), and column 4 and 5 show fgures of ( ) and ( ) afer allocaon of neracon erm ( ) 6. The poson s gven for each sub-sample separaely. For he perod of sudy on an average, 55 percen of he oal relave prce varably s bul of real componen and he res s arbued o nflaon. Ths shows ha proporon of relave prce varance whch can be arbued o real facors such as echnologcal changes, supply shock, real ncome growh ec. s hgher han nflaon componen. Thus n nda 45 percen of oal relave prce varably s buld-up of nflaon componen whch s very hgh. The share of nflaon componen s hghes n he hgh nflaonary perod from Aprl-1994 o June and lowes n low nflaon perod. Also he nflaon componen s greaer han s respecve real componen n hgh nflaon perod and lower n low nflaon. Snce average nflaon rae does no vary oo much across he res of he sample perods so does he shares of ( ) and ( ). Therefore seems ha n hgher nflaonary perods he share of nflaon componen n relave prce varance ncreases and vce-versa. Also he share of neracon componen s hghes n he perod july-1997 o Febuary-001, whch means ha for hs perod relave prce of commodes are ncreasng due o real changes, bu decreasng due o nflaon. 6 resuls are provded only for fve domnan commodes n each subsample, full nformaon can be colleced on he reques. 8

10 Table 4: Decomposon of relave prce varably* COMMODTY ) ( ) ( ) Aprl-1994 o Jun 1996 aw Coon Chlles(Dry) 4 78 Capsule oher han vamn & anbocs Coon Yarn-Cones Comper & Compuer Based Sysem July-1996 o Feb- 001 Kerosene Lqufed peroleum Polyser Yarn Hgh speed desel ol Table execpv &pencl March-001 o November-003 Cakes & Swee oles Packed ea Coffee Urad Avaon urbne fuel December-003 o November-009 ron ore 8 98 Acd all Knds Sugar Bars & rods Poaoes 68 3 *all fgures gven are n percenage 3.4. Mnmum-varance nflaon rae and naural rae of relave prce varably To evaluae he mnmum-varance nflaon rae and naural rae of relave prce varably equaons (.19) and (.0) were used. For each me perod (sub- sample) hese esmaes were calculaed separaely o have an dea of varaons n hese esmaes over he perod of me and across dfferen nflaonary suaons. The esmaed mnmum-varance ( nflaon rae seem o be 4.5 percen for he enre sample perod,whch s conssen wh he offcal hreshold rae mananed by he eserve Bank of nda. The esmae of naural rae of relave prce varably s 0.06 percen for he same perod. However, here are dramac varaons n hese esmaes across varous he subsamples of he sudy perod. The mnmum-varance nflaon rae seems o have been 7 percen for he perod Aprl-1994 o June-1996 whch s characersed by he average nflaon rae of abou 8 percen. For he lower nflaonary perod from March-001 o November-003 he mnmum-varance nflaon rae s 3.5 percen whch s lowes of all. Thus from he emprcal esmaes, s evden ha he mnmum-varance nflaon rae seem o be hgher (lower) for he hgher (lower) nflaonary regme and vares beween 3.5 o 7.1. The naural rae of relave prce varably vares beween 0.11 o s lowes for perod from December-003 o Novermber-009 and s hghes for he perod from July-1996 o Febuary-001. Table 5: Mnmum-varance nflaon rae* Sample Perod ( ) 0 ( *) 1994M4-009M M4-1996M M7-001M M3-003M M1-009M *all fgures gven are n percenage 4. Concluson Applcaon of he mehodology suggesed by Clemens and Nguyen (1981, 198) o ndan daa on WP daa appears o yeld several neresng resuls. The esmaes ndcae ha he ngredens, nflaon componen and real componens of relave prce varably accoun for 45 and 55 percen respecvely. Ths proporon of nflaon and 9

11 real componens vary slghly across he subsamples. However hese proporons seem o be que oppose for hgh and low nflaonary perods. The proporon due o nflaon facors s hghes durng nal monhs of sample perod (Aprl-1996 o June- 1996) whch s characersed by he hgh nflaon rae. Thus he share of nflaon componen seem o ncrease n hgh nflaonary perod and decreases n low nflaonary suaon. Cross commody dsrbuon of relave prce varance n nda shows commodes lke ron ore, Texursed Yarn, Hgh speed desel ol, raw coon, Compuer & Compuer Based Sysem, LPG, TV ses Color and Polyser saple fbre domnae n share over he perod of sudy. Almos 60 percen of oal relave prce varably s arbued o op 30 commodes and res o all oher commodes. Thus we conclude ha major consuen of hgh varably n relave prces s due o hese few domnan commodes and care full aenon should be pad ohe momens of prces of hese commodes n any assessmen of relave prce varably. From he polcy perspecve, he flucuaons n he prces of hese commodes may gve mporan sgnals regardng he movemens n overall relave prce varably. The more crucal nference ha emerges from he emprcal analyss s ha he hreshold nflaon rae ha mnmzes he relave prce varably seems o have been 4.5 percen, whch s broadly conssen wh he offcal hreshold rae mananed by he eserve Bank of nda. These esmaes of mnmum varance nflaon rae and he naural rae of relave prce varably may provde useful benchmarks for monorng he economy and for assessng he effecs of polcy change on srucure of relave prces o eserve Bank of nda. eferences Ba J., Perron P.,(1998), Esmang and Tesng Lnear Models wh Mulple Srucural Changes, Economerca, pp Barro ober J.(1976), aonal Expecaons and he role of Moneary polcy, Journal of Moneary Economcs, pp.1-3. Clemens, Nguyen (1981), nflaon and elave Prces-A Sysem-Wde Approach, Economc Leers, pp Clemens, Nguyen (198), nflaon and elave Prces- A Decomposon Analyss, Economc Leers, pp Cukerman, A. (1983), elaveprce Varably and nflaon: A Survey and Furher esuls,carnege-ocheser Conference Seres on Publc Polcy, 19, Amserdam: Norh-Holland. Fscher, S. (1981), "elave Shocks, elave Prce Varably, and nflaon", Brookngs Papers on Economc Acvy,, pp Hercowz, Z. (1981), Money and he Dsperson of elave Prces,Journal of Polcal Economy, 89, pp Lucas,.E., 1973, Some nernaonal evdence on oupu nflaon radeoffs, Amercan Economc evew 63, pp Parks,.W., 1978, nflaon and relave prce varably, Journal of Polcal Economy 86, pp a am (1990), elave Prce Varably n he Uned Saes, Economcs Leers, pp Vnng, D.L. and T.C. Elwerowsk, 1976, The relaonshp beween relave prces and he general prce level, Amercan Economc evew, 66, pp

12 Workng Paper Seres 3/011 DEPATMENT OF ECONOMCS PONDCHEY UNVESTY V NAGA, KALAPET PUDUCHEY