Data Sources and Descriptions

80 The production of Fresh Fruit Bunches FFB with the fraction of production around 15-17 tons per hectare will also grown significantly. The production of FFB in small holders in year 2010 was 7 millions tons, private owned was 12 millions tons and government owned was 1.5 millions tons. From computer simulation we can predict that the FFB production from small holder in year 2030 is around 61 millions tons, private owned is around 34 millions tons and government is around 1.7 tons millions. Total FFB production is around 97 millions ton. We assumed oil extraction rate of CPO is 20 and palm oil mill productivity rate is 90. Lembito, 2013.In the short term, it seems that land expansion will remain the main strategy increasing CPO production capacity. In the medium term, the trend of industry integration will increase, along with the growth of downstream industry for oleochemicals and biofuel. PWC, 2013 Figure 9 : Growth of FFB Production Source: Lembito, 2013

3.2. CPO Demand and Supply Sub-model

From the data we analyzed from year 2000-2010, we calculate the total demand domestic demand and export demand. The computation the domestic demand mainly come from the assumption of cooking oil with an average 9 kgscapitayear cooking oil or equal to 13.5 kgs CPO capitayear. From BPS we find the Indonesan population around 237 millions in year 2010 with the birth rate is 0.011 and dead rate is 0.007 Kuncoro, 2011. The export demand on year 2010 around 12.8 millions tons of CPO and domestic demand around 3.2 millions tons. With the computer simulation we can predict the demand in year 2030. The Export demand will be around 60 millions tons and domestic will be around 5 millions tons, and the estimated production will be around 80 millions ton . This means that Indonesia can increase the export market since there will be enough stock to sell. And Indonesia will have to handle competition from other country producers especially that of Malaysia. The majority of CPO producers in Indonesia currently sells overseas. Markets in China and India are predominantly for culinary purposes including cooking oil, whilst those in the EU are for biodiesel and confectionery manufacturing. The domestic market is mainly for cooking oil. The export profile for 2011 is dominated by CPO ahead of olein and biofuel, whereas the export markets are dominated by India and China. The drop in the global crude palm oil CPO price in 1999 due to oversupply has suggested palm oil plantations to find alternative uses of palm oil besides edible consumption. At the same time, Indonesia is also facing an increase in domestic fossil fuel consumption, while its oil exports have been decreasing for the past five years. To find an alternative fuel other than fossil based fuels would be beneficial for Indonesia in order to be self sufficient in the energy sector. One alternative is to convert crude palm oil to CPO-Methyl Ester, widely known as biodiesel fuel. 81 Figure 10 : Growth of CPO Production and Demand Domestic + Export Source : Lembito, 2013 3.3. CPO Revenue and Cost Sub-model From the yearly price both in Domestic Market price and export market CIF Rotterdam, the CPO price per ton in year 2000 was Rp. 2 millions domestically, while the CIF Rotterdam price was USD 311. In year 2010 the export price was around US 900, slightly going down compared to the price in year 2008 which was US 948. USDA, 2010. The price drop is due to world economic crisis in year 2008. The production cost was estimated USD 400 per ton and we assumed the logistics cost of USD 60- USD 80 per ton. From the computer simulation we can predict the total revenue for Indonesia Palm Oil Industry is around in year 2030 to be around USD 53 billions or more than 3 times larger than that of the year 2010. And the gross profit excluding export tax seems promising still. Lembito, 2013 Figure 11 : Total Revenue and Total Cost Source : Lembito, 2013

3.4. Model Validation

For model validation we use data Total Growth Area and Total CPO production from year 2000-2010. The actual data taken from statistics and computation were compared with data simulation computed by Powersim Software studio 2005. From the analysis data we find that the forecast error is 3 and 3.4 . And the range of accepted errors in MAE is 5 . Model validations were done to prove the fitness of the model.