Demand analyses of rice in Malaysia

Transcription

1 MPRA Munch Personal RePEc Archve Demand analyses of rce n Malaysa (John) Yeong-Sheng Tey and Mad Nasr Shamsudn and Zanalabdn Mohamed and Amn Mahr Abdullah and Alas Radam 7. August 2008 Onlne at MPRA Paper No , posted 7. May :22 UTC

2 DEMAND ANALYSES OF RICE IN MALAYSIA by Tey (John) Yeong-Sheng*, Mad Nasr Shamsudn, Zanalabdn Mohamed, Amn Mahr Abdullah 1, and Alas Radam 1 Abstract As a typcal developng Asan county, the growth n per capta ncome generally brngs to dversfcaton n Malaysans food basket. The most sgnfcant observaton s the fallng n per capta consumpton of rce wth contnuous growth of demand for wheat based products. The objectve of ths study s to estmate the demand elastctes of rce n Malaysa, focusng whether rce s an nferor good. By usng data from Household Expendture Survey 2004/2005, ths study obtans demand elastctes of rce, as well as for other 11 food tems va Lnear Approxmate Almost Ideal Demand System (LA/AIDS) and Quadratc Almost Ideal Demand System (QUAIDS). The emprcal results ndcate that ncome elastcty of demand for rce (0.7104) s the hghest compared to other food tems n the LA/AIDS model, whle ncome elastcty of demand for wheat (0.5087) s hgher than rce (0.4712). Both of the ncome elastctes of demand for rce suggest that rce s not an nferor good n Malaysa. However, by comparng both estmates of demand elastctes and adjusted R 2 s, the QUAIDS model provdes more plausble results than the LA/AIDS model. Keywords: Rce, Wheat, Inferor Good, Lnear Approxmate Almost Ideal Demand System, Quadratc Almost Ideal Demand System, Income Elastcty JEL code: Q11, D Introducton The growth n per capta ncome generally brngs to dversfcaton n food basket. There have been ncreasng per capta consumpton of wheat and meats (partcularly poultry) and decreasng per capta consumpton of the mportant staple food, rce n Malaysa. Statstcally, annual per capta consumpton of rce has decreased from 121kg n 1960 to 70.8kg n Such phenomenon arouses the concern whether rce s a normal or nferor food. Malaysan agrcultural polcy would be msdrected wthout a thorough study of the characterstc. Instead of the fallng per capta consumpton of rce, Nnth Malaysan Plan s target s to ncrease the producton of paddy from 2400 metrc tonnes n 2005 to 3202 metrc tonnes n Wth such effort, self-suffcency level n rce s expected to ncrease from 72 per cent n 2005 to 90 per cent n However, the goal to have hgher self-suffcency level n rce always has conflct wth other polcy objectves of mantanng low food prces and Insttute of Agrcultural and Food Polcy Studes, Unverst Putra Malaysa, Malaysa. * Correspondng author: tyeong.sheng@gmal.com Faculty of Envronmental Studes, Unverst Putra Malaysa, Malaysa. Department of Agrbusness and Informaton System, Faculty of Agruculture, Unverst Putra Malaysa, Malaysa.

3 hgh farm ncome (Chern, 2000). Therefore, Malaysan government has been subsdzng to lowerng prce of rce. Yet, the prcng strategy has not been good enough n correspond to ncreasng demand for wheat, the closest substtute to rce that Malaysans rely heavly on mports. The objectve of ths study s to estmate the demand elastctes of rce n Malaysa. Ths s n regards to the ncome elastcty of demand for rce that shows as the ncome ncreases, whether per capta rce consumpton goes up or down. Also, ths s to study the own-prce elastcty of demand for rce that shows how consumers react to the prce change of rce. Understandng of these demand elastctes s able to shed more lght for demand assessment of rce and further asssts n drawng agrcultural polcy n Malaysa. 2.0 Background Changes of dets wth economc development and ncreasng per capta ncomes have been well documented n Blandford (1984), Garnaut and Ma (1992), Mtchell et al. (1997) and Wu and Wu (1997). Fgures 1 and 2 llustrate the annual per capta consumpton of rce, wheat and meats n Malaysa from 1960 to As per capta ncome of Malaysans grew from very low levels after ndependence, there was an ncrease n consumpton of the basc staple (rce), whch was to curb the malnutrton assocated wth poverty. Increasng per capta ncome led to dversfcaton n food basket. Accordng to Kumar (1997), dversfcaton n the food basket wll mprove the qualty of lfe by addng to the nutrtonal status and welfare of the populaton. Wth dversfcaton, consumers are exposed to a wder choce of foods and shfts n detary pattern. It s observed that per capta consumpton of rce started to declne and whle per capta consumpton of wheat started to ncrease n 1970 s. In the same perod, the consumpton of cheapest protenrch meat, poultry started to ncrease from very low levels. As per capta ncome approached hgher levels wthn 1980 s-2000 s, the role that rce as the man staple food and calorc provder was offset even more sgnfcantly by growth n per capta consumpton of wheat. Contnuous ncrease n per capta consumpton of poultry experenced ts peak n early 1990 s whle stronger purchasng power (manly because of hgher per capta ncome) has seen steady ncrease n per capta consumpton of hgher value meat product, beef. Fgure 1: Annual per capta consumpton of rce and wheat n Malaysa, kg/year Year Rce Wheat Source: Food and Agrculture Organzaton of the Unted Natons, 2007.

4 Fgure 2: Annual per capta consumpton of meats n Malaysa, kg/year Year Beef Mutton Pork Poultry Source: Food and Agrculture Organzaton of the Unted Natons, Demand Elastctes n Prevous Studes There s sgnfcant dfference n the estmated ncome elastctes for rce by usng tmeseres and cross-sectonal data. Table 1 presents the estmated ncome elastctes obtaned from cross-sectonal data. Usng cross-sectonal data, Ishda et al. (2003) and FAPRI (2007) found that the Engel elastctes for rce demand are postve n Malaysa. Most noteworthy s the study by Ishda et al. (2003) that focused on the changes n food consumpton n Malaysa over tme. The estmated postve Engel elastctes for rce suggest that rce has been a normal good over tme. However, the study by Ishda et al. (2003) only utlzed the data collected n West Malaysa. Omttng the sample populaton n East Malaysa may have the Engel elastctes of rce underestmated. Ths s because the ncome level of resdence n East Malaysa s generally lower than West Malaysa. Probably that s the reason that the estmated Engel elastctes for rce are relatvely low compared to other food tems, whch s always nterpreted n a way that the poston of rce as a staple food s decreasng and substtuted by cereal based products. In order to probe the ndcaton mentoned earler, t s nterestng to nvestgate the actual ncome elastctes rather than expendture or Engel elastctes for the varous food tems. In fact, most of the prevous studes (Ishda et al., 2003; Radam et al., 2005; and Baharumshah and Mohamed, 1993) got the demand elastctes for food tems aganst the hypothess as lad down n Engel s law. Engel s law explans that as ncome rses, the proporton of ncome spent on food falls, even f actual expendture on food rses. In other words, ncome elastcty of demand for food s expected to be less than 1 v. v As explaned by Holcomb et al. (1995), note that w pq / y w, where p s prce of food and q s the quantty of food, respectvely. Accordng to Engel s law, w / y 0. But, w / y ( p / y)( q / y) ( w / y). Then p( q / y) w under the condton that w / y 0. Hence, 1, where s ncome elastcty.

5 Table 1: Estmated expendture elastctes of foods n Malaysa, usng cross-sectonal data Food Item Engel a / Expendture b & c / Income d Elastcty / Cereal b - - Rce 0.34 a 0.42 a a 0.09 d Bread and other cereals 0.74 a 0.68 a a - Meat 1.42 a 1.06 a 1.08 b 0.97 a - Beef c - - Mutton c - - Chcken c - - Pork c - - Fsh 0.67 a 0.53 a 0.70 b 0.49 a - Mlk and eggs 0.96 a 0.75 a 1.22 b 0.66 a - Ols and fats 0.78 a 0.67 a 1.63 b 0.64 a - Butter d Cheese d Fruts and vegetables 0.86 a 0.68 a a - Fruts b - - Vegetables b - - Sugar 0.21 a 0.29 a 1.92 b a - Others 0.88 a 0.75 a 1.62 b 0.95 a - Notes: a Ishda et al., 2003 b Radam et al., 2005 c Baharumshah and Mohamed, 1993 d FAPRI, 2007 Table 2 presents the estmated ncome elastctes obtaned from tme seres data. Lke prevous tme seres studes (Baharumshah, 1980; Ishda, 1995; and Nk Faud, 1993), Asan Development Bank (1988), Ito et al. (1989), and Huang et al. (1991) found that the ncome elastctes for rce demand are negatve n Malaysa. It s observed that the estmates of negatve ncome elastctes for rce are ncreasngly hgher over the years as per capta ncome ncreases. In lne wth ths, Huang and Bous (1996) argued that such estmated elastctes from aggregate tme seres data are smply the correlaton between decreasng per capta consumpton of rce and ncreasng per capta ncome, not a true demand relatonshp. Huang and Bous (1996) ponted out the real cause for the declnng per capta consumpton of rce s the rural-urban mgraton, whch s often related to changng lfestyle that leads to change n food ntake. Other than that, the declnng trend may also have been caused by agng populaton, westernzaton of Malaysans det, health conscousness, awareness of food safety and other demographc and soco-economc factors. Accordng to Chern (2000), f rce s an nferor good, then there should be a tendency for rce consumpton to be negatvely assocated wth the household ncome level at any gven pont n tme. Also, f rce s an nferor good, t can then be observed that rce consumpton becomes zero when ncreasng per capta ncome of Malaysans approaches a certan affluence level. Such expectaton totally defeats the meanng of rce as the most

6 mportant staple food to Malaysans and government s plan to ncrease producton of paddy. Thus, t s ratonalzed that rce s stll a normal good n Malaysa. Table 2: Estmated ncome elastctes of rce n Malaysa, usng tme seres data Year Ito et al. e ADB f Huang et al. g Notes: e Ito et al., 1989 f Asan Development Bank (ADB), 1988 g Huang et al., Revew of Econometrc Models As a summary for prevous secton, to date, several studes have prevously estmated food demand systems n Malaysa usng ether pooled aggregate data (Ito et al., 1989; ADB, 1988; Huang et al., 1991) or cross-sectonal data (Ishda et al., 2003; Radam et al., 2005; Baharumshah and Mohamed, 1993; Baharumshah, 1993; Mustapha, 1994; Mustapha et al., 1999, 2000, 2001; FAPRI, 2007). The analyses usng aggregate tme seres data are dfferent from those usng cross-sectonal data. There are numerous fndngs that show ncorporaton of demographc varables enhances the performance of demand analyss. However, lttle attenton has been pad n the demographc effects n the studes of food demand n Malaysa. The man reason to ncorporate demographc effects nto a demand functon s to acheve better estmates of elastctes (Muellbuaer, 1977). Pollak and Wales (1981) and Lewbel (1985) proposed general methods to ncorporate demographc effects nto theoretcally plausble demand

7 systems. The technques are famously appled by Chern et al. (2003) n all of the demand analyss models. However, t s stll uncertan whch model specfcaton s most preferable n analyzng the food demand system n Malaysa. Started wth the study by Baharumshah (1993) that appled Lnear Approxmate Almost Ideal Demand System (LA/AIDS), the model has been remaned ts popularty n most of the studes (Radam et al., 2005; Baharumshah and Mohamed, 1993; Mustapha, 1994; Mustapha et al., 1999, 2000, 2001) of demand analyss n Malaysa. On another hand, Chern (1997) showed notable dfferences of estmated results between the Lnear Expendture System (LES) and LA/AIDS. Chern (2000) compared the performance of the AIDS and LA/AIDS. Lu and Chern (2001) compared Workng-Leser form, the LES, Quadratc Expendture System (QES), and LA/AIDS and concluded that LES or LA/AIDS are preferred n terms of predcton ablty. Further to such fndngs, Cranfeld et al. (2002) probed the performance of models even deeper by comparng the LES and Almost Ideal Demand System (AIDS) wth several rank three systems (An Implctly Drect Addtve Demand System - AIDADS, Quadratc Almost Ideal Demand System QUAIDS, and the QES) n predctng food demands. The study showed that the full rank QES, AIDADS and QUAIDS do ndeed out-perform the LES and AIDS. Lu (2003) found that QUAIDS s superor to the AIDS. Ths s because QUAIDS has propertes of both a flexble form (Fsher et al., 2001) and a nonlnear Engel functon, whch s more approprate to household data (Banks et al., 1997). Thus, ths study chooses both LA/AIDS and QUAIDS wth ncorporaton of demographc varables to estmate demand elastctes and further determne whch demand system performs better. 5.0 Data Descrpton Ths study utlzes the data n Household Expendture Survey 2004/2005. Household Expendture Survey 2004/2005 conducted by the Department of Statstcs s consumer expendture surveys n Malaysa. The survey conssted of a random sample of 14,084 households throughout Malaysa. Fgure 3 shows the food expendture shares for twelve aggregate food groups at home n Malaysa n 2004/2005. Fsh share n Malaysa were sgnfcantly hghest among all. Ths s manly attrbuted by the hgh prces of fsh and oceanc products n the Malaysan market. Ths s followed by expendture shares on bread and other cereals and meat whle the shares of vegetables and fruts were relatvely low compared wth other major foods. Bread and other cereals share s sgnfcantly hgher than rce share. Ths probably s a drect mplcaton of the decreasng mportance of rse as staple food n Malaysa.

8 Fgure 3: Food expendture shares at home n Malaysa, 2004/2005. Beverage 7% Others 5% Sugar 3% Vegetables 11% Rce 10% Bread and other cereals 14% Fruts 7% Ols and fats 3% Eggs 2% Mlk and dary 6% Fsh 19% Meat 13% Source: Household Expendture Survey 2004/2005. One of the major problems n analyzng demand usng cross-sectonal household expendture data s encounterng zero consumpton. Zero consumpton happened when many households dd not purchase varous foods durng the survey perod. Table 3 presents the percentage of households wth zero consumpton at home n Malaysa n 2004/2005. It shows that many households dd not purchase ols and fats, mlk and dary, and eggs durng the survey perod. Zero consumpton of rce and meat were about the same. Bread and other cereals are sgnfcantly lowest. Ths observaton further llustrates the ncreasng mportance of wheat and cereal based products n Malaysans daly ntake compared to rce. Table 3: Percentage of Households wth zero consumpton, 2004/ Methodology Food Item % Rce Bread and other cereals 0.96 Meat Fsh 5.37 Mlk and dary Eggs Ols and fats Fruts 8.89 Vegetables 4.69 Sugar 7.99 Others 8.82 Beverage 8.09 For estmatng demand elastctes, prevous studes n Malaysa (Radam et al., 2005; Baharumshah and Mohamed, 1993; Mustapha, 1994; Mustapha et al., 1999, 2000, 2001) typcally analyzed a complete demand system usng one-step approach. The most approprate procedure s to estmate the frst-stage demand system, where the household makes decsons on how much of ther total ncome s to be allocated for food and nonfood goods consumpton, condtonal on household characterstcs. Due to data lmtaton, prevous study (Dey, 2000) used expendture of non-food tems as the proxy for the prce

9 ndex of non-food tems n order to consder the substtuton relatonshp between food and non-food tems. However, as consumers averagely allocated bggest share of expendture budget for non-food tems, the substtuton effect by non-food tems for food may have been overestmated. Thus, ths procedure s replaced by an Engel functon, followng the suggeston by Chern (2000). The Engel functon s useful to derve ncome elastcty from expendture elastcty. The Workng-Leser of Engel functon can be expressed as: s log X log P H k k 0 1 (1) The quadratc form of Engel functon can be expressed as: log x k k (2) where s = Expendture share of aggregate food, x = Total expendtures of the aggregate food, X = Total expendtures of food and non-food consumer goods and servces, P = Stone prce ndex for the twelve foods, and s random dsturbances assumed wth zero mean and constant varance. H k ncludes dummy varable where k 8 AGE = age of household head, HHSIZE = household sze, URBAN = dummy varable for household that resded n urban area, EMPLOYED = dummy varable for household head who was employed, MALE = dummy varable for household head who s male, MALAY = dummy varable for household head who s Malay, CHINESE = dummy varable for household head who s Chnese, INDIAN = dummy varable for household head who s Indan, SARAWAK = dummy varable for household that resded n Sarawak, and PENINSUL = dummy varable for household that resded n Pennsular Malaysa. k The quadratc form of Engel functon s also useful to valdate whether the QUAIDS model properly apples to food demand analyss n Malaysa. As suggested by Banks et al. (1997), the Workng-Leser form s chosen snce t satsfes the addng-up property. Followng Deaton and Muellbauer (1980a), equaton (1) and (2) are estmated ndependently utlzng the ordnary least squares estmator (OLS). From equaton (1), followng the formulae and procedures of Chern (2000), the ncome elastcty of demand for aggregate food can be derved as, x X. e LA y (3) X x From Blundell et al. (1993), the responsveness of expendture on aggregate food by ncome change n equaton (2) can be computed as, log X (4) e QU y In order to overcome zero consumpton problems, ths study adopts two-step estmator used by Heen and Wessells (1990). Heen and Wessells (1990) extended Heckman s sample selecton model to evaluate the nverse Mlls rato (IMR). The use of IMRs are also ncorporated nto the model to correct the possble bas created by the presence of zero consumpton. Lnear and quadratc form of probt regressons s computed n order to estmate the probablty that a gven household consumes the food tem n queston. k

10 These regressons are used to estmate the IMRs for each household, whch s used as an nstrument n the second stage LA/AIDS and QUAIDS respectvely. The LA/AIDS model for the 12 food tems can be estmated as follows: w log( p ) 1 log( x / P) H mr (5) j j k k j The QUAIDS model for the 12 food tems can be estmated as follows: 2 w log( p ) 1 log( x / P) 2 (log( x / P)) H mr (6) j j k k j where, j = 1, 2,., 12 food groups; w s the budget share of the th food tem; p s the prce of the th food tem, and other varables are the same as prevously mentoned. The addng up, homogenety and symmetry restrctons are mposed for both LA/AIDS and QUAIDS models. Followng the formulae and procedures of Green and Alston (1990), the demand elastctes of LA/AIDS can be computed at sample means as follows: Expendture elastctes LA / AIDS 1 e 1 w (7) Marshallan measures of prce elastctes j LA AIDS e j j w j w w, j 1..., n (8) where j s the Kronecker delta that s unty f = j and zero otherwse. From Blundell et al. (1993), the demand elastctes of QUAIDS can be computed as, Expendture elastctes QUAIDS 2* 2 log( x) e 1 1 w (9) Prce elastctes j w QUAIDS j ej 1 2 * 2 log( x) j w w, j 1..., n (10) Followng the formulae of Chern (2000), the ncome elastctes of demand for aggregate food from equaton (3) and (4) are useful to convert the expendture elastctes from AIDS and QUAIDS to ncome elastctes for food tems respectvely. Income elastcty on the bass of LA/AIDS model can be computed as, LA / AIDS LA LA / AIDS e * e (11) Income elastcty on the bass of QUAIDS model can be computed as, QUAIDS QU QUAIDS e * e (12) k k 7.0 Emprcal Results Both of the Workng-Leser and quadratc form of Engel functon allow a drect test on the hypotheses of Engel's law. As the dependent varable s the logarthm of monthly expendture on food, quadratc form of Engel functon shows that food expendtures are an ncreasng functon of ncome. Consstent wth the expectaton, the Workng-Leser regresson reported negatve and statstcally sgnfcant coeffcent for the logarthm of

11 monthly ncome. It shows that the shares of ncome spent on food are nversely related to ncome level, where poorer households devote hgher shares of ncome to food than rcher households. The Workng-Leser and quadratc form of Engel functon reported that households of bgger famly sze devoted a hgher share of ncome to food and spent more on food than households of small famly sze respectvely. At the mean tme, quadratc form of Engel functon s useful to determne whether or not the demand system n Malaysa s quadratc n log ncome. Thus, more attenton s pad to the coeffcents, 2, of quadratc n log ncome n ths analyss. Specfcally, t s to test the hypothess of 2 = 0 aganst 2 0. The estmated 2 s statstcally dfferent from zero at the 0.01 level. Ths result shows that the demand functon s a non-lnear Engel curve. As a result, the QUAIDS s approprate to be used n the analyss of food demand n ths study. Table 4: Regresson results for Engel curve analyses Workng-Leser Quadratc form Varable Coeffcent Coeffcent (Std. Error) (Std. Error) Intercept (0.016)*** (0.228)*** Log (Total expendture) (0.001)*** (0.071)*** Log (Total expendture)* Log (Total expendture) (0.006)*** Log (age) (0.003)*** (0.016)*** Log (household sze) (0.002)*** (0.009)** Urban (0.002)*** (0.010)*** Employed (0.002)** (0.013)*** Male (0.002)** (0.013)*** Malay (0.003)*** (0.017)*** Chnese (0.003)*** (0.018)*** Indan (0.005)*** (0.025)*** Pennsular Malaysa (0.003)*** (0.017)*** Sarawak (0.004)*** (0.019) R Note: Sgnfcance levels are denoted by *** for 1%, ** for 5%, and * for 10%. The estmated ncome elastctes for aggregate food expendture are presented n Table 5. It clearly shows that Workng-Leser form yelded hgher elastcty of ncome for aggregate food than quadratc form. The estmated ncome elasttces obtaned from the Engel functons would be used to convert the expendture elastctes for ndvdual food tems, whch to be estmated from the LA/AIDS and QUAIDS.

12 Table 5: Estmated ncome elastcty for total food expendture n Malaysa, 2004/2005 Workng-Leser Quadratc In order to determne whch demand system performs better, appendx tables 1 and 2 present the regresson results for LA/AIDS and QUAIDS respectvely. Generally, all estmatons of QUAIDS yelded hgher R 2 values than LA/AIDS. In LA/AIDS model, the R 2 values vary from for other foods to for mlk and dary. The R 2 values n QUAIDS model are hgher than LA/AIDS model s, varyng from for other foods to for mlk and dary. Another focus s pad to the food expendture varable n QUAIDS. The food expendture varable ( 1 ) and ts square term ( 2 ) are sgnfcant n most of the tems regresson, except other foods. Ths suggests that the responses of these food tems expendtures to changes n food expendture are sgnfcantly non-lnear. Prce and expendture elastctes are the center focus n demand analyss. Table 6 depcts the own-prce and expendture elastctes estmated from LA/AIDS and QUAIDS. Both models produced very smlar estmates of own-prce elastctes. Both models reported that own-prce elastctes of demand for rce ( , ) are elastc whle ownprce elastctes of demand for bread and other cereals ( , ) are nelastc. The estmated expendture elastcty of demand for ndvdual food tem n LA/AIDS ranges from for mlk to dary to for rce, whle QUAIDS estmated expendture elastcty ranges from for ols and fats to for beverage. One dstnct dfference between the LA/AIDS and QUAIDS models s that the LA/AIDS model yelded hgher expendture elastcty of demand for rce (1.2836) than other food tems, especally meat (1.0212) and fsh (0.9685), whle QUAIDS model produced smlar expendture elastctes of demand for rce (0.9810), meat (0.9761), and fsh (0.9772). The LA/AIDS specfcaton produced lower expendture elastcty of demand for bread and other cereals (0.7790) than rce (1.2836). Ths result s not consstent wth the expectaton, whch hstorcal experence has shown that as ncome ncreases, Malaysans would substtute wheat based products for rce. Reasonably, the QUAIDS specfcaton reported hgher expendture elastcty of demand for bread and other cereals (1.0591) than rce (0.9810). Thus, the QUAIDS appears to yeld more plausble food demand elastctes than the LA/AIDS model n Malaysa.

13 Table 6: Estmated expendture and own-prce elastctes for food tems, Malaysa Food LA/AIDS QUAIDS Own-prce Expendture Own-prce Expendture Rce Bread and other cereals Meat Fsh Mlk and dary Eggs Ols and fats Fruts Vegetables Sugar Others Beverage Most of the studes (Ishda et al., 2003; Radam et al., 2005; and Baharumshah and Mohamed, 1993) of food demand n Malaysa used the expendture elastcty as the proxy for ncome elastcty. By dong so, some of the foods were regarded as luxury goods due to the more than unty expendture elastctes. As lad down n the hypothess of Engel s law, foods are normal goods, thus, the ncome elastcty must be less than one. By multplyng the estmated ndvdual expendture elastcty wth ncome elastcty for total food expendture, table 7 presents the estmated ncome elastctes for food tems n Malaysa. All of the estmated ncome elastctes are less than unty. However, the observatons are smlar lke those dscussed n the earler secton of expendture elastctes. It comes to the concern whether t s reasonable to have hgher ncome elastcty of demand for rce (0.7104, ) than meat (0.5652, ) n both models. Chern (2000) suggested that the best way to gan nsght of ths phenomenon s to compare the prce of the foods n the data. From the Household Expendture Survey 2004/2005, the average prce of rce s RM3.11/kg, wth normal rce and fragrant rce prced at RM1.74/kg and RM2.50 respectvely. The average prce of meat s RM3.11/kg, wth beef, poultry and mutton prced at RM16.90/kg, RM5.44 and RM11.00 respectvely. In relevance to the effect of prce and affordablty, t s observed that Malaysans consumed as much as much as 5kg of rce (mostly attrbuted by lower qualty normal rce) and 2.84kg of meat monthly. Wth these statstcs, t s noteworthy that the ncome elastctes are estmated on a bass of at-home consumpton only. As Malaysans tend to consume lesser rce but more meat and fsh on the bass of food away from home, t harmonzes the estmates of hgher ncome elastcty of demand for rce than meat on the bass of athome consumpton n ths study.

14 Table 7: Estmated ncome elastctes for food tems, Malaysa Food LA/AIDS QUAIDS Rce Bread and other cereals Meat Fsh Mlk and dary Eggs Ols and fats Fruts Vegetables Conclusons Sugar Others Beverage By utlzng data from Household Expendture Survey 2004/2005, ths secton frst summarzes the applcablty of the LA/AIDS and QUAIDS models n Malaysa. The adjusted R 2 s show that the performance of the QUAIDS model s better than the LA/AIDS model. The QUAIDS also yelded more reasonable and plausble estmated demand elastctes, especally of hgher ncome elastctes for bread and other cereals than rce that s more consstent wth the researchers expectaton. The postve expendture and ncome elastctes both ndcate that rce s not an nferor good n Malaysa. Thus, hgher per capta ncome wll nduce hgher demand for rce. Gven postve forecasts of healthy growth n Malaysan economc, ncome effect alone may not strong enough to yeld a defnte ncreasng trend of rce consumpton. Other factors, namely urbanzaton and westernzaton n taste and preference are lkely to offset the effect of ncome n shapng consumpton of rce. The decrease n rce consumpton s always accompaned wth an ncrease n demand for wheat based products and meat. As an extenson to the dscusson of urbanzaton mpacts above, the follows dscuss more about the results of dummy urban varable n appendx tables 1 and 2. The regresson results of rce ndcate that Malaysans n urban areas devoted lower share of food expendture on rce compared to those n rural areas. Contrary, the regresson results of bread and other cereals and meat ndcate that Malaysans n urban areas devoted hgher share of food expendture on bread and other cereals and meat compared to those n rural areas. Snce rce s suggested not an nferor good, n order to curb such vulnerable scenaro of decreasng demand for rce due to urbanzaton and taste and preference, rce based agrfood ndustry players may want to consder to offer rce n other processed or convenent forms, rather than ordnary rce as physcally seen rce n Malaysa.

15 REFERENCES Asan Development Bank, Evaluaton of Rce Market Interventon Polces. Manla. Baharumshah, A.Z., The Malaysan rce polcy: Welfare analyss of current and alternatve programs. Ph.D. dssertaton. Unversty of Illnos. Baharumshah, A.Z., Applyng The Almost Ideal Demand System (AIDS) to meat expendture data: Estmaton and specfcaton ssues. Malaysan Journal of Agrcultural Economcs, 10. Baharumshah, A.Z. and Mohamed, Z.A., Demand for Meat n Malaysa: An Applcaton of the Almost Ideal Demand System Analyss. Pertanka Socal Scence and & Humantes, 1 (1): Banks, J., Blundell, R. and Lewbel, A., Quadratc Engel Curves and Consumer Demand. The Revew of Economcs and Statstcs, 79 (4): Blandford, D., Changes n food consumpton Patterns n the OECD Area. European Revew of Agrcultural Economcs, 11(1): Blundell R., Pashardes P. and Weber G., What do we learn about consumer demand patterns from mcro data? Amercan Economc Revew, 83, Chern, W.S., Estmated Elastctes of Chnese Gran Demand: Revew, Assessment and New Evdence, a report submtted to the World Bank. Chern, W.S., Assessment of Demand-Sde Factors Affectng Global Food Securty. In Chern, W.S., Carter, C.A. and She, S.Y. eds. Food Securty n Asa: Economcs and Polces. Cheltenham, UK: Edward Elgar Publshng Lmted. Ch. 6. Chern, W.S., Ishbash, K., Tanguch, K. and Yokoyama, Y., Analyss of Food Consumpton Behavor by Japanese Households. FAO Economc and Socal Development Paper, 152. Cranfeld, J.T., Hertel, J.E. and Preckel, P., Changes n the Structure of Global Food Demand. Amercan Journal of Agrcultural Economcs, 80 (5): Deaton, A, and Muellbauer, J., 1980a. An Almost Ideal Demand System. Amercan Economcs Revew, 70. Dey M.M., Analyss of demand for fsh n Bangladesh. Aquaculture Economcs and Management, 4: Food and Agrcultural Polcy Research Insttute (FAPRI) Demand Elastctes across Countres. [Onlne]. Avalable at: [accessed 10 July 2007] Fsher, D., Flessg, A.R. and Serlets, A., An Emprcal Comparson of Flexble Demand System Forms. Journal of Appled Econometrcs, 16:

16 Food and Agrculture Organzaton of the Unted Natons FAOSTAT. [Onlne]. Avalable at: [accessed 12 July 2007]. Garnaut, R. and Ma, G., Gran n Chna. Report for the East Asan Analytcal Unt, Department of Foregn Affars and Trade, Australan Government Publshng Servce, Canberra. Heen D. and Wessells C.R., Demand system estmaton wth mcrodata: a censored regresson approach. Journal of Busness & Economc Statstcs, 8(1), Holcomb, R., J. Park, and Capps, Jr., Examnng Expendture Patterns for Food at Home and Food Away from Home. Journal of Food Dstrbuton Research, 26:1-8. Huang J. and Bous, H., Structural changes n the demand for food n Asa. Food, Agrculture, and the Envronment Dscusson Paper of Internatonal Food Polcy Research Insttute, 11. Huang, J., Davd, C.C., and Duff, B., Rce n Asa: Is t becomng an nferor good? Comment. Amercan Journal of Agrcultural Economcs, 71: Ishda, A., An econometrc analyss of rce economy n Pennsular Malaysa. Agrcultural Economc Papers of Kobe Unversty, 28-29, Ishda, A., Law, S.H. and Ata, Y., Changes n Food Consumpton Expendture n Malaysa. Agrbusness, 19 (1): Ito, Shoch, E. Wesley F. Peterson, and Warren R. Grant., Rce n Asa: Is t Becomng an Inferor Good? Amercan Journal of Agrcultural Economcs, 71: Kumar, P., Food Securty: Supply and Demand Perspectve. Indan Farmng, 12: 4-9. Lewbel, A., A Unfed Approach to Incorporatng Demographc or Other Effects nto Demand Systems. Revew of Economc Studes, 52: Lu, K.E., Food Demand n Urban Chna: An Emprcal Analyss Usng Mcro Household Data. Ph.D. dssertaton. Oho State Unversty. Lu, K.E. and Chern, W.S., Effects of Model Specfcaton and Demographc Varables on Food Consumpton: Mcrodata Evdence from Jangsu, Chna. In, Internatonal Food and Agrbusness Management Assocaton, World Food and Agrbusness Forum XI, Sydney, Australa. June Mtchell, D.O., Ingco, M.D. and Duncan, R.C., The World Food Outlook. Cambrdge Unversty Press, Cambrdge. UK. Muellbauer, J., Testng the Barten Model of Household Composton Effects and the Cost of Chldren. Economc Journal, 87 (347):

17 Mustapha, R.A., Incorporatng Habt n the Demand for Fsh and Meat Products n Malaysa. Malaysan Journal of Economc Studes, 31 (2): Mustapha, R.A., Azz, A.R.A., Radam, A. and Baharumshah, A.Z., Demand and Prospects for Food n Malaysa. In IDEAL UPM, Repostonng of the Agrculture n the Next Mllennum. July Mustapha R. A., Radam, A. and Ismal, M.M., Household Food Consumpton Expendture n Malaysa. In Malaysan Consumer and Famly Economcs Assocaton, 5th Natonal Semnar on Malaysan Consumer and Famly Economcs. Unverst Tenaga Nasonal, Bang, Selangor. August Mustapha, R.A., Azz, A.R.A., Zubad, B.A. and Radam, A., Demand and Prospects for Food n Malaysa. In Radam, A. and Arshad, F.M. ed. Repostonng of the Agrculture Industry n the Next Mllennum. Unverst Putra Malaysa Press, pp Nk Faud, K., Government polcy mpacts on the Malaysan rce sector. Serdang: MARDI. Pollak, R.A. and Wales, T.J., Demographc Varables n Demand Analyss. Econometrca, 49 (6): Radam, A., Arshad, F.M. and Mohamed, Z.A., The Fruts Industry n Malaysa: Issues and Challenges. UPM, Press. Wu, Y. and Wu, H.X., Household Gran Consumpton n Chna: effects of ncome, prce and urbanzaton. Asan Economc Journal, 11(3):

18 Appendx 1: Maxmum lkelhood estmates of LA/AIDS Rce Bread & other cereals Meat Fsh Mlk & dary Eggs Ols & fats Fruts Vegetables Sugar Others Beverage Intercept (0.0063)* (0.0100)*** (0.0078)** (0.0091)*** (0.0060)*** (0.0021)*** (0.0025)*** (0.0065)*** (0.0054)*** (0.0032)*** (0.0084)*** - log (prce of rce) (0.0029)*** (0.0046)*** (0.0037)*** (0.0043)*** (0.0028)*** (0.0009)*** (0.0011)* (0.0029)*** (0.0014)*** (0.0038)** - log (prce of bread and other cereals) (0.0009)*** (0.0016)** (0.0003)** (0.0009)*** (0.0005)*** (0.0014)*** - log (prce of meat) (0.0020)*** (0.0018)*** - (0.0030)*** (0.0020)*** (0.0008)** (0.0021)*** (0.0017)*** (0.0010)*** (0.0027)*** - log (prce of fsh) (0.0027)*** (0.0043)*** (0.0026)*** (0.0009)*** (0.0011)*** (0.0027)*** (0.0023)*** (0.0013)*** log (prce of mlk and dary) (0.0014)*** (0.0011)*** (0.0007)*** - (0.0003)*** (0.0003)** (0.0009)*** (0.0007)*** (0.0004)*** (0.0012)*** - log (prce of eggs) (0.0021)*** (0.0034)*** (0.0027)*** (0.0032)*** (0.0007)*** - (0.0008)** (0.0022)*** (0.0018)*** (0.0011)** log (prce of ols and fats) (0.0008)*** (0.0013)*** (0.0010)** (0.0012)*** (0.0002)*** - (0.0008)*** (0.0007)** (0.0004)*** (0.0010)*** - log (prce of fruts) (0.0012)*** (0.0020)*** (0.0019)** (0.0004)*** (0.0004)*** log (prce of vegetables) (0.0022)*** (0.0035)*** (0.0028)*** (0.0032)*** (0.0009)*** (0.0014)*** - (0.0011)*** (0.0029)*** - log (prce of sugar) (0.0007)*** (0.0011)*** (0.0010)** (0.0011)*** (0.0003)*** (0.0008)*** (0.0003)*** - (0.0011)*** - log (prce of others) (0.0008)** (0.0014)*** (0.0011)*** (0.0013)*** (0.0003)*** (0.0009)* (0.0007)*** (0.0004)*** - -

19 log (prce of beverage) log (x/p) (0.0009)*** (0.0014)*** (0.0012)** (0.0013)*** (0.0009)*** (0.0003)*** (0.0003)*** (0.0009)*** (0.0008)*** (0.0004)* (0.0012)*** - Log (age) (0.0000)** (0.0001)*** (0.0001)*** (0.0001)*** (0.0000)*** (0.0000)*** 0 (0.0000)*** (0.0000)*** (0.0000)*** Log (household sze) (0.0003)*** (0.0004)*** (0.0003)*** (0.0004)*** (0.0003)*** (0.0001)** (0.0003)*** (0.0002)*** (0.0001)* (0.0003)*** - Urban (0.0012)*** (0.0020)*** (0.0016)*** (0.0019)*** (0.0012)*** (0.0013)*** (0.0011)*** Employed (0.0016)** (0.0026)*** (0.0021)*** (0.0024)*** (0.0016)*** (0.0005)* (0.0016)*** (0.0014)** Male (0.0026)* (0.0021)** (0.0024)* (0.0016)** (0.0005)* (0.0014)*** (0.0008)*** Malay (0.0020)*** (0.0033)*** (0.0026)* (0.0031)*** (0.0020)** (0.0007)*** (0.0008)*** (0.0021)*** (0.0017)*** (0.0010)*** Chnese (0.0021)*** (0.0028)*** (0.0032)*** (0.0021)** (0.0007)*** (0.0008)*** (0.0022)*** (0.0018)*** (0.0011)*** (0.0028)*** - Indan (0.0030)*** (0.0049)*** (0.0039)* (0.0030)*** (0.0010)*** (0.0012)*** (0.0026)*** (0.0040)*** - Pennsular Malaysa (0.0021)*** (0.0028)*** (0.0032)*** (0.0021)*** (0.0007)*** (0.0008)*** (0.0022)*** (0.0018)* (0.0010)*** (0.0028)*** - Sarawak (0.0023)*** (0.0037)*** (0.0030)*** (0.0035)** (0.0007)*** (0.0009)** (0.0024)*** (0.0020)*** (0.0011)*** (0.0030)*** - IMR (0.0061)*** (0.0068)*** (0.0029)*** (0.0040)*** (0.0013)*** (0.0005)*** (0.0008)*** (0.0021)*** (0.0032)*** (0.0012)*** (0.0027)*** - Adjusted R Note: Sgnfcance levels are denoted by *** for 1%, ** for 5%, and * for 10%.

20 Appendx 2: Maxmum lkelhood estmates of QUAIDS Rce Bread & other cereals Meat Fsh Mlk & dary Eggs Ols & fats Fruts Vegetables Sugar Others Beverage Intercept (0.0064)*** (0.0100)*** (0.0079) (0.0093)*** (0.0061)*** (0.0021)*** (0.0025)*** (0.0066)*** (0.0055)*** (0.0032)*** (0.0085)*** - log (prce of rce) (0.0029)*** (0.0046)*** (0.0037)*** (0.0043)*** (0.0028)*** (0.0009)*** (0.0011) (0.0029)*** (0.0024) (0.0014)*** (0.0038)** - log (prce of bread and other cereals) (0.0009)*** - (0.0014) (0.0016)** (0.0010) (0.0003)** (0.0004) (0.0011)* (0.0009)*** (0.0005)*** (0.0014)*** - log (prce of meat) (0.0020)*** (0.0018)*** - (0.0030)*** (0.0020)*** (0.0007) (0.0008)** (0.0021)*** (0.0017)*** (0.0010)*** (0.0027)*** - log (prce of fsh) (0.0027)*** (0.0043)*** (0.0026)*** - (0.0026) (0.0009)*** (0.0011)*** (0.0027)*** (0.0023)** (0.0013)*** (0.0036) - log (prce of mlk and dary) (0.0009) (0.0014)*** (0.0011)*** (0.0007)*** - (0.0003)*** (0.0003)** (0.0009)*** (0.0007)*** (0.0004)*** (0.0012)*** - log (prce of eggs) (0.0021)*** (0.0034)*** (0.0027)*** (0.0032)*** (0.0007)*** - (0.0008)* (0.0022)*** (0.0018)*** (0.0011)** (0.0028) - log (prce of ols and fats) (0.0008)*** (0.0013)*** (0.0010)** (0.0012)*** (0.0008) (0.0002)*** - (0.0008)*** (0.0007)** (0.0004)*** (0.0010)*** - log (prce of fruts) (0.0012)*** (0.0020)*** (0.0016) (0.0019)** (0.0012) (0.0004)*** (0.0004)*** - (0.0011) (0.0006) (0.0016) - log (prce of vegetables) (0.0022)*** (0.0035)*** (0.0028)*** (0.0032)*** (0.0022) (0.0007) (0.0009)*** (0.0014)*** - (0.0011)*** (0.0029)*** - log (prce of sugar) (0.0007)*** (0.0011)*** (0.0010)* (0.0011)*** (0.0007) (0.0002) (0.0003)*** (0.0008)*** (0.0003)*** - (0.0011)*** - log (prce of others) (0.0008)** (0.0014)*** (0.0011)*** (0.0013)*** (0.0008) (0.0003) (0.0003)*** (0.0009)*** (0.0007)*** (0.0004)*** - -

21 log (prce of beverage) log (x/p) (0.0011)*** (0.0017)*** (0.0014)*** (0.0016) (0.0011)*** (0.0004)*** (0.0004)*** (0.0011)*** (0.0009)*** (0.0005) (0.0014)*** - log (x/p)*log (x/p) (0.0001)*** (0.0001)*** )*** (0.0001)*** (0.0001)** (0.0000)*** (0.0000)*** (0.0001)*** (0.0001)*** (0.0000)* (0.0001) - Log (age) (0.0000)* (0.0001)*** (0.0001)*** (0.0001)*** (0.0000)*** (0.0000)*** (0.0000) (0.0000)*** (0.0000)*** (0.0000)*** (0.0001) - Log (household sze) (0.0003)*** (0.0004)*** (0.0003)*** (0.0004)*** (0.0003)*** (0.0001) (0.0001)** (0.0003)*** (0.0002)*** (0.0001)** (0.0003)*** - Urban (0.0012)*** (0.0020)*** (0.0016)*** (0.0019)*** (0.0012)*** (0.0004) (0.0005) (0.0013)*** (0.0011)*** (0.0006) (0.0016) - Employed (0.0016)** (0.0026)*** (0.0021)*** (0.0024)*** (0.0016)*** (0.0005)* (0.0006) (0.0016)*** (0.0014)* (0.0008) (0.0021) - Male (0.0016) (0.0025)** (0.0021)** (0.0024)* (0.0016)** (0.0005) (0.0006) (0.0016) (0.0013)*** (0.0008)*** (0.0021) - Malay (0.0020)*** (0.0032)*** (0.0026)* (0.0031)*** (0.0020)** (0.0007)*** (0.0008)*** (0.0021)*** (0.0017)*** )*** (0.0027) - Chnese (0.0021)*** (0.0034) (0.0028)*** (0.0032)*** (0.0021)** (0.0007)*** (0.0008)*** (0.0022)*** (0.0018)*** (0.0011)*** (0.0028)*** - Indan (0.0030)*** (0.0049)*** (0.0039)* (0.0046) (0.0030)*** (0.0010)*** (0.0012)*** (0.0031) (0.0026)*** (0.0015) (0.0040)*** - Pennsular Malaysa (0.0021)*** (0.0034) (0.0028)*** (0.0032)*** (0.0021)*** (0.0007)*** (0.0008)*** (0.0022)*** (0.0018)** (0.0010)*** (0.0028)*** - Sarawak (0.0023)*** (0.0037)*** (0.0030)*** (0.0035)** (0.0023) (0.0007)*** (0.0009)** (0.0024)*** (0.0020)*** (0.0011)*** (0.0030)*** - IMR (0.0061)*** (0.0068)*** (0.0029)*** (0.0040)*** (0.0013)*** (0.0005)*** (0.0008)*** (0.0021)*** (0.0031)*** (0.0012)*** (0.0027)*** - Adjusted R Note: Sgnfcance levels are denoted by *** for 1%, ** for 5%, and * for 10%.