
JOURNAL OF RESEARCH IN NATIONAL DEVELOPMENT VOLUME 7 NO 2, DECEMBER, 2009
STUDY OF FACTORS AFFECTING DEMAND AND SUPPLY OF SUGAR
IN INDONESIA
Rizal Rustam
State Polytechnic of Jember, Indonesia
Email: rizalsp2001@yahoo.com
Abstract
This study is aimed at determining: (a) the factors influencing the demand of national sugar, and the price elasticity of demand (ep) and; (b) the factors influencing the supply of national sugar, and the price elasticity of supply (es). The data comprised a time series of 19902006, taken from the Central Statistics Bureau, The Indonesian Sugar Firm Research Centre, and the World Sugar Statistics and Logistic Institution (BULOG). The “Double Logarithm Multiple Regression Model” and “Two Step Least Square Method” were the analytical techniques used. Results showed that: (a) the factors that significantly affected national cane sugar demand were the price of coffee, size of population, and level of income. On the other hand, the price of sugar, tea and the price ratio did not have significant effect on the sugar demand in Indonesia. The evaluation showed that except for the population rate (3,866) all of the variables were inelastic (less than 1) with an elasticity range of –0,442 to 0,087; and (b) the factors significantly affecting the national sugar supply were the labour wage and land area, while the factors of land productivity and the price of sugar showed otherwise. The elasticity of all variables was inelastic (less than 1), with the range of elasticity being 0,2159 to 0,8959.
Keywords: Demand, supply, price elasticity of demand, price elasticity of supply
Introduction
Sugar is one of the basic needs of communities in Indonesia. Since 1969 the approach to the marketing system has involved governmental activity. Implementing the basic price has some consequences for the continuation of government, and the protection of the interests of consumers. Government intervention in the chain of commerce is because there is some market imperfection that could harm consumers or producers. However, the government intervention must be done carefully in order not to result in the loss or stability for the operators of the market.
Various efforts to increase the production of sugar have produced an increase from 1.72 million tons in 1990 to 2.37 million tons in 2006. In 2006 production reached 12.44 kg per capita,while consumption was higher at 18.17 kg. per capita. So the import of sugar grew to 1.15 million tons with a foreign exchange value of U.S. $ 195.2 million. The import of sugar in the year 2007 was projected to reach a value of U.S. $ 426 million, an amount which when converted to rupiahs is worth Rp. 4 trillion or more, assuming an exchange rate of $ 1 = Rp 9000 (Kiptiyah 2006). Commercial policies that increase competitiveness and efficiency in sugar production should get priority in order to reduce imports and save foreign exchange.
Based on the above background, the problem addressed in this work can be formulated as follows: (a) how is the demand for sugar arrived at? (b) an investigation of the supply procedure for sugar in Indonesia; and (c) what the dominant factors that affect both demand and supply?
Demand for an agricultural commodity is the number of the agricultural commodity needed and purchased by consumers at various price levels. According Winardi (1976), demand is the number of goods able to be purchased at a place and time with the current price at that time. If an item in the demand is influenced by its own price (P), income level (Y), and the price of other goods (X), and assuming a constant elasticity of demand, the demand can be exponentially mathematically expressed in the form of a demand function as follows (Nicholson, 1995; and Koutsoyiannis, 1979):
QD = a Pb Yc Xd
Tomek and Robinson (1977) show that a static demand curve indicates a change in quantity demanded throughout the curve. In this case, to see the sensitivity of quantity demanded to changes in the price, and assuming other variables constant, a mathematical representation, called Demand Function, is used. Meanwhile if the demand is variable or shifts, and the price of other goods, consumer income,
and the other variables are no longer considered to be constant, the mathematical representation is called the Demand Relation. Demand function, mathematically can be formulated as
Qd = f (P°, Ps°, Pk°, Y°, N°, Q°) where
P°, Ps°, Pk°, Y°, N°, Q° are variables.
The Demand relation can be mathematically formulated as follows:
QD = f (P, Ps, Pk, Y, N, T) where, as above,
P, Ps, Pk, Y, N, T are variables.
Supply is the number of goods offered for sale at various price levels in a market at a particular time (Rosyidi, 1984). Supply curve behaviour pictures producers reactions to price changes as it slopes upwards from the left to the topright. Thus, it can be stated that when the price level is lower, the less the amount offered to the market, and the higher the price of a product that applies on the market, the greater the number of products offered. Based on the theory of bidding, the price changes are always associated with a corresponding response. Along the supply curve, the price and the amount of goods move in the same direction (Stanislaus, 1985).
Research method
Time series data from 1990 to 2006 was obtained from the Central Bureau of Statistics, Research and Development Centre Sugar Indonesia (P3GI), Logistics Affairs Agency (BULOG), and institutions related to this research. To avoid any mistake in the interpretation of the variables used in this research, there is need to explain the definitions and measurements of the variables used in this research as follows:
 National demand for sugar in the domestic market in this research (Qdt) is the total sugar that was sold in Indonesia each year from 1990 until 2006,
 Coffee prices in the domestic market (Pkt), is the price of domestic tea in the market (Ptt) and the price of sugar (Pat) is the average price of coffee, tea, and sugar in the domestic market after deflating with the consumer price index in the country (on yearspecific basis) from 1990 to 2006.
 Price Ratio (Hrt) of Indonesia sugar to sugar imported from other countries (Hrt = Pa/Pd) is the domestic sugar price divided by the price of sugar imported from other countries (1990 to 2006).
 Indonesian per capita income (Yt) is the per capita income of the country in the period 1990 to 2006, stated in dollar. This has been deflated with the consumer price index to get real income per capita.
 Number of population (Dt) is the number of people in Indonesia from 1990 to 2006.
 The supply of sugar in the national study (Qst) is the amount of sugar produced in the country every year of the study period.
 Harvest area (Li) is the sugar harvest area from 1990 to 2006.
 Wage labour (Wnt) is the average wage labour in the sugar cane plantations in Indonesia per year from 1990 to 2006.
 Sugar cane plant productivity (Pr) is the output level per year of production of sugar cane from 1990 to 2006.
To answer the first hypothetical question, the estimation model used was the loglinear regression. The basic assumptions are that sugar is a normal good, the consumption level is still far from the saturation level for a period of time long enough, the consumption level is still low, consumer income varies, and the demand elasticity is constant.
The sugar demand function is as follows:
LnQdt = AO + LnPat Al LnPkt + A2 + A3 LnPttt + A4 LnPYt + LnPDt + A5 + A6 LnHrt ut
Where:
Qdt = the demand for sugar (in units of 1,000 tons).
Pat = domestic price of sugar that has been deflated with the consumer price index (Rp/kg).
Hrt = ratio of the domestic price of sugar to imports.
Pkt = the price of coffee which has been deflated (Rp/kg ).
Ptt = price of tea which has been deflated (Rp/kg).
Yt = household income at constant price.
Dt = the population of Indonesia.
AO = is constant while Al, A2, A3, A4, AS, A6 are the parameters that will be estimated.
U = disturbance term.
t = year.
The preferred sugar function is:
ln Qst = BO + B1 + B2 Pat ln ln ln Lit + B3 + B4 ln Wnt PRT + ut
Where:
Qst = quantity of Indonesian sugar (in 1,000 tons).
Lit = area of land planted with sugarcane (in 1000 ha).
Wnt = wage labour in the sugar cane industry (Rp/day).
PRT = sugarcane plant productivity level.
Pat = domestic price of sugar (that has been deflated with the consumer price index, Rp/kg).
BO = is constant while Bl, B2, B3, B4, B5, B6 are parameters that will be estimated.
u = error term.
t = year.
Tests of Single
Ho: βi = 0 and H1: βi ≠ 0
where: i = 1, 2, 3,. . . . n and tcalculated = βi / S βi
Decision: When t calculated > t table, Ho is rejected; the partial of each variable has effect on the demand and the national sugar. If the calculated t < ttable, Ho is accepted; which means that the partial of each variable has no effect on the demand and supply of national sugar.
Testing simultaneously or in unison
Ho: βi = 0
HI: at least one regression coefficient is not equal to zero.
ESS / (k1)
F measure = 
RSS/(nk)
Description: ESS (explained Sum of Square)
RSS (Residual Sum of Square)
k = the number of parameters, and n= number of pairs of data
Decisions: When the F  calculated > Ftable, Ho is rejected; simultaneously all the independent variables affect the demand for national sugar. If the F calculated < Ftable, Ho is accepted; which means that, simultaneously all variables do not affect the demand.
Research results and discussion
Overview and development of sugar production in Indonesia
Sugar is one of the basic needs of the Indonesian people. The need for sugar will continue to increase in line with the growing population and income in Indonesia. The more detailed data on the sugar industry in Indonesia is presented in Table 1.
Table 1. Sugar production, demand for sugar, sugar import, land area, wage labour, and
sugarcane productivity level in Indonesia (1990 2006).
Year 
Production (ton) 
Demand import (ton) 
Land harvest
(Ha) 
Wage labour
(Rp/day) 
Cane productivity
(Ton/ha) 
1990 
1.725.179 
0 
277.615 
2.175 
75.70 
1991 
2.025.171 
0 
317.090 
2.244 
79,20 
1992 
2.117.710 
179.000 
334.000 
2.428 
77,20 
1993 
2.289.645 
92.000 
323.302 
2.550 
76,60 
1994 
2.435.881 
15..207 
428.736 
2.750 
78,90 
1995 
2.059.576 
687.936 
436.037 
3.250 
76,90 
1996 
2.094.195 
975.830 
446.533 
3.887 
72,93 
1997 
2.191.986 
1.364.000 
386.878 
4.475 
79,19 
1998 
1.488.269 
1.730.473. 
377.089 
5.040 
78,60 
1999 
1.439.933 
1.500.000 
342.211 
.6.750 
71,26 
2000 
1.690.004 
1.500.000 
340.660 
7.800 
71,47 
2001 
1.725.467 
1.500.000 
344.441 
10.500 
71,00 
2002 
1.755.354 
1.500.000 
350 722 
10.450 
72,30 
2003 
1.634.560 
1.500.000 
336.257 
10.250 
72,70 
2004 
2.051.000 
1.348.349 
344.000 
10.765 
72,50 
2005 
2.265.000 
1.245.000 
365.450 
12.750 
72,85 
2006 
2.375.000 
1.150.000 
380.000 
13.500 
73,00 
2007 
2.412.000 
1.147.212 
388.677 
13.768 
73,12 
2008 
2.443.000 
1.443.000 
396.000 
13.876 
74.58 
Source: Several editions of Central Bureau of Statistics (BPS), Research and Development Centre Sugar
Indonesia (P3GI), Logistics Affairs Agency (BULOG)
Data in Table 1 shows that during the period 19902006, the amount of sugar production in Indonesia increased. But the increase could still not meet the national demand in Indonesia . Sugar imports remained the only means of meeting the excess demand. The increase was not followed by increased levels of productivity. Increase in sugar production was merely caused by the increased land area planted with sugar cane. In fact, sugar cane productivity decreased from an average of 75.70 tons per hectare in 1990 an average of 73.00 tons per hectare in 2006.
Even then, production of sugar in Indonesia in the study period concentrated mainly on the island of Java. Currently the island is inhabited by nearly 65 percent of the total population in Indonesia. The island which is estimated to contribute almost 77 percent of the total sugar production in the country, is also the biggest consumer. In the period 2004/2005, of the total production in Indonesia, the output of sugar cane in the island was 27.9 million tons (74.4 percent) and 9.6 million tons (25.6 percent) elsewhere. In 2005/2006, the production in Java decreased to 23.8 million tons (73.2 percent) and outside Java it decreased also to 8.5 million tons (26.3 percent) (P3GI, 2006).
On the other hand, according to research conducted by P3GI, the area planted on both dry land and wet land in the year 2006 was still high in the central island of Java, with an area of 285,025 ha, or approximately 67.25 percent of total planted area. Further according to P3GI, dry land sugar cane records a productivity 31.60 percent lower. According Ratnawati (2006), dry land sugar cane production costs more per kg of sugar based on technology and also because the location of the factory is relatively remote.
Factors influencing demand of sugar in Indonesia
In evaluating the factors that influence demand for sugar in the test model used, the relevant regressors were sugar prices in Indonesia, the ratio of domestic to import prices for sugar, the price of coffee, tea prices, income, and the size of population.
The double Logarithm regression model of the demand function for sugar in Indonesia shows that. :
a. Ftest value of 36.66 is currently ¬ Ftable value of α 0.05 (6.7) = 3.87 so the Ftest> Ftable, that means all variables are statistically independent of pro ¬ same effect on the dependent variable.
b. Determination coefficient value (R2) of 0.87 independent variable shows the model is able to explain variations in changes in the national sugar demand of 87.0% and 13.0% explained by variations in external independent variables in the model.
Our results show that, of the six regressors, three have real effect at the alpha level of 0.05. The evaluation showed that there is some multicollinearity in the model. According to Gujarati (1995), if the goal is forecasting in a regression, the multicollinearity is not a serious problem, because the higher the R2, the better with the assumption that the relationship between variables do occur in the future. The model for sugar demand in Indonesia is as follows:
Ln Qdt = Ao + Al Ln Pat + A2 Ln Pkt + A3 Ln Ptt + A4 Ln Dt + AS Ln W.+ A6 Lo Hrt + ut
Table 2. The regression of sugar demand function In Indonesia with simultaneous
equation model
Independent Variable 
Regression Coefficient 
Tmeasure 
Price of Sugar (Ln Pat) 
0,228ns 
0,892 
Price of Coffee (LnPkt) 
0,442* 
 2,632 
Price of tea (Li Ptt) 
0,216ns 
 1,039 
Size of Population (LnDt) 
3,866* 
3,284 
Income (Yt) 
0,087* 
2,475 
Sugar Price Ratio (Hrt) 
0,023ns 
0,223 
Constant 
5,22. 1040 

Rsquare 
0,87 

Fmeasure 
36,88 

D 
2,066 

* = Significant at alpha (α) = 0,05, ns = not significant
The evaluation of the regression in Table 2 indicates that the value of the coefficient of determination (R2) and the Fvalue computation is very large. Interpretation regression evaluation of each variable is given as follows:
Price of sugar
The price of sugar is the average price per kilogram of sugar each year. The price of sugar is used as the basis to measure the level of costs incurred up to buying sugar. Variable prices have marked negative coefficient, meaning that if the price of sugar increased, the national sugar demand will go down. Variable coefficients of regression for sugar prices shows that the decreasing demand of sugar in Indonesia is due to an increase in the price of sugar. The increased price of sugar each year was caused by increasing cost of production of sugar.
The evaluation shows that this price does not influence significantly the demand for sugar in Indonesia. Based on the statistical test, the price of sugar is an independent variable not significant to the demand of sugar, because the calculated tvalue = 0.892, less than ttable (alpha 0.05, 15) = 2.131. So, the null hypothesis has to be accepted. This means that, statistically, sugar prices do not significantly affect (not significant to) sugar demand in Indonesia. The basic explanation is that the per capita demand for sugar in Indonesia is still low (17.6 kg/year). The need occupied only 3% of overall food consumption of families (Khudori, 2000). This means that even if the price of sugar goes down there is not going to be a steep rise in the consumption; if the price of sugar increased (high), the level of reduction in demand is not necessarily going to be large. These research results point to the same direction as the results of research conducted by Syafrial (1997), Ismanto (1992), Deborah, and Schrader (1990), and Nuhfil (1986).
Price of coffee
The cross price elasticity between the price of coffee (pt) and the demand for sugar in Indonesia came up with a negative sign of 0.442, meaning that if coffee prices rise 1%, the demand for sugar will record a drop of 0.442%. It can be said that coffee and sugar products have relationships that are complementary. An increase in the consumption of one leads to an increase in the consumption of the other. When tested statistically, the independent variable, price of coffee, turned up significant against Indonesia sugar demand at the alpha (α) = 0.05, because the calculated tvalue = 2.632 is greater than the value of ttable (α = 0.05, 15) of  2.131. Results of this research are similar to results of research conducted by Kiptiyah (1994), and Ismanto (1992) that state that the demand for sugar in Indonesia is influenced significantly by price of coffee.
Price of tea
The parameter of the independent variable of tea came up negative 0.216. 0.216 means that if the price of tea increased 1%, the national demand for sugar decreases by 0.216%. The relationship is complementary, meaning that an increase in the price of sugar, in addition to lowering demand for sugar in Indonesia, also decrease the demand for the product tea. Statistically, tea prices are not significant because the calculated tvalue = 1.039 is smaller than the value of ttable (α = 0.05, 15) = 2.131.
Size of population
Parameter value of the independent variable, size of population, is 3.866. It means that if the number of inhabitants increased by 1%, the demand for sugar in Brazil increases by 3.866%. The relationship between sugar consumption size of population is positive. If the number of inhabitants increases, the consumption of sugar will also rise. When the statistical test was performed, the independent variable is significant at the alpha (α) = 0.05, because the calculated tvalue = 3.284 is greater than the value of ttable (α = 0.05, 15) that is 2.131.
Income level
The parameter value of income came to 0.087, meaning that if the level of per capita income increased 1% , the national demand for sugar also increased by 0.087%. The relationship between sugar consumption and the income level is positive, meaning that if the income increased, consumption of sugar also increased. Research results show that increases in income per capita are not entirely used to increase the national demand of sugar. This is confirmed by the fact that the national sugar consumption level is still low, and as well as per capita income. When the statistical test was performed, the independent variable was significant at the alpha (α) = 0.05, with tcalculated = 2.475larger than ttable (α = 0.05, 15) = 2.131. Results are found to be in line with the results obtained by Syafrial (1997), Ismanto (1992), Kiptiyah (1994) and Suharto (1999) which state that the national sugar demand is influenced positively by the level of income, and the effect of the independent variable is significant against the dependent variable.
The ratio of domestic sugar price to import
This is a comparison between the domestic price of sugar and sugar import prices. This variable is used as a basis to measure the level of costs incurred up to buying domestic sugar and imported sugar. The parameter value for sugar price ratio is a negative 0.023, meaning if the price of sugar in comparison to the price of Indonesia sugar imports increased by 1%, the demand for sugar in Indonesia will decrease by 0.023%. This shows that an increase in sugar price ratio will attract a negative response by sugar producers. This is because domestic sugar prices have been low when compared with the price of other crops (especially rice). If statistical testing is done, the price ratio of sugar is not significantly different at the alpha (α) = 0.05, because the calculated tvalue = 0.223 is smaller than the value of ttable (α = 0.05, 15) = 2.131 .
Factors that influence sugar supply in Indonesia
The sugar supply function is as follows:
Ln Qst = Bo + Bl Ln Pat + B2 Ln Lit + B3 Ln Prt + B4 Ln Wnt + ut
The evaluation of the simultaneous equation model of factors that affect the supply of sugar in Indonesia is presented in Table 3.
Table 3: The evaluation of the sugar supply function in Indonesia with
simultaneous equation model
Independent Variable 
Regression Coefficient 
Tmeasure 
Price of Sugar (Ln Pat) 
0,0341ns 
0,735 
Wage Labour (LnWnt) 
0,2159* 
 2,684 
Cane Productivity (Ln Prt) 
0,3347ns 
0,617 
Cane Harvest (Ilt) 
0,8959* 
5,401 
Constant 
5,147 

Rsquare 
0,883 

Fmeasure 
16,861 

D 
2,787 

* = Significant at alpha (α) = 0,05, ns = not significant
Regression analysis results in Table 3 indicate that the estimated supply function is good enough, because the value of the coefficient of determination and calculated Fvalue are high. Interpretations are as follows:
The sugar price
Price is the price of sugar per kilogram every year in Indonesia. The influence of the price of sugar on the sugar supply is not significant, although positive. If the price of sugar increased, the supply will also rise. This means that the sugar price increase will be responded to positively by sugar cane producers (Koestono, 1991). If the test is done statistically, the influence of the price of sugar is not significant, because the calculated tvalue = 0.7345 is smaller when compared with the ttable value (alpha (α) 0.05, 14) = 2.131.
Wage labour
The parameter estimate for wage labour as a regressor is 0.2159. If wage labour increased by 1%, the sugar output will decrease by 0.2159%. This shows that if wage labour increased, the preferred sugar will decrease in supply. Reduction of labour usage will automatically reduce the production of cane each hectare. The statistical test with the alpha (α) = 0.05, shows that the calculated tvalue, 2.684 is greater than the value of ttable, 2.131, so that Ho is rejected. This means that the influence of the independent variable of wage labour on the preferred sugar is significant.
Sugar cane productivity
The parameter of sugar cane productivity is 0.3347, meaning that if the productivity of sugar cane plants increased the preferred national sugar also increased by 0.3347%. Theoretically, the increase in productivity will increase the amount of output. If the test is done statistically, with the alpha (α) = 0.05, the independent variable, sugar cane productivity has no significant effect on the supply sugar, because the calculated tvalue = 0.617 is smaller than the table tvalue = 2.131.
Cane harvest
The parameter estimate in this case is a positive 0.8959, meaning that if cane harvest increased 1%, the preferred national sugar will also increase, by 0.8959%. This shows that enhancing the cropped area will cause increased production. Tested statistically, with the alpha (α) = 0.05, the independent variable of cane harvest has significant effect on preferred sugar, because the value of t = 5.401 is larger than the table tvalue = 2.131.
Conclusion
Based on the results of the data analysis, certain conclusions can be drawn and some suggestions proffered.
1) The factors that affect the demand of sugar are national coffee prices, the size of population, and income, while the price of tea, the price of sugar, and the price ratio between domestic and imported do not affect real demand for sugar in Indonesia. Results obtained from calculations show that the elasticity coefficients of all the variables, except population (3.866), fall within the range of from 0.442 to 0.087. thus, they are inelastic (smaller than one).
2) The factors which significantly exert influence on the preferred sugar are: national wage labour and land area, while the productivity of land and the price of sugar do not affect it significantly. Results obtained from the calculation that the elasticity coefficient of each the variable is less than one, within the
elasticity range 0.2159 to 0.8959. again, these show inelasticity.
3) The evaluation that has been done shows that, in order to increase production, the government should recommend the purchase prices of sugar and also needs to increase the planting area while increasing productivity through improved technology at the level of the farm.
4) It is recommended that further research, with a broader scope, be undertaken that includes the calculation of the competitiveness of sugar production for the regions outside Java.
References
Badan Urusan Logistik. (2006) Statistik Bulog, BULOG. Jakarta
Biro Pusat Statistik. (19902006) Indonesia Year Book, Jakarta
Gujarati, D. N. (1995) Basic Econometrics. International Edition, London: McGrawHill, Inc.
Ismanto, H.S. (1992) Elastisitas Permintaan Gula Pasir di Jawa Timur, Tesis Program Pasca Sarjana Universitas Gajah Mada. Yogyakarta.
Khudori. (2000) Revitalisasi industri gula nasional menuju ekonomi kerakyatan. Makalah Pada Seminar Sehari Strategi Pengembangan Industri Gula Menghadapi Era Pasar Bebas. Politeknik Pertanian Jember.
Kiptiyah, S.M. (2004). Implikasi kebijakan harga gula terhadap kesejahteraan dan pendapatan produsen, konsumen dan pemerintah dalam usaha mencapai swasembada gula. Jurnal Universitas Brawijaya, Vol.6. No.1, Malang.
Kiptiyah, S.M. (2006). Implikasi kebijakan harga gula terhadap kesejahteraan dan pendapatan produsen, konsumen dan pemerintah dalam usaha mencapai swasembada gula. Makalah Pada Seminar Sehari Strategi Pengembangan Industri Gula Menghadapi Era Pasar Bebas. Politeknik Pertanian Jember.
Koutsoyianis, A. (1979) Modern Microeconomics. Second Edition, London: The Macmillan Press Ltd.
Nuhfil, H. (1986) Evaluasi permintaan dan penawaran teh Indonesia di pasaran domestik dan Internasional. Thesis Magister Institut Pertanian Bogor.
Pusat Penelitian Perkebunan. Gula (2006). Ikhtisar Angka Perusahaan Tahun Giling 1998. P3GI. Pasuruan
Ratnawati. (2006) Kajian efisiensi sistem agribisnis dan alternatif kebijakan dalam pengembangan produksi serta distribusi gula tebu di Indonesia. Bogor: Institut Pertanian Bogor..
Schrader, D. and Schrader, L. F. (1990) Cholesterol information and sell egg consumption, American Journal Agricultural Economics, Vol 72, no. 3.
Soeharto, A. (1999) Evaluasi permintaan dan penawaran gula nasional. Penelitian Thesis Universitas Brawijaya. Malang
Stanislaus, S. M. (1985) Farmers Response to Price Changes in Developing Economy, New Delhi: Ashish Publiching House..
Syafrial et. al. (1997). Model semulasi kebijakan efisiensi ekonomi gula pasir di Indonesia dalam rangka menuju era perdagangan bebas. Jurnal Penelitian IlmuIlmu Sosial (Social science), Vol.9,2, Agustus 1997. Universitas Brawijaya Malang.
Tomek, W and Robinson, K (1987) Agricultural produck Price. NewYork Cornel Universty Press Ltd,.

